2017 Top Cited Papers from IWA Publishing Journals by IWA Publishing

IWA World Water Congress & Exhibition 2018 16th-19th September, Tokyo, Japan

Congress Journal Collection 2017 Top Cited Papers from IWA Publishing

iwaponline.com worldwatercongress.org

Copyright ÂŠ 2018 IWA Publishing Except as otherwise permitted under the Copyright, Designs and Patents Act, 1988, this publication may only be reproduced, stored or transmitted, in any form or by any other means, with the prior permission in writing of the publisher, or in the case of xerographic reproduction, in accordance with the terms of a licence issued by the Copyright Licensing Agency. In particular, IWA Publishing permits the making of a single photocopy of an article from this issue (under Sections 29 and 38 of this Act) for an individual for the purposes of research or private study. IWA Publishing is a company registered in the UK, no 3690822

Introducing IWA Publishingâ&#x20AC;&#x2122;s exclusive Congress Journal Collection IWA Publishing journals are industry-leaders in water, wastewater and related environmental fields. Our portfolio of 13 journals publish high-quality, peer-reviewed research, providing comprehensive coverage across all areas of interest for the global water community. To showcase our high-quality content and provide examples of the range of topics covered, we have here compiled the most-cited papers from each journal in 2017 in an exclusive collection for the IWA World Water Congress & Exhibition 2018. We hope that it is of great interest and use to delegates. For more information on any journal, including the aims and scope, editorial boards, instructions for authors and subscription options, please visit our e-content website: iwaponline.com

Page 1

Contents Page AQUA: Synthesis of chitosan zero-valent iron nanoparticles-supported for cadmium removal: Characterization, optimization and modeling approach

Efficient removal of some anionic dyes from aqueous solution using a polymer-coated magnetic nano-adsorbent

H2Open Journal: Prioritization of sub-catchments of a river basin using DEM and Fuzzy VIKOR

Occurrence of trihalomethane in relation to treatment technologies and water quality under tropical conditions

Journal of Hydroinformatics: Transient frequency response based leak detection in water supply pipeline systems with branched and looped junctions

Distribution of mean flow and turbulence statistics in plunge pools

Hydrology Research: A depth-duration-frequency analysis for short-duration rainfall events in England and Wales

101

From (cyber)space to ground: new technologies for smart farming

117

Journal of Water and Climate Change: Assessment of climate change impact on crop yield and irrigation water requirement of two major cereal crops (Rice and wheat) in Bhaktapur district, Nepal

135

Wavelet analyses of western us streamflow with ENSO and PDO

151

Journal of Water and Health: Towards a research agenda for water, sanitation and antimicrobial resistance

169

Safe drinking water and waterborne outbreaks

175

Journal of Water Re-use and Desalination: Heavy metal removal from wastewater using various adsorbents: A review

193

Influence of nitrite on the removal of Mn(II) using pilot-scale biofilters

227

Page 2

Page Journal of Water Sanitation, Hygiene and Development: The elimination of open defecation and its adverse health effects: A moral imperative for governments and development professionals

239

Qualitative comparative analysis for WASH research and practice

251

Water Policy: Linking environmental flows to sediment dynamics

267

Canadian and Australian researchersâ&#x20AC;&#x2122; perspectives on promising practices for implementing Indigenous and Western knowledge systems in water research

285

Water Practice Technology: Removal of pharmaceuticals with ozone at 10 Swedish wastewater treatment plants

307

Aerobic granular biomass technology: advancements in design, applications and further developments

319

Water Quality Research Journal: A novel cloud point extraction method for separation and preconcentration of cadmium and copper from natural waters

333

A computational fluid dynamics analysis of placing UV reactors in series

343

Water Science Technology: Nitrite inhibition and limitation - The effect of nitrite spiking on anammox biofilm, suspended and granular biomass

357

Combined ultrafiltration-electrodeionization technique for production of high purity water

367

Water Supply: Iron based sustainable greener technologies to treat cyanobacteria and microcystin-LR

379

In-stream detection of waterborne priority pollutants, and applications in drinking water contaminant warning systems

387

Page 3

Journal of Water Supply: Research and Technology

AQUA

ISSN 0003-7214 iwaponline.com/aqua

The Journal of Water Supply: Research and Technology - AQUA dealing covers research and development in water supply technology and management covering the complete water cycle. The journal’s scope includes: • Sustainable water resources management: Source water quality, quantity, protection • Applied limnology • Hydraulics of water systems including source waters, treatment and distribution systems • Water treatment processes, residuals treatment and management • Modelling of source waters, treatment and distribution systems • Applied methods to characterize water quality • Distribution systems • Water system management and policy: Legislation, economics, public relations, crisis management • Public health, risk assessment, regulations and standards • Water reclamation and reuse (e.g. for agricultural or industrial use) • Irrigation • Desalination systems for water supply For more details, visit iwaponline.com/aqua

Page 6

116

66.2

2017

Synthesis of chitosan zero-valent iron nanoparticlessupported for cadmium removal: characterization, optimization and modeling approach Mehdi Ahmadi, Majid Foladivanda, Nemat Jaafarzadeh, Zahra Ramezani, Bahman Ramavandi, Sahand Jorﬁ and Babak Kakavandi

ABSTRACT Herein, chitosan (CS) impregnated with nanoparticles of zero-valent iron (NZVI) was fabricated onto a 2þ

magnetic composite of CS@NZVI as an adsorbent for cadmium (Cd

) removal from aqueous solution.

The characteristics of CS@NZVI were analyzed by Fourier transform infrared spectroscopy, X-ray diffraction, transmission electron microscopy, CHONS and Brunauer, Emmett and Teller techniques. The average diameter of NZVI was found to be 50 nm, and it was successfully coated onto the CS. The influential experimental variables such as contact time, solution pH, adsorbent dosage and initial Cd2þ concentration were investigated to determine optimum conditions. Results revealed that with an optimum dosage rate of 0.6 g/L, Cd2þ concentration was reduced from 10 to 0.016 mg/L within 90 min reaction time at pH of 7 ± 0.2. Experimental data were fitted to the Freundlich and pseudo-secondorder models. Maximum adsorption capacity was obtained from the Langmuir monolayer 142.8 mg/g. Desorption experiments showed that the CS@NZVI had good potential with regard to regeneration and reusability, and its adsorption activity was preserved effectively even after three successive cycles owing to its good stability. As a conclusion, CS@NZVI can be considered as an effective adsorbent for heavy metals removal from water and wastewaters, because it can be separated both quickly and easily, it has high efficiency, and it does not lead to secondary pollution. Key words

| adsorption, cadmium, chitosan, magnetic composite, ZVI nanoparticles

INTRODUCTION

Mehdi Ahmadi Nemat Jaafarzadeh Sahand Jorﬁ Babak Kakavandi (corresponding author) Environmental Technologies Research Center, Ahvaz Jundishapur University of Medical Sciences, Ahvaz, Iran E-mail: Kakavandi.b@ajums.ac.ir Zahra Ramezani Nanotechnology Research Center, Ahvaz Jundishapur University of Medical Sciences, Ahvaz, Iran Mehdi Ahmadi Majid Foladivanda Nemat Jaafarzadeh Sahand Jorﬁ Babak Kakavandi Department of Environmental Health Engineering, Ahvaz Jundishapur University of Medical Sciences, Ahvaz, Iran Bahman Ramavandi Department of Environmental Health Engineering, Bushehr University of Medical Sciences, Ahvaz, Iran Babak Kakavandi Student Research Committee, Ahvaz Jundishapur University of Medical Sciences, Ahvaz, Iran

Heavy metals are highly toxic and hazardous elements that

applications such as phosphate fertilizers, batteries, electro-

have a high atomic weight and a density at least 5 times

plating industries, mining, metal production, stabilizers and

greater than that of water. They are widely used in industrial,

alloys and the manufacturing of pigments. It has been classi-

domestic, agricultural, medical and technological appli-

ﬁed as a human carcinogen and teratogen impacting lungs,

cations, which has led to their continuous release into the

kidneys, liver and reproductive organs (Azari et al. ;

environment. Due to their high degree of toxicity, arsenic

Naghizadeh ). The World Health Organization (WHO)

(As), cadmium (Cd), chromium (Cr), lead (Pb) and mercury

has set a maximum guideline concentration of 0.003 mg/L

(Hg) rank among the priority metals that are of public health

for Cd2þ in drinking water (WHO ). Considering the

signiﬁcance ( Jaafarzadeh et al. ; Begum et al. ). Cad-

negative effects, toxicity and stability of heavy metals, their

mium (Cd2þ) is one of the most dangerous pollutants that is

complete removal from water resources and wastewater

released into the environment, mainly via industrial

efﬂuents is deemed necessary.

doi: 10.2166/aqua.2017.027

Page 7

117

M. Ahmadi et al.

Chitosan zero-valent iron nanoparticles-supported for cadmium removal

Journal of Water Supply: Research and Technology—AQUA

66.2

2017

In the last several years, different technologies have

NZVI, due to its extremely small particle size, large speciﬁc

been studied to remove heavy metals from aqueous sol-

surface area and greater reactive sites and capacity, is

utions including adsorption, ion exchange, chemical

notable for this purpose in wastewater treatment to

precipitation, membrane ﬁltration and coagulation–ﬂoccula-

remove heavy metals with a higher efﬁciency (Esfahani

tion (Azari et al. ). However, most of them suffer from

et al. b). Moreover, the magnetic properties of NZVI

several disadvantages such as higher operational and capital

facilitate the rapid separation of nano iron from soil and

costs, more energy and chemicals consumption, and pro-

water via a magnetic ﬁeld (Babaei et al. a). However,

blems regarding sludge disposal (Sobhanardakani et al.

there is a strong tendency of NZVI particles to agglomerate

). Other drawbacks are the requirement for large settling

as well as to become oxidized, resulting in a reduction in

tanks in chemical precipitation, regeneration in the ion

surface area, reactivity and removal efﬁciency (Babaei

exchange process, chemical requirements, low efﬁciency in

et al. a). An effective approach to overcome this pro-

coagulation–ﬂocculation methods and large amounts of

blem is to incorporate NZVI into a porous supporting

sludge in membrane ﬁltration (Gupta et al. b; Kakavandi

material. Recent studies have reported that NZVI particles

et al. ). During the past few years, the adsorption process

can be coated with CS (a protective polymer due to its out-

has been widely applied; also, this process is proven to be

standing chelation behavior) to increase its dispersibility

a suitable method for the treatment of heavy metals

and stability (Liu et al. ). Furthermore, these supports

(Ahmadi et al. ; Amiri et al. ). In this regard, up to

can facilitate the separation of NZVI particles from aqu-

now, a wide variety of adsorbents have been used for Cd

2þ

eous solutions.

removal such as agricultural waste biomass, chitosan–

Herein, we hypothesize that NZVI particle impreg-

silica, microorganisms, biopolymers, zeolites, metal oxides,

nation on the CS surface combines the synergistic effects

ﬂy ash and activated carbon ( Jaafarzadeh et al. ; Lim

of NZVI and CS, which may have a superbly enhanced

& Aris ).

adsorption activity as well as easy separation. The present

However, most of these adsorbents showed a relatively

study therefore aimed to synthesize CS@NZVI using a

low adsorption capacity for Cd2þ under the optimum oper-

liquid phase method. The inﬂuence of operating parameters

ation conditions. In addition, some operational problems

in the adsorptive removal of Cd2þ was evaluated in details in

such as resultant turbidity in the treated water or efﬂuent,

a batch system. Isothermic and kinetic studies were also car-

and consequently the need to ﬁlter or centrifuge, have lim-

ried

ited the application of these adsorbents, particularly nano-

regeneration and reusability of the composite were indeed

sized adsorbents. Magnetic nanoparticles (e.g. NZVI,

evaluated for three consecutive cycles.

out

under

optimum

conditions.

Finally,

the

Fe3O4, α-Fe2O3, γ-Fe2O3 and FeO(OH)) have recently been adopted by researchers in the ﬁeld of adsorption/biosorption for removing pollutants from aquatic environments, which makes separation of both adsorbent and adsorbate

MATERIALS AND METHODS

much easier (Mohseni-Bandpi et al. ). Several authors have magnetized adsorbents such as activated carbon for

Materials and chemicals

Pb2þ and Hg adsorption (Oliveira et al. ; Kakavandi et al. ), carbon nanotubes for Pb2þ, Ni and Sr adsorption

All chemicals were of analytical laboratory grade and used

(Chen et al. ; Hu et al. ), zeolite for Cr, Cu, and Zn

without further puriﬁcation. Sodium borohydride (NaBH4)

adsorption (Oliveira et al. ) and CS for Zn2þ and Pb2þ

was purchased from Sigma-Aldrich. Cadmium nitrate tetra-

adsorption (Fan et al. , ) by magnetic iron nanopar-

hydrate (Cd(NO3)2.4H2O, Merck, Co) was used for

ticles as a magnetic separation technology.

preparing the stock solutions of Cd2þ according to the

Among magnetic nanoparticles, NZVI has been

ASTM D3557-12 (ASTM ) procedure. The pH of the sol-

applied recently for in-situ and ex-situ remediation, due to

utions was adjusted by adding 0.1 M hydrochloric acid

being non-toxic and inexpensive (Esfahani et al. a).

(HCl) and sodium hydroxide (NaOH) solutions. All the

Page 8

118

M. Ahmadi et al.

Chitosan zero-valent iron nanoparticles-supported for cadmium removal

reagents were prepared with de-ionized water (DI-water)

Journal of Water Supply: Research and Technology—AQUA

66.2

2017

Synthesis of the CS@NZVI

and kept in a refrigerator at 4 C prior to experiments. W

CS@NZVI composite was synthesized in the laboratory using a chemical reduction method (reducing Fe3þ to Fe0

CS preparation

using NaBH4). Excess NaBH4 was used to ensure that all In this work, CS was prepared from shrimp shell wastes,

the Fe3þ was reduced. Firstly, 0.25 g of CS was dissolved

which is available in abundance in southern parts of Iran. It

in 50 mL of 0.05 M acetic acid. Due to the poor solubility

was obtained from chitin according to the method reported

of CS, the mixture was vortexed to aid complete dissolution

in the literature with some modiﬁcations (Brown ).

and kept for 2 h at 150 rpm. To this solution, 1 g of

Initially, in order to improve the purity of shrimp shells,

FeCl3.7H2O was added and the solution was stirred quickly

they were washed with DI-water and dried, then ground

in an N2-purged environment for 2 h. Then, to this mixture

and passed through a size 50 mesh. Graded shells were agi-

freshly prepared aqueous solution containing 2% NaBH4

tated in 0.5% NaOH solution for 30 min and ﬁnally washed

was added drop-wise. At this stage, black precipitation was

with hot DI-water several times until pH reached neutral

observed, and evolution of H2. Again, the mixture was stir-

value. This completed the preliminary phase for the prep-

red for another 60 min until the entire reduction of metal

aration of chitin. De-proteinization of the shrimp shells was

salts. The black solid was collected using a magnet (with a

performed using a 1.2N NaOH solution (1:20 w/v) for 3 h

1.5 tesla ﬁled magnet) and washed at least three times

at 90 C and in constant agitation conditions. The residue

with oxygen-free DI-water to get rid of the extra chemicals.

was separated by ﬁltration and washed with hot DI-water sev-

The CS@NZVI composite was dried at 100 C for 4 h, and

eral times. Thereafter, it was demineralized with a 1.6N HCl

stored in a brown sealed bottle under dry conditions for

solution (1:10 w/v) at room temperature (25 ± 2 C) for 2 h.

characterization and future use (Geng et al. ; Gupta

After ﬁltration, the residue was again washed with hot

et al. a).

DI-water until the pH reached 7. Finally, the obtained chitin was decolorized via agitation in acetone solution [(CH3)2CO] for 1 h in order to remove all pigments. Chitin was separated, dried at 60 C for 24 h and then weighed to determine the W

chitin content of the shrimp shells according to Equation (1) (Westergren ). Afterwards, the obtained chitin was deacetylated using a 50% NaOH (1:5 w/v) solution at boiling temperature for 8 h. After de-acetylation, the produced CS was washed with DI-water until the pH reached 7, and dried in an oven at 60 C for 24 h. Thereafter, it was weighed and W

the CS yield was determined using Equation (2), and assayed for degree of de-acetylation. One of the easiest ways to detect CS production is dissolving in a weak acid solution as an indicator test (Westergren ; Brown ).

Characterization of CS@NZVI In order to determine the CS degree of acetylation, the elemental composition was analyzed using a COSTECH ECS 4010, Italia, CHONS equipment. The percentage of N-deacetylated varies from 5.145 in completely N-deacetylated CS (C6H11O4N repeat unit) to 6.861 in chitin, the fully N-acetylated polymer (C8H13O5N repeat unit). The DA % of CS samples was calculated via Equation (3) (Al Sagheer et al. ).

DA(%) ¼

C=N � 5:145 × 100 6:861 � 5:145

(3)

X-ray diffraction (XRD) spectra of CS@NZVI were %Chitin ¼

Product (g) × 100 Shell (g)

(1)

obtained using a Quantachrome, 2000, NOVA X-ray Diffractometer with graphite monochromatic copper radiation (Cu Kα, λ ¼ 1.54 Å) in the range of 10–70 . The patterns W

were compared with the Joint Committee on Powder Dif%Chitosan ¼

Product (g) × 100 Shell (g)

(2)

fraction Standards (JCPDS). The speciﬁc surface area and pore volume of CS@NZVI were measured by the Brunauer, Page 9

119

M. Ahmadi et al.

Chitosan zero-valent iron nanoparticles-supported for cadmium removal

Emmett and Teller (BET, Quantachrome, 2000, NOVA) method using N2 adsorption–desorption isotherms at 77.3 K. A transmission electron microscopy (TEM, PHILIPS, EM) was used to characterize the size and shape of NZVI particles at 100 keV. Fourier transform infrared spectroscopy (FTIR) spectra of the CS@NZVI composite were obtained using BRUKER’s Vertex 70 model to conﬁrm the

Journal of Water Supply: Research and Technology—AQUA

66.2

Ce × 100 Cd removal efficiency (%) ¼ 1 � Ci 2

2017

(5)

where Ci and Ce are the initial and residual concentrations of Cd2þ (mg/L) in the solution, respectively, m is the dry mass of CS@NZVI (g) and V is the volume of the solution (L).

functional groups present on the adsorbent surfaces.

RESULTS AND DISCUSSION Batch adsorption experiments Characterization of CS@NZVI 2þ

on CS@NZVI

Batch experiments for the adsorption of Cd

composite were carried out in 250 mL polytetraﬂuoroethylene 2þ

XRD can provide very useful information about the physical

solutions at

and chemical structures of the magnetic particles embedded

25 ± 1 C. The effects of experimental parameters such as the

in the CS matrix. The XRD spectra of CS@NZVI in the 2θ

pH of the solution, contact time, different CS@NZVI and

range of 0–80

bottles ﬁlled with 50 mL of the pH-adjusted Cd W

2þ

concentrations and solution temperatures on the

removal efﬁciency of Cd2þ were investigated. After adjusting

at 25 C (Cu Kα, λ ¼ 1.54 Å) are shown in W

Figure 1(a). A narrow diffraction peak at 2θ ¼ 44.9

was

observed, belonging to NZVI crystal (JCPDS, No. 06-0696)

the pH of the solution, a speciﬁc amount of composite was

(Fu et al. ; Babaei et al. a). This conﬁrms that the

put in the aqueous solution, having a ﬁxed concentration.

NZVI particles were successfully synthesized. In addition,

Then, bottles were agitated on a rotary shaker at a rate of

the X-ray pattern of CS@NZVI exhibited characteristic crys-

200 rpm and maintained for a certain period of time at a con-

talline peaks belonging to CS at 2θ ¼ 8 and 20.1 ( Jagtap

stant temperature (25 ± 1 C). At appropriate time intervals, W

et al. ; Mohseni-Bandpi et al. ). These results suggest

2 mL of the solution was withdrawn from each bottle and

that the NZVI particles were successfully loaded on either

the composite was magnetically separated using a strong

the outside or inside of CS.

magnet. After that, the remaining Cd2þ concentration in the

The results of TEM analysis, Figure 1(b), showed that

solution was determined according to the ASTM (D3557-90

NZVI had a diameter less than 50 nm and also demon-

method) (ASTM ) using atomic absorption spectropho-

strated that it was successfully synthesized as individual

tometry (Analytikjena, vario 6, Germany) at a wavelength of

nano-sized particles. The speciﬁc surface area, volume,

228.8 nm. Herein, all measurements were performed in an

and average pore diameter of CS@NZVI were measured

air/acetylene ﬂame. The lamp current and slit width were

using the BET method. The surface of the synthesized adsor-

2.0 mA and 1.2 nm, respectively. The instrument was cali-

bent, according to this analysis, was 78.3 m2/g. It is notable

brated with a standard solution (in the range 0.05–2.0 mg/L)

that the speciﬁc surface area of CS decreased after the coat-

within a linear range, and a high correlation coefﬁcient

ing of NZVI, as reported in the literature (Babaei et al.

(R2 > 0.997) was obtained. All experiments were performed

a). This decrease may result from the impregnation pro-

in duplicate and the results were reported as the mean values

cess and/or NZVI presence in the structure of CS. Similar

of measurements.

observations were also reported by other researchers (Kaka-

The amount of the Cd2þ adsorbed on CS@NZVI, qe

vandi et al. , ). The average size and volume pores of

(i. e. adsorption capacity, mg/g), and the removal efﬁciency

the composite were obtained to be 26.57 nm and 0.982 cc/g,

were calculated using the following equations:

respectively. According to the IUPAC classiﬁcation, the

V Cd2 adsorption capacity (mg=g) ¼ (Ci � Ce ) m Page 10

average size of 26.57 nm can be classiﬁed as mesoporous (4)

groups (Depci ). The results of this analysis reveal that the CS@NZVI is porous in structure and could provide

120

Figure 1

M. Ahmadi et al.

Chitosan zero-valent iron nanoparticles-supported for cadmium removal

Journal of Water Supply: Research and Technology—AQUA

66.2

2017

(a) Powder XRD pattern, (b) TEM image of CS@NZVI, (c) FTIR spectra of CS@NZVI before, and (d) after Cd2þ adsorption.

more reactive sites and a good adsorption capacity for con-

adsorption process. To characterize the functional groups

taminants. Because adsorption reactions mainly occurred

on the surfaces of the adsorbent and to measure the binding

on the adsorbent surfaces, the functional groups on the sur-

mechanism of the pollutants, the FTIR spectra of the

faces of the adsorbent can play a signiﬁcant role in the

CS@NZVI before and after adsorption of Cd2þ in the Page 11

121

M. Ahmadi et al.

Journal of Water Supply: Research and Technology—AQUA

Chitosan zero-valent iron nanoparticles-supported for cadmium removal

66.2

2017

range of 400–4,000 cm�1 are shown in Figure 1(c) and 1(d),

vibration, implying that the NZVI nanoparticles were suc-

respectively.

cessfully prepared and introduced into the CS (Yang et al.

The FTIR spectra showed some absorption peaks

).

belonging to various functional groups or different vibration

In the CS@NZVI spectrum after adsorption, a signiﬁ-

modes. A comparison between the FTIR spectrums of the

cant reduction of absorption in this spectral area can be

CS@NZVI before and after the adsorption of Cd

2þ

is given

in Table 1. The absorption bonds at wave number (ν) values at ∼3,354 and ∼3,268 cm

�1

attributed to the formation of CS � Fe bonds. All the aforementioned

peaks were also observed

in the

‘after

indicate the presence of

adsorption’ FTIR spectra with notable changes. These func-

O–H and N–H bond stretching, respectively. The absorption

tional groups may form surface complexes with Cd2þ and

peaks at 2,860 cm�1 are due to the C–H stretching vibration

thus can increase the speciﬁc adsorption of Cd2þ by

of the –CH2 groups in CS (Du et al. ; Mohseni-Bandpi

CS@NZVI. As shown in Figure 1(c) and 1(d) and

et al. ). The peak observed at 1,631 cm�1 may be from

Table 1, the spectra display a number of absorption

the N-H bending vibration, indicating the existence of

peaks, indicating the complex nature of the CS@NZVI.

amide(II) and hydroxyl groups in CS (Liu et al. ). More-

Large changes are clearly observed on the FTIR spectrum

over, the bond at near 1,600 cm�1 that appeared on

of CS@NZVI following Cd2þ adsorption. After Cd2þ

CS@NZVI before and after adsorption of 10 mg/L Cd2þ

adsorption, the FTIR spectrum, Figure 1(d), shows a new

was assigned to the OH bending vibrational mode due to

strong peak at 2,868 cm�1, belonging to the stretching

the adsorption of moisture when FTIR sample disks were

vibration of symmetric and asymmetric –CH2 groups

prepared in an open-air atmosphere (Mohseni-Bandpi

(Ngah et al. ). Furthermore, FTIR spectra of Cd2þ

�1

et al. ). The bands at about 1,363 cm

can be attributed

adsorbed on CS@NZVI indicated that the peaks expected

to C–N stretching vibration (Malkoc & Nuhoglu ). In

at 3,354, 3,268, 2,856, 1,589, 1,363 and 1,147 cm�1 had

�1

can be

shifted, respectively to 3,357, 3,288, 2,868, 1,597, 1,376

apportioned to the C¼ O stretching of ether groups

and 1,151 cm�1 due to Cd2þ sorption. It seems that the

(Malkoc & Nuhoglu ). The peaks at 1,083 cm�1 and

mentioned functional groups inﬂuence the Cd2þ adsorption

the FTIR spectra, the peaks at around 1,140 cm

1,023 cm

�1

correspond to C–OH bond stretching (Reddy �1

& Lee ). The peaks at around 570 cm

in the

CS@NZVI spectrum were attributed to the Fe–O stretching

Table 1

on the CS@NZVI. Generally, the ﬁndings of FTIR studies clearly conﬁrm the existence of CS and NZVI in the CS@NZVI composite.

The FTIR spectral characteristics of CS@NZVI before and after Cd2þ adsorption �1

Frequencies (cm

)

IR peaks

Before adsorption

After adsorption

Differences

Assignment

References

3,354–3,268

3,357–3,288

–3, –20

O–H bond stretching and N–H bond stretching

Babaei et al. (b), Mohseni-Bandpi et al. ()

2,856

2,868

–12

C–H stretching vibration of the –CH2 groups

Viswanathan & Meenakshi (), MohseniBandpi et al. ()

1,589

1,597

–8

OH bending vibrational

Mohseni-Bandpi et al. ()

1,363

1,376

–4

C–N stretching vibration

Malkoc & Nuhoglu ()

1,147

1,151

1,063

1,078

–3

C–OH bond stretching

Reddy & Lee ()

1,024

1,027

–3

C–OH bond stretching

Reddy & Lee ()

572

556

Fe–O stretching vibration

Yang et al. ()

Page 12

þ6

C ¼ O stretching of ether groups

Malkoc & Nuhoglu ()

122

M. Ahmadi et al.

Chitosan zero-valent iron nanoparticles-supported for cadmium removal

Inﬂuence of initial solution pH

Journal of Water Supply: Research and Technology—AQUA

66.2

2017

obtained when pH values of the solution were 6.0, 7.0 and 8.0, respectively, indicating that the changes of the removal

The pH of the solution, affecting the surface functional

efﬁciency is not notable. Hence, to ensure the interference

groups of the adsorbent and adsorbate, is one of the most

from metal precipitation we set the pH of solutions at 7.0

inﬂuential parameters of the adsorption process. The domi-

for the following experiments. This would allow for Cd2þ

nant forms of heavy metals in aqueous solution were also

removal in wastewater without pre-adjustment of pH.

affected by pH (Kakavandi et al. b). It has been pre-

These results are in good agreement with those of previous

viously reported that the efﬁciency of heavy metals

studies for several sorbent-Cd2þ sorption processes (Ünlü

adsorption is highly dependent on the initial pH of the sol-

& Ersoz ; Wang et al. ; Liu et al. ). Azouaou

ution. Similar results were also observed in our work,

et al. () in studying Cd2þ adsorption shows that at pH 7

which are illustrated in Figure 2. As can be seen, Cd2þ

the numbers of competing hydrogen ions are lower and

adsorption percentages enhanced with an increase in the

more ligands are exposed with negative charges, resulting

pH from 4 to 8 for 10 mg/L Cd2þ concentration during

in greater cadmium sorption.

then

Liu et al. () studied the Cd2þ adsorption on CS

decreased signiﬁcantly, when the pH value reached 9.0.

beads-supported Fe0 and showed that when solution pH

Cd2þon

increased, the number of negatively charged sites was

CS@NZVI is favored at around neutral pH values, which

improved, leading to the enhanced attraction force between

can be attributed to the changes of surface properties of

heavy metals (Cu2þ, Cd2þ and Pb2þ) and the beads surface.

the adsorbent and adsorbate. A maximum Cd2þ uptake

Furthermore, Azari et al. () reported that as the pH

was observed at a solution pH of 8.0. At acidic conditions

increased, surface positive charges of the adsorbent

(pH < 5), the surface of the adsorbent is positive and so,

decreased and the more active surface sites can be obtained

electrostatic repulsion occurs between protons (Hþ) and

for Cd2þ, which resulted in lower repulsion of the adsorbing

Cd2þ cations for the adsorption sites. Therefore, competition

metal ions. At alkaline conditions, however, a decrease in

between protons and metal species could be a reason for the

the adsorption efﬁciency can be derived from the formation

weak adsorption in this condition. After 90 min reaction,

of metal hydroxides precipitation and also a decrease in the

removal efﬁciencies of 96.5%, 97.6% and 98% were

concentration of Cd2þ, as reported in the literature (Kaka-

90 min Figure

agitation 2

time.

indicates

that

The the

removal

efﬁciency

adsorption

vandi et al. ). Rao et al. () reported that at pH > 7.5, the predominant species of Cd exists in the hydrolyzed form (i.e. Cd(OH)þ and Cd(OH)02) and Cd2þ ions are present in only very small amounts. Therefore, at the value of pH < 7.0, the main species adsorbed onto the CS@NZVI were predominantly Cd2þ and less amounts of Cd(OH)þ and Cd(OH)2. Based on the aforementioned, at the optimum pH the predominant species of Cd were in ionic form (Cd2þ) and metal hydroxide precipitation does not take place. Inﬂuence of adsorbent dosage The dosage of CS@NZVI as a factor inﬂuencing the adsorption of Cd2þ was also investigated. It was examined in the range of 0.2–0.7 g/L at this condition: 10 mg/L of Cd2þ over 90 min at pH 7.0 ± 0.2. As shown in Figure 3(a), by raisFigure 2

Effect of solution pH on the adsorption of Cd2þ on CS@NZVI (Experimental conditions: adsorbent dose ¼ 0.6 g/L; C0 ¼ 10 mg/L; contact time ¼ 90 min;

and T ¼ 25 ± 1 C). W

ing adsorbent dosage from 0.2 to 0.7 g/L, the removal percentage of Cd2þ ions signiﬁcantly increased from 81.7 Page 13

123

Figure 3

M. Ahmadi et al.

Chitosan zero-valent iron nanoparticles-supported for cadmium removal

Journal of Water Supply: Research and Technology—AQUA

66.2

2017

Effect of (a) adsorbent dose and (b) initial Cd2þ concentration on adsorption capacity and removal of Cd2þ by CS@NZVI. Experimental conditions: pH ¼ 7.0 ± 0.2; contact time ¼ 90 min; and T ¼ 25 ± 1 C; for (a) C0 ¼ 10 mg/L; for (b) adsorbent dose ¼ 0.6 g/L. W

to 99.9%, while the adsorption capacity (the amount

the adsorption reaction (Jafari et al. ; Kakavandi et al.

adsorbed per unit mass of adsorbent) declined from 40.8

a). In addition, some of the particle interactions (e.g.

to 14.3 mg/g. The promotion of sorption efﬁciency can be

aggregation) which result from a high sorbent concentration

explained by the fact that increasing the adsorbent dosage

lead to a signiﬁcant reduction in the active surface area of

increased the accessibility of active sites on the pores of

the adsorbent and, consequently, reduce its adsorption

the CS@NZVI to the Cd2þ ions, which led to an enhanced

capacity. Similar observations have been reported for

removal efﬁciency, as observed by the other researchers

adsorption of Cd2þ onto the different adsorbents in the lit-

(Rao et al. ; Azari et al. ). In other words, for a

erature (Rao et al. ; Shen et al. ; Azari et al. ).

ﬁxed initial adsorbate concentration, increasing adsorbent dosage provides greater surface area or more adsorption

Inﬂuence of initial Cd2þ concentration

sites. As shown in Figure 3(a), the complete removal of Cd2þ was approximately achieved at high adsorbent

The effect of initial concentrations of Cd2þ on its removal

dosage level (0.7 g/L). The experiments also indicated that

efﬁciency by CS@NZVI in the range of 10-300 mg/L is

the removal efﬁciency was faster as the adsorbent dosage

shown in Figure 3(b). It can be seen that the removal efﬁ-

raised from 0.2 to 0.6 g/L. According to Figure 3(a), a

ciency decreased with enhancement of the Cd2þ from 10

removal efﬁciency of 99.8 and 99.9% was obtained in the

to 300 mg/L. So that, with the rise in the initial concen-

presence of 0.6 and 0.7 g/L, respectively, of CS@NZVI in

tration from 10 to 300 mg/L, the removal efﬁciency

2þ

solution, demonstrating that beyond the 0.6 g/L

decreased from 99.8% to 34.4%. This is probably due to

dosage the removal efﬁciency did not change with the adsor-

the ﬁxed number of active sites on the adsorbent versus

bent dose. Hence, 0.6 g/L was chosen as the optimal dosage

the number of metal ion molecules (Teymouri et al. ).

of the adsorbent to conduct further experiments on the

Figure 3(b) also reveals that the adsorption capacity of

adsorption process.

Cd2þ on the CS@NZVI signiﬁcantly enhanced as the initial

the Cd

However, a decrease in adsorption capacity with an

Cd2þ concentration increased. This phenomenon can be

increase in the adsorbent dosage is mainly attributed to

described by the fact that amounts of Cd2þ adsorbed per

the increase in unsaturation of adsorption sites through

unit mass of CS@NZVI increase with an increase in initial

Page 14

124

M. Ahmadi et al.

Journal of Water Supply: Research and Technology—AQUA

Chitosan zero-valent iron nanoparticles-supported for cadmium removal

66.2

2017

Cd2þ concentration in the solution. Moreover, an increase

that the rate of Cd2þ adsorption onto CS@NZVI was initially

in initial concentration dramatically enhanced the inter-

fast; then, the rate slowed down gradually until the equili-

action between the adsorbent and Cd2þ. This can be

brium was reached, beyond which no further adsorption

attributed to the increased force of concentration gradient

could be observed. Thereafter, 90 min was selected for the

(Kumar et al. ).

future experiments as the equilibrium time. The rapid

Generally, at higher initial concentration of metal ions

increase of the adsorption capacity in the initial stages

the available adsorption sites of the CS@NZVI become

might be due to the availability of a large number of vacant

fewer and the percent removal of metal ions is dependent

sites that become saturated over time (Kakavandi et al.

upon the initial concentration. However, the ratio of the

). With further increasing time, the availability of the

initial number of metal ions to the available sorption sites

Cd2þ ions to unoccupied active sites on the surface of the

of the CS@NZVI is decreased at a lower initial concen-

adsorbent diminished; and these sites ultimately become satu-

2þ

, and subsequently the fractional adsorption

rated when the process reaches its equilibrium state. The

of metal ions by the CS@NZVI becomes independent of

adsorption equilibrium is the point at which the concen-

its initial concentration (Rao et al. ; Yang et al. ;

tration of the adsorbate in the bulk solution is in a dynamic

Azari et al. ).

balance with that of the interface (Kakavandi et al. b).

tration of Cd

In Table 2, the values of the kinetic model parameters of Inﬂuence of contact time and adsorption kinetics

2þ

adsorption onto CS@NZVI are listed. In this study, we

used four widely used kinetic models: pseudo-ﬁrst-order, The effect of contact time and adsorption kinetics were

pseudo-second-order, Elovich, and intraparticle diffusion

studied at a period of 3 h under optimum conditions (i.e.

models to estimate overall sorption rates. Further details of

pH of 7.0 ± 02 and 0.6 g/L of CS@NZVI) for 10 mg/L

these models (i.e. equations and parameters) are given in

Cd2þ. As shown in Figure 4(a), the adsorption capacity of

the supplementary data, Table S1 (available with the

2þ

increased rapidly during the ﬁrst 60 min and then

online version of this paper). The correlation coefﬁcients

reached the equilibrium point after 90 min. It was observed

were found to be less than 0.96, 0.85 and 0.67 for the

Figure 4

(a) Kinetic and (b) isotherm models and experimental data of Cd2þ adsorption on CS@NZVI under optimum conditions (pH ¼ 7.0 ± 0.2; adsorbent dose ¼ 0.6 g/L and T ¼ 25 ±

1 C; for (a) C0 ¼ 10 mg/L; for (b) C0 ¼ 10-300 mg/L and contact time ¼ 90 min). W

Page 15

125

Table 2

M. Ahmadi et al.

Chitosan zero-valent iron nanoparticles-supported for cadmium removal

The values of kinetics and isotherms of Cd2þ adsorption on CS@NZVI

Models

Parameters

Value

Kinetic qe,cal (mg/g)

11.25

kf (min�1)

0.056

0.9519

Pseudo-second-order t/qt ¼ t/qe þ 1/k2 q2e

qe,cal (mg/g)

ks (g/mg min)

0.016

0.9999

ki (mg/gmin0.5)

0.6415

Intraparticle diffusion qt ¼ ki t0.5

Ci (mg/g) R2

12.61 0.66

Elovich qt¼ β ln(αβ)þ β lnt

α (mg/g min)

12.73

β (g/mg)

2.44

0.8479

qe,exp (mg/g)

19.52

Isotherm kF (mg/g(Lmg)1/n)

31.5

3.63

0.9998

Langmuir Ce/qe ¼ Ce/q0 þ 1/kLq0

q0 (mg/g)

142.8

kL (L/mg)

0.062

0.05 - 0.61

0.9816

Temkin qe ¼ B1 ln KT þ B1 ln Ce

qm (mg/g)

15.96

10.5

0.9113

D-R ln qe ¼ ln qm-Dε2

66.2

2017

model was more than 0.999. This suggests that the pseudosecond-order is a better fit to the experimental data of Cd2þ adsorption with a significantly high coefficient of corthe pseudo-second-order model it is also strongly confirmed that the calculated qe values are in good agreement with the experimental qe values, indicating that this model better explains the adsorption process of Cd2þ on the CS@NZVI than the other models. The confirmation of this model demonstrates that the concentrations of both adsorbent and adsorbate are associated with the rate-determining step of the adsorption process ( Jafari et al. ). It also suggests that chemisorption was the rate-limiting step in the adsorption process of Cd2þ onto the CS@NZVI, and there was no mass transfer reaction (Kakavandi et al. ). In the previous studies conducted, the same model for the adsorption of Cd2þ on various adsorbents, such as magnetic activated carbon (Azari et al. ), activated carbon (Rao et al. ) (Dong et al. ), and clarified sludge (Naiya et al. ) were reported. Other models (i.e. pseudo-first-order, Elovich and intraparticle diffusion) present lower R2 values, indicating that

Freundlich lnqe ¼ lnkF þ n�1 lnCe

relation (R2) (>0.99), compared to other kinetic models. For

Pseudo-ﬁrst-order ln(qe-qt) ¼ lnqe-k1 t

Journal of Water Supply: Research and Technology—AQUA

qm (mol/g) D (mol2/kJ2) E (kJ/mol) R

215.1 0.0023 14.74 0.9865

these models could not properly ﬁt the experimental kinetic data. Based on the results, it was found that the intraparticle diffusion model plays a less signiﬁcant role in the adsorption process. According to Table 2, for intraparticle diffusion the y-intercept (Ci) is not zero, illustrating that the intraparticle diffusion is part of the adsorption but not the only rate-controlling step in this process, as reported previously by Boparai et al. (). Therefore, it can be stated that other mechanisms (i.e. complexes or ion-exchange) could also control the rate of the adsorption of Cd2þ on CS@NZVI. Adsorption equilibrium and isotherm study In this study, Langmuir, Freundlich, Temkin and Dubinin– Radushkevich (D-R) equilibrium as the four most common isotherm models were employed to predict the behavior of Cd2þ adsorption onto the CS@NZVI surfaces. The equations and corresponding parameters of the aforementioned models are represented in Table S1. The adsorption isotherm exper-

pseudo-ﬁrst-order, Elovich and intraparticle diffusion kin-

iments were conducted using 10 to 300 mg/L Cd2þ under

etic models, respectively; whereas the corresponding

the optimum conditions (i.e. pH 7.0 ± 0.2, 0.6 g/L adsorbent

amount calculated for the pseudo-second-order kinetic

and 90 min contact time) at 25 ± 1 C. Table 1 shows the

Page 16

126

M. Ahmadi et al.

Chitosan zero-valent iron nanoparticles-supported for cadmium removal

obtained values of equilibrium isotherm parameters of the Cd

2þ

adsorption onto the CS@NZVI surfaces. Based on the 2

Journal of Water Supply: Research and Technology—AQUA

66.2

2017

capacity of other adsorbents applied in previous research. The observed differences in the adsorption capacities for

correlation coefﬁcients (R ), the adsorption isotherm

the listed adsorbents can be due to the structure, surface

models ﬁtted the experimental data in accordance to the fol-

area and the properties of the functional groups in each

Freundlich > D-R > Langmuir > Temkin.

adsorbent. As presented in Table 3, α-ketoglutaric acid-

Considering this result, the Freundlich model is a better ﬁt

modiﬁed magnetic CS provides a high adsorption capacity

lowing

order:

adsorption by

for Cd2þ compared with other adsorbents, which can be

CS@NZVI than the other three models. In addition, we

attributed to its textural characteristics, high porosity and

observed the best ﬁt for the Freundlich model by employing

surface area and functional groups. Moreover, it is worth

to the experimental data of the Cd

2þ

a nonlinear method, as plotted in Figure 4(b). This model

noting from this table that the CS@NZVI had a positive

suggests that the heterogeneous functional sites are distribu-

effect on Cd2þ removal and can be considered as one of

ted uniformly on the surfaces of CS@NZVI and the

the most effective adsorbents for Cd2þ adsorption. Never-

adsorption of Cd2þ ions onto non-energetically equivalent

theless, in order to enhance the adsorption capacity of

sites of the CS@NZVI (Kakavandi et al. ; Rezaei Kalantry

CS@NZVI, further studies can be conducted on its modiﬁ-

et al. ). Meanwhile, the value of 1/n (less than unity) in

cation through increasing the surface area and changing of

the Freundlich isotherm model implies the favorable adsorp-

functional groups.

2þ

tion of Cd

onto CS@NZVI. In addition, as presented in

Table 2, the values for the dimensionless separation parameter RL (RL ¼ 1/(1 þ kLC0)), which were related to the

Langmuir model, fell between 0 and 1. Since RL > 1, RL ¼

1, RL ¼ 0 and 0 < RL < 1 indicate unfavorable, linear, irre-

versible and favorable adsorption, respectively, it can be concluded that the simultaneous adsorption of Cd2þ onto CS@NZVI is favorable. For the D-R model, the mean free energy of adsorption

(E ¼ 1/(-2D)0.5) per mole of the adsorbate is the energy

needed to transfer one mole of an adsorbate to the adsor-

bent surfaces from inﬁnity in solution. It gives information about either chemical or physical adsorption. With the magnitude of E, between 8 and 16 kJ/mol, the adsorption mechanism follows chemical ion-exchange, while for the values of E < 8 kJ/mol, the adsorption process is of a physical nature (Azouaou et al. ; Kakavandi et al. ). As shown in Table 2, the value of the mean free energy of adsorption, E, for Cd2þ on CS@NZVI, was found to be between 8 and 16 kJ/mol, indicating that the adsorption process follows a chemical mechanism. The chemisorption nature of Cd2þ adsorption on different types of adsorbents

Regeneration and reusability of CS@NZVI The regeneration of the adsorbent and the restoration of adsorption are crucial factors in the applicability of a typical adsorbent. In this study, regeneration and the reusability experiments of Cd2þ on the CS@NZVI were assessed under the optimum conditions through three successive cycles. To regenerate the spent CS@NZVI at the end of each adsorption cycle for the next adsorption, the used adsorbent was collected magnetically and stirred in DI-water for 90 min. The adsorbent was subsequently ﬁltered and dried overnight for the next use. For desorption experiments, 0.10 g of CS@NZVI loaded with Cd2þ was then shaken at 200 rpm for 90 min with 5 mL of DI-water at 25 ± 1 C. The regenerated adsorbent was then dried in W

an oven at 100 C for 2 h and used for the next adsorpW

tion–desorption cycle, in order to test the reusability of CS@NZVI for Cd2þ removal. At the end of each adsorption/desorption cycle, the desorption percentage (%) was calculated using Equation (6).

has been reported previously (Ünlü & Ersoz ; Naiya et al. ; Boparai et al. ). The

maximum

adsorption

capacity,

qm,

the

Desorption (%) ¼

Amount of Cd2þ desorbed Amount of Cd2þ adsorbed

× 100

(6)

CS@NZVI was compared with the other adsorbents (see Table 3). It is worth mentioning that the CS@NZVI

As can be seen from Figure 5, the adsorption percen-

poses a better adsorption capacity, compared with the

tages of Cd2þ by CS@NZVI slightly dropped from 99.8% Page 17

127

Table 3

M. Ahmadi et al.

Chitosan zero-valent iron nanoparticles-supported for cadmium removal

Journal of Water Supply: Research and Technology—AQUA

66.2

2017

Maximum adsorption capacity and optimum conditions of different adsorbents for Cd2þ removal qm (mg/g)

Adsorbent

Isotherm

Kinetic

Ion-imprinted carboxymethyl CS-functionalized silica gel

5.0

Pseudo-second-order

20.7

Crosslinked CS/poly(vinyl alcohol) beads

6.0

Langmuir and Freundlich

Pseudo-second-order

142.9

Kumar et al. ()

α-Ketoglutaric acid-modiﬁed magnetic CS

6.0

Langmuir

Pseudo-ﬁrst-order

255.7

Yang et al. ()

from Phaseolus aureus hulls

8.0

Freundlich

Pseudo-second-order

15.7

Rao et al. ()

Magnetic activated carbon

5.7

Langmuir

Pseudo-second-order

63.52

Azari et al. ()

Clariﬁed sludge

5.0

Langmuir

Pseudo-second-order

36.23

Naiya et al. ()

References

Lü et al. ()

Activated carbon prepared

Dithiocarbamated-sporopollenin

7.0

Langmuir

Pseudo-second-order

7.1

Ünlü & Ersoz ()

Untreated coffee grounds

7.0

Freundlich

Pseudo-second-order

15.65

Azouaou et al. ()

Untreated Pinus halepensis sawdust

9.0

Freundlich

Pseudo-second-order

5.36

Oxidized granular activated carbon

6.0

Langmuir

Pseudo-second-order

NaCl-treated Ceratophyllum demersum

6.0

Langmuir

Pseudo-second-order

35.7

5.73

CS@NZVI

7.0

Freundlich

Pseudo-second-order

142.8

Semerjian () Huang et al. () Jaafarzadeh et al. () This study

desorbent solutions such as HCl, NaCl, NaOH and methanol could provide a good potential for regeneration of CS@NZVI.

CONCLUSIONS Results revealed that CS@NZVI has a high potential and adsorption capacity for Cd2þ ion removal from aqueous solutions. At a pH of 7 ± 0.2, the adsorption efﬁciency was enhanced by an increase in the contact time and adsorbent Figure 5

Reusability and regeneration results for the adsorption of Cd2þ by CS@NZVI composite in aqueous solution.

dosage and a decrease in the initial Cd2þ concentration. The equilibrium adsorption data were found to ﬁt best using a Freundlich isotherm and pseudo-second-order kinetic models. The maximum adsorption capacity obtained was

to 83.9%. This suggests that the CS@NZVI can be reused

142.8 mg/g based on the Langmuir isotherm. The adsorp-

for at least three successive cycles while maintaining high

tion process of Cd2þ onto the synthesized composite was

adsorption efﬁciency. As implied in Figure 5, however,

chemisorption. Moreover, the adsorbent was successfully

the desorption percentage of DI-water is very low for all

recycled for three cycles with a little decrease of variation

studied cycles. This means that DI-water is not suitable

in adsorption ability. The CS@NZVI provides very promis-

to be used as a desorbent solution for the regeneration of

ing results for cost-effective treatment of wastewaters

CS@NZVI loaded with Cd2þ ions. It seems that some

contaminated by Cd2þ, as well as high adsorption capacities,

Page 18

128

M. Ahmadi et al.

Chitosan zero-valent iron nanoparticles-supported for cadmium removal

good and rapid separations and an efﬁcient technology for heavy metals removal.

ACKNOWLEDGEMENTS The present work was ﬁnancially supported by the Environmental

Technologies

Research

Center,

Ahvaz

Jundishapur University of Medical Sciences (Grant No. ETRC-9112). The authors are grateful for the support of Iranian Nano Technology Initiative Council.

REFERENCES Ahmadi, M., Teymouri, P., Setodeh, A., Mortazavi, M. S. & Asgari, A.  Adsorption of Pb(II) from aqueous solution onto Lewatit FO36 nano resin: equilibrium and kinetic studies. Environmental Engineering & Management Journal (EEMJ) 10, 1579–1587. Al Sagheer, F., Al-Sughayer, M., Muslim, S. & Elsabee, M. Z.  Extraction and characterization of chitin and chitosan from marine sources in Arabian Gulf. Carbohydrate Polymers 77, 410–419. American Society for Testing and Material (ASTM)  Standard Test Methods for Cadmium in Water, Methods D3557-12 and D3557-90. ASTM, Pennsylvania, USA. Amiri, H., Jaafarzadeh, N., Ahmadi, M. & Martínez, S. S.  Application of LECA modified with Fenton in arsenite and arsenate removal as an adsorbent. Desalination 272, 212–217. Azari, A., Kakavandi, B., Kalantary, R. R., Ahmadi, E., Gholami, M., Torkshavand, Z. & Azizi, M.  Rapid and efficient magnetically removal of heavy metals by magnetite-activated carbon composite: a statistical design approach. Journal of Porous Materials 22, 1083–1096. Azouaou, N., Sadaoui, Z., Djaafri, A. & Mokaddem, H.  Adsorption of cadmium from aqueous solution onto untreated coffee grounds: equilibrium, kinetics and thermodynamics. Journal of Hazardous Materials 184, 126–134. Babaei, A. A., Azari, A., Kalantary, R. R. & Kakavandi, B. a Enhanced removal of nitrate from water using nZVI@ MWCNTs composite: synthesis, kinetics and mechanism of reduction. Water Science and Technology 72 (11), 1988–1999. Babaei, A. A., Khataee, A., Ahmadpour, E., Sheydaei, M., Kakavandi, B. & Alaee, Z. b Optimization of cationic dye adsorption on activated spent tea: equilibrium, kinetics, thermodynamic and artificial neural network modeling. Korean Journal of Chemical Engineering 33, 1352–1361. Begum, S. A., Golam Hyder, A. H. M. & Vahdat, N.  Adsorption isotherm and kinetic studies of As(V) removal

Journal of Water Supply: Research and Technology—AQUA

66.2

2017

from aqueous solution using cattle bone char. Journal of Water Supply: Research and Technology – Aqua 65, 244–252. Boparai, H. K., Joseph, M. & O’Carroll, D. M.  Kinetics and thermodynamics of cadmium ion removal by adsorption onto nano zerovalent iron particles. Journal of Hazardous Materials 186, 458–465. Brown, D.  Method for Chitosan Production. US Patent 20,150,167,03, USPTO, Alexandria, VA. Chen, C., Hu, J., Shao, D., Li, J. & Wang, X.  Adsorption behavior of multiwall carbon nanotube/iron oxide magnetic composites for Ni(II) and Sr(II). Journal of Hazardous Materials 164, 923–928. Depci, T.  Comparison of activated carbon and iron impregnated activated carbon derived from Gölbaşı lignite to remove cyanide from water. Chemical Engineering Journal 181, 467–478. Dong, C., Chen, W. & Liu, C.  Preparation of novel magnetic chitosan nanoparticle and its application for removal of humic acid from aqueous solution. Applied Surface Science 292, 1067–1076. Du, Y., Pei, M., He, Y., Yu, F., Guo, W. & Wang, L.  Preparation, characterization and application of magnetic Fe3O4-CS for the adsorption of orange I from aqueous solutions. PLoS One 9 (10), e108647. Esfahani, A. R., Firouzi, A. F., Sayyad, G. & Kiasat, A. a Transport and retention of polymer-stabilized zero-valent iron nanoparticles in saturated porous media: effects of initial particle concentration and ionic strength. Journal of Industrial and Engineering Chemistry 20, 2671–2679. Esfahani, A. R., Hojati, S., Azimi, A., Farzadian, M. & Khataee, A. b Enhanced hexavalent chromium removal from aqueous solution using a sepiolite-stabilized zero-valent iron nanocomposite: impact of operational parameters and artiﬁcial neural network modeling. Journal of the Taiwan Institute of Chemical Engineers 49, 172–182. Fan, L., Luo, C., Lu, Z., Lu, F. & Qiu, H.  Preparation of magnetic modiﬁed chitosan and adsorption of Zn2þ from aqueous solutions. Colloids and Surfaces B: Biointerfaces 88, 574–581. Fan, L., Luo, C., Sun, M., Li, X. & Qiu, H.  Highly selective adsorption of lead ions by water-dispersible magnetic chitosan/graphene oxide composites. Colloids and Surfaces B: Biointerfaces 103, 523–529. Fu, F., Ma, J., Xie, L., Tang, B., Han, W. & Lin, S.  Chromium removal using resin supported nanoscale zero-valent iron. Journal of Environmental Management 128, 822–827. Geng, B., Jin, Z., Li, T. & Qi, X.  Preparation of chitosan-stabilized Fe0 nanoparticles for removal of hexavalent chromium in water. Science of the Total Environment 407 (18), 4994–5000. Gupta, A., Yunus, M. & Sankararamakrishnan, N. a Zerovalent iron encapsulated chitosan nanospheres – a novel adsorbent for the removal of total inorganic arsenic from aqueous systems. Chemosphere 86, 150–155. Gupta, V., Agarwal, S. & Saleh, T. A. b Chromium removal by combining the magnetic properties of iron oxide with

Page 19

129

M. Ahmadi et al.

Chitosan zero-valent iron nanoparticles-supported for cadmium removal

adsorption properties of carbon nanotubes. Water Research 45, 2207–2212. Hu, J., Shao, D., Chen, C., Sheng, G., Li, J., Wang, X. & Nagatsu, M.  Plasma-induced grafting of cyclodextrin onto multiwall carbon nanotube/iron oxides for adsorbent application. The Journal of Physical Chemistry B 114 (20), 6779–6785. Huang, X., Gao, N.-y. & Zhang, Q.-l.  Thermodynamics and kinetics of cadmium adsorption onto oxidized granular activated carbon. Journal of Environmental Sciences 19, 1287–1292. Jaafarzadeh, N., Amiri, H. & Ahmadi, M.  Factorial experimental design application in modification of volcanic ash as a natural adsorbent with Fenton process for arsenic removal. Environmental Technology 33, 159–165. Jaafarzadeh, N., Teymouri, P., Babaei, A. A., Alavi, N. & Ahmadi, M.  Biosorption of cadmium (II) from aqueous solution by NaCl-treated Ceratophyllum demersum. Environmental Engineering & Management Journal (EEMJ) 13, 763–773. Jafari, A. J., Kakavandi, B., Kalantary, R. R., Gharibi, H., Asadi, A., Azari, A., Babaei, A. A. & Takdastan, A.  Application of mesoporous magnetic carbon composite for reactive dyes removal: process optimization using response surface methodology. Korean Journal of Chemical Engineering 33 (10), 2878–2890. Jagtap, S., Yenkie, M., Labhsetwar, N. & Rayalu, S.  Defluoridation of drinking water using chitosan based mesoporous alumina. Microporous and Mesoporous Materials 142 (2), 454–463. Kakavandi, B., Rezaei Kalantary, R., Farzadkia, M., Mahvi, A. H., Esrafili, A., Azari, A., Yari, A. R. & Javid, A. B.  Enhanced chromium (VI) removal using activated carbon modified by zero valent iron and silver bimetallic nanoparticles. Journal of Environmental Health Science and Engineering 12, 1–10. Kakavandi, B., Kalantary, R. R., Jafari, A. J., Nasseri, S., Ameri, A., Esrafili, A. & Azari, A.  Pb(II) adsorption onto a magnetic composite of activated carbon and superparamagnetic Fe3O4 nanoparticles: experimental and modeling study. CLEAN – Soil, Air, Water 43, 1157–1166. Kakavandi, B., Jahangiri-rad, M., Rafiee, M., Esfahani, A. R. & Babaei, A. A. a Development of response surface methodology for optimization of phenol and p-chlorophenol adsorption on magnetic recoverable carbon. Microporous and Mesoporous Materials 231, 192–206. Kakavandi, B., Jonidi Jafari, A., Rezaei Kalantary, R., Nasseri, S., Esrafili, A., Gholizadeh, A. & Azari, A. b Simultaneous adsorption of lead and aniline onto magnetically recoverable carbon: optimization, modeling and mechanism. Journal of Chemical Technology & Biotechnology 91, 3000–3010. Kumar, M., Tripathi, B. P. & Shahi, V. K.  Crosslinked chitosan/polyvinyl alcohol blend beads for removal and recovery of Cd(II) from wastewater. Journal of Hazardous Materials 172, 1041–1048. Lim, A. P. & Aris, A. Z.  A review on economically adsorbents on heavy metals removal in water and wastewater. Reviews in Environmental Science and Bio/Technology 13, 163–181.

Page 20

Journal of Water Supply: Research and Technology—AQUA

66.2

2017

Liu, T., Wang, Z.-L., Zhao, L. & Yang, X.  Enhanced chitosan/ Fe0-nanoparticles beads for hexavalent chromium removal from wastewater. Chemical Engineering Journal 189–190, 196–202. Liu, T., Yang, X., Wang, Z.-L. & Yan, X.  Enhanced chitosan beads-supported Fe(0)-nanoparticles for removal of heavy metals from electroplating wastewater in permeable reactive barriers. Water Research 47, 6691–6700. Lü, H., An, H. & Xie, Z.  Ion-imprinted carboxymethyl chitosan–silica hybrid sorbent for extraction of cadmium from water samples. International Journal of Biological Macromolecules 56, 89–93. Malkoc, E. & Nuhoglu, Y.  Fixed bed studies for the sorption of chromium(VI) onto tea factory waste. Chemical Engineering Science 61, 4363–4372. Mohseni-Bandpi, A., Kakavandi, B., Kalantary, R. R., Azari, A. & Keramati, A.  Development of a novel magnetite-chitosan composite for the removal of fluoride from drinking water: adsorption modeling and optimization. RSC Advances 5, 73279–73289. Naghizadeh, A.  Comparison between activated carbon and multiwall carbon nanotubes in the removal of cadmium (II) and chromium (VI) from water solutions. Journal of Water Supply: Research and Technology – Aqua 64, 64–73. Naiya, T., Bhattacharya, A. & Das, S.  Removal of Cd(II) from aqueous solutions using clarified sludge. Journal of Colloid and Interface Science 325, 48–56. Ngah, W. W., Hanafiah, M. & Yong, S.  Adsorption of humic acid from aqueous solutions on crosslinked chitosan– epichlorohydrin beads: kinetics and isotherm studies. Colloids and Surfaces B: Biointerfaces 65, 18–24. Oliveira, L. C., Rios, R. V., Fabris, J. D., Garg, V., Sapag, K. & Lago, R. M.  Activated carbon/iron oxide magnetic composites for the adsorption of contaminants in water. Carbon 40, 2177–2183. Oliveira, L. C., Petkowicz, D. I., Smaniotto, A. & Pergher, S. B.  Magnetic zeolites: a new adsorbent for removal of metallic contaminants from water. Water Research 38, 3699–3704. Rao, M. M., Ramana, D., Seshaiah, K., Wang, M. & Chien, S.  Removal of some metal ions by activated carbon prepared from Phaseolus aureus hulls. Journal of Hazardous Materials 166, 1006–1013. Reddy, D. H. K. & Lee, S.-M.  Application of magnetic chitosan composites for the removal of toxic metal and dyes from aqueous solutions. Advances in Colloid and Interface Science 201, 68–93. Rezaei Kalantry, R., Jonidi Jafari, A., Esrafili, A., Kakavandi, B., Gholizadeh, A. & Azari, A.  Optimization and evaluation of reactive dye adsorption on magnetic composite of activated carbon and iron oxide. Desalination and Water Treatment 57, 6411–6422. Semerjian, L.  Equilibrium and kinetics of cadmium adsorption from aqueous solutions using untreated Pinus halepensis sawdust. Journal of Hazardous Materials 173, 236–242. Shen, Y., Tang, J., Nie, Z., Wang, Y., Ren, Y. & Zuo, L.  Preparation and application of magnetic Fe3O4 nanoparticles

130

M. Ahmadi et al.

Chitosan zero-valent iron nanoparticles-supported for cadmium removal

for wastewater purification. Separation and Purification Technology 68, 312–319. Sobhanardakani, S., Zandipak, R., Fili, Z., Ghoochian, M., Sahraei, R. & Farmany, A.  Removal of V(V) ions from aqueous solutions using oxidized multi-walled carbon nanotubes. Journal of Water Supply: Research and Technology – Aqua 64, 425–433. Teymouri, P., Ahmadi, M., Babaei, A. A., Ahmadi, K. & Jaafarzadeh, N.  Biosorption studies on NaCl-modified ceratophyllum demersum: removal of toxic chromium from aqueous solution. Chemical Engineering Communications 200, 1394–1413. Ünlü, N. & Ersoz, M.  Removal of heavy metal ions by using dithiocarbamated-sporopollenin. Separation and Purification Technology 52, 461–469. Viswanathan, N. & Meenakshi, S.  Enriched fluoride sorption using alumina/chitosan composite. Journal of Hazardous Materials 178, 226–232.

Journal of Water Supply: Research and Technology—AQUA

66.2

2017

Wang, J., Zheng, S., Shao, Y., Liu, J., Xu, Z. & Zhu, D.  Aminofunctionalized Fe3O4@SiO2 core–shell magnetic nanomaterial as a novel adsorbent for aqueous heavy metals removal. Journal of Colloid and Interface Science 349, 293– 299. World Health Organisation  Desalination for Safe Water Supply: Guidance for the Health and Environmental Aspects Applicable to Desalination. Public Health and the Environment, World Health Organization, Geneva. Westergren, R.  Arsenic Removal Using Biosorption with Chitosan: Evaluating the Extraction and Adsorption Performance of Chitosan from Shrimp Shell Waste. MSC Thesis, Rotal Institute of Technology (KTH), Sweden. Yang, G., Tang, L., Lei, X., Zeng, G., Cai, Y., Wei, X., Zhou, Y., Li, S., Fang, Y. & Zhang, Y.  Cd(II) removal from aqueous solution by adsorption on α-ketoglutaric acid-modiﬁed magnetic chitosan. Applied Surface Science 292, 710–716.

First received 29 April 2016; accepted in revised form 26 October 2016. Available online 30 January 2017

Page 21

239

66.4

2017

Efﬁcient removal of some anionic dyes from aqueous solution using a polymer-coated magnetic nano-adsorbent Armin Kiani, Pouya Haratipour, Mazaher Ahmadi, Rouholah Zare-Dorabei and Ali Mahmoodi

ABSTRACT For the efﬁcient removal of some anionic dyes, a novel adsorbent was developed. The adsorbent was prepared by coating a synthetic polymer on magnetite nanosphere surface as a magnetic carrier. The synthesized nano-adsorbent was fully characterized using Fourier transform infrared spectroscopy (FT-IR), vibrating sample magnetometer, X-ray diffractometer, scanning electron microscope, and transmission electronic microscopy measurements. The synthesized nano-adsorbent showed high adsorption capacity towards removal of some anionic dyes (221.4, 201.6, and 135.3 mg g�1 for reactive red 195, reactive yellow 145, and reactive blue 19 dye, respectively) from aqueous samples. The dye adsorption was thoroughly studied from both kinetic and equilibrium points of view. It was found that the Langmuir isotherm showed a better correlation with the experimental data. The kinetic data showed that the process was very fast, and the adsorption process followed pseudo-second order kinetic models for the modiﬁed magnetic nano-adsorbent. Furthermore, the results showed that a stable and reusable (up to 20 cycles) nano-adsorbent for dye removal purposes was synthesized. Key words

| adsorption, anionic dyes, dye removal, magnetite nanospheres, polymeric adsorbent

Armin Kiani Research Center for Analytical Chemistry, KAVA Research Institute, Tehran, Iran Pouya Haratipour Department of Chemistry, Sharif University of Technology, Tehran, Iran Mazaher Ahmadi (corresponding author) Young Researchers and Elite Club, Hamedan Branch, Islamic Azad University, Hamedan, Iran E-mail: m.ahmadi@iauh.ac.ir Rouholah Zare-Dorabei Research Laboratory of Spectrometry & Micro/ Nano Extraction, Department of Chemistry, Iran University of Science and Technology, Narmak, Tehran, Iran Ali Mahmoodi Department of Textile Engineering, Textile Engineering Faculty, Isfahan University of Technology, Isfahan, Iran

INTRODUCTION The presence of organic contaminants in water causes some

colored wastewater discharge as well as developing more

serious problems to aquatic life and human health disorders

efﬁcient treatment technologies.

even in trace amounts (Chen & Wu ). Among various

Various methods, such as adsorption, advanced oxi-

organic contaminants, discharge of synthetic dyes into the

dation processes, biodegradation, coagulation, and the

hydrosphere possess a signiﬁcant source of pollution due

membrane process, have been suggested to handle dye

to their recalcitrant nature. This gives an undesirable color

removal from water. All these processes have some advan-

to water bodies which will reduce sunlight penetration and

tages or disadvantages over the other methods (Khataee &

disturbs photochemical and biological cycles of aquatic life

Kasiri ; Chen & Wu ). A balanced approach is,

(Wong et al. ). Synthetic dyes are widely used in

therefore, needed to look into the worthiness on choosing

many ﬁelds of advanced technology, e.g., in various kinds

an appropriate method which can be used to remove the

of the textile, paper, leather tanning, food processing, plas-

dye in solution. The adsorption method is the most applied

tics, cosmetics, rubber, printing, and dye manufacturing

in the removal of organic dyes and pigments from waste-

industries. The release of synthetic dyes to the environment

waters since it can produce high-quality water, and it can

poses challenges to environmental scientists. These con-

be employed as a process that is economically feasible

cerns have led to new and strict regulations concerning

(Madrakian et al. ). Many textile industries use

doi: 10.2166/aqua.2017.029

Page 23

240

A. Kiani et al.

Journal of Water Supply: Research and Technology—AQUA

Removal of anionic dyes using a magnetic nano-adsorbent

commercial activated carbon for the treatment of dye

66.4

2017

EXPERIMENTAL

waste. Although activated carbon is commonly used as an adsorbent for color removal, its main disadvantage is

Reagents and materials

its high production and treatment costs (Afkhami et al. ; Madrakian et al. ). Thus, many researchers

Reactive yellow 145 (RY145), reactive blue 19 (RB19), and

throughout the world have focused their efforts on optimiz-

reactive red 195 (RR195) anionic dyes were purchased

ing adsorption and developing novel alternative adsorbents

from Sigma-Aldrich Company (St Louis, MO, USA).

with higher adsorption capacities and lower costs. In this

Table 1 shows some chemical information of the investi-

regard,

gated dyes. All of the other chemicals used were of

much

attention

has

recently

been

paid

nanotechnology.

analytical reagent grade and were purchased from Merck

Nanometer-sized materials are widely used for the effec-

Company (Darmstadt, Germany). Double distilled water

tive adsorption of different chemical species from water

(DDW) was used throughout the work. The investigated

samples (Madrakian et al. a; Kyzas & Matis ). The

dyes’ stock solutions were prepared in DDW from their

magnetic nanoparticle as an efﬁcient adsorbent with a

sodium salts and the working standard solutions of different

large speciﬁc surface area and small diffusion resistance

dyes’ concentrations were prepared daily by diluting the

has been recognized (Ngomsik et al. ). The magnetic

stock solution with DDW. Britton–Robinson (B-R) universal

separation provides a suitable route for online separation,

buffer was used for pH adjustment of the working solutions.

where particles with affinity to target species are mixed with the heterogeneous solution. Upon mixing with the solution, the particles tag the target species. External magnetic fields are then applied to separate the tagged particles from the solution. Iron oxide nanoparticles (i.e., magnetite, maghemite, etc.) are attractive examples of magnetic nanoparticles. The synthesis of magnetite nanoparticles has been intensively developed not only for its high fundamental scientific interest but also for many technological applications in biology (Xie et al. ), medical applications (Ahmadi et al. ), bioseparation (Bucak et al. ); and separation and preconcentration of various anions and cations (Afkhami & Norooz-Asl ; Madrakian et al. ), due to their novel structural, electronic, magnetic, and catalytic properties. Recently, employing magnetite nanoparticles with modified surfaces has attracted the high attention of researchers for removal of cationic and anionic dyes from water (Ambashta & Sillanpää ; Madrakian et al. c). Herein, a novel magnetic adsorbent has been developed for dye removal and wastewater treatment purposes. In this regard, magnetite nanospheres were synthesized using the

Apparatus The size, morphology, and structure of the synthesized nanospheres were characterized by transmission electronic microscopy (TEM, Philips-CMC-300 KV) and scanning electron microscope (SEM, MIRA FEG-SEM, and TESCAN). The crystal structure of the synthesized nanospheres was determined by an X-ray diffractometer (XRD, 38066 Riva, d/G. via M. Misone, 11/D (TN) Italy) at ambient temperature. The magnetic properties of the synthesized nanospheres were measured using a vibrating sample magnetometer (VSM, 4 in. Daghigh Meghnatis Kashan Co., Kashan, Iran). The midinfrared spectra of the synthesized nanospheres in the region of 4,000–400 cm�1 were recorded by a Fourier transform infrared spectrometer (FT-IR, Perkin-Elmer model Spectrum GX) using KBr pellets. A single beam ultraviolet (UV)-miniWPA spectrophotometer was used for the determination of dye concentration in the solutions. A Metrohm model 713 pH meter was used for pH measurements. A 40 kHz universal ultrasonic cleaner water bath (RoHS, Korea) was used.

solvothermal method and further surface modiﬁcation was performed using a synthetic polymer. The adsorbent was

Synthesis of amidoamine monomer (AAM)

successfully utilized for removal of some anionic dyes (i.e., reactive yellow 145, reactive blue 19, and reactive red 195)

The AAM monomer was synthesized according to a pre-

from aqueous samples.

viously reported procedure (Madrakian et al. c).

Page 24

241

Table 1

A. Kiani et al.

Journal of Water Supply: Research and Technology—AQUA

Removal of anionic dyes using a magnetic nano-adsorbent

66.4

2017

Chemical information of the investigated anionic dyes �1

λmax (nm)

Name

Molecular formula

Molecular weight (g mol

RY145

C28H20ClN9Na4O16S5

1,026.25

421

RB19

C22H16N2Na2O11S3

626.55

602

RR195

C31H19ClN7Na5O19S6

1,136.32

542

)

Structure

Brieﬂy, the amidoamine monomer was synthesized by

In order to prepare the AMNSs, the amidoamine mono-

the slow addition of 1 g (0.01 mol) maleic anhydride to

mer was polymerized in the presence of MNSs (0.5 g, as the

the solution of 1 mL (0.015 mol) ethylenediamine in

magnetic core), ammonium persulfate (0.1 g, as the

20 mL DDW. The solution was heated at 120 C for 1 h

initiator), and ethylene glycol dimethacrylate (50 μL, as the

until all the water was removed and ethylenediamine

cross-linking monomer) in 30 mL DDW at 85 C for 12 h.

reacted with maleic anhydride through ring opening

Then, the product was separated using a magnet and

(Figure 1(a)).

washed with methanol and DDW to remove unreacted

reagents. Synthesis of magnetite nanospheres (MNSs) and polymer-coated magnetite nanospheres (AMNSs)

Point of zero charge (pHPZC) of AMNS nanospheres

MNSs were synthesized by the solvothermal reduction

The pHPZC of the AMNSs was determined in degassed

method with minor modiﬁcations (Deng et al. ). Typi-

0.01 mol L�1 NaNO3 solution at room temperature. Ali-

cally, FeCl3.6H2O (1.35 g) was dissolved in ethylene glycol

quots of 30.0 mL 0.01 mol L�1 NaNO3 were mixed with

(40.0 mL) to form a clear solution, followed by the addition

0.03 g of the nanospheres in several beakers. The initial

of sodium acetate (3.6 g) and polyethylene glycol (1.0 g). The

pH of the solutions was adjusted to 3.0, 4.0, 5.0, 6.0,

mixture was ultrasonicated vigorously for 30 min and then

7.0, 8.0, 9.0, and 10.0 using 0.01 mol L�1 of HNO3 and

reﬂuxed at 180 C for 8 h, and then allowed to cool down

NaOH solutions as appropriate. The initial pHs of the sol-

to room temperature. The black products were washed sev-

utions were recorded, and the beakers were covered with

eral times with ethanol and DDW water and then dried at

paraﬁlm and shaken for 24 h. The ﬁnal pH values were

60 C for 6 h (Figure 1(b)).

recorded and the differences between the initial and

Page 25

242

Figure 1

A. Kiani et al.

Removal of anionic dyes using a magnetic nano-adsorbent

Journal of Water Supply: Research and Technology—AQUA

66.4

2017

Overall routes for the synthesis of (a) AAM monomer and (b) AMNS nanospheres.

ﬁnal pH (ΔpH) of the solutions were plotted against their

5.0 using a B-R pH buffer. The mixed solution was then

initial pH values. The pHPZC corresponds to the pH

shaken at room temperature for 20 min. Then, the dye

where ΔpH ¼ 0 (Madrakian et al. b). The pHPZC for

loaded AMNSs were separated by magnetic decantation.

AMNSs was determined using the above procedure and

The concentration of the dye in the solution was measured

was obtained as 6.7. The results are shown in Figure 2.

spectrophotometrically at the wavelength of the maximum absorbance of each dye (Table 1). The concentration of dyes decreased with time due to their adsorption by

Dye removal experiments

AMNSs. The adsorption percent for each dye, i.e., the dye Adsorption studies were performed by adding 0.02 g of �1

AMNSs to 50.0 mL solution of 50 mg L

of dyes in a

removal efﬁciency, was determined using the following expression:

100 mL beaker, and the pH of the solution was adjusted at %Re ¼

(C0 � Ct ) × 100 C0

(1)

where Co and Ct represent the initial and ﬁnal (after adsorption) concentration of dye in mg L�1, respectively.

Adsorption kinetics Adsorption is a physicochemical process that involves the mass transfer of a solute from the liquid phase to the adsorFigure 2

Page 26

Point of zero charge (pHpzc) of AMNS nanospheres.

bent surface. The adsorption capacities of adsorbents were

243

A. Kiani et al.

Journal of Water Supply: Research and Technology—AQUA

Removal of anionic dyes using a magnetic nano-adsorbent

66.4

2017

calculated from the difference between the initial and the

Results showed that 0.1 M sodium hydroxide aqueous sol-

ﬁnal concentration at any intermediate time. The sorption

ution was an effective eluent for desorption of the dyes. It is

dynamics of the adsorption by AMNSs were tested with

notable that the equilibrium of desorption was achieved

the pseudo-ﬁrst order and the pseudo-second order kinetic

within about 10 min, which was fast, similar to the adsorption

models (Madrakian et al. a). To study the adsorption kinetics of the investigated dyes on AMNSs, 50.0 mg L�1 initial concentration of corresponding dye solutions which had been stirred in the presence of 0.02 g adsorbents at pH ¼ 5.0 and for different time ranges (0–50 min) were used at room temperature. The solution was separated by magnetic decantation to remove adsorbent and analyzed spectrophotometrically. Adsorption isotherms The capacity of the adsorbent is an important factor that determines how much sorbent is required to remove quantitatively a speciﬁc amount of the dye from solution. For

Figure 3

Magnetization curves obtained by VSM at room temperature: ( ) bare MNS and ( ) AMNS nanospheres.

measuring the adsorption capacity of AMNSs, the absorbent was added into dye solutions at various concentrations (under optimum condition), and the suspensions were stirred at room temperature until the equilibrium was reached, followed by magnetic removal of the absorbent. An adsorption isotherm describes the fraction of the sorbate molecules that are partitioned between the liquid and the solid phase at equilibrium. Adsorption of the dyes by AMNS adsorbent was modeled using Freundlich (Freundlich & Heller ) and Langmuir (Langmuir ) adsorption isotherm models. The remaining dye in the supernatants was measured spectrophotometrically at the wavelength of the maximum absorbance of each dye, and the results were used to plot the isothermal adsorption curves.

Figure 4

The FT-IR spectra of (—) MNS and (---) AMNS nanospheres.

Figure 5

XRD pattern of AMNS nanospheres.

Reusability and stability of the adsorbent To evaluate the possibility of regeneration and the reuse of AMNS adsorbent, desorption experiments were performed. Dye desorption from the AMNSs was conducted by washing the dyes loaded on AMNSs using 2.0 mL of pure methanol, sodium hydroxide aqueous solution (0.1 M) and acetonitrile. For this purpose, 2.0 mL of eluent was added to 0.02 g of dye loaded AMNSs in a beaker. Then, the AMNSs were collected magnetically from the solution. The concentration of dyes in the desorbed solution was measured spectrophotometrically.

Page 27

244

Figure 6

A. Kiani et al.

Journal of Water Supply: Research and Technology—AQUA

Removal of anionic dyes using a magnetic nano-adsorbent

66.4

2017

SEM and TEM (insets) images of (a) MNS and (b) AMNS nanospheres.

equilibrium. This was due to the absence of internal diffusion

band of Fe-O in Fe3O4 (around 600 cm�1) was observed in

resistance. After elution of the adsorbed dyes, the adsorbent

FT-IR spectra of MNSs and AMNSs. Two new absorption

was washed with DDW and vacuum dried at 50 C overnight

peaks at 1,730 cm�1 and 1,440 cm�1 in FT-IR spectra of

and reused for the dye removal.

AMNSs are assigned to C ¼ O and C-N bands in the

AMNSs, respectively. Moreover, new absorption peaks at 2,820 and 2,860 cm�1 are related to the stretching modes

RESULTS AND DISCUSSION

of aliphatic C-H bonds in the ﬁnal product. Based on the

The synthesized AMNSs were fully characterized using

cedure was successfully performed.

XRD, SEM, TEM, VSM, and FT-IR measurements. Then, batch experiments were used for evaluation and optimization affecting various parameters such as pH, contact time, and nanosphere dosage.

Characterization of the investigated nanospheres

above results, it can be concluded that the fabrication proThe XRD pattern of AMNSs (Figure 5) shows diffraction peaks that are indexed to (2 2 0), (3 1 1), (4 0 0), (4 2 2), (5 1 1), (4 4 0), and (5 5 3) reﬂection characteristics of the cubic spinel phase of Fe3O4 (Joint Committee on Powder Diffraction Standards (JCPDS) powder diffraction data ﬁle no. 79–0418), revealing that the resultant nanospheres are mostly Fe3O4. The average crystallite size of the AMNSs was estimated to be 13 nm from the XRD data according

The magnetization curves of the bare MNSs and AMNSs

to the Scherrer equation (Madrakian et al. a).

recorded with VSM are illustrated in Figure 3. As shown in Figure 3, the magnetization of the samples approach the saturation values when the applied magnetic ﬁeld increases to 10,000 Oe. The saturation magnetizations of the MNSs and AMNSs were 55.20 and 40.05 emu/g, respectively. These results show that the AMNSs retain approximately 75% of the magnetization of the bare MNSs. A magnetization reduction of about 27.44% was observed between the bare MNSs and AMNSs. This may be related to the nanospheres’ size effect, the increased surface disorder, and the diamagnetic contribution of the polymer layer. The FT-IR spectra of the products were recorded to verify the formation of the expected products. The related spectra are shown in Figure 4. The characteristic absorption Page 28

Figure 7

Removal efﬁciencies of ( ) RR195, ( ) RY145, and ( ) RB19 at different pHs � (conditions: 0.01 g of AMNSs, 50.0 mL of 50.0 mg L 1 of dyes, agitation time of 45 min, N ¼ 3).

245

Table 2

A. Kiani et al.

Journal of Water Supply: Research and Technology—AQUA

Removal of anionic dyes using a magnetic nano-adsorbent

66.4

exp

2017

Adsorption kinetics parameters of the investigated dyes’ adsorption on AMNSs Pseudo-second order

Pseudo-ﬁrst order

Kinetics models Dye

qe,

RR195

191.16

RY145

134.56

RB19

72.18

cal

�1

(mg g

)

�3

�1

RMS

qe,

0.61

0.996

1.32

146.01

1.24

0.989

1.05

107.54

1.05

0.987

2.10

57.91

k2 × 10

(g mg

)

cal

�1

(mg g

)

�1

RMS

qe,

0.025

0.974

3.21

192.30

0.153

0.961

2.85

133.34

0.127

0.909

3.64

75.79

k1 (h

)

�1

(mg g

)

The TEM and SEM images of the MNSs in Figure 6(a)

concentration of 50.0 mg L�1 and a stirring time of 45 min,

indicate that spherical monodispersed nanoparticles with

where the pH was adjusted with B-R buffer (Figure 7).

an average diameter of about 110 nm were synthesized.

Figure 7 indicates that the adsorbent provides the highest

Figure 6(b) indicates that MNSs were successfully coated

afﬁnity to the dyes’ molecules at pH 3–5. This is reasonable,

with a layer of the polymer. This ﬁgure shows that after

because at this pH, the dyes are negatively charged and, on

the polymer layer coating process, thickness and morpho-

the other hand, the adsorbent surface charge at pH < 6.7

logical properties were, to some extent, changed.

(pHPZC ¼ 6.7) is positive and electrostatic attraction force

is responsible for the high dye removal efﬁciencies (Yagub Effect of pH One of the important factors affecting the removal of the dyes from aqueous solutions is the pH of the solution. The dependence of dyes’ molecules sorption on pH is related to both the dyes’ chemistry in the solution and the ionization state of the functional groups of the sorbent which affects the availability of binding sites (Madrakian et al. ). All of the investigated dyes are anionic dyes. In the case of the adsorbent, the responsible parameter is the point of zero charge (pHPZC). The point of zero charge is a characteristic of metal oxides (hydroxides) and is of fundamental importance in surface science. It is a concept relating to the phenomenon of adsorption and describes the condition

et al. ). Effect of nanosphere dosage The dependence of the adsorption of the dyes on the modiﬁed nanospheres’ amount was studied at room temperature and pH 5.0 by varying the adsorbent amount from 0.01 to 0.05 g in contact with 50.0 mL solution of 50.0 mg L�1 of the dyes with agitation time of 45 min. The results showed that increasing the amount of AMNSs increases the removal efﬁciencies of the dyes due to the availability of higher adsorption sites. The adsorption reached a maximum with 0.02 g of the adsorbent, and maximum percentage removal was about 98%.

when the electrical charge density on a surface is zero. The surface charge of AMNSs with primary amine groups (belong to the functional monomer) and hydroxyl groups (belong to MNSs) is largely dependent on the pH of the solution. The pHPZC is caused by the amphoteric behavior of hydroxyl and surface amino groups, and the interaction between surface sites and the electrolyte species. When brought into contact with aqueous solutions, hydroxyl groups of surface sites can undergo protonation or deprotonation, depending on the solution pH, to form charged surface species. The effect of pH on the dyes’ removal efﬁciencies was investigated in the range of 3.0–10.0 using an initial dye

Figure 8

Isothermal adsorption curves of ( ) RR195, ( ) RY145, and ( ) RB19 on AMNS adsorbent under optimum condition.

Page 29

246

Table 3

A. Kiani et al.

Journal of Water Supply: Research and Technology—AQUA

Removal of anionic dyes using a magnetic nano-adsorbent

66.4

2017

Adsorption isotherm parameters of Langmuir and Freundlich models for the adsorption of the dyes’ molecules on AMNS adsorbent Langmuir �1

Isotherm models

KL (L mg

RR195

0.26

RY145

0.14

RB19

0.65

Freundlich )

�1

RMS

1/n

RMS

221.4

0.9927

0.97

70.27

0.25

0.8991

3.14

201.6

0.9968

1.14

55.25

0.26

0.7846

2.11

135.3

0.9944

1.08

66.89

0.15

0.8608

1.79

qmax (mg g

)

Adsorption kinetics

adsorption, that there are no interactions between adsorbed molecules, and that the adsorption energy is distributed

The removal of the dyes by adsorption on AMNSs was

homogeneously over the entire coverage surface. This sorp-

found to be rapid at the initial period (in the ﬁrst ≈5th

tion model serves to estimate the maximum uptake values

min) and then to become slow and stagnate with the

where they cannot be reached in the experiments.

increase in the contact time (≈5th to ≈15th min), and

According to the results (Table 3), the maximum

nearly reached a plateau after approximately the 20th min

amounts of the dyes that can be adsorbed by AMNSs were

of the experiment. Different kinetic parameters of the

found to be 221.4, 201.6, and 135.3 mg g�1 at pH 5.0 in

dyes’ adsorption onto AMNSs are shown in Table 2. All

the case of RR195, RY145, and RB19, respectively. As the

the experimental data showed better compliance with the

results show, the capacity factor for RR195 and RY145 is

pseudo-second order kinetic model regarding higher corre-

higher than that for RB19. The difference in capacity may

lation coefﬁcient value (R > 0.98) and lower root mean

be due to the difference in the structure of dyes and the

square (RMS) value. Moreover, the q values (qe,

number of the anionic functional groups.

cal ) calcu-

lated from the pseudo-second order model were more consistent with the experimental q values (qe,

exp)

than

with those calculated from the pseudo-ﬁrst order model. Hence, it could be found that the pseudo-second order kinetic model was more valid to describe the adsorption behavior of the dyes onto AMNSs.

Reusability and stability of the adsorbent The reusability and stability of AMNSs for the removal of the investigated dyes were assessed by performing 25 consecutive separations/desorption cycles under the optimized conditions (Figure 9). The results showed that there was no signiﬁcant change in the performance of the adsorbent

Adsorption isotherms The isothermal adsorption curves are shown in Figure 8. The adsorption equilibrium data were fitted to Langmuir and Freundlich isotherm models by nonlinear regression. The resulting parameters are summarized in Table 2. The higher correlation coefficient obtained for the Langmuir model, for all of the investigated dyes, and lower RMS values indicates that the experimental data are better fitted to this model, and adsorption of the investigated anionic dyes on AMNS adsorbent is more compatible with Langmuir assumptions, i.e., adsorption takes place at specific homogeneous sites within the adsorbent. The Langmuir model is based on the physical hypothesis that the maximum Page 30

adsorption

capacity

consists

monolayer

Figure 9

The reusability and stability of AMNSs for the removal of 50.0 mL of 50.0 mg �

L 1 of ( ) RR195, ( ) RY145, and ( ) RB19 (conditions: AMNS dosage: 0.02 g, pH: 5, adsorption time: 45 min, eluent: 2 mL of 0.1 M sodium hydroxide solution, desorption time: 10 min).

247

Table 4

A. Kiani et al.

Removal of anionic dyes using a magnetic nano-adsorbent

Journal of Water Supply: Research and Technology—AQUA

66.4

2017

Comparison of the calculated capacity factor for some synthetic adsorbents with the proposed one �1

Capacity factor (mg g Adsorbent

RR195

) RY145

RB19

Ref.

Wheat bran

103.4

125.0

97.1

Çiçek et al. ()

P. oxalicum pellets

–

137

159

Zhang et al. ()

Chitosan ﬂake

–

188

–

Wu et al. ()

Nano-MgO

–

166.7

Moussavi & Mahmoudi ()

Magnetic nanoparticle/ polyethyleneimine

–

121

Liao et al. ()

AMNSs

221.4

201.6

135.3

This work

during the ﬁrst 20 cycles, indicating that the fabricated

adsorbent providing a large surface area and good afﬁnity

adsorbent is a reusable solid phase adsorbent for the

for the facile and fast adsorption of dye molecules. The

removal of the investigated dyes during these 20 cycles. Fur-

Langmuir isotherm model well ﬁtted the adsorption data.

thermore, the results showed that the efﬁciencies of the

As the calculated capacity factors of AMNSs show, they

recycled adsorbent for removing the investigated dyes are

are a very good adsorbent for removing the investigated

nearly the same as those for the fresh ones even after 20

anionic dyes. The results of this study suggest that the devel-

times recycling. The removal efﬁciencies decreasing at

oped adsorbent can be considered as an alternative

higher cycles might be due to washing the polymer from

adsorbent for wastewater treatments and controlling

the magnetic nanospheres during the adsorbent regener-

environmental pollution.

ation process. To evaluate this theory, the bare MNSs were used for the removal of the same concentration of the investigated dyes under the optimized conditions. The results showed that only 20.6, 18.2, and 12.4% of RR195, RY145, and RB19 could be removed using the bare MNSs at the ﬁrst cycle, and conﬁrming the critical role of the coated polymer to increase the adsorption capacity of the AMNSs toward the investigated dyes. In Table 4, we compared the ability of our inexpensive adsorbent with other adsorbents in the removal of the dyes from aqueous solutions. The results show that AMNSs are a

ACKNOWLEDGEMENTS The authors gratefully acknowledge the ﬁnancial support provided by Centre for Analytical Chemistry-KAVA Research Institute (CAC-KRI, Project No. Y34CS024/2013). The authors also acknowledge the Research Laboratory of Spectrometry & Micro/Nano Extraction of Iran University of Science and Technology (RLSMNE-IUST) for their support.

better adsorbent compared to some of the adsorbents.

REFERENCES CONCLUSION In this work, a magnetic adsorbent was developed for dye removal purposes. The prepared magnetic adsorbent is well dispersed in the water medium and be easily separated magnetically from the medium after the dyes’ adsorption process. The rapid adsorption rate is mainly attributed to the polymer structure and functional groups on the

Afkhami, A. & Norooz-Asl, R.  Removal, preconcentration and determination of Mo(VI) from water and wastewater samples using maghemite nanoparticles. Colloids and Surfaces A: Physicochemical and Engineering Aspects 346, 52–57. Afkhami, A., Madrakian, T., Karimi, Z. & Amini, A.  Effect of treatment of carbon cloth with sodium hydroxide solution on its adsorption capacity for the adsorption of some cations. Colloids and Surfaces A: Physicochemical and Engineering Aspects 304, 36–40.

Page 31

248

A. Kiani et al.

Removal of anionic dyes using a magnetic nano-adsorbent

Ahmadi, M., Madrakian, T. & Afkhami, A.  Solid phase extraction of doxorubicin using molecularly imprinted polymer coated magnetite nanospheres prior to its spectrofluorometric determination. New Journal of Chemistry 39, 163–171. Ambashta, R. D. & Sillanpää, M.  Water purification using magnetic assistance: a review. Journal of Hazardous Materials 180, 38–49. Bucak, S., Jones, D. A., Laibinis, P. E. & Hatton, T. A.  Protein separations using colloidal magnetic nanoparticles. Biotechnology Progress 19, 477–484. Chen, Q. & Wu, Q.  Preparation of carbon microspheres decorated with silver nanoparticles and their ability to remove dyes from aqueous solution. Journal of Hazardous Materials 283, 193–201. Çiçek, F., Özer, D., Özer, A. & Özer, A.  Low cost removal of reactive dyes using wheat bran. Journal of Hazardous Materials 146, 408–416. Deng, H., Li, X., Peng, Q., Wang, X., Chen, J. & Li, Y.  Monodisperse magnetic single-crystal ferrite microspheres. Angewandte Chemie International Edition 44, 2782–2785. Freundlich, H. & Heller, W.  The adsorption of cis- and transazobenzene. Journal of the American Chemical Society 61, 2228–2230. Khataee, A. R. & Kasiri, M. B.  Photocatalytic degradation of organic dyes in the presence of nanostructured titanium dioxide: influence of the chemical structure of dyes. Journal of Molecular Catalysis A: Chemical 328, 8–26. Kyzas, G. Z. & Matis, K. A.  Nanoadsorbents for pollutants removal: a review. Journal of Molecular Liquids 203, 159–168. Langmuir, I.  The constitution and fundamental properties of solids and liquids. Part I. Solids. Journal of the American Chemical Society 38, 2221–2295. Liao, M.-H., Chen, W.-C. & Lai, W.-C.  Magnetic nanoparticles assisted low-molecular weight polyethyleneimine for fast and effective removal of reactive blue 19. Fresenius Environmental Bulletin 15, 609–613. Madrakian, T., Afkhami, A., Ahmadi, M. & Bagheri, H.  Removal of some cationic dyes from aqueous solutions using magnetic-modified multi-walled carbon nanotubes. Journal of Hazardous Materials 196, 109–114. Madrakian, T., Afkhami, A., Zolfigol, M., Ahmadi, M. & Koukabi, N.  Application of modified silica coated magnetite

Journal of Water Supply: Research and Technology—AQUA

66.4

2017

nanoparticles for removal of iodine from water samples. Nano-Micro Letters 4, 57–63. Madrakian, T., Afkhami, A. & Ahmadi, M. a Simple in situ functionalizing magnetite nanoparticles by reactive blue-19 and their application to the effective removal of Pb2þ ions from water samples. Chemosphere 90, 542–547. Madrakian, T., Afkhami, A., Mahmood-Kashani, H. & Ahmadi, M. b Superparamagnetic surface molecularly imprinted nanoparticles for sensitive solid-phase extraction of tramadol from urine samples. Talanta 105, 255–261. Madrakian, T., Ahmadi, M., Afkhami, A. & Soleimani, M. c Selective solid-phase extraction of naproxen drug from human urine samples using molecularly imprinted polymer-coated magnetic multi-walled carbon nanotubes prior to its spectroﬂuorometric determination. Analyst 138, 4542–4549. Moussavi, G. & Mahmoudi, M.  Removal of azo and anthraquinone reactive dyes from industrial wastewaters using MgO nanoparticles. Journal of Hazardous Materials 168, 806–812. Ngomsik, A.-F., Bee, A., Draye, M., Cote, G. & Cabuil, V.  Magnetic nano- and microparticles for metal removal and environmental applications: a review. Comptes Rendus Chimie 8, 963–970. Wong, Y. C., Szeto, Y. S., Cheung, W. H. & McKay, G.  Adsorption of acid dyes on chitosan–equilibrium isotherm analyses. Process Biochemistry 39, 695–704. Wu, F.-C., Tseng, R.-L. & Juang, R.-S.  Enhanced abilities of highly swollen chitosan beads for color removal and tyrosinase immobilization. Journal of Hazardous Materials 81, 167–177. Xie, X., Zhang, X., Yu, B., Gao, H., Zhang, H. & Fei, W.  Rapid extraction of genomic DNA from saliva for HLA typing on microarray based on magnetic nanobeads. Journal of Magnetism and Magnetic Materials 280, 164–168. Yagub, M. T., Sen, T. K., Afroze, S. & Ang, H. M.  Dye and its removal from aqueous solution by adsorption: a review. Advances in Colloid and Interface Science 209, 172–184. Zhang, S. J., Yang, M., Yang, Q. X., Zhang, Y., Xin, B. P. & Pan, F.  Biosorption of reactive dyes by the mycelium pellets of a new isolate of Penicillium oxalicum. Biotechnology Letters 25, 1479–1482.

First received 22 April 2016; accepted in revised form 23 December 2016. Available online 7 April 2017

Page 32

H2Open Journal

ISSN 2616-6518 iwaponline.com/h2open

H2Open Journal is an open access journal that covers all aspects of 21st century water research: fundamental science, application of science and technology and impact of societal nd political and economic factors. The journal’s scope includes: • • • • • • • •

• •

Integrated catchment management Provision of alternative water sources Emerging contaminants Provision of safe water and sanitation in developing communities Development of water sensitive cities Circular economy investigations that demonstrate advances in the minimisation of waste/pollution from the water cycle Water infrastructure Development and optimisation of advanced technologies including treatment and resource recovery processes, process control and water quality monitoring technologies and data management systems Building capacity and capability in the water industry in all communities All aspects of fundamental science underpinning the water cycle (natural and engineered) For more details, visit iwaponline.com/h2open

Page 36

H2 Open Journal doi: 10.2166/h2oj.2017.001

Prioritization of sub-catchments of a river basin using DEM and Fuzzy VIKOR K. Srinivasa Rajua, D. Nagesh Kumarb,c,* and Anmol Jalalic a

Department of Civil Engineering, Birla Institute of Technology and Science, Pilani-Hyderabad Campus, Hyderabad 500 078, India

Department of Civil Engineering, Indian Institute of Science, Bangalore 560 012, India

Centre for Earth Sciences, Indian Institute of Science, Bangalore 560 012, India

*Corresponding author. E-mail: nagesh@iisc.ac.in

Abstract Fuzzy VIKOR, a decision making technique, is applied to prioritize 224 sub-catchments of Mahanadi Basin, India. Seven geomorphology based criteria viz., drainage density, bifurcation ratio, stream frequency, texture ratio, form factor, elongation ratio and circulatory ratio are estimated from ﬁve digital elevation models (DEMs). Triangular membership functions were formulated for each criterion for each sub-catchment which are based on individual values obtained from individual DEM’s. Entropy method is employed for estimation of weights of criteria and a similar mechanism is followed while formulating triangular membership function for weights. Eight groups are formulated with a number of sub-catchments in each group as 5, 26, 69, 65, 29, 11, 12, 7 for taking up conservation measures. Effect of varying strategy weight, (ν) on the ranking pattern is also studied and found that ν value effects ranking pattern signiﬁcantly. Key words: digital elevation models, entropy, Fuzzy VIKOR, prioritization, sub-catchments

INTRODUCTION Water is the most precious natural resource available on Earth. It is unequivocally one of the most important factors for the life to thrive and prosper. The steady rise of human and livestock population, urbanization, demands from other sectors and erratic rainfall have put pressure on this scarce resource and this pressure is likely to grow in the near future. It becomes imperative to form a strategy for effective, efficient and sustainable improving/development of catchments which are basis for water resources and land management. Alarmingly, problem of erosion is becoming more complex due to increasing human activities, deforestation, inadequate and poor farming practices and effects both quantity and quality of soil, accelerates sediment deposition in reservoirs, floodplains, and even impacts agriculture significantly. All these factors are eventually leading to deterioration of the quality of catchments in developing countries. Keeping this in view, catchments are expected to be improved such that expectations from them can be met. However, due to financial and other limitations, improvement, maintenance and management strategies cannot be implemented simultaneously for all the catchments necessitating prioritization. Accordingly, catchments which require earlier soil This is an Open Access article distributed under the terms of the Creative Commons Attribution Licence (CC BY 4.0), which permits copying, adaptation and redistribution, provided the original work is properly cited (http://creativecommons.org/licenses/by/4.0/).

Page 37

H2 Open Journal doi: 10.2166/h2oj.2017.001

and water conservation treatment can be first improved and other catchments can follow as per their priority for improvements. Another lacuna for improvement is inadequate data availability of the catchments. Water resources planners in the absence of adequate/precise/gauged data are employing geomorphological parameters (Rai et al. 2001; Kumar et al. 2017) i.e., linear parameters (Bifurcation Ratio, Drainage Density, Stream Flow Frequency and Texture Ratio) and shape parameters (Circulatory Ratio, Form Factor and Elongation Ratio) for characterizing the catchment and can be used as the basis for prioritizing the catchments. Kumar et al. (2017) mentioned that linear and shape parameters have a direct and inverse relationship with erodibility respectively. A higher value of linear parameters and low value of shape parameters represents higher erodibility (Raju & Nagesh Kumar 2013). Here prioritization or ranking is the process of arranging the catchments in the order of their importance, employing decision making algorithms facilitating the process of prioritization (Lee et al. 2015). Incorporating effective improvement strategies will not only lead to improvement of catchments over time, but will also enable to improve the socio-economic aspect of the area through the sustained generation of employment for the local population. Complimentarily, digital elevation models’ (DEMs) capability to yield more precise terrain information with much ease, accelerated the application of DEM based geomorphic models (Wolock & Price 1994; Williams et al. 2000). Noman et al. (2001) extensively reviewed delineation of flood plain from digital terrain models with various perspectives. Manfreda et al. (2011) highlighted the role of DEMs in detecting flood prone areas. They employed DEMs such as the ASTER global, Shuttle Radar Topography Mission (SRTM), and national elevation data to assess their sensitiveness to the chosen problem. They found that SRTM DEM is suitable for delineation of flood-prone areas. Yan et al. (2014) highlighted the role of DEMs as a main data source in the field of geomorphology. Papaioannou et al. (2015) analyzed the role of DEM derived geomorphological and hydrological attributes for identification of flood prone areas. On the geomorphology aspects, Thakkar & Dhiman (2007) performed morphometric analysis and prioritization of eight watersheds of Mohr watershed, Gujarat, India. They also compared various morphological parameters. Rudraiah et al. (2008) studied part of Kagna river basin, Karnataka, India using remote sensing and geographical information systems (GIS). Javed et al. (2009) applied morphometric analysis for prioritization of sub-watersheds for Kanera watershed, Madhya Pradesh, India. Out of the seven sub-watersheds (SW1 to SW7), SW1 and SW6 qualified for high priority, whereas SW7 was categorized as medium priority. Deshmukh et al. (2011) analyzed eight watersheds (W1 to W8) adjacent to Narmada and Sher rivers for analysis of erodibility. It was found that the watersheds W5 and W6 were high and least degraded respectively. Javed et al. (2011) prioritized fourteen sub-watersheds (SW1 to SW14) of Jaggar watershed, Eastern Rajasthan based on morphometric analysis and land use/land cover categories. It was observed that only SW7 and SW10 fall under very high priority. Kanth & Hassan (2012) prioritized nineteen watersheds of Wular Catchment, India and a compound value was calculated for identifying highest, medium and low priority zones. Yasmin et al. (2013) performed morphometric analysis for Milli watershed, Karnataka, India using GIS. They found that GIS was useful for similar situations. Uniyal & Gupta (2013) prioritized twenty micro-watersheds (MW1 to MW20) of Bhilangana watershed of Uttarakhand, India and classified into high, medium and low priority for conservation and management. Raju & Nagesh Kumar (2013) applied TOPSIS for prioritizing twenty two micro-watersheds of Kherthal catchment, Rajasthan, India using seven geomorphological parameters. Entropy method was used to compute weights of geomorphological parameters. It was observed that the methodology adopted was found to be effective. Aher et al. (2014) identified critical and priority sub-watersheds in water scarce region of India and applied weighted sum analysis approach for ranking each hydrological unit. They found that 51.66% of sub-watersheds were in the moderately to highly susceptible zones. Page 38

H2 Open Journal doi: 10.2166/h2oj.2017.001

Iqbal & Sajjad (2014) prioritized five watersheds, D1A, D1B, D1C, D2A, D2B of Dudhganga catchment. It was found that D1C and D1A fall under high and medium priority respectively. Jaiswal et al. (2014) prioritized thirty six sub-watersheds of Benisagar dam catchment of Bundelkhand region, Madhya Pradesh, India and applied Saaty’s analytical hierarchy process with nine erosion hazards for identification of environmentally stressed sub-watersheds. Similar studies were also reported by Jaiswal et al. (2015) using fuzzy Analytic Hierarchy Process. Patel et al. (2015) identified suitable sites for thirteen mini-watersheds of Hathmati for identifying water harvesting structures and found that watershed number 2 was of maximum priority. Makwana & Tiwari (2016) prioritized nineteen sub-watersheds in the semi-arid middle region of Gujarat, India using the compound parameter. They used remote SRTM data for the analysis. They opined that prioritization helps to implement soil conservation measures. Chandniha & Kansal (2017) performed prioritization of nine sub-watersheds of Piperiya watershed, Hasdeo river basin and classified them into high, medium and low priorities. Singh & Singh (2017) made an effort to prioritize sub-watersheds of Dangri River watershed, Haryana, India based on Snyder’s synthetic unit hydrograph and grouped them as high, medium and low soil-erosive. They compared the outcome with land use/ land cover and morphometric analysis. Patel et al. (2012), Zhang et al. (2015), Khanday & Javed (2016) and Kumar et al. (2017) performed similar studies. It is observed that (a) most of the studies used geomorphological parameters for ranking of the watersheds without assigning any weightage to them (b) no study was reported in fuzzy environment for ranking of the watersheds in geomorphological perspective. In other words, no study was reported in Indian conditions where DEM data from five sources were used for computing geomorphological parameters and on the basis of which ranking of sub-catchments were performed in fuzzy environment. Keeping the above observations from the literature review and practical aspects into consideration, the objectives of the present study are formulated as follows:

• To estimate geomorphological parameters, namely, Drainage Density, Bifurcation Ratio, Stream Frequency, Texture Ratio, Form Factor, Elongation Ratio and Circulatory Ratio for all the 224 sub-catchments of Mahanadi Basin, India using ﬁve different DEM sources, namely, GMTED2010 7.5 arc-sec, SRTM (30 m & 90 m), ASTER and CARTOSAT-1. • To explore the applicability of (a) Entropy method for obtaining weights for the parameters (b) Fuzzy VIKOR (Vise Kriterijumska Optimizacija I Kompromisno Resenje), a decision making technique, to prioritize 224 sub-catchments of Mahanadi Basin in India. Present paper covers introduction, case study, description of methods, results and discussion followed by conclusions.

CASE STUDY Mahanadi basin lies between East longitudes 80° 300 and 86° 500 , and North latitudes 19° 150 and 23° 350 . The basin is broadly divided into three sub-basins; Upper, Middle and Lower consisting of 91, 88, 48 sub-catchments (totalling to 227) (Figure 1). The climate in the basin is predominantly sub-tropical. The annual rainfall trend based on 34 years of India Meteorological Department (IMD) grid data shows a trend towards an increase of about 100 mm of rainfall since 1971. The annual variability of rainfall in the basin indicates that the year 1994 had the highest annual rainfall of ∼1,780 mm whereas 1979 had the least rainfall in the past 34 years (∼900 mm). The climate is predominantly sub-tropical. April and May are the hottest months. Maximum temperature hovers upto 40 °C ( Jalali 2015). Page 39

H2 Open Journal doi: 10.2166/h2oj.2017.001

Figure 1 | Mahanadi Basin.

The Land Use/Land Cover (LULC) study of the basin for the year 2005–2006 shows 23 LULC classes. DEM data from GMTED2010 DEM (7.5 arc-sec; spatial resolution 231.525), SRTM DEM (1 arc-sec, 30.87 m; 3 arc-sec, 92.61 m), ASTER Global DEM (30.87 m) and CARTOSAT-1 DEM (30.87 m) are used for the analysis. Three sub-catchments, 21, 53 and 183 are not considered due to lack of data resulting in only 224 sub-catchments taken up for the present study. The Mahanadi basin has varying topography with the lowest elevation in coastal reaches and highest elevation found in Northern hills. The basin is divided into 11 elevation zones based on SRTM DEM. Major part of the plain region of the Mahanadi basin falls under the 200–400 m elevation zone. The middle Mahanadi sub-basin comprises of both high hilly terrain in its North-Eastern part and central table land which divides the Mahanadi middle and lower sub-basins. The elevation of middle Mahanadi sub-basin ranges between 500–1,000 m. Major part of the basin is covered with agricultural land and accounts for around 54.27% of the total basin area.

METHODS EMPLOYED AND METHODOLOGY GIS analysis

GIS analysis was performed on all the DEM datasets for delineating sub-catchments which include Georeferencing, shape ﬁle creation, joining of DEM tiles and terrain pre-processing (repeated for 224 sub-catchments and for all DEM datasets). Page 40

H2 Open Journal doi: 10.2166/h2oj.2017.001

Counting of streams

The number of streams present in each sub-catchment were counted for the estimation of geomorphological parameters (Refer Table 1). A program was developed for auto-counting of streams present in each sub-catchment. The output of the model gives information about each stream segment present in a catchment along with its stream order. This information was utilized to count number of streams present in each stream order. The point at which streams join another stream is called a node. In modern days, GIS programs can efﬁciently assign a unique number to each node present in stream network. ‘From node number’ is the point from which stream segment has originated. ‘To node number’ is the point at which a stream segment is terminating.

Table 1 | Mathematical expressions of geomorphological parameters (Raju & Nagesh Kumar 2013) Parameter

Mathematical expression

Units

Basin length (Lb )

1:312A0:568

Drainage density (Dd )

L A Nu Nuþ1

km km�1

Bifurcation ratio (Rb ) Stream frequency (Fu ) Texture ratio (T ) Form factor (Rf ) Elongation ratio (Re ) Circulatory ratio (Rc )

No units km�2

N0 A N1 P A L2b 1:128 12:57

km�1 No units

A0:5 Lb A

No units No units

A ¼ Area of catchment (Km2); P ¼ Perimeter of catchment (Km); L ¼ Total length of stream segments of all orders (Km); Nu & Nuþ1 ¼ Number of streams of a given order u and u þ 1; N0 ¼ Total number of stream segments of all orders; N1 ¼ Number of stream segments of ﬁrst order.

Entropy method

Entropy method is employed to obtain weights of the geomorphological criteria (Raju & Nagesh Kumar 2014). Steps of the methodology are as follows: 1. Formulation of payoff matrix (Array of sub-catchments and geomorphological criteria) and computation of normalized payoff matrix (pij ); i and j respectively represent sub-catchments (1,2,…m) and criteria (1,2,…n) 2. Entropy value for each geomorphological criteria j,

Ej ¼ �

m 1 X pij ln (pij ) ln (m) i¼1

(1)

3. Computation of degree of diversiﬁcation of criteria Dj ¼ 1 � Ej

(2) Page 41

H2 Open Journal doi: 10.2166/h2oj.2017.001

4. Computation of weights of criteria Dj wj ¼ Pn

j¼1

(3)

Fuzzy VIKOR

The ﬁrst priority sub-catchment is obtained through Fuzzy VIKOR. Brief methodology of fuzzy VIKOR is as follows: (‘f’ was added before the variable to represent it as fuzzy variable) (Wu et al. 2016): 1. Input the fuzzy payoff matrix, fxij in triangular membership function form (lij , mij , uij ) consisting of sub-catchments and criteria. 2. Identify fuzzy best value ffj� and worst value ffj�� for each criterion; for example in case of maximi(lij , mij , uij ) and zation, such as beneﬁt perspective, ffj� ¼ (l�j , m�j , u�j ) ¼ Maximum �� ffj�� ¼ (l�� j , mj , uj ) ¼ minimum (lij , mij , uij ); In case of minimization, such as cost perspective,

�� ffj� ¼ (l�j , m�j , u�j ) ¼ Minimum (lij , mij , uij ) and ffj�� ¼ (l�� j , mj , uj ) ¼ maximum (lij , mij , uij ) 3. Computation of normalized fuzzy difference

ffj� � fxij

fdij ¼

u�j � l�� j

fxij � ffj�

fdij ¼

� u�� j � lj

ðMaximization perspectiveÞ

(4)

ðMinimization perspectiveÞ

(5)

u l m u 4. Computation of index values fSi (Sli , Sm i , Si ) and fRi (Ri , Ri , Ri ) representing the separation measures for sub-catchment Ai from the best and worst values (Lee et al. 2015).

fSi ¼

n X j¼1

(wl , wm , wu ) � (dijl , dijm , diju )

(6)

fRi ¼ Max (wl , wm , wu ) � (dijl , dijm , diju )

(7)

5. Computation of values of summation operator fQi, using the Equation (8) "

fSi � fSmin fQi ¼ v u Smax � Slmin

fRi � fRmin þ (1 � v) l Ru max � Rmin

(8)

where fSmin ¼ Min fSi ¼ (Smin l , Smin m , Smin u )

(9)

fRmin ¼ Min fRi ¼ (Rmin l , Rmin m , Rmin u )

(10)

u l l Su max ¼ MaxSi ; Smin ¼ MinSi

(11)

u l l Ru max ¼ MaxRi ; Rmin ¼ MinRi

(12)

v represents maximum group utility strategy weight and (1-v) is the weight of the individual regret Page 42

H2 Open Journal doi: 10.2166/h2oj.2017.001

function. ν varies from 0 to 1. Defuzziﬁcation of fQi yields Qi ¼

(fQil þ fQim þ fQiu ) 3

(13)

which provides crisp value. Lower Qi value based sub-catchment is preferred for analysis and can be given priority for taking up soil and conservation improvements. The ﬂowchart of the approach developed is presented in Figure 2.

Figure 2 | Flow chart of Fuzzy VIKOR Methodology.

RESULTS AND DISCUSSION Estimation of geomorphological parameters and weights

Total number of pixels present in DEM raster was estimated using GIS software and procedure mentioned above is used for ﬁnding the area, perimeter and length of each sub-catchment. MatLab (www. mathworks.com) based program was developed for computation of the seven geomorphological parameters for all 224 sub-catchments based on the information in Table 1. Minimum and maximum values obtained among 5 DEMs for Drainage density, Bifurcation ratio, Stream Frequency, Texture Ratio, Form Factor, Elongation Ratio and Circulatory Ratio respectively are (0.053, 0.107), (2, 13), (0.002, 0.051), (0.008, 0.23), (0.211, 0.263), (0.519, 0.579), (30.25, 108.3) and corresponding Page 43

H2 Open Journal doi: 10.2166/h2oj.2017.001

differences are (0.054, 11, 0.049, 0.222, 0.052, 0.06, 78.05). Signiﬁcant variation is observed for some geomorphological parameters across all DEM sets. The present study aims at handling the variation in fuzzy environment for better modeling of the case study and suggests a methodology where variation is observed in similar situations elsewhere. Weights of criteria related to each DEM are computed using entropy method (Equations (1)–(3)). Table 2 presents weights of the various parameters for the 5 DEM sources. It is observed that texture ratio, bifurcation ratio and stream frequency contribute around 85% of total weightage whereas remaining four criteria contribute around 15% while ranking sub-catchments. In all the DEM sources, Texture ratio, bifurcation ratio and stream frequency are occupying ﬁrst three positions. Table 2 | Weights of geomorphological criteria Parameter

GMTED2010 7.5 arc-sec

SRTM90

SRTM30

ASTER

CARTOSAT-1

Triangular membership function perspective

Drainage density

0.0472

0.0420

0.0429

0.0502

0.0445

(0.042, 0.0445, 0.0502)

Bifurcation ratio

0.2890

0.2907

0.2697

0.2564

0.3009

(0.2564, 0.2890, 0.3009)

Stream frequency

0.2043

0.2186

0.2337

0.2058

0.2020

(0.2020, 0.2058, 0.2337)

Texture ratio

0.3576

0.3578

0.3609

0.3789

0.3563

(0.3563, 0.3578, 0.3789)

Form factor

0.0046

0.0041

0.0042

0.0049

0.0044

(0.0041, 0.0044, 0.0049)

Elongation ratio

0.0012

0.0010

0.0011

0.0012

0.0011

(0.001, 0.0011, 0.0012)

Circulatory ratio

0.0962

0.0857

0.0875

0.1025

0.0908

(0.0857, 0.0908, 0.1025)

Membership function formulation

Triangular membership functions are proposed to handle the deviation of geomorphological parameters obtained from the 5 DEMs. For example, bifurcation ratios obtained for DEMs 1 to 5 for sub-catchment 2 are 3.1667, 3.1667, 3.2407, 3.119, 3.4643 and these values are arranged in the ascending order 3.119, 3.1667, 3.1667, 3.2407, 3.4643. While formulating in triangular membership form, ﬁrst, third and last values are chosen as the elements representing lower, middle and upper (l, m, u) i.e., (3.119, 3.1667, 3.4643). Similar process is repeated for all the seven parameters for all the 224 sub-catchments respectively. Similar procedure is adopted for weights of criteria for formulation of triangular membership. These are presented as part of Table 2. Ranking/grouping of the sub-catchments

MatLab based Fuzzy VIKOR code is developed for ranking the 224 sub-catchments based on the formulated payoff matrix in a fuzzy environment. Various steps employed in ranking/grouping the subcatchments are as follows: Main aim of normalized fuzzy difference matrix is to make the data dimensionless. This is required when different features are simultaneously considered. High values of ﬁrst four criteria, Drainage Density, Bifurcation Ratio, Stream Frequency, Texture Ratio are preferred whereas low values of Form Factor, Elongation Ratio and Circulatory Ratio are preferred (Kumar et al. 2017). Accordingly, normalized fuzzy difference matrix values are computed (Equations (4) and (5)). Bifurcation Ratio values for sub-catchment 1 are (2.556, 2.6111, 2.6111). Ideal (fj� ) and anti-ideal �� (fj ) values for each Bifurcation Ratio are found to be (11, 12, 13) and (2, 2, 2) and accordingly u�j and l�� j values are chosen as 13 and 2. Based on Equation (4), normalized fuzzy difference value is: fdij ¼ Page 44

ffj� � fxij u�j � l�� j

ð11; 12; 13Þ � ð2:556; 2:6111; 2:6111Þ ð11 � 2:6111; 12 � 2:6111; 13 � 2:556Þ ¼ 13 � 2 11

H2 Open Journal doi: 10.2166/h2oj.2017.001

or (0.7626, 0.8535, 0.9495). Similar computations yield normalized fuzzy difference matrix values for other parameters. Equations (6)–(13) were applied for 224 sub-catchments using the described methodology for computing fSi , fRi , fQi . Lower Qi is preferred and ranking is performed accordingly. Table 3 presents ranking pattern/grouping of the sub-catchments. Table 3 | Ranking pattern/grouping of the sub-catchments (for v ¼ 0.5) Number of Group

catchments

Improvement List of Sub-catchments

Q value

Priority ranking

26, 44, 47, 158, 179

0.03–0.09

2, 5, 11, 18, 30, 34, 36, 37, 38, 40, 78, 96, 108, 130, 141, 142, 148, 149, 153, 174, 175, 183, 190, 197, 200,220

0.10–0.15

3, 4, 6, 7, 9, 14, 16, 20, 22, 23, 28, 29, 35, 39, 45, 48, 52, 55, 59, 61, 63, 69, 71, 73, 74, 79, 82, 85, 86, 88, 93, 98, 99, 100, 103, 104, 106, 109, 111, 113, 117, 118, 122, 123, 126, 128, 135, 137, 138, 143, 156, 159, 160, 161, 163, 164, 167, 169, 172, 178, 180, 185, 186, 195, 196, 201, 203, 211, 212

0.16–0.20

1, 8, 10, 12, 13, 15, 19, 21, 24, 25, 27, 31, 46, 49, 50, 53, 54, 56, 57, 60, 67, 68, 72, 76, 80, 81, 83, 89, 91, 94, 102, 105, 107, 114, 120, 121, 125, 129, 133, 140, 144, 145, 151, 154, 157, 166, 168, 170, 171, 173, 177, 181, 184, 191, 192, 193, 194, 198, 204, 205, 206, 207, 219, 221, 222

0.21–0.25

17, 43, 51, 62, 65, 77, 84,90, 95, 101, 110, 115, 116, 119, 124, 127, 132, 136, 146, 165, 176, 182, 189, 199, 202, 208, 209, 218, 223

0.26–0.30

33, 41, 58, 87, 134, 147, 150, 152, 188, 214, 216

0.31–0.35

66, 70, 75, 92, 112, 131, 139, 155, 187, 210, 213, 224

0.36–0.40

32, 42, 64, 97, 162, 215, 217

0.41–0.45

It is observed that Qi values for most of the catchments are almost same with minute differences. Keeping this in view, grouping of the catchments is performed instead of ranking based on the range of Qi values. A total of eight groups are formulated with number of catchments in each group as 5, 26, 69, 65, 29, 11, 12, 7 respectively. Highest number of catchments are falling in group 3 and 4 with a Q value range of 0.16–0.20 & 0.21–0.25. It is observed that group 1 can be explored for improvement on a priority basis and accordingly other groups can be improved as noted in Table 3. Ranking method proposed here facilitates prioritization of sub-catchments. These sub-catchments based on their priority can be provided suitable conservation measures which ultimately are expected to provide sustainable water management practices in the Mahanadi river basin. Some of the conservation measures that can be explored are check dams, initiation of woody plants, masonry stone bunds construction, gullies reforestation, ponds and embankments. However, precise information on the magnitude and rates of erosion and sedimentation and socioeconomic and environmental effects are key to success in implementing sustainable soil conservation programs. Sensitivity analysis

Effect of strategy weight (v) in fuzzy VIKOR on the ranking pattern is also studied and presented in Table 4. Values of ‘v’ are varied from 0 to 1. It is found that sub-catchment 158 occupied first position (for v values 0 to 0.6) whereas sub-catchment 179 occupied first position (for v values 0.7 to 0.8) and sub-catchment 11 in case of v values of 0.9 to 1. In case of second position, these are sub-catchment 47 (v values from 0 to 0.3) and 179 (from 0.4 to 0.6). To our knowledge, this is the first application of fuzzy VIKOR for ranking sub-catchments in Mahanadi Basin using morphological data explored from five DEM sources. Page 45

H2 Open Journal doi: 10.2166/h2oj.2017.001

10 Table 4 | Effect of strategy weight on the top 5 sub-catchments Rank

ν¼0

ν ¼ 0.1

ν ¼ 0.2

ν ¼ 0.3

ν ¼ 0.4

ν ¼ 0.5

ν ¼ 0.6

ν ¼ 0.7

ν ¼ 0.8

ν ¼ 0.9

ν ¼ 1.0

158

179

158

179

158

179

130

179

200

179

130

200

128

130

175

190

175

130

200

158

100

175

200

175

190

175

141

126

128

197

200

141

The methodology proposed in the present study utilizes only the topographic information to prioritize the sub-catchments. This method can be easily applied to areas which do not have sufﬁcient data for detailed hydraulic studies.

CONCLUSIONS In this study, data from ﬁve DEM sources i.e., GMTED2010 7.5 arc-sec, SRTM90, SRTM30, ASTER and CARTOSAT-1 were used to calculate the 7 geomorphological parameters for 224 sub-catchments of Mahanadi basin. Fuzzy VIKOR, was utilized for prioritizing the sub-catchments. Eight groups of sub-catchments were formulated for possible implementation of conservation measures for the chosen strategy weight of 0.5. However, careful selection of strategy weight is essential for meaningful inferences from the present study. Present study is preliminary work initiated to evaluate the study area in terms of sub-catchments prioritization. This will be followed by ﬁeld validation which is targeted as further study.

REFERENCES Aher, P. D., Adinarayana, J. & Gorantiwar, S. D. 2014 Quantiﬁcation of morphometric characterization and prioritization for management planning in semi-arid tropics of India: a remote sensing and GIS approach. J. Hydrol. 511, 850–860, DOI:10. 1016/j.jhydrol.2014.02.028. Chandniha, S. K. & Kansal, M. L. 2017 Prioritization of sub-watersheds based on morphometric analysis using geospatial technique in Piperiya watershed, India. Appl. Water Sci. 7(1), 329–338, DOI:10.1007/s13201-014-0248-9. Deshmukh, D. S., Chaube, U. C., Tignath, S. & Pingale, S. M. 2011 Geomorphological analysis and distribution of badland around the conﬂuence of Narmada and Sher river, India. Eur. Water 36, 15–26. Iqbal, M. & Sajjad, H. 2014 Watershed prioritization using morphometric and land use/land cover parameters of Dudhganga catchment Kashmir valley India using spatial technology. J. Geophys. Remote Sens. 3(1), 1–12, DOI:10.4172/2169-0049. 1000115. Jaiswal, R. K., Thomas, T., Galkate, R. V., Ghosh, N. C. & Singh, S. 2014 Watershed prioritization using Saaty’s AHP based decision support for soil conservation measures. Water Resour. Manage. 28(2), 475–494, DOI:10.1007/s11269-013-0494-x. Jaiswal, R. K., Ghosh, N. C., Lohani, A. K. & Thomas, T. 2015 Fuzzy AHP based multi criteria decision support for watershed prioritization. Water Resour. Manage. 29(12), 4205–4227, DOI:10.1007/s11269-015-1054-3. Jalali, A. 2015 Prioritization of Sub-Catchments of Mahanadi Basin Based on Geomorphology. M.Tech Dissertation, Indian Institute of Science, Bangalore. Javed, A., Khanday, M. Y. & Ahmed, R. 2009 Prioritization of sub-watersheds based on morphometric and land use analysis using remote sensing and GIS techniques. J. Indian Soc. Remote Sens. 37(2), 261–274, DOI:10.1007/s12524-009-0016-8. Javed, A., Khanday, M. Y. & Rais, S. 2011 Watershed prioritization using morphometric and land use/land cover parameters: a remote sensing and GIS based approach. J. Geol. Soc. India 78(1), 63–75, DOI:10.1007/s12594-011-0068-6.

Page 46

H2 Open Journal doi: 10.2166/h2oj.2017.001

Kanth, T. A. & Hassan, Z. 2012 Morphometric analysis and prioritization of watersheds for soil and water resource management in Wular catchment using geo-spatial tools. Int. J. Geol. Earth Environ. Sci. 2(1), 30–41. Khanday, M. Y. & Javed, A. 2016 Prioritization of sub-watersheds for conservation measures in a semi-arid watershed using remote sensing and GIS. J. Geol. Soc. India 88(2), 185–196, DOI:10.1007/s12594-016-0477-7. Kumar, L., Khalkho, D., Katre, P. & Nigam, G. K. 2017 Prioritization of sub watersheds using morphometric analysis: a remote sensing and GIS perspective. Int. J. Innov. Res. Sci. Eng. Technol. 6(1), 135–146, DOI:10.15680/IJIRSET.2017.0601017. Lee, G., Jun, K. S. & Chung, E. S. 2015 Group decision-making approach for flood vulnerability identification using the fuzzy VIKOR method. Nat. Hazards Earth Syst. Sci. 15, 863–874, DOI:10.5194/nhess-15-863-2015. Makwana, J. & Tiwari, M. K. 2016 Prioritization of agricultural sub-watersheds in semi- arid middle region of Gujarat using remote sensing and GIS. Environ. Earth Sci. 75, 137, DOI:10.1007/s12665-015-4935-0. Manfreda, S., Leo, M. D. & Sole, A. 2011 Detection of flood-prone areas using digital elevation models. J. Hydrol. Eng. 16(10), 781–790, DOI:10.1061/(ASCE)HE.1943-5584.0000367. Noman, N. S., Nelson, E. J. & Zundel, A. K. 2001 A review of automated flood plain delineation from digital terrain models. J. Water Resour. Plann. Manage. 127(6), 394–402, DOI:10.1061/(ASCE)0733-9496(2001)127:6(394). Papaioannou, G., Vasiliades, L. & Loukas, A. 2015 Multi-Criteria analysis framework for potential flood prone areas mapping. Water Resour. Manage. 29, 399–418, DOI:10.1007/s11269-014-0817-6. Patel, D. P., Dholakia, M., Naresh, N. & Srivastava, P. K. 2012 Water harvesting structure positioning by using geo-visualization concept and prioritization of mini-watersheds through morpho-metric analysis in the lower Tapi basin. J. Indian Soc. Remote Sens. 40(2), 299–312, DOI:10.1007/s12524-011-0147-6. Patel, D. P., Srivastava, P. K., Gupta, M. & Nandhakumar, N. 2015 Decision support system integrated with geographic information system to target restoration actions in watersheds of arid environment: a case study of Hathmati watershed, Sabarkantha district, Gujarat. J. Earth Sys. Sci. 124(1), 71–86, DOI:10.1007/s12040-014-0515-z. Rai, S. P., Kumar, V. & Goyal, V. C. 2001 Geomorphology and Soil Erosion in Juni Nadi Watershed, District Udhampur, J& K. National Institute of Hydrology, Roorkee, Report No. CS/AR-4/2000-2001, pp. 1–29. Raju, K. S. & Nagesh Kumar, D. 2013 Prioritization of micro-catchments based on morphology. Proc. Inst. Civil Eng. Water Manage. 166(WM7), 367–380. DOI/10.1680/wama.11.00076. Raju, K. S. & Nagesh Kumar, D. 2014 Multicriterion Analysis in Engineering and Management. Prentice Hall of India, New Delhi. Rudraiah, M., Govindaiah, S. & Vittala, S. S. 2008 Morphometry using remote sensing and GIS techniques in the sub-basins of Kagna river basin, Gulburga district, Karnataka, India. J. Indian Soc. Remote Sens. 36(4), 351–360, DOI:10.1007/s12524008-0035-x. Singh, N. & Singh, K. K. 2017 Geomorphological analysis and prioritization of sub-watersheds using Snyder’s synthetic unit hydrograph method. Appl. Water Sci. 7(1), 275–283, DOI:10.1007/s13201-014-0243-1. Thakkar, A. K. & Dhiman, S. D. 2007 Morphometric analysis and prioritization of miniwatersheds in Mohr watershed, Gujarat using remote sensing and GIS techniques. J. Indian Soc. Remote Sens. 35(4), 313–321, DOI:10.1007/BF02990787. Uniyal, S. & Gupta, P. 2013 Prioritization based on morphometric analysis of Bhilangana watershed using spatial technology. Int. J. Remote Sens. Geosci. 2(1), 49–57. Williams, W. A., Jensen, M. E., Winne, J. C. & Redmond, R. L. 2000 An automated technique for delineating and characterizing valleybottom settings. Environ. Monit. Assess. 64(1), 105–114, DOI:10.1023/A:1006471427421. Wolock, D. M. & Price, C. V. 1994 Effects of digital elevation model map scale and data resolution on a topography-based watershed model. Water Resour. Res. 30(11), 3041–3052, DOI:10.1029/94WR01971. Wu, Z., Ahmad, J. & Xu, J. 2016 A group decision making framework based on fuzzy VIKOR approach for machine tool selection with linguistic information. Appl. Soft Comput. 42, 314–324, DOI:10.1016/j.asoc.2016.02.007. Yan, K., Tarpanelli, A., Balint, G., Moramarco, T. & Baldassarre, G. D. 2014 Exploring the potential of SRTM topography and radar altimetry to support flood propagation modeling: Danube case study. J. Hydrol. Eng. 20(2), 04014048, DOI:10.1061/ (ASCE)HE.1943-5584.0001018. Yasmin, K., Polisgowdar, B. S., Kumar, U. S., Ayyangoudar, M. S. & Rao, K. N. 2013 Morphometric analysis of Milli watershed of Raichur district using GIS techniques. Karnataka J. Agric. Sci. 26(1), 92–96. Zhang, H. Y., Shi, Z. H., Fang, N. F. & Guo, M. H. 2015 Linking watershed geomorphic characteristics to sediment yield: evidence from the Loess Plateau of China. Geomorphology 234, 19–27, DOI:10.1016/j.geomorph.2015.01.014.

Page 47

H2 Open Journal doi: 10.2166/h2oj.2018.007

Occurrence of trihalomethane in relation to treatment technologies and water quality under tropical conditions A. A. G. D. Amarasooriyaa,*, S. K. Weragodab, M. Makehelwalab and R. Weerasooriyac a

Postgrduate Institute of Science, University of Peradeniya, Peradeniya, Sri Lanka

National Water Supply and Drainage Board, Advanced instrumental Laboratory, Kandy, Sri Lanka

National Institute of Fundamental Studies, Hantana Road, Kandy, Sri Lanka

*Corresponding author. E-mail: gayanamarasooriya@gmail.com

Abstract Distribution of most prevalent disinfection by-products, trihalomethanes (THMs) in relation to treatment technology and common water quality parameters (turbidity, conductivity, color, pH, and residual chlorine) was examined for two water supply schemes (WSS) in Sri Lanka (locations: Greater Kandy-WSS (GKWSS) (80.56– 80.66 °E, 7.28–7.38 °N) and Kandy South-WSS (KSWSS) (80.49–80.63 °E, 7.21–7.30 °N). In both treatment plants, only CHCl3 and CHCl2Br were detected in appreciable concentrations and total THMs (TTHMs) values were well below the WHO limits (80 μg/L). TTHMs variations ranged from 0 to 16 μg/L and 0 to 54 μg/L in GKWSS and KSWSS, respectively. Highest TTHM value (54 μg/L) was found in KSWSS which employs pulsation treatment technology. Correlations between CHCl3 and CHCl2Br in both water schemes are noteworthy, but THM levels relate to most of the water quality parameters ambiguously. However, a distinct relationship is observed between THM levels and degree of chlorination, resident time, pipeline corrosion, and temperature. THM formation increased towards the boundaries of most of the sub-water supply schemes (SWSS). Key words: disinfection by-products, Sri Lanka, THM, TTHM, water quality

INTRODUCTION Chlorination is the most common disinfection method used to destroy pathogenic microorganisms in potable water (Morris & Levin 1995). Although the exact chemical structures of natural organic matter (NOM) are unresolved to date, they are ubiquitous in natural waters. It is well known that upon chlorination, the NOM often acts as a precursor in the formation of disinfection by-products (DBPs) (Rook 1974; Li & Mitch 2018). The most prevalent DBP classes are CHCl3, CHCl2Br, CHClBr2, and CHBr3; the sum of them are designated as total trihalomethanes (hereafter TTHMs) (Rook 1974; Krasner et al. 2006). During the past four decades, many researchers have examined DBP levels in water (Hu et al. 2010), their formation pathways (Wang et al. 2017), biotoxicity, and mitigation methods (Ashbolt 2004). Epidemiologic studies have shown a relationship between long-term exposure to DBPs and increased cancer risks and adverse reproductive effects (IARC 1991; Singer 1999; Gordon et al. This is an Open Access article distributed under the terms of the Creative Commons Attribution Licence (CC BY 4.0), which permits copying, adaptation and redistribution, provided the original work is properly cited (http://creativecommons.org/licenses/by/4.0/).

Page 49

H2 Open Journal doi: 10.2166/h2oj.2018.007

2008). Thus, the US Environmental Protection Agency (EPA) has promulgated the maximum contaminant level of THMs as 80 μg/L (Fooladvand et al. 2011). Operational procedures typically implemented in water treatment plants (WTP) such as chlorine dosage, resident time, pH, total organic carbon content, etc., have a marked impact on THM formation (Sadiq & Rodriguez 2004; Navalon et al. 2008). THM formation is also varied with seasonal fluctuations and geography of the water resources (Williams et al. 1997). In different geographical locations, such as Spain, China, South Korea, Greece, the US, and Iran, the average THMs in WTP vary widely, namely, 9–177 mg/L and sometimes exceeding regulatory limits (Krasner & Wright 2005; Platikanov et al. 2012; Hladik et al. 2014; Ramavandi et al. 2015). Particularly in tropical regions, research focus to examine the effects of waterworks’ management practices on THM formation are limited (Abdullah et al. 2003; Panyapinyopol et al. 2005; Baytak et al. 2008; Hasan et al. 2010; Amjad et al. 2013). None of these studies specifically assessed the effects of different treatment technologies on THM formation. To address this issue, for the first time in Sri Lanka, we carried out research to compare THM levels in water that resulted from different treatment methods upon chlorination. Two major WTP were selected to monitor THM levels and their impact on treatment technology. The selection of treatment plants was made due to the following reasons. Greater Kandy treatment plant (80.6203 °E, 7.3166 °N) (hereafter GKWTP) follows conventional technology whereas Kandy South treatment plant (80.5943 °E, 7.2487 °N) (hereafter KSWTP) operates under pulsation technology. Both plants receive water from the same surface water source (intake locations; GKWTP 80.6220 °E, 7.3064 °N and KSWTP 80.5946 °E, 7.2487 °N). Both plants received ISO: 9001 accreditation under a common source to tap water quality. THM monitoring and speciation was carried out using ECD-GC coupled with automated headspace analyzer systems.

MATERIALS AND METHODS Materials

Methanol (HPLC grade), THMs certiﬁed standards (reference number: 4S8746), and Na2S2O3 (analytical grade) were obtained from Sigma-Aldrich (USA). The chemicals used for free Cl2 and total Cl2 measurements were purchased from HACH (USA). Analytical method

THMs were analyzed by head space method as proposed in Kuivinen & Johnsson (1999). Dedicated head space gas chromatography coupled with an electron capture detector (ECD) and a built-in auto sampler (Thermo trace 1300 GC-ECD and TRIPlus RSH auto sampler) was used for THMs analyses under split/ split-less mode (TRACE – TR5). The auto sampler consists of agitation and incubation steps that automatically convert samples into headspace. Data processing was carried out using dedicated quality assured software (Choromelen 7, version 7.2, USA). Free chlorine was measured using DPD standard colorimetric method 4500-Cl F (APHA 2005) with DR 5000 colorimeter (HACH, USA). The pH, EC, and turbidity measurements were determined using a pH meter (Hansen’s IONþ pH3, USA), conductivity meter (Model: ELE 470, EU), and a turbidity meter (HACH 2100P, USA), respectively. Sampling sites were located using Gramin global positioning system GPS (Graminetrex, USA). Quality assurance and quality control

In compliance with QA & QC protocols, quality of the THM analysis was controlled utilizing ten external standards with different THM concentrations; the relative standard error was always less Page 50

H2 Open Journal doi: 10.2166/h2oj.2018.007

than 9% with an excellent linearity of calibration (R2 . 0.99). To evaluate possible matrix effects, samples were spiked with 20 μg/L THMs at each 20th point; in all cases, over 95% spike recoveries for all THMs examined were observed. The THMs were analyzed in replicates. Field and laboratory blank analyses were used for background corrections. Water treatment systems

Process flow charts of GKWTP and KSWTP are shown in Figure 1(a) and 1(b), respectively. Both plants utilize the same surface water from the River Mahaweli (intake locations: GKWTP 80.6220 °E, 7.3064 °N and KSWTP 80.5946 °E, 7.2487 °N). GKWTP is located 10.8 km downstream along the river from KSWTP. In the GKWTP, the raw surface water was pumped directly into the treatment plant. At the chemical mixing point, poly-aluminum chloride (PAC) was added as a coagulant. Chemically mixed water was then transferred to the flocculation basin under gravity. Baffle walls of the basin increase residence time by increasing water flow paths that are essential to enhance coagulation and

Figure 1 | Process ﬂow diagram of (a) GKWTP and (b) KSWTP.

Page 51

H2 Open Journal doi: 10.2166/h2oj.2018.007

flocculation. Most of the settled flocs in the sedimentation basin were regularly removed by scraping and transferred to a lagoon for dehydration. The water with fine unsettled flocs then flowed into sand filters for removal of fine particulates. After sand filtration, to raise water pH, lime was added. The treated water after chlorination was transferred to a clear water tank. In the KSWTP, before entering into the pulsator-clarifier, water was pumped into a cascade aerator. PAC coagulant was added to the raw water at the inlet point of the cascade aerator. The coagulated water was then transferred to a vacuum chamber located in the pulsator clarifier. By the vacuum pressure created by the vacuum fan, the water level was raised to a pre-determined level. At this point, a vacuum breaker was opened automatically which surged water into bottom perforated distribution pipes. As water was distributed through the stilling plates, gently stirring turbulence created by ‘pulses’ enhanced coagulation. The flocs accumulated on the plates as a sludge blanket overflowed to sludge concentrators for periodic removal. The treated water was collected from the top part of the clarifier. When the water level was lowest in the vacuum chamber, the automatic vacuum breaker closed, repeating the cycle. The treated water was filtered by rapid sand filters and disinfected, and stored in the clear water tank for distribution.

Sample collection and preservation

Figure 2(a) and 2(b) show the coordinates of WTP, service reservoirs, and distribution networks used for sampling. Fixation of residual chlorine was carried out in the ﬁeld by adding sodium thiosulfates (10 mg per 40-mL sample for up to 5 mg/L chlorine) to empty amber-colored bottles prior to sampling. Water samples of distribution pipe lines were taken from the tap at the nearest possible point to the main pipeline. Before sampling, water was allowed to run for 5 min and then sample bottles were ﬁlled with water without leaving a headspace. Sampling details are shown in Table 1. A total of 56 samples were collected from GKWSS and 73 samples from KSWSS in the year 2014.

Data processing

Statistical analysis was carried out using public domain R statistical software (R Foundation for Statistical Computing, version 1.14.4). Spatial variation maps of THM and TTHM concentrations were developed using Surfer surface mapping software (Golden Software Inc., Version 11.0.642).

RESULTS AND DISCUSSION The variations of THMs, TTHMs, and other water quality parameters, namely, pH turbidity, conductivity, residual chlorine, and color, at different locations of GKSSS and KSWSS are shown in Figures 3 and 4. For both plants, a common water source is used; this ensues similarity in water composition, particularly with respect to major constituents. Therefore, upon chlorination, we expect a similar formation mechanism/s of THMs. Out of the THMs (CHCl3, CHCl2Br, CHClBr2, and CHBr3) examined, only CHCl3 and CHCl2Br were detected in service reservoirs (SR) and sub-water supply schemes (SWSS). In the presence of NOM particularly enriched with phenolic groups, bromide readily converts HOCl → HOBr forming CHCl2Br by bromination of CHCl3 (Heeb et al. 2014; Criquet et al. 2015). The provenance of Br� in the receiving water is inconclusive to date; however, natural bromide is often concentrated in top layers of soils (upper 60 cm) that can be leached into surface water upon intense precipitation. Further enhanced soil erosion resulting from farming may also accelerate the migration of Br� into natural waters (Wang et al. 2010). Page 52

H2 Open Journal doi: 10.2166/h2oj.2018.007

Figure 2 | Layout of water treatment plants, service reservoirs, and distribution networks used for sampling (a) GKWSS and (b) KSWSS.

Variation of CHCl3 and CHCl2Br in GKSS and KSWSS

In the GKWSS, the CHCl3 concentration range is 13.9–16.2 (GKWTP), 0–19.5 (NG), 0.00–18.3 (KH), 8.79–18.8 (KL), 0.00–18.0 (AS), and 3.76–18.2 (KN) μg/L and in the KSWSS it is 11.5–13.8 (KSWTP), Page 53

H2 Open Journal doi: 10.2166/h2oj.2018.007

6 Table 1 | Number of samples collected from GKWSS and KSWSS Locations and SWSS in GKWSS

Number of samples

Locations and SWSS in KSWSS

Number of samples

Nugawela-SWSS(NG)

Maligathanna-SWSS(ML)

Kahalla-SWSS(KH)

Angunawala-SWSS(AG)

Kulugammana-SWSS(KL)

Mahakanda-SWSS(MH)

Asgiriya-SWSS(AS)

Service reservoirs (SR)

Kondadeniya-SWSS(KN)

KSWTP

Service reservoirs (SR)

GKWTP

Figure 3 | Variations of THMs, pH, turbidity, conductivity, residual chlorine, and color in GKWSS.

Page 54

H2 Open Journal doi: 10.2166/h2oj.2018.007

Figure 4 | Variations of THMs, pH, turbidity, conductivity, residual chlorine, and color in KSWSS.

0–42.5 (ML), 8.46–25.7 (AG), and 7.74–32.1(MH) μg/L. As shown in Figures 3 and 4, the highest concentration of CHCl3 is detected at NG (19.5 μg/L; GKSS), ML (42.5 μg/L; KSWSS), and MH (54.9 μg/L; KSWSS). Distance between NG/ML or MH and corresponding WTP is higher than Page 55

H2 Open Journal doi: 10.2166/h2oj.2018.007

other SWSS; therefore, the water residence time is increased with a concomitant increase of CHCl3 formation. At different localities in the GKWSS, the CHCl2Br concentration range is 2.43–3.76 (GKWTP), 0.0–5.99 (NG), 0.0–6.16 (KH), 0.0–5.64 (KL), and 0.0–6.17 (KN) μg/L. Similarly in the KSWSS, the CHCl2Br concentration range is 2.67–3.73 (KSWTP), 0.0–12.3 (ML), 0.0–13 (AG), and 0.0–6.15 (MH) μg/L. The highest concentration of CHCl2Br is detected at KN (6.17 μg/L; GKWSS) and ML (12.3 μg/L; KSWSS). However, for the KSWTP, KN is not located at the highest distance; due to low water demand of the plant, the water stagnation resulted in high residence time yielding high CHCl2Br concentration. Variations of TTHM and THM in water supply schemes

Similar assessments are performed for TTHMs in the KSWSS and GKWSS. In the GKWSS, the concentration of TTHMs range is 16.7–20.0 (GKWTP), 0.0–25.5 (NG), 0.0–24.5 (KH), 8.89–25.0 (KL), and 0.0–23.6 (KN) μg/L and in the KSWSS it is 14.1–17.5 (KSWTP), 0.0–54.9 (ML), 8.46–38.7 (AG), and 7.74–38.3 (MH) μg/L. As expected, the highest TTHMs concentration is found in KH (25.5 μg/L; GKSS) and in MH (54.9 μg/L; KSWSS). However, KH (GKWSS) is not located at the highest distance from the plant. As argued earlier, the high TTHM concentration is accounted for by the enhanced residence time of water due to stagnation. Variations of TTHM and THM in water supply plants

Average concentrations of THMs in the GKWTP (CHCl3 14.5 μg/L, CHCl2Br 3.27 μg/L, TTHMs 18.1 μg/L) are higher than KSWTP (CHCl3 6.31 μg/L, CHCl2Br 1.48 μg/L, TTHMs 7.79 μg/L). In both plants (GKWTP and KSWTP), booster chlorination points are absent. Therefore, the minimum residual chlorine concentration of 0.2 mg/L at the consumers’ end point is maintained introducing high chlorine doses. When compared to KSWSS, GKWSS has a lengthy distribution system (Figure 2(a) and 2(b)). Therefore, the formation of THMs is more favored in the GKWSS than in the KSWSS. Additionally, the technologies adapted to the treatment plants differ, which can also be considered as a contributing factor for THM formation. Variation of THM/TTHM and water distribution

Temperature also exerts an important role on THM retention in the water phase. Even ambient conditions, namely, 25 °C, the loss of THMs and chlorine gas from aqueous phase is marked (Kuivinen & Johnsson 1999; Gordon et al. 2005; Danileviciute et al. 2012). THM levels increase with the temperature and chlorine residuals (Nikolaou et al. 1999). The annual temperature of the locations of treatment plants ranges from 30 to 23 °C with an average of 24.5 °C. The annual ﬂuctuation of temperature is around 2 °C (Climate-Kandy 2017). Therefore, essentially, constant outﬂux of THMs from treatment plants into the gaseous phase is envisaged. Further, none of the treatment plants is operated continuously; hence, the creation of gaseous headspace in pipe networks also favors TTHM and chlorine residual evaporation. Variation of TTHM/THM and water quality

Variation of pH, turbidity, conductivity, residual chlorine, and color in the GKWSS and KSWSS is shown by Figures 3(d)–3(h) and 4(d)–4(h), respectively. In the GKWSS, pH and turbidity values ranged from 6.64 to 7.50 and 0.08 to 0.84 NTU, respectively. In the KSWSS, pH and turbidity values ranged from 5.90 to 7.50 and 0.00 to 1.70 NTU, respectively. In some SWSS localities, turbidity, conductivity, and color levels are high both in the GKWSS and KSWSS. This is due to the Page 56

H2 Open Journal doi: 10.2166/h2oj.2018.007

suspension of the deposited mud on pipelines by turbulence flow, leakages, cross-connection, and pipe material corrosions. Residual chlorine values ranged from 0.4 to 1.2 mg/L and 0.0 to 0.8 mg/L in the GKWSS and KSWSS, respectively. Usually, high water demand reduces water retention time by minimizing water stagnation which reduces chlorine decay. Natural decays of chlorine in the distribution networks have been described extensively (Clark 1998; Ozdemir & Ger 1998; Powell et al. 2000). When compared to the KSWSS, in the GKWSS the variations of pH as a function of sampling locations are marked. However, along the same water flow lines, in the GKWSS the residual chlorine concentration shows a minimal variation, but in the KSWSS it shows rapid fluctuations. However, in both plants, the variation of water color follows opposite trends. Both turbidity and conductivity show intermediate variations with sampling localities. Strikingly, along with the same water flow paths, in the GKWSS the variation of CHCl3, CHCl2Br, and TTHM are minimal whereas in the KSWSS some fluctuations occur with respect to CHCl3 and TTHMs concentrations, in particular. It is important to note that in both cases the composition of inlet water is essentially the same; therefore, the observed variations in water quality parameters and occurrence of TTHMs can largely be ascribed to the different treatment technologies adopted. The results so far presented indicate a complex behavior of THM and TTHM formation even in the inlet water from a common source. The situation is further complicated when different treatment technologies are introduced. Therefore, particular attention is given below to analyze the effects of the aforementioned parameters on THM/TTHM formation. Correlation of water quality parameters and THM/TTHM

Average concentrations of CHCl3 and CHCl2Br in water strongly correlate in most of the sampling localities (locations: KH, NG, AS, KL; GKWTP) and (locations: ML, AG, MH; KSWTP) (Figures 5 and 6). In the presence of precursory NOM, it appears that CHCl3 is formed ﬁrst, which shows subsequent conversion into CHCl2Br via bromine addition (Kumar & Margerum 1987; Krasner 1999; Hua et al. 2006). Correlations with turbidity depend on the particular species of TTHM. The formation of THM and CHCl3 seems to reduce with the increase of turbidity (location GKWSS; NG, KL, KH) whereas in some locations (e.g., GKWTP; KSWSS, KSWTP) an opposite trend is observed. This implies an intimate association between NOM and turbidity. Turbidity measurement peak response is between 400 and 600 nm (APHA 2005). Therefore, the turbidity spectral peak overlaps with humic acids (Hua et al. 2014). However, when pipes are ruptured or cracked, the release rates of residual chlorine offset TTHM formation which results in a negative relationship with the turbidity.

Figure 5 | Graphical illustration of correlation matrix for GKWSS. TTHMs: total trihalomethane, RCl: residual chlorine, WTP: water treatment plant, SR: service reservoir, Avg.: averaged.

Page 57

H2 Open Journal doi: 10.2166/h2oj.2018.007

Figure 6 | Graphical illustration of correlation matrix for KSWSS. TTHMs: total trihalomethane, RCl: residual chlorine, WTP: water treatment plant, SR: service reservoir, Avg.: averaged.

As shown in Figures 3(d) and 4(d), the variations of pH in water throughout sampling sites of GKWSS and GKWTP show minimal fluctuations. The increase of pH is favorable for enhanced THM formation (Rook 1976; El-Dib & Ali 1995). However, our correlation data suggest that even a small change of pH exerts a marked effect on THM formation. Therefore, even minute pH fluctuations may result in wide variations of THMs, CHCl3, and CHCl2Br levels rendering data instability in correlation calculations. The same arguments can be made to explain the apparent variability of TTHM formation with the residual chlorine. In the GKWSS, the variations of residual chlorine levels at different localities are essentially constant (Figures 3(g) and 5). As expected, the levels of CHCl3, CHCl2Br, and TTHM also show indifferent variations. In the KSWSS, the variations of the concentrations of residual chlorine shown are high; this corresponds to fluctuations of CHCl3, CHCl2Br, and TTHM concentrations throughout sampling locations (Figures 4(g) and 6). In both plants, the conductivity exhibits strong negative correlations with THM concentration. Although the exact explanation for this observation is not possible to date, it seems that conductivity plays an indirect role in THM formation. One of the favorable precursors for THMs is hydrophobic NOM (Guanghui & David 2007; Jagatheesan et al. 2008). When the conductivity of the aqueous phase is increased (by increasing conductivity), the NOM seems to be aggregated into large moieties due to its water repellency. However, in low conductivity water, this effect seems to be minimized, thus dispersing small grained NOM moieties in the aqueous phase. The surface reactivity of small NOM moieties are expected to be higher than large ones due to its enhanced reactivity sites (Gang et al. 2003). Therefore, THM formation is expected to be high in low conductivity water. In most locations of GKWSS and KSWSS, negative correlations between color and THM levels are observed. Light absorption by dissolved organic carbon has a strong influence on the penetration of ultraviolet and photo-chemically active radiation (Morris et al. 1995). High molecular weight fractions of NOM, commonly known as humic acids, contribute to the intense coloration of water. Although water coloration is low, the low molecular NOM fraction favors enhanced THM formation (Xu et al. 2015). Therefore THM formation is expected to be inhibited in colored water. Both in GKWSS and KSWSS, inverse relationships are shown between TTHM concentrations and the distance between the service reservoirs and treatment plants. Residual chlorine concentration is reduced with the distance which retards THM formation. Possibly the increased residence time seems Page 58

H2 Open Journal doi: 10.2166/h2oj.2018.007

to implicate THM formation patterns which result in a weak direct relationship. To understand the trends of THM variations, spatial maps are developed for both WSS (Figures 7–10). Spatial variation of THMs in GKWSS

Spatial distribution of CHCl2Br, CHCl3, and TTHMs in the GKWSS based on service reservoirs’ THMs are shown in Figure 7. According to these ﬁgures, service reservoirs’ CHCl3 and TTHMs levels were higher in the north than the south. A possible reason could be the high retention time. Variation of CHCl3 was similar to TTHMs, but CHCl2Br did not show a similar trend since CHCl2Br concentrations were changed slightly. Figure 8 shows THM variation in individual SWSS. According to these ﬁgures, the trend of variation of major components CHCl3 and CHCl2Br were similar to TTHMs. Towards the AS and KN boundary, CHCl3, CHCl2Br, and TTHM values were decreased slightly. In NG, THMs were decreased towards the south-west boundary and increased toward the north-east boundary. In KL,

Figure 7 | CHCl2Br, CHCl3, and TTHM variation in GKWSS based on THM levels in service reservoirs.

Figure 8 | CHCl2Br, CHCl3, and TTHMs variation in SWSS in GKWSS.

Page 59

H2 Open Journal doi: 10.2166/h2oj.2018.007

Figure 9 | CHCl2Br, CHCl3, and TTHM variation in KSWSS based on THM levels in service reservoirs.

Figure 10 | CHCl2Br, CHCl3, and TTHM variation in SWSS in KSWSS.

THM values increased towards its boundary. Therefore, each SWSS has its own tendency of variation of THMs. The possible reasons could be the water stagnation and water demand. However, a slight decrease of THMs was observed in the samples which were away from the service reservoirs. Page 60

H2 Open Journal doi: 10.2166/h2oj.2018.007

Spatial variation of THMs in KSWSS

As shown in Figure 9, the spatial patterns of the THMs, TTHMs, CHCl2Br, and CHCl3 variations in KSWSS service reservoirs are somewhat similar. They clearly indicate a relation with the retention time on CHCl2Br, CHCl3, and TTHMs formation. However, the spatial variations of CHCl2Br, CHCl3, and TTHM SWSS are not distinct as in KSWSS (Figure 10). Each SWSS displays its own pattern. In order to understand the THMs variation in SWSS, variation maps for each subscheme were developed. According to the ﬁgure, no clear trend of variation of THMs is observed. In ML, THM levels are decreased toward the north boundary. However, in AG, THM concentrations are increased with the distance from the service reservoir. The ML scheme showed a slight decrease of THMs with the increasing distance from its service reservoir. However, each scheme displays its own pattern of spatial distribution of THMs. This could be due to multifactorial reasons such as water demand, TOC level, temperature, as well as residual chlorine concentrations on THM formation.

CONCLUSIONS For the ﬁrst time in Sri Lanka, variations of THM and TTHM formation as a function of treatment technology, residence time, pH, residual chlorine, conductivity, turbidity, and color were conducted. In both water schemes, only CHCl2Br and CHCl3 were detected which were below WHO and USEPA guidelines. When compared to conventional treatment, the water treated by pulsation technology produced lower THMs. The directional increase of THMs in SWSS is due to increased residence time of residual chlorine. The occurrence of THMs/TTHMs is dependent on temperature, water quality, water demand, and water distribution mechanics.

ACKNOWLEDGEMENTS National Research Council Sri Lanka supported this work (Grant No. NRC 12-116). National Water Supply and Drainage Board, Sri Lanka provided laboratory facilities. The contributions of Environmental Engineering Research laboratory, University of Peradeniya, Sri Lanka is also acknowledged.

REFERENCES Abdullah, M. P., Yew, C. H. & Ramli, M. S. 2003 Formation, modeling and validation of trihalomethanes (THM) in Malaysian drinking water: a case study in the districts of Tampin, Negeri Sembilan and Sabak Bernam, Selangor, Malaysia. Water Research 37(19), 4637–4644. Amjad, H., Hashmi, I., Rehman, M. S., Ali Awan, M., Ghaffar, S. & Khan, Z. 2013 Cancer and non-cancer risk assessment of trihalomethanes in urban drinking water supplies of Pakistan. Ecotoxicology and Environmental Safety 91, 25–31. APHA 2005 Standard Methods for the Examination of Water and Wastewater. American Public Health Association/American Water Works Association/Water Environment Federation, Washington DC. Ashbolt, N. J. 2004 Risk analysis of drinking water microbial contamination versus disinfection by-products (DBPs). Toxicology 198, 255–262. Baytak, D., Sofuoglu, A., Inal, F. & Sofuoglu, S. C. 2008 Seasonal variation in drinking water concentrations of disinfection byproducts in IZMIR and associated human health risks. Science of the Total Environment 407(1), 286–296. Clark, R. M. 1998 Chlorine demand and TTHM formation kinetics: a second-order model. Journal of Environmental Engineering 124(1), 16–24. Climate-Kandy 2017 Climate-data.org. Criquet, J., Rodriguez, E. M., Allard, S., Wellauer, S., Salhi, E., Joll, C. A. & von Gunten, U. 2015 Reaction of bromine and chlorine with phenolic compounds and natural organic matter extracts – electrophilic aromatic substitution and oxidation. Water Research 85, 476–486.

Page 61

H2 Open Journal doi: 10.2166/h2oj.2018.007

Danileviciute, A., Grazuleviciene, R., Vencloviene, J., Paulauskas, A. & Nieuwenhuijsen, M. J. 2012 Exposure to drinking water trihalomethanes and their association with low birth weight and small for gestational age in genetically susceptible women. International Journal of Environmental Research and Public Health 9(12), 4470–4485. El-Dib, M. A. & Ali, R. K. 1995 THMs formation during chlorination of raw Nile River water. Water Research 29(1), 375–378. Fooladvand, M., Ramavandi, B., Zandi, K. & Ardestani, M. 2011 Investigation of trihalomethanes formation potential in Karoon River water, Iran. Environmental Monitoring and Assessment 178(1–4), 63–71. Gang, D., Clevenger, T. & Banerji, K. 2003 Relationship of chlorine decay and THMs formation to NOM size. Journal of Hazard Material 96(1), 1–12. Gordon, S. M., Brinkman, M. C., Ashley, D. L., Blount, B. C., Lyu, C., Masters, J. & Singer, P. C. 2005 Changes in breath trihalomethane levels resulting from household water-use activities. Environmental Health Perspectives 114(4), 514–521. Gordon, B., Callan, P. & Vickers, C. 2008 WHO guidelines for drinking-water quality. WHO Chronicle 38(3), 564. Available at: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=4053895&tool=pmcentrez&rendertype=abstract\nhttp:// www.who.int/water_sanitation_health/dwq/fulltext.pdf\nhttp://whqlibdoc.who.int/publications/2011/ 9789241548151_eng.pdf?ua=1. Guanghui, H. & David, R. 2007 Characterization of disinfection by product precursors based on hydrophobicity and molecular size. Environmental Science and Technology 41(9), 3309–3315. Hasan, A., Thacker, N. P. & Bassin, J. 2010 Trihalomethane formation potential in treated water supplies in urban metro city. Environmental Monitoring and Assessment 168(1–4), 489–497. Heeb, M. B., Criquet, J., Zimmermann-Steffens, S. G. & von Gunten, U. 2014 Oxidative treatment of bromide-containing waters: formation of bromine and its reactions with inorganic and organic compounds – A critical review. Water Research 48(1), 15–42. Hladik, M. L., Focazio, M. J. & Engle, M. 2014 Discharges of produced waters from oil and gas extraction via wastewater treatment plants are sources of disinfection by-products to receiving streams. Science of the Total Environment 466–467, 1085–1083. Hu, J., Song, H. & Karanﬁl, T. 2010 Comparative analysis of halonitromethane and trihalomethane formation and speciation in drinking water: the effects of disinfectants, pH, bromide, and nitrite. Environmental Science and Technology 44(2), 794–799. Hua, G., Reckhow, D. A. & Kim, J. 2006 Effect of bromide and iodide ions on the formation and speciation of disinfection byproducts during chlorination. Environmental Science and Technology 40(9), 3050–3056. Hua, G., Kim, J. & Reckhow, D. A. 2014 Disinfection byproduct formation from lignin precursors. Water Research 63, 285–295. IARC 1991 International Agency for Research on Cancer Monographs on the Evaluation of Carcinogenic Risks to Humans, Vol. 53. Occupational exposures in insecticide application and some pesticides. International Agency for Research on Cancer, Lyon, p. 612. Available at: ,Go to ISI. ://BIOSIS:PREV199243035734. Jagatheesan, V., Weragoda, S. & Visvanathan, C. 2008 Rapid water quality characterization for chlorine demand and THM formation in drinking water. Journal of Water Supply: Research and Technology-AQUA 57(4), 259–272. Krasner, S. W. 1999 Chemistry of disinfection by product formation. In: Formation and Control of Disinfection By Products in Drinking Water (Singer, P. C., ed.). American Water Works Association, Denver, CO. Krasner, S. W. & Wright, J. M. 2005 The effect of boiling water on disinfection by-product exposure. Water Research 39, 855–864. Krasner, S. W., Weinberg, H. S., Richardson, S. D., Pastor, S. J., Chinn, R., Sclimenti, M. J., Onstad, G. D. & Thruston Jr., A. D. 2006 Occurrence of a new generation of disinfection byproducts. Environmental Science & Technology 40(23), 7175–7185. Kuivinen, J. & Johnsson, H. 1999 Determination of trihalomethanes and some chlorinated solvents in drinking water by headspace technique with capillary column gas-chromatography. Water Research 33(5), 1201–1208. Kumar, K. & Margerum, D. W. 1987 Kinetics and mechanism of general-acid-assisted oxidation of bromide by hypochlorite and hypochlorous acid. Inorganic Chemistry 26(16), 2706–2711. Li, X.-F. & Mitch, W. A. 2018 Drinking water disinfection byproducts (DBPs) and human health effects: multidisciplinary challenges and opportunities. Environmental Science & Technology 52(4), 1681–1689, p.acs.est.7b05440. Morris, R. D. & Levin, R. 1995 Estimating the incidence of waterborne infectious disease related to drinking water in the United States. IAHS Publications-Series of Proceedings 233, 75–88. Available at: http://hydrologie.org/redbooks/a233/ iahs_233_0075.pdf. Morris, D. P., Zagarese, H., Williamson, C. E., Balseiro, E. G., Hargreaves, B. R., Modenutti, B., Moeller, R. & Queimalinos, C. 1995 The attenuation of solar UV radiation in lakes and the role of dissolved organic carbon. Limnology and Oceanography 40(8), 1381–1391. Navalon, S., Alvaro, M. & Garcia, H. 2008 Carbohydrates as trihalomethanes precursors. Inﬂuence of pH and the presence of Cl(-) and Br(-) on trihalomethane formation potential. Water Research 42(14), 3990–4000. Nikolaou, A. D., Kostopoulou, M. N. & Lekkas, T. D. 1999 Organic by-products of drinking water chlorination. Global Nest: the International Journal 1(3), 143–156. Ozdemir, O. N. & Ger, A. M. 1998 Realistic numerical simulation of chlorine decay in pipes. Water Research 32(11), 3307– 3312. Panyapinyopol, B., Marhaba, T. F., Kanokkantapong, V. & Pavasant, P. 2005 Characterization of precursors to trihalomethanes formation in Bangkok source water. Journal of Hazardous Materials 120(1–3), 229–236. Platikanov, S., Martín, J. & Tauler, R. 2012 Linear and non-linear chemometric modeling of THM formation in Barcelona’s water treatment plant. Science of the Total Environment 432, 365–374.

Page 62

H2 Open Journal doi: 10.2166/h2oj.2018.007

Powell, J. C., Hallam, N. B., West, J. R., Forster, C. J. & Simms, J. 2000 Factors which control bulk chlorine decay rates. Water Research 34, 117–126. Ramavandi, B., Farjadfard, S., Ardjimand, M. & Dobaradaman, S. 2015 Effect of water quality and operational parameters on trihalomethanes formation potential in Dez River water, Iran. Water Resources and Industry 11, 1–12. Rook, J. J. 1974 Formation of haloforms during chlorination of natural waters. Water Treatment and Examination 23, 234–243. Rook, J. J. 1976 Haloforms in drinking water. Journal of the American Water Works Association 68(3), 168–172. Sadiq, R. & Rodriguez, M. J. 2004 Disinfection by-products (DBPs) in drinking water and predictive models for their occurrence: a review. Science of the Total Environment 321(1–3), 21–46. Singer, P. C. 1999 Humic substances as precursors for potentially harmful disinfection by-products. Water Science and Technology 40(9), 25–30. Wang, H., Ju, X., Wei, Y., Li, B., Zhao, L. & Hu, K. 2010 Simulation of bromide and nitrate leaching under heavy rainfall and high-intensity irrigation rates in North China Plain. Agricultural Water Management 97(10), 1646–1654. Wang, X., Zhang, H., Zhang, Y., Shi, Q., Wang, J., Yu, J. & Yang, M. 2017 New insights into trihalomethane and haloacetic acid formation potentials: correlation with the molecular composition of natural organic matter in source water. Environmental Science and Technology 51(4), 2015–2021. Williams, D. T., LeBel, G. L. & Benoit, F. M. 1997 Disinfection by-products in Canadian drinking water. Chemosphere 34(2), 299–316. Xu, T., Cui, C. & Ma, C. 2015 Color composition in a water reservoir and DBPs formation following coagulation and chlorination during its conventional water treatment in northeast of China. Desalination and Water Treatment 54(4–5), 1375–1384.

Page 63

Journal of

Hydroinformatics

ISSN 1464-7141 iwaponline.com/jh

Journal of Hydroinformatics is devoted to the application of information technology in the widest sense to problems of the aquatic environment. It promotes hydroinformatics as a cross-disciplinary field of study, combining technological, human-sociological and more general environmental interests, including an ethical perspective. The journal’s scope includes: • • • • • • • • • • • • • • • • • •

Physically-based simulation modelling Numerical methods Data-driven modelling and management Artificial neural networks Evolutionary methods Cellular automata Modelling systems Geographic information systems (GIS) and virtual imaging Ecology and water quality modelling Environmental impact assessment Knowledge engineering and management Socio-economic framework Intelligent decision support, negotiation and management Education and training Internet-based applications Optimisation and control Risk analysis, fuzzy logic and management of uncertainty Tools, environments and languages For more details, visit iwaponline.com/jh

Page 66

19.1

2017

Transient frequency response based leak detection in water supply pipeline systems with branched and looped junctions Huan-Feng Duan

ABSTRACT The transient frequency response (TFR) method has been widely developed and applied in the literature to identify and detect potential defects such as leakage and blockage in water supply pipe systems. This type of method was found to be efﬁcient, economic and non-intrusive for pipeline condition assessment and diagnosis, but its applications so far are mainly limited to single and simple pipeline systems. This paper aims to extend the TFR-based leak detection method to relatively

Huan-Feng Duan Department of Civil and Environmental Engineering, The Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong SAR, China E-mail: hf.duan@polyu.edu.hk

more complex pipeline connection situations. The branched and looped pipe junctions are ﬁrstly investigated for their inﬂuences to the system TFR, so that their effects can be characterized and separated from the effect of other components and potential leakage defects in the system. The leak-induced patterns of transient responses are derived analytically using the transfer matrix method for systems with different pipe junctions, which thereafter are used for the analysis of pipe leakage conditions in the system. The developed method is validated through different numerical experiments in this study. Based on the analytical analysis and numerical results, the applicability and accuracy as well as the limitations of the developed TFR-based leak detection method are discussed for practical applications in the paper. Key words

| leak detection, pipe junction, transfer matrix, transient frequency response, transient tests, water pipeline system

INTRODUCTION The problem of potential leaks in water supply pipelines has

the mean pipe ﬂows. Infrared thermography technique is

raised great interest for a long time to both academic

another common method and involves the use of infrared

researchers and practical engineers in this ﬁeld. Pipe leak-

imaging to analyze the ground temperature characteristics

age may cause waste for water and energy resources and

surrounding water pipes. Other common methods include

can also provide entry points for contaminants in urban

ﬂuoride testing and tracer gas analysis. While useful, these

water supply systems (Lee et al. ). Various leak detection

methods are limited to large leaks and can only work

methods have been developed in the past decades and

when the operator happens to be in the vicinity of the leak

widely used in urban water pipeline systems. The most

(Wang ; Lee ). Particularly, the fact that over 30%

common leak location technique is acoustic analysis. This

of portable water is lost from pipes around the world is a

method involves the use of a special listening device (i.e.

clear testimony that current methods are far from satisfac-

geophone) to listen to the sounds emanating from a pipeline.

tory (Duan et al. ).

Acoustic analysis relies on the fact that sound emanating

Recent research activities have intensiﬁed the transient-

from a leak has well-deﬁned characteristics, which enables

based leak detection methods that utilize the hydraulics of

leak-induced noise to be distinguished from the noise of

the transient ﬂows to detect leaks in the pipeline (e.g. Liggett

doi: 10.2166/hydro.2016.008

Page 67

H.-F. Duan

Journal of Hydroinformatics

Leak detection in pipeline systems with branched and looped junctions

19.1

2017

& Chen ; Brunone ; Vítkovský et al. ; Mpesha

effort in going around and isolating pipes is bewildering

et al. ; Wang et al. ; Lee et al. , ; Duan et al.

given that the total length of water supply lines in a

, , ). The tenet of this kind of method is that a

modern city attains to an order of 1,000 km or more (e.g.

pressure wave with appropriate bandwidth and amplitude

about 8,000 km in Hong Kong). Therefore, an extension of

is intentionally injected into the pipeline (Lee et al. ).

such transient-based methods to more realistic and complex

The system response (e.g. pressure head) is then measured

pipelines is urgently required and practically signiﬁcant to

at speciﬁed location(s) in the pipeline and analyzed for

reduce leakage in urban water supply systems.

leak detection (Duan et al. ). Such transient-based

Recently, few researchers in this ﬁeld have attempted

methods have become popular for the advantages of their

to extend the transient-based method to relatively more

fast speed, ability to work online and large operational

complex pipeline systems. Particularly, the TWR method

range (Colombo et al. ).

based on wavelet analysis has been applied to simple

A leak in a pipeline system results in an increased tran-

branched pipeline systems (e.g. Ferrante et al. ; Meni-

sient damping rate and the creation of new leak reﬂected

coni et al. ). The ITA method has been applied to small-

signals within the time traces (Tang et al. ; Duan et al.

scale real-life pipe networks (e.g. Soares et al. ). For the

detection

TFR method, Duan et al. () recently studied the possi-

methods have been developed by researchers and applied

bility of leak detection in relatively complex pipeline

to water piping systems relying on these two effects. The

systems which consist of multiple pipes in series. Both

developed leak detection methods vary greatly in their

the leak-induced and series-pipe-junctions induced transi-

modes of operation, but may be divided into four main cat-

ent effects were investigated analytically and numerically

egories according to their utilized transient information

in that study. Using the TFR-based method, an analytical

(Duan et al. ), namely: (1) transient wave reﬂection

expression was derived for the single leak-induced transi-

(TWR) based method, such as Brunone (), Brunone &

ent

Many

).

different

transient-based

leak

‘pattern’

series-pipeline

systems.

The

results

Ferrante (), Meniconi et al. (, ) and Covas et al.

conﬁrmed that the leak-induced transient behaviors could

(); (2) transient wave damping (TWD) based method

be separated from those by the connecting junctions of

by Wang et al. () and Nixon et al. (); (3) transient

series pipes as long as the original intact (leak-free) pipe

frequency response (TFR) based method by Mpesha et al.

system is well-deﬁned for its conﬁguration and boundaries

(), Ferrante & Brunone (), Covas et al. (), Lee

and the change extent of pipe diameters at junctions is not

et al. (), Sattar & Chaudhry (), Duan et al. (,

too large to violate the linear assumptions made in the

) and Ghazali et al. (); and (4) inverse transient

analytical derivation. In addition, the analysis indicated

analysis (ITA) based method studied in Liggett & Chen

that the pipe connecting junctions with different diameters

(), Vítkovský et al. (), Stephens (), Covas &

can cause the shifting of the system resonant frequencies

Ramos () and Soares et al. ().

but leaks do not, which gives the possibility of separating

While these different types of transient leak detection

the leak-induced effect from the junctions. This result was

methods have been proposed and applied to many simple

consistent with many experimental observations in pre-

pipe systems in the literature, it was found from many ﬁeld

vious works such as Ferrante & Brunone (), Lee

studies that these methods encountered difﬁculties in deal-

() and Brunone et al. (), and thereafter conﬁrmed

ing

in relevant studies by the author and his partners (e.g.

with

systems

with

complex

conﬁgurations

commonly seen in practical water pipeline systems (Ste-

Duan et al. ; Lee et al. ).

phens et al. ). Currently, the transient-based methods

Compared with other methods, the TFR method has the

have been largely applied to simple pipelines that could be

additional advantage of increased tolerance to system noises

isolated by valves from the rest of the network (Stephens

and ﬂow instabilities (Lee et al. , ; Duan et al. ).

et al. ; Lee ; Stephens ). Even then, the solution

However, only the cases of single and simple series pipelines

would probably fail if this pipeline happens to have continu-

are considered for the TFR-based method in previous

ous changes in diameters (non-uniform). In addition, the

studies; and for the cases of branched and looped pipelines

Page 68

H.-F. Duan

Journal of Hydroinformatics

Leak detection in pipeline systems with branched and looped junctions

that commonly exist in practical systems, an extension of this method is highly required in both method and application, which is the scope of this study. In this paper, the influences of typical pipe branched and looped junctions to the transient responses are firstly examined by numerical applications. The method and principles for TFR-based leak detection in branched and simple looped pipeline systems are then derived and developed, which are thereafter applied for different numerical cases. In the end, the results and findings of this study are analyzed and the limitations and future improvements of the developed method are discussed for practical applications in this field.

19.1

2017

result in the frequency domain becomes: 2 � �2 q cosh (μl) ¼4 h Y sinh (μl)

3 � �1 1 sinh (μl) 5 q , Y h cosh (μl)

(3)

or in a matrix form: O ¼ UI ,

(4)

where I, O, U ¼ input of transient information (e.g. the

upstream end), output of transient information (e.g. the downstream end), and the transfer matrix; q, h ¼ transient

discharge and pressure head in the frequency domain; l ¼ length of pipe section; the superscripts ‘1’ and ‘2’ represent quantities at the two ends/sides of the pipe section

MODELS AND METHODS

or system element under investigation respectively; μ and The one-dimensional (1D) waterhammer model and its equivalent form in the frequency domain based on the transfer matrix are used in this study, which are described in this section. The classic 1D waterhammer model is expressed as

Y ¼ propagation factor and impedance coefﬁcient, and: ω μ¼ a

rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi gAR 1�i ; ω

Y ¼�

a gA

rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi gAR 1þ , iω

(5)

follows (Chaudhry ; Wylie et al. ): gA @H @Q þ ¼ 0, a2 @t @x

in which ω ¼ frequency, i ¼ imaginary unit, R ¼ friction (1)

related coefﬁcient and R ¼ fQs =gDA2 with Qs being steady

(pre-transient) state discharge. Equations (3) or (4) are called the transfer matrix equation that represent the modi-

@Q @H f þ gA þ QjQj ¼ 0, @t @x 2DA

(2)

ﬁcation effect of the given element (e.g. pipeline, junction, and valve) on hydraulic responses from one end/side to the other. With this result, the frequency response of a

where H ¼ pressure head, Q ¼ pipe discharge, A ¼ pipe

cross-sectional area, D ¼ pipe diameter, a ¼ acoustic wave speed, t ¼ time, x ¼ spatial coordinate along pipeline, g ¼

gravitational acceleration, ρ ¼ ﬂuid density and f ¼ pipe fric-

tion factor. The method of characteristics is applied to solve

whole transient pipe system can then be obtained by multiplying the relevant transfer matrices of all the system elements in the order of connections (Lee ; Duan et al. ). This method is used later in this study for deriving the TFR results of the branched and looped pipe systems.

the waterhammer model (Chaudhry ). Note that only steady friction effect is considered in the analytical derivation and the unsteady friction effect will be included

TRANSIENT INFLUENCES OF PIPE JUNCTIONS

and validated in the numerical simulations. The frequency domain equivalents of the 1D mass and

Prior to developing the detection methods for relatively com-

momentum equations in Equations (1) and (2) above can

plex pipeline systems, it is necessary to understand and

be obtained by applying the linear transfer matrix method

investigate the impacts of different pipe connecting junctions

for describing the transient system behaviors in the fre-

on the transient responses. For illustration, three test cases of

quency domain (Chaudhry ; Lee et al. ; Duan

systems with single and uniform pipeline (without junction)

et al. ). After linearization and transformation, the

and multiple pipes with simple branched and looped Page 69

H.-F. Duan

Leak detection in pipeline systems with branched and looped junctions

Journal of Hydroinformatics

19.1

2017

junctions shown in Figure 1(a)–1(c) respectively are used

large bandwidth of wave injection for transient system

herein for comparative study in both the time and frequency

analysis (e.g. defect detection), the transients in all test

domains (denoted as systems no. 1, no. 2, no. 3 in this

cases are generated by the side-discharge valve with oper-

study). Speciﬁcally, the main pipelines for these three systems

ations of fast closure-open-closure as given in previous

(i.e. from node a to node b) are assumed to be the same so as to

studies (e.g. Duan et al. , , ; Lee et al. ).

fairly analyze the impacts of junctions on the system transient

The numerical results of transient pressure traces collected

responses through result comparisons. The details of system

at the just upstream of the inline valve are used for analysis.

settings and parameters are provided in Figure 1. In each test system in Figure 1, the side-discharge valve at the downstream (V2 in the ﬁgure) is used for generating

Time domain transient responses

transients and the inline valve (V1 in the ﬁgure) is used for controlling the initial steady state discharge (Qs) in the

The obtained transient pressure head responses in the time

system. For simplicity of analysis and to highlight the transi-

domain are shown in Figure 2(a) for the three systems. For

ent behaviors (separated from steady state), initially both

comparison, the axial coordinate of the ﬁgure is dimension-

valves (V1 and V2) are fully closed (i.e. Qs ¼ 0). That is, the

less time with regard to wave period of single pipeline case

transient ﬂows are generated on the basis of initial static

(i.e. 4L0/a0), and the vertical coordinate is normalized by

ﬂow condition. The effect of initial non-static ﬂow con-

the ﬁrst peak amplitude of transient head at side-discharge

ditions will be included in the analytical and numerical

valve (i.e. Joukowsky head, a0ΔVd/g with ΔVd being the

analyses later in this study. In order to provide a preferably

velocity change through the valve operation).

Figure 1

Page 70

Sketch for test pipeline systems: (a) no. 1: single and uniform pipeline system; (b) no. 2: branched pipeline system; (c) no. 3: looped pipeline system.

H.-F. Duan

Figure 2

Journal of Hydroinformatics

Leak detection in pipeline systems with branched and looped junctions

19.1

2017

Results of test pipeline systems with/without pipe junctions in: (a) the time domain results; (b) the frequency domain.

The results in Figure 2(a) clearly show the differences of

it is very difﬁcult to clearly characterize the transient wave

the transient wave traces for the pipeline systems with/with-

behaviors in the time domain for such relatively complex

out pipe junctions. Particularly, more frequent reﬂections

pipeline systems. Meanwhile, it has been demonstrated that

are caused by the junctions, which results in complex (e.g.

this selected type of injected signal with relatively large band-

non-monotonic) wave amplitude envelope attenuations with

width (high frequencies) could provide more accurate results

time. Moreover, different pipe junctions (e.g. the simple

of leak detection in the pipeline (Lee et al. ). Therefore,

branched and looped junctions here) may induce different

current transient-based time domain methods (i.e. TWR and

extent and frequency of wave reﬂections from the result com-

TWD), which depend mainly on the wave reﬂection and

parison of systems no. 2 and no. 3 in Figure 2(a). In this regard,

damping information, may become inapplicable or inaccurate Page 71

H.-F. Duan

Journal of Hydroinformatics

Leak detection in pipeline systems with branched and looped junctions

19.1

2017

for using this preferable signal injection with relatively large

TFRs under intact (leak-free) and leakage conditions. That

bandwidth for the leak detection in complex pipeline systems.

is, the leak-induced patterns are required to be explored

This result has also been conﬁrmed in the previous study for

and derived for the TFRs of pipeline systems with different

series-pipeline systems in Duan et al. (). Based on these

pipe connecting junctions (Duan et al. ). Two typical

ﬁndings here and from previous studies, the frequency

junctions of three-pipe branch and simple two-pipe loop

domain transient response is examined in the following

shown in Figure 1(b) and 1(c) are considered in this study.

study, with its features used for characterizing and diagnosing

For simplicity and illustration, only the single leakage situ-

relatively complex pipe systems.

ation is considered in this study, and for multiple leaks, the similar derivation and analysis procedure can be extended and applied. The main results of TFR for these

Frequency domain transient responses

two cases of branched and looped pipe systems are summarThe TFRs can be obtained from the Fourier transform of the

ized in this section, with the derivation details presented in

time domain traces in Figure 2(a), and the results of the

the appendix (available with the online version of this paper).

three systems are shown in Figure 2(b) for analysis. As

For the intact case of branched pipeline system shown in

expressed in Figure 2(a), the axial and vertical coordinates

Figure 1(b), the following resonant condition is obtained by

of Figure 2(b) are non-dimensionalized by the fundamental

the transfer matrix method as given in Equation (A10) in the

frequency

(a0/4L0)

and

the

ﬁrst

peak

amplitude

(Max_ΔH0) of single pipeline case respectively. As indicated similarly from the time domain results in Figure 2(a), obvious differences between the results of pipe systems with and without pipe junctions are observed in the frequency domain. With the existence of different pipe junc-

appendix: 2

3 Y3 Y2 sin (μ3 l3 ) cos (μ2 l2 ) cos (μ1 l1 ) 4 �Y3 Y1 sin (μ3 l3 ) sin (μ2 l2 ) sin (μ1 l1 ) 5 ¼ 0, þY2 Y1 cos (μ3 l3 ) cos (μ2 l2 ) sin (μ1 l1 )

(6)

tions, both the resonant frequency shifts and amplitude

where subscript numbers are pipe numbers described in

changes of the TFRs are caused with different extents by

Figure 1(b). This result has been validated and used in pre-

these two junctions. This result is consistent with various

vious studies by the author for dead-end side branch

numerical and experimental observations in the previous

detection (e.g. Duan & Lee ). Under single pipe leakage

studies (e.g. Brunone et al. ; Duan et al. , ; Duan

condition, after mathematical manipulations and essential

& Lee ). However, compared to time domain results, the

simpliﬁcations, a general form of the converted transient

inﬂuences of pipe junctions to the TFRs become relatively

pressure response in the frequency domain can be obtained

simple and independent for different resonant peaks, which

as (see Equations (A14)–(A16) in the appendix):

have similar impact complexities that are not superimposed or accumulated with frequency. From this perspective, it might be easier to use the frequency domain results for characterizing the inﬂuences of pipe junctions to the transient system

^ B ¼ KL �1 � cos (2μ xLn þ φB )�, h n Ln n CnB

(7)

responses than the time domain results. Consequently, the

^ where h Ln is the converted TFR based on the difference

TFR-based method is adopted as the investigation tool for

between the intact and leakage situations; n is the number

the development of leak detection method in the typical branched and simple looped pipeline systems in this study.

of pipe that the potential leakage is located (n ¼ 1, 2, 3 in

this study); xLn is the distance of leakage location from the upstream end of the pipeline n; KL is the impendence factor for describing the leakage size; the subscript L is

TFR RESULTS FOR DIFFERENT PIPE JUNCTIONS

used for quantity for leaking pipe system; the superscript B indicates the quantity for branched pipeline system, and C,

To develop the leak detection method, it is necessary to

φ are intact system based known coefﬁcients with their

understand and characterize the difference of the system

expressions provided in the appendix. The result of

Page 72

H.-F. Duan

Journal of Hydroinformatics

Leak detection in pipeline systems with branched and looped junctions

Equation (7) indicates that the leak-induced pattern for

19.1

2017

TFR-BASED LEAK DETECTION

TFRs is dependent on the system conﬁguration as well as the location of the leaking pipe section in the system. More-

It is known from Equation (7) or (8) that the leak-induced

over, for given branched pipeline system, the leak-induced

pattern is dependent on the potential leaking pipe location

pattern relies only on the potential leak information

(pipe number) in the above-mentioned branched or looped

(location and size), which therefore can be used inversely

pipeline system, which is different from the result of single

to identify and detect pipe leakage in the system.

or simple series-pipeline system (e.g. Duan et al. ). There-

Similarly, for the simple looped pipeline system in

fore, a traversal calculation and comparison of all the

Figure 1(c), the leak-induced patterns for different leaking

possible leak-induced patterns and leak detection processes

conditions can be derived and expressed as follows (see

is required for evaluating such relatively complex pipe sys-

Equations (B11) and (B12) in the appendix):

tems to ﬁnd out the most likely or optimal results of the pipe leakage information in the system. For the simple

^O h Ln

� � q�ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi � � � � � KL O O 2 þ T O 2 sin μ l � 2μ x ¼ O RO þ S þ φ , n n n Ln n n n n Cn

(8)

branched and looped pipeline systems focused in this study (e.g. the total number of pipes is less than 6), an enumeration method is used for such calculation and comparison. To obtain accurate and globally optimal results

where the superscript O indicates the quantities obtained for

for each leak-induced pattern analysis, the GA-based optim-

the looped pipeline system; the expressions of known coefﬁ-

ization procedure developed in Duan & Lee () is used

cients C, R, S, T, φ are given in the appendix. Therefore, there

here for the inverse analysis of Equation (7) or (8). The

are four possible leak-induced patterns in the system of

detailed formulation and steps for applying this GA-based

Figure 1(c) for analyzing the leak information by using

method in water pipeline systems refer to Duan & Lee

Equation (8). Again, these leak-induced patterns are only

(). Figure 3 shows the main application principle and

dependent on the leak information for the speciﬁed

procedure of the proposed TFR-based leak detection

looped pipeline system. The detailed principle and pro-

method in this study.

cedures of applying Equations (7) and (8) for leak detection are stated in the following section.

Figure 3

It is also noted that, in this proposed method and procedure, the potential leakage information is identiﬁed

Flowchart of TFR-based leak detection.

Page 73

H.-F. Duan

Journal of Hydroinformatics

Leak detection in pipeline systems with branched and looped junctions

19.1

2017

through the ﬁtness comparison of different leak-induced pat-

also listed in the table for reference. The system transient

terns in the given pipeline system. Therefore, as in other

responses are obtained by the 1D numerical simulations in

transient-based method for pipe defects detection (e.g.

the time domain (i.e. Equations (1) and (2)). The transient

Duan et al. , , ; Lee et al. , ), the applica-

pressure head at the just upstream of the inline valve are col-

bility and accuracy of this method may be affected by the

lected and then converted by Fourier transform into the

model bias/errors (e.g. linear approximation and turbu-

frequency domain for the analysis. The results of leakage

lence) and system uncertainties (e.g. input and output

detection based on the proposed method and procedure in

measurements). The accuracy and limitations of this

this study are obtained and listed in Table 2. The accuracy

method are discussed through the applications later in the

of the method is evaluated by the difference between the

paper.

real and predicted values of the leakage information, which is deﬁned as the relative error (ε) by:

NUMERICAL VALIDATIONS AND RESULTS ANALYSIS

ε(%) ¼

predictedvalue � realvalue × 100: realvalue

(9)

The system conﬁgurations in Figure 1(b) and 1(c) are ﬁrstly

Based on Equation (9), the prediction errors for the test

used for numerical validations of the proposed TFR-based

cases are also given in Table 2. The results demonstrate the

leak detection method, with the system parameter settings

validity and accuracy of the proposed method for the leak

and information given in Table 1. Different leakage cases

detection (location and size) in the simple branched and

(location and size) are considered for each test system and

looped pipeline systems considered in this study. Speciﬁ-

shown in Table 2, with tests no. 1 to no. 3 for the branched

cally, the maximum relative errors of the prediction are 13

pipe system and tests no. 4 to no. 7 for the simple looped

and 28% respectively for locating and sizing the leakage.

pipe system. For clarity, the relative leak effective area,

That is, this proposed method is more accurate to locate

AL * ¼ CdAL/Ap with CdAL being leaking area and Ap the

the pipe leakage than to size the leakage, which is similar

cross-sectional area of leaking pipe, for each test case is

Table 1

Settings and information of test pipeline systems

System

Pipe length (m)

No. 2 (branched) No. 3 (looped)

Table 2

with the results applied for single and series pipeline systems

l1 ¼ 500, l2¼240; l3 ¼ 200

l1 ¼ 500, l2¼300; l3 ¼ 200; l4 ¼ 350

Pipe size (mm)

Wave speed (m/s)

Pipe friction

D1 ¼ 500, D2¼300; D3 ¼ 60

a1 ¼ 1,000, a2¼1,100; a3 ¼ 1,200

f1 ¼ f2 ¼ f3¼0.01

D1 ¼ 500, D2¼400; D3 ¼ 500; D4 ¼ 200

a1 ¼ 1,000, a2¼1,100; a3 ¼ 1,000; a4 ¼ 1,200

Leakage detection results for branched and looped systems Real leakage information

Results of leakage detection

System

Test no.

xLn (m)

KL (10–4 m2/s)

AL* (10

Branched pipeline system

1 2 3

1.0 3.0 0.2

Looped pipeline system

4 5 6 7

150 (n ¼ 1) 100 (n ¼ 2) 160 (n ¼ 3)

3.0 1.0 4.0 0.8

Page 74

f1 ¼ f2 ¼ f3¼f4¼0.01

300 (n ¼ 1) 120 (n ¼ 2) 150 (n ¼ 3) 100 (n ¼ 4)

xPLn (m)

ϵ (%)

KPL (10–4 m2/s)

ϵ (%)

1.6 13.6 22.6

146 101 167

–2.7 1.0 4.4

0.83 2.84 0.19

–17 –5.3 –5.0

4.9 2.5 6.5 8.1

281 124 169 113

–6.3 3.3 12.7 13

2.68 0.97 3.69 0.58

–10.7 –3 –7.8 27.5

�3

)

H.-F. Duan

Leak detection in pipeline systems with branched and looped junctions

Journal of Hydroinformatics

19.1

2017

(Lee et al. ; Duan et al. ). This is mainly because of

prediction of the leak locations in Table 2. However, the

the linear approximations made for the derivations, which is

results also reveal overall that the analytical result of Equation

discussed later in this study.

(7) or (8) has underestimated the amplitudes of the leak-

To further demonstrate the detection process and results,

induced patterns due to the simpliﬁcations of the nonlinear

the leak-induced patterns of tests no. 1 and no. 4 from the

effects of friction term during the derivations, which also

numerical simulations by 1D models and theoretical predic-

results in the relatively large and negative errors of the leak

tion by Equation (7) or (8) are plotted in Figure 4 for

size prediction in Table 2. In this regard, the inclusion of non-

comparison. Both the results in Table 2 and Figure 4 indicate

linearities of transient effects in the system (e.g. friction or

the good agreements of the phase changes between the leak-

turbulence or wave-structure interactions) is required to

induced patterns by numerical simulations and analytical

improve the accuracy of the leak detection results for the pro-

analysis, which results in the relatively small errors in the

posed method. This aspect may become the next-step work in

Figure 4

Leak-induced patterns of system TFR results for: (a) test no. 1; (b) test no. 4.

Page 75

Figure 5

H.-F. Duan

Journal of Hydroinformatics

Leak detection in pipeline systems with branched and looped junctions

19.1

2017

Test pipeline system with two branched junctions.

the future for the improvement of the TFR-based defect detec-

Table 3

tion method.

Leakage detection results for the system with two branched pipe junctions

Real leakage information KL

AL*

xPLn

(10–5 m2/s)

(10–3)

(m)

8.13

346 15.3 1.36

–32.0

150 (n ¼ 2) 1.2

1.22

141 –6.0 0.78

–35.0

2.26

216 2.9

–14.0

140 (n ¼ 4) 5.0

12.20 157 12.1 4.65

–7.0

2.03

–17.0

Test

FURTHER APPLICATION AND DISCUSSION

no.

xLn (m)

300 (n ¼ 1) 2.0

The application results and analysis above have validated

and conﬁrmed the applicability and accuracy of the pro-

posed method and application procedure for pipe leak

detection in the single branched and simple looped pipeline

Results of leakage detection

210 (n ¼ 3) 0.5 80 (n ¼ 5)

3.0

KPL ϵ (%)

5.0

(10–5 m2/s)

0.43 2.49

ϵ (%)

systems considered in this study. These successful validations provide the possibility of the extension of the TFR-based method for leak detection to relatively more

The TFR-based leak detection results by the proposed

complex pipe systems consisting of multiple branched and

method and procedure in this study are shown in Table 3

looped junctions. From this perspective, and based on the

and the obtained leak-induced patterns for tests no. 8 and

similar procedures of this study, the TFR results can also

no. 10 are plotted in Figure 6, which demonstrate again the

be derived and applied for such pipeline systems with mul-

applicability and accuracy of the TFR-based method for iden-

tiple junctions (branched and looped), which actually

tifying and detecting pipe leakage in relatively more complex

results in a similar form of leak-induced patterns given in

pipe systems with multiple pipe branches. Compared to the

this study, but with different expressions of the known-

single branched pipe system in Figure 1(b), the detection accu-

system based coefﬁcients (e.g. C, R, S, T, and φ). For demon-

racy of the TFR-based method becomes decreased with the

stration in this study, a typical pipeline system with two

increase of the connection complexities of the system. How-

branched pipe junctions shown in Figure 5 is adopted for

ever, the relative errors are still within 16 and 35% for

investigation. The information of system conﬁgurations

leakage location and size respectively, which may also pro-

and parameters are plotted in Figure 5, with different leak-

vide useful information and signiﬁcant implications for the

age test cases (no. 8 to no. 12) listed in Table 3.

pipe leakage detection and diagnosis in practice. From this

Page 76

H.-F. Duan

Leak detection in pipeline systems with branched and looped junctions

Journal of Hydroinformatics

19.1

2017

point of view, the TFR-based leak detection method is extend-

the obvious increase of the detection errors with system

able and applicable to relatively more complex pipeline

complexities (e.g. number of junctions), especially for pre-

systems with multiple branches and simple loops, as long as

dicting the leakage size. This result and trend may be

the pipe system under investigation has been pre-deﬁned for

attributed to the assumptions and simpliﬁcations made for

the topological conﬁgurations and the system properties and

the method development as follows:

operation parameters are well known for the analysts under the original and intact conditions (before the occurrence of leakage). While the successful applications of the developed TFRbased method for leakage detection in pipeline systems with

1. Linearization of steady friction term, which requires the relatively small transient ﬂow perturbation to the steady state discharge (Duan et al. ; Lee et al. ).

single and multiple branched junctions and simple looped

2. Neglect of unsteady friction effect, which is frequency

junctions respectively, the application results also reveal

dependent and could be included in the developed

Figure 6

Leak-induced patterns of system TFR results for: (a) test no. 8; (b) test no. 10.

Page 77

H.-F. Duan

Journal of Hydroinformatics

Leak detection in pipeline systems with branched and looped junctions

method by considering the simpliﬁed form given in Lee et al. () and Duan et al. (). 3. Assumption of relatively small leakage capacity to main pipeline discharge, so that the linearized oriﬁce equation (as indicated by KL) can be applied to simulate the leakage effect (Lee et al. , ).

19.1

2017

(c) Robustness of the inverse analysis algorithm for obtaining optimal and physical solutions of the derived leakinduced patterns, especially for the applications of large-scale and complex pipe systems. Finally, it is important to point out that only the numerical applications are conducted in this paper for the

Meanwhile, different system inﬂuence factors may also

preliminary validations of the developed TFR-based leak

contribute to the discrepancies of leakage prediction results,

detection method. In the future work, further experimental

including the following: 1. Errors of data collections and treatment: such as the sample frequency of the time-domain data; trace cutting length of time-domain data (e.g. number of wave periods for analysis); and the discrete Fourier transform for

tests (laboratory and ﬁeld) are required and designed to validate and verify the accuracy, tolerance and sensitivity of this developed method for practical cases under the inﬂuences of inevitable noises and uncertainties in practical water pipe systems.

frequency data analysis. 2. Inaccuracy of the inverse analysis process (e.g. GA-based optimization in this study): such as the convergence and

SUMMARY AND CONCLUSIONS

error of inverse analysis process; and the non-uniqueness of solutions to the leak-induced patterns for complex pipe

This paper investigates the possibility of the application of

systems.

the TFR-based leak detection method in pipeline systems

3. Uncertainties and complexities of initial and boundary

with different pipe junctions. The systems of simple

conditions in practical pipeline systems: such as the

branched and looped pipeline systems are considered and

external noises and instabilities in water piping systems;

investigated in this study. The inﬂuence of different pipe

and the complex interactions of transient wave, ﬂow tur-

junctions to system transient responses (TFRs) is ﬁrstly

bulence and system components (e.g. junctions and

examined by numerical simulations in the time and fre-

devices).

quency domains, which highlights the merits of using the

With these limitations and inﬂuence factors, it is necessary to improve the transient model and methods for the accurate extension and application of the developed TFRbased leak detection in practical situations, for example, through the following aspects:

frequency domain responses for characterizing the transient system behaviors. The system TFRs are then derived by the linear transfer matrix method for both the pipe systems with single branch and loop connections, which are then used for the detection of pipe leakage information in this study.

(a) Improvement of 1D transient models (in time and fre-

The analytical results indicate that both the typical

quency domains) to accurately represent the physics

branched and looped pipe junctions may have great inﬂu-

and process of transient pipe ﬂows in complex pipe sys-

ences to the system TFRs but have little impacts on the

tems such as unsteady friction and turbulence, wave-

leak-induced patterns. The GA-based optimization is then

junction interaction, and wave-leak interaction.

proposed for solving the analytically derived leak-induced

(b) Selection of optimal injected transient signals to capture

patterns to obtain the leakage information in the system.

the full picture of the leakage characteristics, for

The developed TFR-based method and application pro-

example appropriate bandwidth of signals as suggested

cedure are validated through different numerical tests for

in Duan et al. (, ); and meanwhile, multiple

pipe systems with single branched, single looped and two

signal injections and response collections may also be

branched pipe junctions respectively. The results demon-

helpful to improve the accuracy of the proposed

strate the applicability and accuracy of the developed

method (Lee et al. ).

method for leakage identiﬁcation and detection in these

Page 78

H.-F. Duan

Leak detection in pipeline systems with branched and looped junctions

multiple-pipeline systems. However, the results also imply that this method is more accurate to locate the pipe leakage than to size the leakage in these applications. The results analysis and discussion of this study provide the evidences and conﬁrmations for the extension of the TFR-based method to pipe systems with different connection junctions. It is also noted that extensive experimental tests (laboratory and ﬁeld) are demanded for further validating the accuracy and sensitivity of the proposed method in practical

applications.

Furthermore,

the

feasibility

and

applicability of the TFR-based method for practical water distribution networks still need more investigations in future work.

ACKNOWLEDGEMENTS This paper was supported by research grants from: (1) the Hong Kong Polytechnic University (HKPU) under projects with numbers 1-ZVCD, 1-ZVGF, 3-RBAB and G-YBHR; and (2) the Hong Kong Research Grant Council (RGC) under project numbers 25200616 and T21-602/15-R. The author would like to thank Mr T. C. Che for his kind help in plotting the ﬁgures in this paper.

REFERENCES Brunone, B.  Transient test-based technique for leak detection in outfall pipes. J. Water Resour. Plan. Manage. 125 (5), 302– 306. Brunone, B. & Ferrante, M.  Detecting leaks in pressurized pipes by means of transients. J. Hydraul. Res. 39 (4), 1–9. Brunone, B., Ferrante, M. & Meniconi, S.  Discussion of ‘Detection of partial blockage in single pipelines’ by Mohapatra et al. J. Hydraul. Eng. 134 (6), 872–874. Chaudhry, M.  Applied Hydraulic Transients. Van Nostrand Reinhold, New York. Colombo, A. F., Lee, P. J. & Karney, B. W.  A selective literature review of transient-based leak detection methods. J. Hydro-Environ. Res. 2 (4), 212–227. Covas, D. & Ramos, H.  Case studies of leak detection and location in water pipe systems by inverse transient analysis. J. Water Resour. Plan. Manage. 136 (2), 248–257. Covas, D., Ramos, H., Graham, N. & Maksimovic, C.  Application of hydraulic transients for leak detection in water supply systems. Water Sci. Technol. Water Supply 4 (5–6), 365–374.

Journal of Hydroinformatics

19.1

2017

Covas, D., Ramos, H. & Almeida, A. B.  Standing wave difference method for leak detection in pipeline systems. J. Hydraul. Eng. 131 (12), 1106–1116. Duan, H. F. & Lee, P. J.  Transient-based frequency domain method for dead-end side branch detection in reservoirpipeline-valve systems. J. Hydraul. Eng. 142 (2), 04015042. Duan, H. F., Lee, P. J., Ghidaoui, M. S. & Tung, Y. K.  Essential system response information for transient-based leak detection methods. J. Hydraul. Res. 48 (5), 650–657. Duan, H. F., Lee, P. J., Ghidaoui, M. S. & Tung, Y. K.  Leak detection in complex series pipelines by using system frequency response method. J. Hydraul. Res. 49 (2), 213–221. Duan, H. F., Lee, P. J., Ghidaoui, M. S. & Tung, Y. K.  System response function based leak detection in viscoelastic pipeline. J. Hydraul. Eng. 138 (2), 143–153. Duan, H. F., Lee, P. J., Ghidaoui, M. S. & Tuck, J.  Transient wave-blockage interaction and extended blockage detection in pressurized pipes. J. Fluids Struct. 46 (2014), 2–16. Ferrante, M. & Brunone, B.  Pipe system diagnosis and leak detection by unsteady-state tests-1: harmonic analysis. Adv. Water Resour. 26 (1), 95–105. Ferrante, M., Brunone, B. & Meniconi, S.  Leak detection in branched pipe systems coupling wavelet analysis and a Lagrangian model. J. Water Supply Res. Technol. AQUA 58 (2), 95–106. Ghazali, M., Beck, S., Shucksmith, J., Boxall, J. & Staszewski, W.  Comparative study of instantaneous frequency based methods for leak detection in pipeline networks. Mech. Syst. Signal Process. 29, 187–200. Lee, P. J.  Using System Response Functions of Liquid Pipelines for Leak and Blockage Detection. PhD thesis, The University of Adelaide, Adelaide, SA, Australia. Lee, P. J., Vítkovský, J. P., Lambert, M. F., Simpson, A. R. & Liggett, J. A.  Frequency domain analysis for detecting pipeline leaks. J. Hydraul. Eng. 131 (7), 596–604. Lee, P. J., Lambert, M. F., Simpson, A. R., Vítkovský, J. P. & Liggett, J.  Experimental verification of the frequency response method for pipeline leak detection. J. Hydraul. Res. 44 (5), 693–707. Lee, P. J., Duan, H. F., Ghidaoui, M. S. & Karney, B. W.  Frequency domain analysis of pipe fluid transient behaviors. J. Hydraul. Res. 51 (6), 609–622. Lee, P. J., Duan, H. F., Tuck, J. & Ghidaoui, M. S.  Numerical and experimental study on the effect of signal bandwidth on pipe assessment using fluid transients. J. Hydraul. Eng. 141 (2), 04014074. Liggett, J. A. & Chen, L. C.  Inverse transient analysis in pipe networks. J. Hydraul. Eng. 120 (8), 934–954. Meniconi, S., Brunone, B., Ferrante, M. & Massari, C.  Potential of transient tests to diagnose real supply pipe systems: what can be done with a single extemporary test. J. Water Resour. Plan. Manage. 137 (2), 238–241. Meniconi, S., Brunone, B., Ferrante, M., Capponi, C., Carrettini, C. A., Chiesa, C., Segalini, D. & Lanfranchi, E. A.  Anomaly pre-localization in distribution-transmission mains

Page 79

H.-F. Duan

Leak detection in pipeline systems with branched and looped junctions

by pump trip: preliminary ﬁeld tests in the Milan pipe system. J. Hydroinform. 17 (3), 377–389. Mpesha, W., Gassman, S. L. & Chaudhry, M. H.  Leak detection in pipes by frequency response method. J. Hydraul. Eng. 127 (2), 134–147. Nixon, W., Ghidaoui, M. S. & Kolyshkin, A. A.  Range of validity of the transient damping leakage detection method. J. Hydraul. Eng. 132 (9), 944–957. Sattar, A. M. & Chaudhry, M. H.  Leak detection in pipelines by frequency response method. J. Hydraul. Res. 46 (EI1), 138–151. Soares, A. K., Covas, D. I. C. & Reis, L. F. R.  Leak detection by inverse transient analysis in an experimental PVC pipe system. J. Hydroinform. 13 (2), 153–166. Stephens, M.  Transient Response Analysis for Fault Detection and Pipeline Wall Condition Assessment in Field Water Transmission and Distribution Pipelines. PhD thesis, University of Adelaide, Australia. Stephens, M., Lambert, M., Simpson, A., Vitkovsky, J. & Nixon, J.  Field tests for leakage, air pocket, and discrete blockage detection using inverse transient analysis in water distribution pipes. In: World Water and Environmental

Journal of Hydroinformatics

19.1

2017

Resources Congress, ASCE, June 27–July 1, 2004, Salt Lake City, Utah, USA. Stephens, M., Lambert, M., Simpson, A. & Vitkovsky, J.  Calibrating the water hammer response of a ﬁeld pipe network by using a mechanical damping model. J. Hydraul. Eng. 137 (10), 1225–1237. Tang, K. W., Brunone, B., Karney, B. & Rossetti, A.  Role and characterization of leaks under transient conditions. In: Building Partnerships-Proc. ASCE Joint Conf. Water Resource Engineering and Management, Minneapolis, MN, pp. 7–30. Vítkovský, J. P., Simpson, A. R. & Lambert, M. F.  Leak detection and calibration using transients and genetic algorithms. J. Water Resour. Plan. Manage. 126 (4), 262–265. Wang, X. J.  Leakage and Blockage Detection in Pipelines and Pipe Network Systems Using Fluid Transients. PhD thesis. The University of Adelaide, Adelaide, Australia. Wang, X. J., Lambert, M. F., Simpson, A. R., Liggett, J. A. & Vítkovský, J. P.  Leak detection in pipeline systems using the damping of ﬂuid transients. J. Hydraul. Eng. 128 (7), 697–711. Wylie, E. B., Streeter, V. L. & Suo, L. S.  Fluid Transients in Systems. Prentice-Hall, Englewood Cliffs, New Jersey.

First received 11 January 2016; accepted in revised form 18 July 2016. Available online 23 August 2016

Page 80

173

19.2

2017

Distribution of mean ﬂow and turbulence statistics in plunge pools Luis G. Castillo, José M. Carrillo and Fabián A. Bombardelli

ABSTRACT When the capacity of the spillway of a dam is exceeded for a given flood, overtopping occurs; in such cases potentially dangerous hydrodynamic actions and scour downstream of the dam need to be foreseen. Detailed studies of jets impinging in plunge pools from overflow nappe flows are scarce. This work addresses plunge pool flows, and compares numerical results against our own experiments. The energy dissipation is larger than 75% of the impingement jet energy. Instantaneous velocities and air entrainment were obtained with the use of an Acoustic Doppler Velocimeter and optical fibre probe, respectively. Mean velocity field and turbulence kinetic energy profiles were determined. To identify the level of reliability of models, numerical simulations were carried out by using the ‘homogeneous’ model of ANSYS CFX, together with different turbulence closures. The numerical results fall fairly close to the values measured in the laboratory, and with expressions for

Luis G. Castillo (corresponding author) José M. Carrillo Department of Civil Engineering, Universidad Politécnica de Cartagena, UPCT Paseo Alfonso XIII, 52, Cartagena 30203, Spain E-mail: luis.castillo@upct.es Fabián A. Bombardelli Department of Civil and Environmental Engineering, University of California, Davis, 2001 Ghausi Hall, One Shields Ave., Davis, CA 95616, USA

submerged hydraulic jumps and horizontal wall jets. The observations can be well predicted for characterized proﬁles at a minimum distance of 0.40 m downstream from the stagnation point, horizontal velocities greater than 40% of the maximum velocity in each proﬁle, and when the ratio of the water cushion depth to the jet thickness is lower than 20. Key words

| air entrainment jet, computational ﬂuid dynamics (CFD), energy dissipation, impingement jets, plunge pools, velocity proﬁle

INTRODUCTION There is growing consensus regarding the fact that climate

When the rectangular jet or nappe ﬂow occurs due to

change will lead to enhanced extreme ﬂooding in certain

overtopping, the design considerations need to ensure that

areas of the world. This situation must be confronted with

most energy is dissipated, and that there is minimal to no

spillway capacity and special operational scenarios for

erosion downstream of the dam. In other words, we need

large dams. If the capacity of the spillway is insufﬁcient,

to estimate the hydrodynamic actions on the bottom of the

the dam might be overtopped, thus generating new loading

basin where the jet discharges, as a function of the charac-

scenarios, and raising questions about potential risky hydro-

teristics of the jet (Annandale ).

dynamic actions and scour downstream of the dam (Wahl et al. ).

The energy dissipation mechanisms that occur in the jetbasin structure can be grouped into the following: (a) aera-

Spain is the fourth country in the world according to the

tion and disintegration of the jet in its fall, (b) air

number of large dams – it has over 1,200. Fifty percent of

entrainment and diffusion of the jet into the basin, (c)

these dams were built ﬁfty years ago. In this sense, numer-

impact on the basin bottom, and (d) recirculation in the

ous dams need to be re-evaluated in their safety with

basin (Figure 1).

respect to potential overﬂow, in line with what is being done in the USA (FEMA ).

Two of the variables needed to be deﬁned in the design of the jets are the issuance conditions and the impingement

doi: 10.2166/hydro.2016.044

Page 81

174

Figure 1

L. G. Castillo et al.

Distribution of mean ﬂow and turbulence statistics in plunge pools

Journal of Hydroinformatics

19.2

2017

Schematic of falling rectangular jets and receiving basin.

conditions (Castillo , ). The issuance conditions

When the jet falls through a long-enough distance, the jet

correspond to the ﬂow conditions at a location where the

becomes fully developed (see Lb in Figure 1). Castillo et al.

jet leaves the spillway and starts falling freely (z ¼ –h,

() established different equations to calculate the jet

where z is the vertical coordinate with origin in the crest

energy dissipation in the air and in the water cushion, as a func-

weir, and h is the weir head). The impingement conditions

tion of the Y/Bj and H/Lb ratios (where Y and H denote the

correspond to the jet section before the impact with the

depth of the water cushion at the exit and the total head,

water surface of the basin. In this location, the mean vel-

respectively, and Lb is the break-up length). During the falling,

ocity, Vj, and the impingement jet thickness, Bj, must be

the energy dissipation is due to the air entrainment into the fall-

deﬁned. This jet thickness must include the basic thickness

ing jet and the depth of water upstream of the jet. Energy

due to gravity Bg, and the symmetric jet lateral spreading

dissipation in the basin by diffusion effects can only be pro-

due to turbulence and aeration effects, ξ (Castillo et al. ): pffiffiffi�pffiffiffiffiffiffiffi pffiffiffi� q Bj ¼ Bg þ 2ξ ¼ pffiffiffiffiffiffiffiffiffiffi þ 4φ h 2H � 2 h 2gH

duced if there is an effective water cushion (Y/Bj > 5.5 for the rectangular jet case (Castillo et al. )).

(1)

In Figure 2, the velocity Vj and the jet thickness Bj at the impingement conditions, the core depth or minimum depth for effective water cushion and the two principal eddies that

where q is the speciﬁc ﬂow, H the fall height, and h is the

produce the dominant frequencies in the plunge pool (large

energy head at the crest weir. φ ¼ KφTu, with Tu being the

scale eddies and shear layer structures) are sketched. The

turbulence intensity and Kφ an experimental parameter

lowest frequencies correspond to large scale eddies that

(1.14 for circular jets and 1.24 for the three-dimensional

have a dimension on the order of the plunge pool depth

nappe ﬂow case).

(see Bombardelli & Gioia , ; Gioia & Bombardelli

Page 82

175

L. G. Castillo et al.

Distribution of mean ﬂow and turbulence statistics in plunge pools

•

Journal of Hydroinformatics

19.2

2017

Extension of models to two-ﬂuid theories to include air entrainment and jet break-up. Validate impact pressures against experimental data.

•

Develop parametric studies with turbulence models to identify the level of reliability of the computed pressure.

Castillo et al. (, ) and Carrillo () have followed the above suggestions and, specifically, they have developed laboratory research on the velocity, air, and pressure fields in the jet and the basin. They undertook experimental observations of jet velocities, and air concentrations with an optical fibre probe, and of pressures in the Figure 2

Schematic of eddy structures in effective water cushion (Y � 5.5 Bj): large

scale eddies size ∼Y and shear layer structures size ∼De (adapted from Ervine et al. 1997).

). Then, the recirculating velocity for large plunge pools is about Vr ∼ 0.035 Vj and the corresponding Strouhal number of the dominant eddies is S ¼ fY/Vj ¼ (Vr/πY )

(Y/Vj) ∼ 0.01 (Ervine & Falvey ; Falvey ; Ervine

et al. ). The following dominant frequency corresponds to eddy sizes contained in half of the shear-layer width and is proportional to the entry jet velocity; then, the Strouhal number of the shear-layer eddies is equal to a constant Ss ¼ ( fsY/Vj) ¼ K3 ∼ 0.25, and it coincides with the spread

of the jet into the water cushion as shown in Figure 2 (Ervine & Falvey ; Ervine et al. ). Within the plunge pool downstream of the impingement point, the flow resembles a flow in a submerged hydraulic jump and a wall jet. However, the situation is complicated here by the air entrainment. Several formulas have been put forward to express the horizontal velocity distribution in the vertical direction. We revisit some of these formulas later in the paper. There are only a few well-documented references on the numerical simulations of free overflow spillways. Ho & Riddette () analysed the different applications of computational fluid dynamics (CFD) code to hydraulic structures, and identified some limitations that do not allow a comprehensive analysis of the two-phase phenomena. They suggested the following principal lines of future research:

•

Validation of the air entrainment along chutes and freefalling jets.

plunge pool with pressure transducers. Then, they applied ANSYS CFX to simulate overflow nappe impingement jets in a general way, and investigated the different turbulence closures which better represent the data. In those works, the emphasis was put on the turbulence of the falling jet, the pressure distribution near the stagnation point in the water cushion, and the horizontal distance to the stagnation point. A good agreement among numerical simulations and laboratory data was obtained. In Castillo et al. (), the so-called ‘homogeneous’ theoretical model of CFX was employed. It was shown that this model is able to reproduce correctly the jet water velocity, and the averaged pressures in the plunge pool. There is always a challenge in modelling two-phase flows to discern which level of complexity is needed to represent different aspects of the flow (Bombardelli ; Bombardelli & Jha ; Jha & Bombardelli ). One of the objectives of this paper is to determine whether this theoretical model is sophisticated enough to represent velocities in the plunge pool. Continuing the line of research, this work presents a systematic study which considers specific flows and water cushions in the plunge pool. New laboratory data were obtained and new three-dimensional simulations were specifically performed for this work. ANSYS CFX was again selected due to the variety of turbulence closures available in the code, the previous experience with it and, more importantly, due to the diverse two-phase flow models embedded in the package, which can allow us to expand the research further in the near and long-term future. From an engineering point of view, knowing the parameters analysed, designers will be able to estimate the Page 83

176

L. G. Castillo et al.

Journal of Hydroinformatics

Distribution of mean ﬂow and turbulence statistics in plunge pools

19.2

2017

scour effects and the stability of the dam with greater

inlet channel were obtained with an acoustic Doppler velo-

certainty.

cimeter (ADV); mean velocities and air concentrations in different sections of the falling jet were acquired with optical ﬁbre instrumentation; and instantaneous pressure values

LABORATORY EXPERIMENTS

were measured with piezoresistive transducers located on the basin bottom. In addition, ADV and optical ﬁbre

Turbulent jet experimental facility

probe were used in the basin to obtain velocity and air concentration

The experimental facility was constructed at the Hydraulics

proﬁles,

respectively,

downstream

the

impingement point.

Laboratory of the Universidad Politécnica de Cartagena,

The ﬂow volumetric rate was measured with a V-notch

and was described in detail in Castillo & Carrillo (,

weir and was tested with ADV measurement upstream

) and Carrillo & Castillo (). Here, we revisit its

from the weir. Differences were smaller than 5%.

main features. The facility consists of a mobile mechanism

Figure 3 shows a picture of the experimental device in

which permits variation of the weir height between 1.7

which sizable values of air concentration are apparent. Air

and 3.5 m and ﬂows from 10 to 150 L/s. It has an inlet chan-

incorporation and transport are more important for shal-

nel with a length of 4.0 m and width of 0.95 m. The

lower water cushions.

discharge is produced through a sharp-crested weir with a width of 0.85 m and height of 0.37 m.

The ADV settings

The plunge pool, in which different water cushions may be simulated, is a 1.3-m high, 1.1-m wide and 3-m long

The setting characteristics of the ADV employed in these

methacrylate box. Turbulent kinetic energy values at the

tests were determined by considering that the main objective

Figure 3

Page 84

Air entrainment in the plunge pool observed in the laboratory device with q ¼ 0.082 m2/s, H ¼ 2.19 m, and Y ¼ 0.32 m.

177

L. G. Castillo et al.

Journal of Hydroinformatics

Distribution of mean ﬂow and turbulence statistics in plunge pools

19.2

2017

was to characterise the turbulence. The velocity range was

and the departure of the gas phase at the tip of the sensor.

selected as ±4.00 m/s (the maximum available in the equip-

The thresholding values were set to 1.0 and 2.5 V (Boes &

ment). With this setting, the ADV was able to measure

Hager ). The void fraction was deﬁned as the ratio of

horizontal velocities up to 5.25 m/s and vertical velocities

the total time the probe is in the gas (ΣtGi) to the experiment

up to 1.50 m/s. Due to the fact that the Doppler equipment

duration time t.

needs to be totally submerged into water, the ﬁrst 5–6 cm of

According to Stutz & Reboud (a, b), the equipment enables measurement in water up to 20 m/s ﬂow

the water column could not be measured. We used the turbulent kinetic energy measured 0.50 m

velocity and the relative uncertainty concerning the void

upstream of the weir in the experimental facility to specify

fraction is estimated at about 15% of the measured value.

inlet boundary conditions in the numerical simulations in

One source of error in the estimation of air presence in

that location. This distance guarantees that the stream lines

the ﬂow is due to the counting statistic of the number of

are horizontal upstream of the sharp weir crest (0.50 m > 5 h).

air bubbles in contact with the tips of the probe (Stutz

The plunge pool was surveyed in cross sections spaced

). Therefore, a short duration of the sequence would

0.10 m horizontally. Four speciﬁc ﬂows were tested (0.023,

contribute to an increased inaccuracy of the result. In

0.037, 0.058 and 0.082 m2/s) with different water cushion

order to evaluate the minimal measuring duration, André

depths. This covers 24 different conﬁgurations generated

et al. () considered the stabilization of the mean value

downstream of a rectangular free-falling jet.

during the measuring sequence and the quasi-stationary of

In order to characterize the macro turbulence of the

the air concentration signal as statistically representative

ﬂow in the plunge pool, 5,000 values were recorded in

for the air concentration. Based on the sensitivity study of

each measured point by using a frequency of 10 Hz (more

the probe behaviour, the authors recommend a sampling

than 8 minutes of observation). In this way, 2006 points in

sequence of 60 s as a good compromise between accuracy

the symmetrical vertical plane of the basin were obtained.

and time consumption.

As the ﬂows are highly turbulent, the values obtained with

Boes & Hager () carried out experiments with 4,000

ADV may be affected by spurious signals or ‘spikes’. Each

air bubbles and sampling sequences of 30 seconds. The

time series must be ﬁltered with the velocity and accelera-

authors considered the accuracy of the air concentration

tion

this

and velocity measurements is related to the variation of

particular case, the air may also affect the signal of the

thresholds

(Wahl

the phase Nb, air to water or inversely, rather than to the

ADV. Frizell () studied the air effect measuring concen-

sampling duration t.

).

Furthermore,

trations varying from 0 to 3.61%. As the air concentrations

Following those ideas, in this study a sampling sequence

increase and bubble sizes increase, correlation values drop

of 90 was considered. Figure 4 shows the void fraction evol-

dramatically as the acoustic signals used by the probe are

ution until a relative uncertainty of around 1% is reached

absorbed and reﬂected by the two-phase ﬂow mixture.

and the bubbles number detected in the measurement.

Matos et al. () also found that air bubbles affect the

Figure 5 shows the air concentration obtained by means

accuracy of velocity measurements taken with the ADV.

of the optical ﬁbre probe in different sections downstream of

However, their experimental results suggest that the ADV

the jet stagnation point. The equipment measures the air-

can provide reasonable estimates of the velocity for low

water ratio, β (entrained air discharge rate to water ﬂow

air concentrations up to 8%.

rate)

Optical ﬁbre probe

and,

from

this

value,

the

air

concentration,

C ¼ β=ð1 þ β Þ, was calculated.

The maximum air concentration is around 12% (at a dis-

tance of 21% from the bottom) for the ﬁrst sections. To measure the air concentration at the falling jet and at the

However, from the section 0.30 m and a distance from the

basin, an RBI-instrumentation dual-tip probe optical ﬁbre

bottom smaller than 70%, the air concentration is below

phase-detection instrument was used. The rise and fall of

10%. Concentrations remain high still at the upper portion

the probe signal corresponds, respectively, to the arrival

of the water depth in the basin. Page 85

178

Figure 4

L. G. Castillo et al.

Journal of Hydroinformatics

Distribution of mean ﬂow and turbulence statistics in plunge pools

19.2

2017

Temporal convergence of the void fraction (left) and bubble number detected (right).

distribution theory which says that for n independent, identically distributed, standard, normal, random variable ξi, the expected absolute maximum is: � � pffiffiffiffiffiffiffiffiffiffiffiffi E jεi jmax ¼ 2ln n ¼ λU

(2a)

where λU is denominated the Universal threshold (Donoho & Johnstone ; Goring & Nikora ). For a normal, random variable whose standard deviation is estimated by σ and zero mean, the expected absolute maximum is: Figure 5

Air concentration in the basin for different sections downstream of the jet stagnation point. Measurements obtained by means of an optical ﬁbre probe (q ¼ 0.082 m2/s, H ¼ 2.19 m, and Y ¼ 0.32 m).

λU σ ¼

pffiffiffiffiffiffiffiffiffiffiffiffi 2ln nσ

(2b)

Nonetheless, it should be noted that turbulence is Filtering the velocity records

not normally distributed and, therefore, the theoretical

Considering highly turbulent and aerated ﬂow that occurs in

approximately.

result the basin of energy dissipation, the Phase-Space Thresholding filter (Goring & Nikora , modified by Castillo ) was used. The spikes were replaced on each record by the mean value of the twelve closest points. This filter is based on the fact that the numerical derivative of a signal enhances its high frequency components, i.e. it enhances the spikes. The method uses the concept of a three-dimensional Poincaré map or phase-space plot in which the variable and its derivatives are plotted against each other. The points are enclosed by an ellipsoid and

concerning

from a theoretical result from the normal probability

Page 86

constant

λU

applies

only

The main steps of the method are as follows: 1. Calculate surrogates for the ﬁrst, Δu, and second, Δ2 u, derivatives using a centered differences scheme. 2. Calculate the standard deviations of all three variables, σ Δu , σ Δu , and σ Δ2u , and thence the expected maxima using the Universal criterion λU σ: 3. Calculate the rotation angle of the principal axis of σ Δ2u versus ui , using an expression for the cross correlation. Instead of using the expression by Goring & Nikora ():

the points outside the ellipsoid are designated as spikes (Castillo ). The threshold that is usually applied arises

the

θ ¼ tan

�1

! P ui Δ2 ui P 2 , ui

(3a)

179

L. G. Castillo et al.

Journal of Hydroinformatics

Distribution of mean ﬂow and turbulence statistics in plunge pools

Castillo () proposed a new relation obtained by a

19.2

2017

MATHEMATICAL AND NUMERICAL MODELLING

Gauss’ ﬁt: As can be seen from Figures 3 and 5, the ﬂow conditions

2 3 P P P 2 2 (n u Δ u � u u ) Δ i i i i 5 � P � θ ¼ tan�1 4 P n u2i � ( ui )2

(3b)

in the plunge pool are such that the air concentrations are relatively elevated at the point of jet impingement and nearby areas and in the top layer of the water depth. In

4. For each pair of variables, calculate the ellipse that has

these areas, there is a mostly non-dilute, two-phase ﬂow.

maxima and minima from point 3.

However, as we move far from the impingement point,

• For Δu

the ﬂow conditions tend to become quasi-dilute. That is

versus ui , the major axis is λU σ u and, the minor

axis is λU σ Δu .

• For Δ u 2

versus Δui , the major axis is λU σ and, the

minor axis is λU σ Δ2u .

• For Δ u 2

versus ui , the major and minor axes, a and b,

respectively, are the solutions of (Goring & Nikora ): ðλU σ u Þ2 ¼ a2 cos2 θ þ b2 sin2 θ �

(4a)

why we decided to solve the equations for the conservation of mass and momentum for the mixture, which may be written in compact form (ANSYS CFX Manual ) as: � � @ ðρ∅Þ @ @∅ ¼S ρUj ∅ � Γ þ @t @xj @xj

(5)

where ∅ is the transported quantity, i and j are indices which range from 1 to 3, xi represents the coordinate directions (1 to 3

�2

λU σ Δ2 u ¼ a sin θ þ b cos θ

(4b)

Castillo () proposed the following system of equations instead: ðλU σ u Þ2 ¼ a2 cos2

for x, y, z directions, respectively), and t the time. In turn, PNp PNp 1 XNp rk ρk , Uj ¼ rk ρk Ukj , and Γ ¼ k¼1 rk Γk , ρ ¼ k¼1 k¼1 ρ with rk indicating the volume fraction of kth ﬂuid, Γk denoting

the diffusion coefﬁcient associated with the transported quantity for phase k, Np denoting the number of phases

� � � � θ λU σ u θ sin2 þ b2 λU σ Δ2u 2 2

� � � � � �2 λU σ Δ2u θ θ sin2 þ b2 cos2 λU σ Δ2u ¼ a2 λU σ u 2 2

and S indicating the sources/sinks for the transported (4c)

quantity (ANSYS CFX Manual ). In this model, phases share the same velocity ﬁeld. When ϕ ¼ 1, S ¼ 0, and Γ ¼ 0,

the mass conservation equation is recovered, and when (4d)

ϕ ¼ Ui , the momentum equation is recovered, with its

corresponding source terms to account for the Reynolds stresses.

Figure 6 shows the original and ﬁltered signals

The theoretical model (1) comes as a result of the

measured at a point located 0.40 m downstream from the

addition of the equations of the two phases (Drew & Pass-

jet stagnation point and 12.34% of the water depth. We

man ; ANSYS CFX Manual ). Further, ∅K ¼ ϕ.

can also see the three ellipsoids deﬁned by the Universal cri-

Rigorously speaking, models like this have been found to

terion. The points outside of the ellipsoids are designated as

provide adequate predictions only in relatively-dilute mix-

spikes and, in this particular case, the number of points have

tures. Jha & Bombardelli () established that dilute

been 690. The mean velocity of the ﬁltered signal,

mixtures can extend up to the range 2–5% in the context

117.10 cm/s

signal,

of sediment-laden ﬂows; something similar could be

124.00 cm/s; however, the standard deviation is reduced

assumed in bubble ﬂows. For larger concentrations, Jha &

from 106.47 to 79.93 cm/s.

Bombardelli () found that the velocity distribution

very

similar

the

original

From these plots it can be concluded that the ﬁltering

could not be well predicted relatively far from the wall

attenuates the standard deviation. Further ﬁltering affects

with mixture models. Thus, we expect the ‘homogeneous’

the values of the turbulent kinetic energy.

model to be able to represent rather adequately those Page 87

180

Figure 6

L. G. Castillo et al.

Distribution of mean ﬂow and turbulence statistics in plunge pools

Journal of Hydroinformatics

19.2

2017

Original (a) and ﬁltered (b) signals corresponding to the point measured to 0.40 m downstream from jet stagnation point (12.34% of the water depth); (c), (d) and (e) are the three ellipsoids deﬁned by the Universal criterion (q ¼ 0.082 m2/s, H ¼ 2.19 m, and Y ¼ 0.32 m).

areas in which air concentrations are not that high

quasi-steady conditions. Once the quasi-steady conditions are

(cf. Figure 5).

reached, there are small ﬂuctuations around the mean value.

As said, the code ANSYS CFX has been used, which is based on an element-oriented, ﬁnite-volume method. It

Turbulence models

allows different types of volumes, including tetrahedral and hexahedral volumes. Solution variables are stored at

In this work, some of the most usual two-equation turbulence

the nodes (mesh vertices). More details are given in the

models have been tested for the free falling jet and basin case.

ANSYS CFX Manual ().

These models use the gradient diffusion hypothesis to relate

Finally, the instantaneous values in the numerical simu-

the Reynolds stresses to the mean velocity gradients and the

lations show large variations during the establishment of the

turbulent viscosity. The eddy viscosity hypothesis considers

Page 88

181

L. G. Castillo et al.

Journal of Hydroinformatics

Distribution of mean ﬂow and turbulence statistics in plunge pools

19.2

2017

that eddies behave like molecules and the Boussinesq model

time of 60 seconds, using a constant time-step of 0.05 s. The tran-

assumes that the Reynolds stresses are proportional to the

sient statistics were obtained by considering that quasi-steady-

mean velocity gradients, as follows (Pope ):

� ρu0i u0j ¼ μt

@Ui @Uj þ @xj @xi

state conditions are reached after the ﬁrst 20 seconds of simulation (Castillo et al. ). Hence, statistics were obtained

2 � δ ij ðρkÞ 3

(6)

from time-step 400 to time-step 1,200 (800 data in 40 s). Boundary conditions

with μt being the eddy viscosity or turbulent viscosity, ui is the mean square ﬂuctuation of the velocity in i direction,

The model boundary conditions corresponded to the turbu-

k ¼ 1=2(u0 1 u0 1 ) is the turbulent kinetic energy and δij the Kro-

lence kinetic energy at the inlet obtained with ADV (located

The Standard k-ε model is considered to be the traditional

water levels and their hydrostatic pressure distributions.

necker delta.

turbulence model and it is embedded in the majority of the

0.50 m upstream of the weir), the upstream and downstream Figure 7 shows the computational domain employed.

CFD programs. The Re-Normalisation Group (RNG) k-ε

ANSYS CFX has different treatments near the wall.

model is based on the renormalisation group analysis of the

ω-based turbulence models (e.g. SST) use automatic wall

Navier–Stokes equations (Yakhot & Orszag ; Yakhot &

treatments which switch between regular wall functions

Smith ). The transport equations for turbulent kinetic

(Pope ) and low-Reynolds wall treatment (Menter

energy and dissipation rate are similar as those for the standard

). Wall functions are used when the wall adjacent ver-

model, although their respective constants are different.

tices are in the log-law layer (yþ ≈ 20–200). The low-

The k-ω based Shear-Stress Transport (SST) model (Menter

Reynolds wall treatment is used when the wall adjacent ver-

) assumes that the eddy viscosity is linked to the turbulence k kinetic energy, k, and the turbulent frequency, ω, as μt ¼ ρ . ω The SST model takes into account the accuracy of the k-ω

tices are in the viscous sublayer (ANSYS CFX Manual ).

model in the near-wall region and the free stream indepen-

in the rest of the domain. Values of yþ were smaller than 40.

dence of the k-ε model in the outer part of the boundary layer. It is considered as a hybrid model (see Rodi et al. ).

Considering the wall treatment used by ANSYS CFX, the mesh sizes close to the solid boundary were smaller than For simplicity, only a half part of the model was simulated. The symmetry condition in the longitudinal plane of the plunge pool was used.

Free surface modelling The free surface model assumes that each control volume contains three possible conditions:

• • •

rk ¼ 0 if cell is empty (of the k-th ﬂuid); rk ¼ 1 if cell is full (of the k-th ﬂuid);

0 < rk < 1 if cell contains the interface between the ﬂuids. Tracking of the interface between ﬂuids is accomplished

by the solution of the volume fraction equation (see ANSYS CFX Manual ).

The inlet condition considers the volumetric ﬂow rate with a normal direction to the boundary (q ¼ 0.082 m2/s, q ¼ 0.058 m2/s, q ¼ 0.037 m2/s, q ¼ 0.023 m2/s), the turbu-

lent kinetic energy (5.1 × 10–4 m2/s2 for q ¼ 0.082 m2/s, 3.6 × 10–4 m2/s2 for q ¼ 0.058 m2/s, 1.9 × 10–4 m2/s2 for

q ¼ 0.037 m2/s, and 1.1 × 10–4 m2/s2 for q ¼ 0.023 m2/s),

and the water level height at the upstream deposit (2.313 m for q ¼ 0.082 m2/s, 2.285 m for q ¼ 0.058 m2/s, 2.263 m for q ¼ 0.037 m2/s, 2.237 m for q ¼ 0.023 m2/s).

The outlet condition was considered with ﬂow normal

to the boundary and hydrostatic pressure. The water level height at the outlet was modiﬁed according to the water cushion depth, Y, in the laboratory device. For all walls of

MODEL IMPLEMENTATION IN THREE DIMENSIONS

the upper deposit, the weir and the dissipation basin, no slip smooth wall conditions were considered. The roughness

Like in the simulation of the pressure ﬁeld on the basin bottom

of methacrylate was indicated in the walls. In the transverse

(Castillo et al. ), all scenarios were obtained by a calculation

direction, wall boundary conditions were used. Page 89

182

Figure 7

L. G. Castillo et al.

Journal of Hydroinformatics

Distribution of mean ﬂow and turbulence statistics in plunge pools

19.2

2017

Schematic of computational domain and boundary conditions, mesh blocks and water volume fraction.

Mesh-independence tests

Table 1

Number of volumes as a function of mesh size

In Figure 8, the simulations results for the different mesh

Mesh size (mm)

Number of volumes

sizes (5, 7.5, 10 and 12.5 mm) in the free falling jet, obtained

5.0

1,088,896

as a function of the vertical distance to the weir in terms of

7.5

457,810

the ﬂow velocities in the jet, are shown. Differences in vel-

10.0

255,776

ocities with the optical ﬁbre probe measurement are

12.5

134,785

smaller than 2% in all the cases (Castillo et al. ). In Table 1, the size of elements and the corresponding number of volumes required in the different simulations

are indicated. From the analysis of Figure 8, it can be concluded that mesh-independence is reached with an element size of 10 mm. The results are in agreement with previous results obtained on pressures at the stagnation point (Castillo et al. ). In this way, the mesh size of 10 mm seems to be valid for the ﬂow rates analysed. Figure 9 shows a view of the free surface in the threedimensional numerical model. We can see that the solution correctly reproduces the expected features of the jet and the plunge pool observed in the experimental facility. Convergence criteria To judge the convergence of iterations in the numerical solution, we monitored the residuals (Wasewar & Vijay Sarathi

Figure 8

Velocities in the falling jet as a function of the mesh size: q ¼ 0.058 m2/s, h ¼ 0.095 m.

Page 90

). The solution is said to have converged in the iterations if the scaled residuals are smaller than ﬁxed values

183

L. G. Castillo et al.

Journal of Hydroinformatics

Distribution of mean ﬂow and turbulence statistics in plunge pools

Figure 10

19.2

2017

Velocity distribution sketch for submerged jumps (adapted from Wu & Rajaratnam 1996).

To characterize the non-dimensional velocity distribution in hydraulic jumps, some authors have proposed different expressions which we include in the Appendix. Using the results obtained by diverse researchers, Wu & Rajaratnam () considered the length scale δl for the submerged jumps as a function of the distance to the beginning Figure 9

Free surface in the falling jet for the case of a mesh of 10 mm.

ranging between 10–3 and 10–6. In this work, the residual values were set to 10–4 for all the variables. Each time-step required around 12 iterations to reach the convergence criterion. With this choice and for 791,354 elements (255,776 in the falling jet), the mean computational time was 7.2 × 105 seconds (≈8.3 days), using a Central Processing Unit (CPU) with sixteen processors (Intel® Xeon® E5-2699 v3 @ 2.30 GHz).

EMPIRICAL FORMULAS In this section, we present some formulas obtained from the literature to interpret the different hydraulic parameters of plunge pools. To that end, we selected expressions for submerged hydraulic jumps and horizontal wall jets.

Velocity distribution in the plunge pool Following Rajaratnam (), we can compare the velocity

of the jump. They found that most of the observations for submerged jumps are contained within one standard deviation of the mean value of the wall jet, and only the data points near the end of the jump show an accelerated growth rate. Energy dissipation in the plunge pool In a horizontal channel, the total energy variation between the sections located upstream and downstream of the submerged hydraulic jump are, by deﬁnition:

HL ¼

Vj2 2g

þ y3

�

V42 þ y4 � 2g

�

(7)

where y3 and y4 are the water depths upstream and downstream of the submerged hydraulic jump generated by the jet. By using Equation (7) with the continuity equation, the energy dissipation may be obtained as (Ohtsu et al. ):

HL ¼ H0

� � y3 y4 þ 2 � y0 y0

1�

ðy4 =y0 Þ2 2 2ðy3 =y0 Þ þ Fr0

2 Fr0

(8)

proﬁles in the forward ﬂow if they are normalized with a vel-

where H0 is the energy upstream of the hydraulic jump, and y0

ocity scale equal to the maximum velocity, Vmax, at any

and Fr0 the water depth and Froude number in the upstream

section, and with a length scale δl equal to the elevation y

section of the hydraulic jump. When y3 ¼ y0 and y4 ¼ y2, the

from the bottom where the local velocity V ¼ Vmax/2, and the velocity gradient is negative (see Figure 10).

free hydraulic jump expression is recovered. For the plunge pool case, H0 may be assumed as the energy upstream the Page 91

184

L. G. Castillo et al.

Distribution of mean ﬂow and turbulence statistics in plunge pools

Journal of Hydroinformatics

19.2

2017

weir, while y0 may be estimated with Equation (1) or the

the plunge pool. For each horizontal velocity proﬁle

Bernoulli equation without considering the losses.

measured with the ADV, the length scale δl was obtained in each section (X distance to the stagnation point). Data have been classiﬁed as a function of the ratio water depth

RESULTS AND DISCUSSION

in the plunge pool/impingement jet thickness of each test. For ratios Y/Bj up to 20, the behaviour is similar to that

Laboratory velocity proﬁles

found in wall jets (albeit with another slope, as expected).

Figure 11 shows the velocity proﬁles obtained in the labora-

fall within one standard deviation of the mean value. How-

tory, together with equations proposed by diverse authors for

ever, for larger water cushion depths, the characteristic

horizontal wall jets (Görtler ; Rajaratnam ; Ohtsu et

length is higher. In this type of submerged hydraulic jump

al. ; Liu et al. ; Lin et al. ) (see Appendix). The

where the falling jet enters almost vertical, an equation

overall behaviour of the observations can be predicted rather

may be obtained with the data that fall within one standard

well by existing equations up to y=δ 1 ≈ 1:5. Disagreements

deviation of the mean value:

The values for water cushion depths Y/Bj up to 30 tend to

appear when ratio Vx/Vmax < 0.4. This seems to be related to the angle of impingement of the jet. In hydraulic jump studies, the wall jet is horizontal, while the impingement jet enters

δl X ¼ 0:465 þ 2:415 Bj Bj

(9)

almost vertical in these tests. The higher scatter occurs when the water cushion depth is Y/Bj > 20. In this way, the self-similarity disappears when the velocity proﬁles are close to the

Energy dissipation in the plunge pool

stagnation point and when a very submerged condition is obtained for the hydraulic jump. Following Wu & Rajaratnam (), Figure 12 shows the results of the characteristic length obtained through

Figure 11

pation and the Froude number at the jet impingement pffiffiffiffiffiffiffi condition, Fj ¼ Vj = gBj , obtained from experiments. In

Results of laboratory measurements of the non-dimensional proﬁles of the horizontal velocity in the central vertical plane of the plunge pool (X � 0.40 m). Proﬁles encompass

all discharges and water cushion depths tested.

Page 92

Figure 13 shows the contrast between the relative energy dissi-

185

L. G. Castillo et al.

Journal of Hydroinformatics

Distribution of mean ﬂow and turbulence statistics in plunge pools

Figure 12

Characteristic length δl in submerged hydraulic jumps.

Figure 13

Relative energy dissipation in the plunge pool as a function of the impingement Froude number.

addition, results coming from the use of Equation (8) have

Figure 14

19.2

2017

Relative energy dissipation in the plunge pool as a function of the ratio y3/Bj for the cases Bj ¼ 0.015 m and Fj ¼ 13–20.

Velocity and turbulent kinetic energy distributions

been included as a function of the ratio between the upstream water depth and the impingement jet thickness (y3/Bj). Lab-

Velocities at different cross sections of the plunge pool

oratory data are well represented by Equation (8). In the

located downstream of the stagnation point were measured

laboratory device, the impingement Froude number is between

with ADV. Results for the same cross sections were obtained

13 and 20 for the impingement jet thickness of 0.015 m plotted

from the CFD simulations. The velocities and turbulent kin-

in Figure 14. In general, tests carried out show an energy dissi-

etic energies have been made dimensionless by using the

pation larger than 75% of the impingement jet energy. This

maximum horizontal values, Vmax,x and kmax, respectively

ratio increases when the ratio y3/Bj decreases.

(see Figure 15). Page 93

186

L. G. Castillo et al.

Figure 15

Journal of Hydroinformatics

Distribution of mean ﬂow and turbulence statistics in plunge pools

19.2

2017

Horizontal velocity and turbulent kinetic energy proﬁles in the plunge pool downstream of the stagnation point (q ¼ 0.082 m2/s, H ¼ 1.993 m, Y ¼ 0.32 m, X � 0.40 m): (a) SST

velocities; (b) RNG k-ϵ velocities; (c) standard k-ϵ velocities; and (d) SST turbulent kinetic energy.

In general, the measured horizontal velocities, Vx, are

Like in the mean velocity cases, the best results of turbu-

greater than the calculated counterparts. The bellies

lent kinetic energy proﬁles were obtained with the SST

which appear in some proﬁles seem to be related with

model. In general, the results from the numerical simu-

the three-dimensional ﬂow features of the ﬂow. This

lations show the same behaviour as the results obtained

level of agreement is understandable given the observed

in the laboratory. However, the differences are very

air concentrations, close to 10%. As said with this level

important close to the stagnation point, where the numeri-

of concentrations of the disperse phase, Jha & Bombar-

cal model may not obtain accurate results due to the

delli () found that homogeneous-type models can

relatively important air entrainment into the plunge pool

only

(Figures 3 and 5).

provide

approximate

values

velocity.

Figure 15(a)–15(c) we can see that the choice of the turbu-

Figure 16 shows the non-dimensional velocity proﬁles

lence model has a signiﬁcant effect over the result of the

obtained from the numerical simulations as well as the lab-

horizontal ﬂow proﬁles. The less accurate results were

oratory measurements. Results are quite similar to the

obtained with the RNG k-ε turbulence model. However,

proﬁles obtained with the above-mentioned formulae. Due

the SST and the standard k-ε turbulence models produce

to the strong recirculation with air entrainment, in the

velocities fairly close to the values measured in the

upper side of the submerged hydraulic jumps there is a

laboratory.

bigger scatter for ratios Vx/Vmax < 0.40.

In addition to the mean velocities, the turbulent kin-

With all data, a new regression is proposed for sub-

etic energy proﬁles were also compared (Figure 15(d)).

merged hydraulic jumps downstream of the impingement

Page 94

187

Figure 16

L. G. Castillo et al.

Journal of Hydroinformatics

Distribution of mean ﬂow and turbulence statistics in plunge pools

19.2

2017

Velocity distribution downstream of the stagnation point (X � 0.40 m) obtained through numerical simulations and laboratory experiments.

In all cases, data collapse for ratios Vx/Vmax � 0.40. Data

point:

with smaller ratios do not match with a single law. As 1=7 Vx y y ¼ 1:48 1 � erf 0:66 Vmax δl δl

(10)

This proposed function is the separation line between the profiles in which there is negative recirculation flow and the profiles in which the flow is moving towards downstream. For the range of flows and water cushions

before, the larger differences appear when big water cushions are analyzed (ratios Y/Bj > 20).

CONCLUSIONS Observing and predicting two-phase ﬂows in hydraulic

analyzed, the limit between both behaviours seems to

structures is very complicated due to the rather non-dilute

be located at 0.2–0.3 m downstream of the stagnation

nature of the ﬂow. Under non-dilute conditions, both exper-

point.

iments and simulations cannot be expected to lead to clean

For the extreme negative and positive ﬂow proﬁles, two regressions (valid for ratios Vx/Vmax < 0.40) are also proposed, respectively:

comparisons. In this work, mean velocity and turbulent kinetic energy proﬁles have been analyzed in a plunge pool located downstream of a rectangular free-falling jet. The energy

1=7 Vx y y ¼ 1:65 1 � erf 0:72 � 0:10 Vmax δl δl 1=7 Vx y y þ 0:27 ¼ 1:10 1 � erf 0:80 Vmax δl δl

dissipation in a plunge pool is very high. For the test carried (11)

out, the dissipation of the impingement jet energy was between 75 and 95%. This ratio increases when the ratio water cushion/impingement jet thickness decreases.

(12)

In general, the CFD simulations provided results fairly close to the values measured in the laboratory, and to the Page 95

188

L. G. Castillo et al.

Distribution of mean ﬂow and turbulence statistics in plunge pools

formulas proposed by diverse authors, in spite of having used a simple two-phase ﬂow model. ‘Homogeneous’ model seems to be able to predict rather well areas in which air concentration is not very high. However, in the highly aerated regions rather strong differences appear. Regarding the modelling of turbulence, three closures were tested. SST turbulence model offered results closer to the laboratory measurements. The RNG k-ε model tends to underestimate the turbulent kinetic energy. It was possible to propose a single mean velocity distribution law for ratios Vx/Vmax 0.40. For smaller values, there are necessarily diverse distribution laws.

In order to develop this work further, we plan to examine air entrainment in the stilling basin. Comparison of results with diverse CFD codes (open source and commercial ones) against data will be considered.

ACKNOWLEDGEMENTS The first two researchers express their gratitude for the financial aid received from the Ministerio de Economía y Competitividad and the Fondo Europeo de Desarrollo Regional (FEDER), through the Natural Aeration of Dam Overtopping Free Jet Flows and its Diffusion on Dissipation Energy Basins project (BIA2011-28756-C03-02). The first author

acknowledges

the

support

Ministerio

Educación, Cultura y Deporte of España through Estancias de Movilidad de Profesores e Investigadores Senior program (PRX 14/00367), which allowed him to develop a stay as a Visiting Scholar Researcher at the Department of Civil and Environmental Engineering of the University of California, Davis, from April to October 2015.

REFERENCES André, S., Boillat, J. L. & Schleiss, A. J.  Discussion of ‘Twophase ﬂow characteristics of stepped spillways’ by Robert M. Boes and Willi H. Hager. J. Hydraul. Eng. 131 (5), 423–427. Annandale, G. W.  Scour Technology. Mechanics and Engineering Practice. McGraw-Hill, New York. ANSYS, Inc.  ANSYS CFX. Reference Guide. Release 16.0. ANSYS Inc., Southpoint, Canonsburg, PA, USA.

Page 96

Journal of Hydroinformatics

19.2

2017

Boes, R. & Hager, W. H.  Fiber-optical experimentation in two-phase cascade flow. In: Proceedings of the International RCC Dams Seminar (K. Hansen, ed.). Denver, CD, USA. Boes, R. & Hager, W. H.  Two-phase flow characteristics of stepped spillways. J. Hydraul. Eng. 129 (9), 661–670. Bombardelli, F. A.  Turbulence in Multi-Phase Models for Aeration Bubble Plumes. PhD Dissertation, Department of Civil and Environmental Engineering, University of Illinois at Urbana-Champaign. Bombardelli, F. A. & Gioia, G.  Towards a theoretical model for scour phenomena. In: Proc. RCEM 2005, 4th IAHR Symposium on River, Coastal, and Estuarine Morphodynamics (G. Parker & M. García, eds). Urbana, IL, USA, 2, pp. 931–936. Bombardelli, F. A. & Gioia, G.  Scouring of granular beds by jet-driven axisymmetric turbulent cauldrons. Phys. Fluids 18, 088101. Bombardelli, F. A. & Jha, S. K.  Hierarchical modeling of the dilute transport of suspended sediment in open channels. Environ. Fluid Mech. 9, 207–235. Carrillo, J. M.  Metodología numérica y experimental para el diseño de los cuencos de disipación en el sobrevertido de presas de fábrica. PhD Thesis. Universidad Politécnica de Cartagena, Spain (in Spanish). Carrillo, J. M. & Castillo, L. G.  Laboratory measurements and numerical simulations of overtopping nappe impingement jets. In: Proc. Int. Conf. 1st International Seminar on Dam Protection against Overtopping and Accidental Leakage, 24–26 November 2014, Madrid, Spain, pp. 219–230. Castillo, L. G.  Aerated jets and pressure fluctuation in plunge pools. In: Proceedings of the 7th International Conference on Hydroscience and Engineering, Philadelphia, PA, pp. 1–23. Castillo, L. G.  Pressure characterisation of undeveloped and developed jets in shallow and deep pool. In: Proceedings of the 32nd IAHR Congress, 1–6 July 2007, Venice, Italy, pp. 645–655. Castillo, L. G.  Measurement of velocities and characterization of some parameters inside of free and submerged hydraulic jumps. In: Proceedings of 33rd International Association of Hydraulic Engineering & Research Congress, Vancouver, Canada. Castillo, L. G. & Carrillo, J. M.  Hydrodynamics characterization in plunge pools. Simulation with CFD methodology and validation with experimental measurements. In: Proc. Int. Conf. 2nd IAHR European Congress, 27–29 June 2012, Munich, Germany. Castillo, L. G. & Carrillo, J. M.  Analysis of the scale ratio in nappe flow case by means of CFD numerical simulation. In: Proceedings of the 2013 IAHR Congress, 8–13 September 2013, Chengdu, China. Castillo, L. G., Carrillo, J. M. & Sordo-Ward, A.  Simulation of overflow nappe impingement jets. J. Hydroinform. 16 (4), 922–940. Castillo, L. G., Carrillo, J. M. & Blázquez, A.  Plunge pool mean dynamic pressures: a temporal analysis in nappe flow case. J. Hydraul. Res. 53 (1), 101–118.

189

L. G. Castillo et al.

Distribution of mean ﬂow and turbulence statistics in plunge pools

Donoho, D. L. & Johnstone, I. M.  Ideal spatial adaptation by wavelet shrinkage. Biometrika 81 (3), 425–455. Drew, D. A. & Passman, S. L.  Theory of Multicomponent Fluids. Applied Mathematical Sciences. Vol. 135. Springer, New York. Ervine, D. A. & Falvey, H. R.  Behaviour of turbulent water jets in the atmosphere and plunge pools. In: Proceedings of the International Conference of the Institutions of Civil Engineers, 83, pp. 295–314. Ervine, D. A., Falvey, H. T. & Withers, W. A.  Pressure fluctuations on plunge pool floors. J. Hydraul. Res. 35 (2), 257–279. Falvey, H. T.  Cavitation in Chutes and Spillways. Engineering Monograph No. 42. Bureau of Reclamation, Denver, CO, USA. FEMA  Technical Manual: Overtopping Protection for Dams. Federal Emergency Management Agency. FEMA P-1014, May. US Department of Homeland Security, USA. Frizell, K. W.  Effects of aeration on the performance of an ADV. In: 2000 Joint Conf. on Water Resources Engineering and Water Resources Planning & Management (R. H. Hotchkiss & M. Glade, eds). ASCE, Minneapolis, USA, (CD-ROM). Gioia, G. & Bombardelli, F. A.  Localized turbulent flows on scouring granular beds. Phys. Rev. Lett. 95, 014501. Goring, G. & Nikora, V.  Despiking acoustic Doppler velocimeter data. J. Hydraul. Eng. 128 (1), 117–126. Görtler, H.  Berechnung von aufgaben der freien turbulenz auf grund eines neuen näherungsansatzes (Calculation of free turbulence ratios on the basis of a new approximation). J. Appl. Math. Mech./Z. Ang. Math. Mech. ZAMM 22 (5), 244– 254 (in German). Ho, D. K. H. & Riddette, K. M.  Application of computational fluid dynamics to evaluate hydraulic performance of spillways in Australia. Aust. J. Civil Eng. 6 (1), 81–104. Jha, S. K. & Bombardelli, F. A.  Toward two-phase flow modeling of non-dilute sediment transport in open channels. J. Geophys. Res. Earth Surf. 115 (F3), 2003–2012. Lin, C., Hsieh, S.-C., Lin, I.-J., Chang, K.-A. & Raikar, R. V.  Flow property and self-similarity in steady hydraulic jumps. J. Exp. Fluids 53, 1591–1616. Liu, P., Gao, J. & Li, Y.  Experimental investigation of submerged impinging jets in a plunge pool downstream of large dams. Sci. China 41 (4), 357–365.

Journal of Hydroinformatics

19.2

2017

Matos, J., Frizell, K. H., Andre, S. & Frizell, K. W.  On the performance of velocity measurement techniques in air-water flows. In: Hydraulic Measurements and Experimental Methods Conference 2002 (T. L. Wahl, C. A. Pugh, V. A. Oberg & T. B. Vermeyen, eds). ASCE, Estes Park, CO, USA. Menter, F. R.  Two-equation eddy-viscosity turbulence models for engineering applications. AIAA J. 32 (8), 1598–1605. Ohtsu, F., Yasuda, Y. & Awazu, S.  Free and Submerged Hydraulic Jumps in Rectangular Channels. Report of the Research Institute of Science and Technology, 35. Nihon University, Tokyo, Japan. Pope, S. B.  Turbulent Flows. Cambridge University Press, Cambridge. Rajaratnam, N.  The hydraulic jump as wall jet. Proc. ASCE J. Hydraul. Div. 91 (HY5), 107–132. Rajaratnam, N.  Turbulent Jets. Elsevier Scientific, Development in Water Science, 5, New York, USA. Rodi, W., Constantinescu, C. & Stoesser, T.  LargeEddy Simulation in Hydraulics, IAHR Monograph, CRC Press. Stutz, B.  Analyse de la structure diphasique et instationnaire de poches de cavitation (Analysis of the two Phases and NonPermanent Structure of Cavitation Bubbles). PhD Thesis, Institut National Polytechnique de Grenoble, France (in French). Stutz, B. & Reboud, J. L. a Experiment on unsteady cavitation. Exp. Fluids 22 (1997), 191–198. Stutz, B. & Reboud, J. L. b Two-phase flow structure of sheet cavitation. Phys. Fluids 9 (12), 3678–3686. Wahl, T. L.  Analyzing ADV data using WinADV. In: Joint Conference on Water Resources Engineering and Water Resources Planning & Management, July 30–August 2, Minneapolis, MN. Wahl, T. L., Frizell, K. H. & Cohen, E. A.  Computing the trajectory of free jets. J. Hydraul. Eng. 134 (2), 256–260. Wasewar, L. & Vijay Sarathi, J.  CFD modelling and simulation of jet mixed tanks. Eng. Appl. Comp. Fluid Mech. 2 (2), 155–171. Wu, S. & Rajaratnam, N.  Transition from hydraulic jump to open channel flow. J. Hydraul. Eng. 122 (9), 526–528. Yakhot, V. & Orszag, S. A.  Renormalization group analysis of turbulence. I. Basic theory. J. Sci. Comput. 1 (1), 3–51. Yakhot, V. & Smith, L. M.  The renormalisation group, the εexpansion and derivation of turbulence models. J. Sci. Comput. 7 (1), 35–61.

First received 28 April 2016; accepted in revised form 19 October 2016. Available online 19 December 2016

Page 97

Hydrology Research

ISSN 0029-1277 iwaponline.com/hr

It is the principal aim of the Hydrology Research to conform to the highest possible scientific and technical standards. The journal publishes articles within hydrology in its widest sense, with an emphasis on water quantity and quality aspects of the hydrological cycle. Hydrology Research is intended to be a link between basic hydrological research and practical application of scientific results within the broad field of water management. The journal’s scope includes: • • • • •

methodological innovation contribution to theory improved understanding of processes or principles practical implications for water management science of hydrology For more details, visit iwaponline.com/hr

Page 100

1624

48.6

2017

A depth–duration–frequency analysis for short-duration rainfall events in England and Wales Ilaria Prosdocimi, Elizabeth J. Stewart and Gianni Vesuviano

ABSTRACT This study presents a depth–duration–frequency (DDF) model, which is applied to the annual maxima of sub-hourly rainfall totals of selected stations in England and Wales. The proposed DDF model follows from the standard assumption that the block maxima are generalised extreme value (GEV) distributed. The model structure is based on empirical features of the observed data and the assumption that, for each site, the distribution of the rainfall maxima of all durations can be characterised by common lower bound and skewness parameters. Some basic relationships between the location and scale parameters of the GEV distributions are enforced to ensure that frequency estimates for different durations are consistent. The derived DDF curves give a good ﬁt to the observed data. The rainfall depths estimated by the proposed model are then compared with the

Ilaria Prosdocimi (corresponding author) Department of Mathematical Sciences, University of Bath, Claverton Down, Bath BA2 7AY, UK E-mail: prosdocimi.ilaria@gmail.com Elizabeth J. Stewart Gianni Vesuviano Centre for Ecology & Hydrology, Maclean Building, Crowmarsh Gifford, Wallingford, Oxfordshire OX10 8BB, UK

standard DDF models used in the United Kingdom. The proposed model performs well for the shorter return periods for which reliable estimates of the rainfall frequency can be obtained from the observed data, while the standard methods show more variable results. Although the standard methods used no or little sub-hourly data in their calibration, they give fairly reliable estimates for the estimated rainfall depths overall. Key words

| depth–duration–frequency, intensity–duration–frequency, short-duration rainfall, statistical modelling

INTRODUCTION Estimates of the magnitude of rainfall events of a given dur-

model, it is required that frequency curves for different dur-

ation with an expected annual exceedance probability p, are

ations do not cross, meaning that the rainfall depth that is

an important component of current methods of ﬂood fre-

exceeded with probability p should increase monotonically

quency estimation, used in the design and assessment of

with increasing event duration. The probability p is typically

ﬂood defence schemes, bridges and reservoir spillways, as

expressed as a return period T, with p ¼ 1/T, as events larger

well as urban drainage systems. Rainfall frequency estimates

than those corresponding to the quantile that is expected to

are also a key input to mapping studies of the risk of surface

be exceeded with probability p should happen, on average,

water ﬂooding. The estimates can be obtained from depth–

every T years.

duration–frequency (DDF) models, in which the relation-

DDF models, which are often referred to as intensity–

ship between the rainfall depth, event duration and event

duration–frequency models, can then serve two purposes:

rarity is integrated in a unique framework. In a DDF

to estimate the rainfall depth of a hypothetical event with a given duration and rarity, and to assess the rarity of a

This is an Open Access article distributed under the terms of the Creative Commons Attribution Licence (CC BY 4.0), which permits copying,

storm event with known rainfall depth and duration. Svens-

adaptation and redistribution, provided the original work is properly cited

son & Jones () give an overview of different DDF

(http://creativecommons.org/licenses/by/4.0/).

models used in several countries, showing the large array

doi: 10.2166/nh.2017.140

Page 101

1625

I. Prosdocimi et al.

A DDF analysis for short-duration rainfall events in England and Wales

Hydrology Research

48.6

2017

of possible approaches to rainfall frequency estimation.

focused on the estimation of the frequency of long-duration

Many of the countries included in the review use some

events. Considering that the FEH13 model aimed to

form of index rainfall approach combined with regional esti-

improve rainfall frequency estimates for rare events with

mation of growth curves for different durations, although

durations longer than 1 hour, it is not yet clear how it will

some countries were reported to use the linear regression

perform for the more frequent events of very short duration

approach. The idea behind the latter approach is to ﬁt a stat-

which are of interest in this study.

istical distribution separately to the single series of block

The FSR and FEH99 DDF models are based on an

maxima of different accumulation periods and then to ﬁt

index-rainfall approach and were developed with the

regression models across the different durations or frequen-

scope of providing nationwide rainfall frequency estimates.

cies, so that increasing rainfall depths are estimated for

The FEH99 method was calibrated on a larger network of

increasing durations given a certain frequency. See Kout-

stations with longer records than the FSR method and,

soyiannis et al. () for a discussion of the mathematical

unlike the FSR method, incorporated a spatial model in

formulation of the relationship between the duration and

which data from nearby stations were used for rainfall fre-

frequency of rainfall events, and a general discussion of

quency estimation at a given location. On the other hand,

DDF modelling. Although the relationship between rainfall

the FEH99 method was calibrated using data with an

depths and frequencies has been studied for several decades,

accumulation period of at least 1 hour while, in the develop-

there is still much interest in identifying methods to derive

ment of the FSR method, some data with an accumulation

DDF curves (e.g., Overeem et al. ) and in the actual

period of 1 minute were also used. Compared to the FSR

derivation of DDF curves to be used at different sites of

method, the FEH99 method has been found to give much

interest (e.g., Jiang & Tung ).

larger estimates of rainfall depth for the very long return

One interesting ﬁnding of the review in Svensson &

periods required for reservoir safety assessment (Babtie

Jones () is that, in several countries, different models

Group in association with CEH Wallingford & Rodney

are used depending on the duration and rarity of the rainfall

Bridle Ltd ; MacDonald & Scott ). As a result,

events of interest. The need for different models for different

the FSR and FEH99 methods are both still used, but for

durations and frequencies stems from the difﬁculty of devel-

different cases that depend on the duration and rarity of

oping models that can provide reliable results across several

the design event to be estimated (ICE ). As Svensson

rainfall durations and frequencies. One country where

& Jones () report, the FSR method can be used to esti-

several DDF models are currently in use is the UK: the

mate return periods of rainfall events with accumulation

main models are presented below and are the main focus

periods between 1 minute and 25 days and return periods

of this study.

longer than 1,000 years, and is recommended for the esti-

In the UK, the most widely used DDF models are those

mation of rainfall depths associated with return periods up

presented in volume II of the Flood Studies Report (FSR,

to 10,000 years The FEH99 method provides estimates of

Natural Environment Research Council ) and in

rainfall accumulations between 1 hour and 8 days, with

volume 2 of the Flood Estimation Handbook (FEH99, Faul-

return periods shorter than 1,000 years and, although rain-

kner ), which mostly superseded the FSR methods.

fall frequencies up to return periods of 10,000 years can

Recently, a new model (FEH13, Stewart et al. ) has

technically be estimated, their use is not recommended.

been developed, with the speciﬁc aim of overcoming the

The newly developed FEH13 might replace the FSR and

issues encountered when the original FEH99 model is

the FEH99 as the recommended model to use to estimate

used to estimate rare events. Estimates from the FEH13

the magnitude of very rare events, but the ofﬁcial guidelines

model have only been available to practitioners since

have not yet been amended. The FEH99 method can also be

November 2015, and have therefore not yet been widely

extended to estimate the frequency of rainfall events with

used in practice. Furthermore, the performance of the

accumulation periods shorter than 1 hour, although, as no

FEH13 model for short duration events (i.e., under 1 hour)

sub-hourly data were used in the calibration of the

is still being assessed, since most of the model evaluation

method, extrapolation to durations below 30 minutes is

Page 102

1626

I. Prosdocimi et al.

A DDF analysis for short-duration rainfall events in England and Wales

strongly discouraged. The coexistence of the FEH99 and

Hydrology Research

48.6

2017

DATA

FSR methods results in uncertainty when estimates are needed for sub-hourly rainfall events. These cases go

From the large number of tipping bucket rain gauges mana-

beyond the range of reliable estimates for the FSR, a rela-

ged by the EA and Natural Resources Wales (NRW) and

tively old model that was calibrated on fairly short records

providing sub-hourly rainfall data, a subset with sufﬁciently

with very limited sub-hourly data, and beyond the intended

long records was identiﬁed that could allow for good spatial

use of the FEH99, a more complex and structured model

coverage of the area. Sub-hourly data for the rainfall stations

that was calibrated using a dense network of stations but

are available as time of tip (ToT) at some sites series and as

no sub-hourly data at all.

aggregated 15-minute accumulation series at other sites. In 2

Small catchments (i.e., smaller than 25 km ) and plot-

the selection of stations to be included in this scoping

sized areas are expected to be particularly vulnerable to

study, priority was given to those for which ToT data were

short, intense cloudbursts, due to their short response

available, to allow very short durations to be investigated.

times. As Faulkner et al. () emphasise, reliable estimates

It appears that long ToT series are more readily available

of sub-hourly design rainfalls are therefore needed to allow

in some regions (the English Midlands and Wales), hence

credible ﬂow and hydrograph estimates for the smallest

the ﬁnal subset of stations included in the study is a compro-

catchments using rainfall–runoff techniques. The sugges-

mise between the competing needs of having long series and

tions in Faulkner et al. () motivated the second phase

maintaining a good coverage of England and Wales (E&W).

of the Environment Agency’s (EA) Estimating Flood Peaks

In particular, the sites were chosen to be at least 35 km

and Hydrographs for Small Catchments project. The project

apart. The ﬁnal selected stations are shown in Figure 1.

aims to improve the estimation of ﬂood frequencies in

The shortest series in the dataset is 15 years long; the longest

small catchments and encompasses, among other things,

two are each 46 years long. A total of nine ToT series and

an assessment of the most appropriate methods to estimate

ten 15-minute series are included in the study dataset. The

the frequency of very short duration rainfall, which this

analysis was performed on the annual and seasonal

study is concerned with. A novel at-site DDF modelling

maxima of the different accumulations, with two six-

strategy is discussed and an application of the proposed

month seasons included in the study. The ﬁnal dataset was

model is presented using data series available at selected

compiled from the ToT and 15-minute series, following

sites that give a reasonable geographical coverage of

two slightly different workﬂows as outlined below.

England and Wales, for which relatively long records of sub-hourly rainfall are available. The proposed model

•

were composed. From these, 1-minute monthly maxima

does not follow the traditional approaches and uses

were extracted and, by cumulating successive data-

instead the data across all durations to ﬁt a unique

points, monthly maxima for 2-, 5-, 10-, 15-, 30-, 45-, 60-,

model. Rainfall frequency curves estimated with the proposed method are compared to those estimated with the FSR and FEH99 DDF models, and to empirical return level estimates. The stations and datasets used in the study are intro-

From the original ToT data, 1-minute accumulation series

90- and 120-minute accumulations were extracted.

•

From the 15-minute accumulation data, monthly maxima for the 15-, 30-, 45-, 60-, 90- and 120-minute accumulations were extracted.

duced in the next section. Subsequently, a uniﬁed generalised extreme value (GEV) model is proposed and

For all series, a month was considered complete if at

its performance for the stations under study is discussed.

least 75% of the data in the month were non-missing.

The performance of the uniﬁed GEV, FSR and FEH99

Finally, the annual and seasonal maxima series were con-

models for short-duration rainfall frequency estimation are

structed from the monthly maxima series. A year or

compared in the section Comparisons of the uniﬁed GEV

season was considered complete if no more than one

results to current methods. The ﬁnal section of the paper

monthly record within that year or season was incomplete.

contains the conclusions and ﬁnal remarks.

Approximately 89% of the station-seasons have at least Page 103

1627

Figure 1

I. Prosdocimi et al.

A DDF analysis for short-duration rainfall events in England and Wales

Hydrology Research

48.6

2017

Location of the 19 stations included in the study. The record length of the annual maxima series is indicated in the location of each station; numbers in italics indicate ToT stations, numbers in roman indicate 15-minute stations.

99% of valid data points, 98% of the station-seasons have at

extracted as the maximum value recorded in the months

least 90% of valid data points and in just one instance is the

from November to April inclusive.

percentage of valid data points in a season lower than 80%

The availability of the raw ToT information for the tip-

(summer rainfall series of 1995 at Victoria Park, which has a

ping bucket stations allows for the extraction of series at a

total of 79.3% valid data points). Overall, for all stations, for

1-minute resolution and additionally at coarser or even

the series across all years and seasons, more than 99% of the

ﬁner resolutions. However, the level of precision that can

total number of data points are recorded as valid, giving

be reached in high resolution series depends greatly on the

reasonable conﬁdence in the quality of the available data

tip volume of the instrument, a property that might change

and conﬁdence that the maxima were captured. Annual

slightly in time (e.g., due to sediment collecting in the

maxima were extracted as the maximum single value

bucket) or more signiﬁcantly over time (e.g., if the speciﬁc

recorded in each calendar year. Summer maxima were

gauge used at a station is replaced by a different model). Fur-

extracted as the maximum value recorded in the months

thermore, the tip volume might be different at different

from May to October inclusive. Winter maxima were

stations, thus creating inconsistencies in the precision

Page 104

1628

I. Prosdocimi et al.

Hydrology Research

A DDF analysis for short-duration rainfall events in England and Wales

48.6

2017

across different stations. The discrete nature of the tipping

considered, the alignment between the rainfall event and

bucket measurements is also the underlying reason why, in

the station clock is less important, as the depths of rainfall

a number of months, the recorded 1-minute and 2-minute

at the tail ends of the storm, which are difﬁcult to capture

maxima have the same value, and why several annual and

exactly, become less and less important to the storm depth

seasonal maxima are identical across a number of years.

as a whole.

These issues are more common in the earlier years of the

To adjust the maxima extracted from the 15-minute

record, during which time the data were measured at a coar-

stations so that they are closer to the higher values that

ser resolution. The issues connected to systematic errors in

would be attained using sliding windows, correction factors

tipping buckets are known (Molini et al. , and refer-

were introduced. For each ToT record, ﬁxed-period

ences therein). In particular, lower intensities tend to be

(15-minute) annual and seasonal maxima were extracted

overestimated and higher intensities tend to be underesti-

for durations of 15, 30, 45, 60, 90 and 120 minutes. These

mated. Methods to quantify the systematic error of each

series correspond to the maxima that would be obtained if

station are beyond the scope of this study, and the data

the data for the ToT stations were stored as 15-minute

extracted from the original series are used in all subsequent

series (ﬁxed window) rather than ToT series (sliding

analysis without further adjustment. The issues connected

window). The average ratio between the sliding window

with the original data series should, nevertheless, be

maxima and the ﬁxed window maxima at each duration,

acknowledged as they can have an impact on the estimation

shown in Table 1, is used as a sliding window correction

procedures discussed in the section Results for the at-site

factor for that duration. In the rest of this work, the

analysis and in the comparisons discussed in the section

maxima extracted from the 15-minute series are multiplied

Comparisons of the uniﬁed GEV results to current method.

by the appropriate correction factor to give estimates of

Due to differences in the underlying data collection

the equivalent sliding window maxima. Due to the different

methods, the series of maxima extracted from the ToT and

ranges of time resolution present in the two different data

the 15-minute series do not provide the same information

sources, two separate analyses are carried out: one which

for accumulations of 15 minutes or greater. The ToT

uses only the series extracted from the ToT stations and

maxima are computed using a sliding window, so the

covers the range of durations from 1 to 120 minutes; and

15-minute annual maximum value (for example) corre-

one in which data from all stations are included, covering

sponds to the actual largest amount of rainfall recorded in

the range of durations from 15 to 120 minutes.

any 15-minute interval in the year. However, the maximum obtained from the 15-minute records instead corresponds to the maximum amount recorded in one predeﬁned 15-minute

THE UNIFIED GEV DDF MODEL

interval, which is likely to be lower than the actual maximum amount of rainfall that could have been recorded in

The FSR, FEH99 and FEH13 DDF models build on a large

a 15-minute interval without a ﬁxed start time. The true

set of available gauges and allow the estimation of frequency

maximum rainfall is most likely to be under-recorded

curves for a number of durations across the whole UK. In

when its duration is the same as the ﬁxed-duration recording

particular, the FEH99 and the FEH13 have complex spatial

unit, as the rainfall event is very unlikely to align neatly with

model components so that estimates for rainfall frequencies

the station clock. However, when longer durations are

at one point are built incorporating information from nearby

Table 1

The correction factors applied to the maxima obtained from 15-minute series, for different seasons and event durations 15 minutes

30 minutes

45 minutes

60 minutes

90 minutes

120 minutes

Annual

1.15

1.05

1.03

1.02

1.01

Winter

1.14

1.05

1.04

1.03

1.02

Summer

1.15

1.06

1.03

1.02

1.01

Page 105

1629

I. Prosdocimi et al.

Hydrology Research

A DDF analysis for short-duration rainfall events in England and Wales

48.6

2017

gauges. Such complex spatial structures are unattainable

that, denoting by x(F,D) the rainfall depths of durations D

with the subset of stations available in this study. Given

associated with a certain non-exceedance probability F, for

the exploratory scope of this work, a simpler model is pro-

d0 < d1 one should have x(F,d0) < x(F,d1). The proposed

posed: the model allows the estimation of a station’s DDF

model uses the relationship between the GEV parameters

curves based solely on the data series available for that

shown in Equation (2) and stems from some empirical prop-

station; it does not have a component to include information

erties observed via visual explorations of the estimated

from nearby stations.

parameters for the different durations at each station (see

The proposed model builds on extreme value theory

the section Results for the at-site analysis). The GEV distri-

(Coles ), assuming that block (e.g., annual or seasonal)

bution can be shown to be the asymptotic distribution of

maxima follow a GEV distribution: X ∼ GEV(ξ, α, κ) where

sample maxima (see Coles ) and has often been used

X indicates the random variable that describes rainfall

as an underlying distribution in the development of DDF

block maxima and ξ, α and κ are the location, scale and

models (among others, Overeem et al. ; Jiang & Tung

skewness parameters of the GEV distribution, respectively.

). According to the goodness of ﬁt test presented in

The cumulative distribution function of a GEV distributed

Kjeldsen & Prosdocimi (), the GEV distribution was

random variable X ∼ GEV(ξ, α, κ) is deﬁned as:

deemed acceptable for a large majority of the series analysed

( � � ) 1 � κ (x � ξ) 1=κ F(x) ¼ exp � α

in the study. When estimating frequency curves, it is (1)

expected that no upper limit should be attainable by the rainfall values at any duration, so the skewness parameter is constrained in the proposed model to be negative. The

The set on which the variable X is deﬁned, e.g., the

model development is presented below only for the case in

values that might be observed in a sample from a population

which κ < 0, although similar ideas would apply for κ > 0:

with underlying distribution X, is governed by the skewness

It is assumed that the skewness parameter κ is constant

parameter as follows:

across all durations, while the location and scale parameters

8 α > �∞ < x � ξ þ > > κ < �∞ < x < ∞ > > > : ξþ α <x<∞ κ

are dependent on the duration D: ξ(D) and α(D). Taking ‘ to be the lower bound of the distribution, and assuming this to

if κ > 0 if κ ¼ 0

(2)

if κ < 0

The distribution is bounded for the case in which κ ≠ 0,

be the same for all durations, the following relationship is obtained from the inequality in Equation (2): α(D) ¼ (‘ � ξ(D))κ:

(4)

with the lower and upper bound being a linear combination

The quantile function shown in Equation (3) can then be

of the distribution parameters. The skewness parameter

updated to a quantile function xD (F), which depends on the

therefore deﬁnes whether an upper or lower bound for the

event duration D via the location parameter ξ(D):

values of X exists. The quantile function for the GEV distribution, which is used to build frequency curves, is derived as:

x(F) ¼

(

α [1 � (�log F)κ ] κ ξ � α log (�log F)

ξþ

if κ ≠ 0 if κ ¼ 0

(3)

xD (F) ¼ ξ(D) þ

α(D) [1 � (�log F)κ ] κ

¼ ξ(D) þ (‘ � ξ(D)) [1 � (�log F)κ ] ¼ ‘[1 � (�log F)κ ] þ ξ(D)(�log F)κ

(5)

Provided that ξ(D) is monotonically increasing, the func-

where F is the non-exceedance probability, corresponding to

tion xD (F) is a monotonically increasing function of D, so

F ¼ 1 � 1=T for the T-year event. The desired property of a

that the estimated frequency curves give consistent results

DDF model is that the quantile functions for increasing dur-

for increasing durations. For the case of the British rain

ations of rainfall accumulation, D, do not cross. This means

gauges under study, the following relationship is proposed

Page 106

1630

I. Prosdocimi et al.

Hydrology Research

A DDF analysis for short-duration rainfall events in England and Wales

48.6

2017

to model the location as a function of the event duration,

accommodate different data behaviours: the uniﬁed GEV is

based on the observed properties of the location parameter

an addition to the possible modelling approaches used for

for a GEV distribution ﬁtted separately for each different

at-site estimation of DDF curves.

duration across all stations (see Figure 2 in the next section): ξ(D) ¼ a þ b D þ c (1 � exp {�g D})

(6)

RESULTS FOR THE AT-SITE ANALYSIS

which is an increasing function of D provided that its ﬁrst

For each station separately, the parameters of the uniﬁed

derivative is positive:

GEV model (a, b, c, g, ‘, κ) are estimated via maximum likelihood,

b þ c g exp {�g D} > 0

(7)

which

ensures

some

optimal

properties

for

parameter estimates (Coles ). The uniﬁed GEV model is ﬁtted to the data from all the ToT stations and to all the

The scale function is determined by a combination of

series with accumulations of at least 15 minutes, in two

the lower bound (‘), the skewness parameter (κ) and the

different ﬁtting procedures. The location function, shown

location function (ξ(D)) according to Equation (4). The pro-

in Equation (6), and the relationship between the scale

posed uniﬁed GEV model then requires the estimation of a

and other parameters, shown in Equation (4), are used in

total of six parameters (a, b, c, g, ‘, κ), a relatively parsimo-

the two ﬁtting procedures.

nious model which, given some constraints in the location

To illustrate the challenges relating to the model ﬁtting

function, allows for consistent frequency estimates for differ-

procedure and to show some of the features of the ﬁtted

ent durations. It is possible that an even simpler formulation

models, the location (ξ(D)) and scale (α(D)) functions,

could be used for Equation (6), but the suggested function

together with the skewness (κ) and lower bound (‘) par-

originates from the methods discussed in Stewart et al.

ameters, all as estimated by ﬁtting the uniﬁed GEV model

() and seems to give reasonable results.

to the ToT annual maxima series, are shown in Figure 2.

The proposed uniﬁed GEV model uses a different strat-

As a reference, the plot also shows estimates for the GEV

egy to obtain consistent estimates for rainfall frequencies

parameters obtained by applying an L-moments ﬁtting pro-

than many published works, which use approaches based

cedure (Hosking ) to the series of each duration

on linear regression across estimates for the different dur-

separately for all stations. L-moment estimates are fre-

ations. The uniﬁed GEV model presented in this paper

quently used in hydrology due to their good performance

instead seeks to ﬁt a unique model to all series at once, so

when applied to relatively short series, such as the dur-

that all available information is used to estimate the DDF

ation-speciﬁc rainfall series analysed here. The scatter of

curves. The development of the model is inspired by some

the duration-speciﬁc estimates inspired the use of an expo-

of the discussion in Stewart et al. (), on the development

nential function to describe the location of the GEV

of the statistical framework used in the FEH13 model.

distribution as a function of the rainfall duration shown in

The basic novel idea behind the proposed model is to

Equation (6) and there is, indeed, a general agreement

ensure that monotonic quantile functions are obtained by con-

between the duration-speciﬁc estimates and the location

straining some of the parameters of the rainfall distribution to

functions estimated within the uniﬁed GEV model. Note

have common properties across different durations. It is poss-

that the GEV ﬁtted to each duration separately would lead

ible that for a different set of durations, or a new set of gauging

to non-consistent return curves across the different dur-

stations that exhibit different properties, the assumptions of

ations, unlike the uniﬁed GEV model: although it is

which common distributional properties are to be shared

desirable for the uniﬁed GEV parameter functions to

across durations might be different. Furthermore, the func-

resemble the estimates obtained for each duration separ-

tional relationship between the location and the duration

ately, the differences in the estimates are needed to ensure

shown in Equation (6) might not be valid. Nevertheless, the

the consistency of the estimated frequency curves. More-

building blocks of the proposed model could be adapted to

over, the relatively large difference seen between the Page 107

1631

Figure 2

I. Prosdocimi et al.

A DDF analysis for short-duration rainfall events in England and Wales

Hydrology Research

48.6

2017

Estimated parameters obtained from an L-moment estimation for each duration separately (dots) and from the proposed uniﬁed GEV distribution (lines) annual series. Colours and symbols indicate the different ToT stations. Distribution parameters in each panel are (clockwise from top left): the location, scale, lower bound and skewness. To make the ﬁgure readable, lower bound estimates below �40 are not shown.

uniﬁed GEV estimates and the separate-duration GEV par-

plotting positions. The frequency curves seem to ﬁt the

ameter estimates at some stations (e.g., Victoria Park) are

data reasonably well. Due to the constraints in the model

partially the consequence of the model structure, in which

structure that ensures that the location function is monoto-

the skewness parameter, which is constrained to be nega-

nically increasing for increasing durations, the frequency

tive, regulates the curvature of the scale function. For

curves computed from the formula in Equation (5) tend to

Victoria Park, for example, the raw estimate of the skewness

fan out. A noticeable feature of the data is that the winter

parameter for many durations is positive or very close to

maxima tend to be much smaller than the summer

zero, as shown in the lower left panel of Figure 2. The

maxima, which also appear to be the annual maxima.

ﬁnal estimated values for the uniﬁed GEV parameters maxi-

Results for the other ToT stations have similar properties

mise the overall likelihood for all durations within the

to the ones shown in Figure 3 and are shown in Prosdocimi

constraints of the model: this could lead to large discrepan-

et al. ().

cies between the estimates obtained under the uniﬁed GEV

Figure 4 shows the estimated location and scale func-

and those obtained from the GEV parameters estimated for

tions, together with the skewness and lower bound

each duration separately. The results of ﬁtting the uniﬁed

parameters of the uniﬁed GEV model, for annual data at

GEV to winter and summer maxima show a similar pattern.

all 19 stations, considering accumulations of 15 to 120 min-

Estimated rainfall DDF curves for the annual, winter

utes. As in Figure 2, the original estimates for the GEV

and summer series for the ToT station at Dowdeswell are

parameters obtained from an L-moment estimation pro-

shown in Figure 3, together with the block maxima

cedure ﬁtted to each duration separately are also shown.

extracted from the original series plotted using Gringorten

Again, the ﬁtted location functions seem to be mostly in

Page 108

1632

Figure 3

I. Prosdocimi et al.

A DDF analysis for short-duration rainfall events in England and Wales

Hydrology Research

48.6

2017

Estimated frequency curves for the station at Dowdeswell for the annual (left panel), winter (central panel) and summer (right panel) series, superimposed on the Gringorten plotting positions for each duration, starting from 1 minute.

Figure 4

Estimated parameters obtained from an L-moment estimation for each duration separately (dots) and from the proposed uniďŹ ed GEV distribution (lines) annual series. Colours and symbols indicate the different stations with series of at least 15-minute accumulations. Parameters in each panel are (clockwise from top left): the location, scale, lower bound and skewness. To make the ďŹ gure readable, lower bound estimates below ďż˝40 are not shown.

Page 109

1633

I. Prosdocimi et al.

A DDF analysis for short-duration rainfall events in England and Wales

Hydrology Research

48.6

2017

agreement with the estimates obtained from the different

curves shown in Figure 5 have similar properties to those

durations, while more variability can be seen in the esti-

shown in Figure 3 – the curves have a tendency to fan out

mation of the scale function in the top right panel. In

and the annual extremes appear to be mostly driven by

particular, the scale functions for Victoria Park and Otter-

summer rather than winter events.

bourne are very ﬂat, as a result of the estimates for the

Seasonal differences are not explored further in this

skewness parameters at these stations being very close to

analysis, but the estimates obtained from the different

zero. The estimated lower bounds for these two stations

stations could be employed in the future to develop correc-

are also very small: �37.7 at Otterbourne and �124.6 at Vic-

tion factors to obtain seasonal estimates from sub-hourly

toria Park (censored in Figure 4). The fact that the skewness

annual estimates, similarly to Kjeldsen et al. (). The uni-

parameter for these stations is estimated to be very close to

ﬁed GEV proved to be a ﬂexible and reliable modelling

zero in the uniﬁed GEV model is likely to be connected to

approach which could give reasonable estimates across

the fact that some series in these stations appear to have a

different seasons.

finite upper bound (e.g., positive skewness) for some durations. In the unified GEV model, the skewness parameter is required to be negative and to be unique for all durations, so that the final estimate is a summary of the properties of all

COMPARISONS OF THE UNIFIED GEV RESULTS TO CURRENT METHODS

durations. If the behaviour of the series at a station differs across durations, the ﬁnal estimates need to be a compro-

The estimated depths obtained with the methods currently

mise between the different tendencies of each series.

in use (FSR and FEH99) and the proposed uniﬁed GEV, cor-

Nevertheless, the ﬁnal ﬁt of the estimated frequency

responding to some pre-speciﬁed frequencies, are compared

curves compared to the annual maxima shown in Figure 5

to the empirical estimates obtained from the recorded data

seem to indicate that overall an acceptable ﬁt is obtained

series at each station. Since reliable estimates of very rare

for the series at Otterbourne. The estimated frequency

events cannot be obtained from the relatively short records

Figure 5

Estimated frequency curves for the station at Otterbourne for the annual (left panel), winter (central panel) and summer (right panel) series, superimposed on the Gringorten plotting positions, for durations from 15 to 120 minutes.

Page 110

1634

I. Prosdocimi et al.

A DDF analysis for short-duration rainfall events in England and Wales

Hydrology Research

48.6

2017

available (the median record length for the observed series is

to the FSR and FEH99 models for the 2-year events. In par-

24 years), the comparison is limited to the 2-, 5- and 10-year

ticular, the FSR appears to give consistently positively

return periods. The empirical estimates are obtained as the

biased estimates for the 2-year events (Figure 6), with

median (50th percentile), 80th percentile and 90th percen-

lower variabilities in the error for longer durations. The

tile of the recorded data series. For some series, the record

FEH99 seems to perform well on average, although the

length would be less than 2T years long when estimating

results are more variable than the uniﬁed GEV. The results

the 10-year event: these empirical estimates might be less

for the longer return periods show more variation, with the

precise. The comparison is performed on every station for

uniﬁed GEV performing slightly better in terms of the varia-

durations of at least 15 minutes, and the ﬁtted uniﬁed

bility of the error. The uniﬁed GEV, an at-site model ﬁtted

GEV models shown in Figure 4 are used to estimate the rain-

directly to the observed data, performs quite well for most

fall depths.

stations. Among the models currently used in the UK for

Figures 6–8 display the relative differences between the

rainfall frequency estimation, the FEH99 seems to give

rainfall depths, as estimated with the different methods, and

acceptable results, across all return periods, with an error

the empirical quantile corresponding to the speciﬁc frequen-

variability comparable to the FSR estimate.

cies for the 2-, 5- and 10-year return periods, respectively.

These comparisons are based only on empirical esti-

For example, the left panel of Figure 6 shows, for each

mates of events with a relatively short return period, and it

station and each duration, the value (R2FSR–R2Observed)/

is not clear how the different models differ for the estimation

R2Observed, where R2FSR and R2Observed indicate the esti-

of rare events, for which no reliable empirical estimates can

mated 2-year rainfall of the given duration at a station and

be obtained from the observed series. An assessment of the

the empirical 2-year event estimated from the observed

accuracy of the different estimation methods for longer

data, respectively. The uniﬁed GEV model is the only

return periods would, in fact, require reliable information

method directly ﬁtted to the observed data, which explains

on the real frequency of short-duration rainfall events,

the much better performance of that model in comparison

which cannot be easily retrieved. The overall relative

Figure 6

Relative difference between the 2-year rainfall depths estimated by different methods and the 50th percentile of the recorded series. Larger symbols correspond to stations with longer records.

Page 111

1635

Figure 7

I. Prosdocimi et al.

A DDF analysis for short-duration rainfall events in England and Wales

Hydrology Research

48.6

2017

Relative difference between the 5-year rainfall depths estimated by different methods and the 80th percentile of the recorded series. Larger symbols correspond to stations with longer records.

difference between the FEH99 and FSR, which were devel-

GEV model, estimated using only at-site data is investigated

oped with the purpose of allowing DDF estimation for the

in Figure 9. The ďŹ gure shows, for a large range of return

whole UK, and the estimates obtained from the uniďŹ ed

periods, the relative difference between the design events

Figure 8

Relative difference between the 10-year rainfall depths estimated by different methods and the 90th percentile of the recorded series. Larger symbols correspond to stations with longer records.

Page 112

1636

Figure 9

I. Prosdocimi et al.

A DDF analysis for short-duration rainfall events in England and Wales

Hydrology Research

48.6

2017

Relative differences between the FSR (left panel) and FEH99 (right panel) estimates and the uniﬁed GEV estimates for all stations and all durations. Thick lines indicate the average differences for all durations across all stations.

estimated by FSR and the uniﬁed GEV (left panel) and the

long return periods, it appears that the average difference

difference between the design events estimated by FEH99

between the uniﬁed GEV and the FEH99 estimates is smal-

and the uniﬁed GEV (right panel), for all stations and all

ler for events of long duration.

durations. The average relative differences across all stations for each duration are also shown. It should be noted that a large difference between the standard methods and the uni-

DISCUSSION AND CONCLUSIONS

ﬁed GEV estimates does not necessarily indicate poor performance of the standard methods: the uniﬁed GEV

An exploration of rainfall frequency estimation for short-

models are ﬁtted to the recorded data series, which are, at

duration events is presented. A new general at-site model,

most, 46 years long. It is therefore very likely that uniﬁed

the uniﬁed GEV, is proposed. The model is successfully

GEV estimates would be more accurate for shorter rather

used to estimate consistent annual and seasonal rainfall fre-

than longer return periods. Nevertheless, what is visible in

quency curves for a number of stations in England and

the plots is that the variability is much larger for the

Wales for which sub-hourly rainfall records exist. The pro-

longer return periods for all durations. Furthermore, the

posed model builds on the standard assumption that block

FSR seems to give consistently larger results than the uniﬁed

maxima follow a GEV distribution: the properties of the

GEV for short return periods, but the difference between the

GEV distribution are exploited to construct a uniﬁed

two estimates become smaller for return periods longer than

model which is ﬁtted to the data of different duration sim-

10 years. For the 15-minute events the difference is more

ultaneously. The structure of the proposed model is

marked and the FSR seems to give much smaller estimates

indeed quite innovative and different from most of the

than the uniﬁed GEV for longer return periods. The differ-

DDF models currently used in practice. The consistency

ence between the FEH99 and the uniﬁed GEV results

of the frequency curves is ensured by assuming that the

instead appears to increase for longer return periods,

lower bound and the skewness parameter are the same

although for shorter return periods (up to 5 years) the differ-

across all durations and by enforcing some basic relation-

ence in the two estimates is on average very small. At very

ships between the location and scale parameter and the Page 113

1637

I. Prosdocimi et al.

A DDF analysis for short-duration rainfall events in England and Wales

Hydrology Research

48.6

2017

event duration. The effect of the assumptions, enforced to

biased, especially in less recent years, which would under-

ensure the consistency of the frequency curves, is that

mine the quality of any estimation procedure.

curves for different durations diverge more and more as return period increases. The model might therefore give extremely large rainfall depth estimates for very long return periods. The model is designed to be fitted to block maximum series of sub-hourly data at single stations and does not have a procedure to integrate information from nearby stations in the rainfall frequency estimation. The estimation of the model parameters was carried out by maximum likelihood estimation, a procedure that attains some optimal properties when applied to large samples. The final model parameter estimates are influenced by properties of the observed data series and issues might arise when the actual properties of the observed series do not match well with the properties that are assumed

ACKNOWLEDGEMENTS This work was funded by the Environment Agency Project SC090031 – Estimating ﬂood peaks and hydrographs for small catchments (Phase 2). The original data series were provided by the Environment Agency. The authors thank Steven Cole at CEH Wallingford for providing code to process the raw time-of-tip series and David Jones for the fruitful discussions in the initial stage of the study. Part of the work presented in this study was carried out while the ﬁrst author was employed at CEH Wallingford, the support of which is gratefully acknowledged.

during model building. Nevertheless, the proposed model gives overall satisfactory results and ﬁts the empirical data quite well, using a relatively small number of parameters.

The

new

estimated

frequency

curves

are

compared to those obtained using the FSR and FEH99 methods currently employed in the UK. Although no subhourly data were used in the model calibration, the FEH99 method seems to give acceptable results for all of the sub-hourly durations under study, at least for the return periods for which reliable empirical rainfall frequencies can be estimated. The FSR estimates seem to overestimate the rainfall depths for short return periods, although the bias is less marked for longer return periods. In addition, the difference between the FEH99 and the FSR estimates becomes larger for rarer events. However, the comparisons could only be carried out on a small set of stations, and a more in-depth analysis would be needed to give a robust indication of the behaviour of the different models. Potentially, it could be useful to develop a full DDF model for short duration rainfall events at a national scale, in which information from different stations could be used in a unique framework. The relative scarcity of long and precise records of sub-hourly data would be the major obstacle to overcome in the potential development of a DDF model for the whole UK. Most of the available ToT records are fairly short and most are located in only a few areas of the UK. Due to the nature of tipping buckets, the measurement of very short duration events is likely to be Page 114

REFERENCES Babtie Group in association with CEH Wallingford and Rodney Bridle Ltd  Reservoir Safety – Floods and Reservoir Safety: Clariﬁcation on the use of FEH and FSR Design Rainfalls. Final report to the Department of the Environment, Transport and the Regions (DETR). Babtie Group, Glasgow, UK, 36 pp. Coles, S. G.  An Introduction to Statistical Modeling of Extreme Values. Springer, London, UK. Faulkner, D.  Rainfall Frequency Estimation. Volume 2 of the Flood Estimation Handbook. Institute of Hydrology, Wallingford, UK. Faulkner, D., Kjeldsen, T. R., Packman, J. C. & Stewart, E. J.  Estimating Flood Peaks and Hydrographs for Small Catchments: Phase 1. Environment Agency, Bristol, UK. Hosking, J. R. M.  L-moments: analysis and estimation of distributions using linear combinations of order statistics. Journal of the Royal Statistical Society B 52, 105–124. ICE  Floods and Reservoir Safety, 4th edn. Thomas Telford, London, UK. Jiang, P. & Tung, Y.  Establishing rainfall depth-durationfrequency relationships at daily raingauge stations in Hong Kong. Journal of Hydrology 504, 80–93. doi:10.1016/ j.jhydrol.2013.09.037. Kjeldsen, T. R. & Prosdocimi, I.  A bivariate extension of the Hosking and Wallis goodness-of-ﬁt measure for regional distributions. Water Resources Research 51 (2), 896–907. doi: 10.1002/2014WR015912. Kjeldsen, T. R., Prudhomme, C., Svensson, C. & Stewart, E. J.  A shortcut to seasonal design rainfall estimates in the UK. Water and Environment Journal 20, 282–286. doi:10.1111/ j.1747-6593.2006.00028.x.

1638

I. Prosdocimi et al.

A DDF analysis for short-duration rainfall events in England and Wales

Koutsoyiannis, D., Kozonis, D. & Manetas, A.  A mathematical framework for studying rainfall intensityduration-frequency relationships. Journal of Hydrology 206 (1), 118–135. doi:10.1016/S0022-1694(98)00097-3. MacDonald, D. E. & Scott, C. W.  FEH vs FSR rainfall estimates: an explanation for the discrepancies identiﬁed for very rare events. Dams Reservoirs 11, 280–283. Molini, A., Lanza, L. & La Barbera, P.  Improving the accuracy of tipping-bucket rain records using disaggregation techniques. Atmospheric Research 77, 203–217. doi:10.1016/ j.atmosres.2004.12.013. Natural Environment Research Council  Flood Studies Report. Natural Environment Research Council, Swindon, UK.

Hydrology Research

48.6

2017

Overeem, A., Buishand, A. & Holleman, I.  Rainfall depthduration-frequency curves and their uncertainties. Journal of Hydrology 348, 124–134. doi:j.jhydrol.2007.09.044. Prosdocimi, I., Stewart, L., Svensson, C. & Vesuviano, G.  Depth–Duration–Frequency Analysis for Short-Duration Rainfall Events. Environment Agency, Bristol, UK. Stewart, E. J., Jones, D. A., Svensson, C., Morris, D. G., Dempsey, P., Dent, J. E., Collier, C. G. & Anderson, C. A.  Reservoir Safety – Long Return Period Rainfall. Project FD2613 WS194/ 2/39 Final Report (two volumes). London, UK. Svensson, C. & Jones, D. A.  Review of rainfall frequency estimation methods. Journal of Flood Risk Management 3, 296–313. doi:10.1111/j.1753-318X.2010.01079.x.

First received 9 May 2016; accepted in revised form 11 September 2016. Available online 6 May 2017

Page 115

656

48.3

2017

From (cyber)space to ground: new technologies for smart farming Giovanni Ravazzani, Chiara Corbari, Alessandro Ceppi, Mouna Feki, Marco Mancini, Fabrizio Ferrari, Roberta Gianfreda, Roberto Colombo, Mirko Ginocchi, Stefania Meucci, Daniele De Vecchi, Fabio Dell’Acqua and Giovanna Ober

ABSTRACT Increased water demand and climate change impacts have recently enhanced the need to improve water resources management, even in those areas which traditionally have an abundant supply of water, such as the Po Valley in northern Italy. The highest consumption of water is devoted to irrigation for agricultural production, and so it is in this area that efforts have to be focused to study possible interventions. Meeting and optimizing the consumption of water for irrigation also means making more resources available for drinking water and industrial use, and maintaining an optimal state of the environment. In this study we show the effectiveness of the combined use of numerical weather predictions and hydrological modelling to forecast soil moisture and crop water requirement in order to optimize irrigation scheduling. This system combines state of the art mathematical models and new technologies for environmental monitoring, merging ground observed data with Earth observations from space and unconventional information from the cyberspace through crowdsourcing. Key words

| crowdsourcing, hydrological model, irrigation management, satellite observations, soil moisture, weather forecast

Giovanni Ravazzani (corresponding author) Chiara Corbari Alessandro Ceppi Mouna Feki Marco Mancini Department of Civil and Environmental Engineering, Politecnico di Milano, Piazza Leonardo da Vinci 32, Milano 20133, Italy E-mail: giovanni.ravazzani@polimi.it Fabrizio Ferrari Roberta Gianfreda Terraria srl, via Melchiorre Gioia 132, Milano 20125, Italy Roberto Colombo Mirko Ginocchi Remote Sensing of Environmental Dynamics Laboratory, DISAT, University of Milano Bicocca, Piazza della Scienza 1, Milano 20126, Italy Stefania Meucci MMI srl, via Daniele Crespi 7, Milano 20123, Italy Daniele De Vecchi Fabio Dell’Acqua Department of Industrial and Information Engineering, University of Pavia, via Ferrata 5, Pavia 27100, Italy Giovanna Ober CGS S.p.A., Via Gallarate 150, Milano 20151, Italy

INTRODUCTION Despite growing slower than in the recent past, the world

demand for water and food – not only due to a higher

population is projected to increase by more than one billion

number of people, but also to trends towards more water

people within the next 15 years, reaching 8.5 billion in 2030,

demanding lifestyles and diets. The agricultural sector is

and to increase a further 9.7 billion in 2050 and 11.2 billion

going to face enormous challenges in order to sustain food

by 2100 (United Nations Department of Economic and

production, which is required to increase by 70% by 2050.

Social Affairs Population Division ). Growth in popu-

Additional factors, such as climate change, will further

lation and income will imply a substantial increase in

contribute to affect water availability. Changes of average

doi: 10.2166/nh.2016.112

Page 117

657

G. Ravazzani et al.

Hydrology Research

New technologies for smart farming

48.3

2017

precipitation will not be uniform, with some regions experi-

each pixel of the domain. This can be achieved with those

encing increases, and others decreases, or not much change

models based on energy and water balance algorithms in

at all (Ravazzani et al. a). According to climate projec-

combination with remotely sensed data, in particular of

tions, the Mediterranean area should be affected by a

land surface temperature (LST) (Corbari & Mancini ).

decrease of total precipitation with the exception of the

The information content of satellite images may be

Alps in winter (Coppola & Giorgi ). With earlier snow

useful not only in providing a temporal dataset to calibrate

melting and rainfall variation, inter-annual run-off is chan-

and validate hydrological models, but even for assessment

ging towards less water during summer but more water

of biophysical attributes, such as leaf area index (LAI)

during winter (Dedieu et al. ; Gaudard et al. ).

(Colombo et al. ) and surface albedo (Corbari et al.

This would negatively alter the current seasonal cycle of

), and their temporal variation (Mattar et al. ).

runoff, even in those areas where mean annual precipitation

Thus, the combined use of physically raster-based hydrologi-

is expected to remain steady with negative impacts on agri-

cal models and satellite data may be an answer to the

cultural production.

question about extending prediction to larger ungauged

The increase in consumption of water resources, com-

areas.

bined with climate change impacts, calls for new sources

However, not all necessary information can be derived

of water supply (Ravazzani et al. ) and/or different man-

from satellite-based Earth observation. For example, veg-

agements of available resources in agriculture. One way to

etation height is an important piece of information, but it

increase the quality and quantity of agricultural production

is rarely used due to challenges in its extraction. Therefore,

is using modern technology to make farms more ‘intelligent’,

crowdsourcing becomes a valid, integrative source of infor-

the so-called ‘precision agriculture’ also known as ‘smart

mation, leveraging on the popularity of smartphones and

farming’.

tablets. Examples of applied crowdsourcing can be found

The scientiﬁc literature provides some studies focused

in different topics, from ﬁre mapping (Goodchild &

on ‘smart farming’ by coupling meteorological and hydrolo-

Glennon ) to risk management purposes (Bevington

gical models (Gowing & Ejieji ; Cai et al. ). Ceppi

et al. ).

et al. () demonstrated that in-advance prediction of

Sophisticated physically based hydrological models

soil moisture (SM) and crop water requirement allows a pre-

need more meteorological variables to compute water and

cise irrigation scheduling with beneﬁts on farmer income in

energy ﬂuxes. Besides the fact that full meteorological obser-

terms of reduction of water consumption and increase of

vations are not always available with sufﬁcient spatial

crop yield. However, their investigation was funded on

density, questions arise about reliability of meteorological

local analysis in one single cultivated site where ground

prediction by weather forecast models that are needed for

measurements of meteorological and hydrological variables

SM and crop water requirement forecast (Ceppi et al.

were acquired hourly. Moreover, they used only temperature

). Many studies have been devoted to analyze the accu-

forecast to predict evaporation by applying an empirical

racy of precipitation forecast and performance of hydro-

model (Ravazzani et al. ). Open issues still remain

meteorological coupled systems, mainly for the purpose of

about how to extend application to larger areas, and how

ﬂood forecasting (Amengual et al. ; Rabuffetti et al.

physically based methods that are fed the complete set of

; Ceppi et al. ; Pianosi & Ravazzani ; Senatore

meteorological variables can improve SM forecast.

et al. ; Arnault et al. ; Larsen et al. ). However,

Spatially distributed, physically based hydrological

accuracy of the forecast of other meteorological variables

models, with their ability to estimate energy and water

except precipitation and performance of meteo-hydrological

ﬂuxes at the agricultural district scale, are invaluable tools

systems for SM forecast still need in-depth investigation.

for water resources management for agricultural water use

The aim of this paper is to assess how mathematical

(Corbari et al. ). Satellite data, for their intrinsic raster

models for weather and hydrological simulations, together

structure, can be effectively used for the internal cali-

with new technologies in the ﬁeld of Earth observation

bration/validation of distributed hydrological models in

from space and technologies for getting information from

Page 118

658

G. Ravazzani et al.

Hydrology Research

New technologies for smart farming

48.3

2017

cyberspace (crowdsourcing), can help managing irrigation

). Winter is generally cold with a mean monthly tempera-

scheduling in a rich cultivated area in northern Italy. This

ture of 2 C in January and summer is hot and humid with a

work was part of the SEGUICI project, an Italian acronym

mean monthly temperature of 23.4 C in July (ERSAF ).

that stands for smart technologies for water resources man-

Evapotranspiration (ET) amounts can reach up to 500 mm

agement for civil consumption and irrigation.

during the summer season, therefore, most of the water

supply for agriculture comes from the irrigation network. In the period 2010–2012 one test-site of the Pre.G.I. pro-

MATERIALS AND METHODS

ject (Prediction and Guiding Irrigation) (Ceppi et al. ) was located in the central area of the basin at Cascina

Study area

Nuova farm, in Livraga town. Here, one meteorological and one eddy-covariance station and time-domain reﬂecto-

The studied area is the Muzza Bassa Lodigiana (MBL) con-

metry (TDR) probes were installed to measure mass and

sortium in the middle of the Po Valley, close to the city of

energy exchanges between soil, plant and atmosphere (Mas-

Lodi. The territory of the MBL covers an area of 740 km2

seroni et al. , ; Corbari et al. ).

where there are over 150 irrigation basins and thousands

In 2015, the monitoring station was moved from Livraga

of irrigation sub-basins with individual ﬁelds of landowners

to Secugnago site (Figure 1). In both the monitored ﬁelds,

(Figure 1).

farmers cultivated corn and ﬂood irrigation was scheduled

The Muzza canal (about 40 km long) derives water from

by the MBL consortium according to planned water allot-

the Adda river at Cassano d’Adda and it ﬂows back into the

ments that were determined in advance. Landowners

Adda river close to Castiglione d’Adda. Along the canal

cannot irrigate their ﬁelds on days other than the scheduled

there are 38 intakes and many more hydraulic nodes; the

ones (the Italian name of this irrigation scheduling method

entire Muzza network is composed by open earth canals.

is turno irriguo). On average, farmers can irrigate ﬁelds

The Muzza is both the largest irrigation canal by capacity

once every 2 weeks.

and the ﬁrst artiﬁcial canal built in northern Italy.

Speciﬁc ﬁeld campaigns were performed in order to

Average annual rainfall in the MBL consortium ranges

characterize soil properties. Soil water retention curve par-

between 800 (southern area) and 1,000 mm (northern area)

ameters for Livraga and Secugnago are reported in

with two peaks in spring and autumn (Ceriani & Carelli

Table 1. Sampling points were selected randomly within

Figure 1

The Lombardy region in the north of Italy (left) and the Muzza basin with its irrigation sub-basins (right).

Page 119

659

G. Ravazzani et al.

Hydrology Research

New technologies for smart farming

48.3

2017

both study sites. Samples were collected from different

used, daily provided at a 250-m spatial resolution. As regards

depths. The parameters presented in Table 1 are average

MOD09GA, surface reﬂectance data from Bands 1 to 7

values. Particle size distribution for each soil sample was

(from visible to infrared spectral range), daily provided at a

determined by wet sieving and hydrometer method. Sand,

500-m spatial resolution, and a speciﬁc dataset of reﬂectance

silt and clay contents together with soil texture were identiﬁed

data state quality assurance (to generate cloud-cover masks),

according to the United States Department of Agriculture

daily provided at a 1,000-m spatial resolution, were used.

system of soil classiﬁcation. Undisturbed soil samples were

First, by means of a binary mask (0-1) identifying the study

used to measure the saturated hydraulic conductivity follow-

area and a land-cover map of Lombardy region, a time-invar-

ing the falling-head method (Lee et al. ). Soil water

iant mask referred to Muzza basin was created, wherein a

retention curve parameters were deﬁned through evaporation

numerical value from 1 to 3 was assigned to every pixel, thus

method experiments (Wendroth et al. ) using the Hydrau-

carrying information about the corresponding land-use class

lic Property Analyzer device. Results were afterwards ﬁtted to

(i.e., crops, grasslands and agro-foresty areas; pixel not falling

the Brooks & Corey () parametric equation.

under these three classes were set to NaN). MOD09GQ and MOD09GA products were initially converted from their original sinusoidal projection to UTM

Satellite observations

Zone 32N WGS-84, with a pixel size of 250 m, by using In order to obtain a spatial estimation of some biophysical

MODIS Reprojection Tool in batch mode.

index,

In order to improve parameter estimation quality, a

NDVI; LAI; fractional cover, FC; albedo) and of the hydro-

time-variant cloud-cover mask was daily created from the

logical model variable LST over the Muzza basin, remote

reﬂectance data state quality assurance dataset. Then,

sensing data acquired from the moderate resolution imaging

spatial maps of NDVI, FC, LAI and albedo were created

spectroradiometer (MODIS) were chosen, and in particular,

for that day-of-year (DOY) throughout the study area.

parameters

(normalized

difference

vegetation

two types of surface reﬂectance data (from Collection-5

Reﬂectance (ρ) data in Bands 1 (R) and 2 (NIR) of

MODIS/Terra Land Products) used, namely, MO09GQ

MOD09GQ product were used for NDVI calculation over

and MOD09GA, both already atmospherically corrected

the study area, according to the classic formula:

for vegetation parameters’ retrieval. Data from the MODIS near real-time (NRT) context were used. As regards MOD09GQ, surface reﬂectance data in Bands 1 (red spectral range) and 2 (near infrared spectral range) were

NDVI ¼

ρNIR � ρR ρNIR þ ρR

(1)

The resulting matrix was weighed with the cloud-cover mask (resampled at 250-m pixel size) generated for that DOY; each pixel of NDVI matrix maintained its value only

Table 1

Soil water retention curve parameters for Livraga and Secugnago sites

Parameter

Livraga

Secugnago

Saturated water content [m3/m3]

0.389

0.379

Residual water content [m3/m3]

0.015

0.051

Field capacity [m3/m3]

0.33

0.301

Wilting point [m3/m3]

0.133

0.179

Saturated conductivity [m/s]

2.36 × 10�7

6.79 × 10�6

Brooks and Corey pore size index [�]

0.234

0.509

% Sand

32.73

71.94

% Silt

48.08

22.43

% Clay

19.19

5.63

Soil texture

Loam

Sandy loam

Page 120

if the corresponding pixel of cloud-cover mask was 1 (i.e., cloud-clear and cloud-shadow free), otherwise NDVI value was set to NaN. All the following analyses were carried out only if the percentage of cloud-pixels was lower than 50%, otherwise no map was created for the given DOY. Moreover, for every class, minimum and maximum NDVI values (ndviMIN and ndviMAX ) were computed by selecting (through frequency histogram calculation, assuming a uni-modal distribution) the lowest and the highest NDVI values, respectively, with a frequency of more than a certain threshold (e.g., 0.5% for crops class, which is the largest one). Then, maps of FC were calculated for every class, according to the empirical formula proposed by Richter &

660

G. Ravazzani et al.

Hydrology Research

New technologies for smart farming

FC ¼ 1 �

ndviMAX � ndvi ndviMAX � ndviMIN

48.3

2017

domain has been created with higher resolution, 3 × 3 km

Timmermans ():

(25 × 25 grid cells) nested within the national domain (Figure 2). (2)

The quite small size of the second domain was selected in order to keep the computational time within acceptable limits, and still provide satisfactory modelling results.

with p set to 0.9 (Campbell & Norman ); ndviMAX and

The model was set up using single-moment 6-class

ndviMIN (calculated as described above) are assumed to be

microphysics scheme (WSM6) containing ice, snow and

NDVI values of a surface fully covered and completely

graupel processes (Hong & Lim ), the Noah land sur-

uncovered by vegetation, respectively.

face model scheme (Tewari et al. ), the PBL Yonsei

By using FC values thus obtained, maps of LAI were cal-

University (YSU) scheme (Hong et al. ) and the CAM

culated for every class, according to Choudhury ():

scheme for radiation (Collins et al. ).

ln (1 � fc) LAI ¼ � 0:5

vations were assimilated in the model using WRFDA

To improve the estimation of the initial values, obser(3)

For the albedo, reﬂectance data of bands from 1 to 7 were used (Liang ):

þ 0:101ρB7

et al. ). Data taken in were derived from NCEP database and include: satellite radiances in BUFR format and conventional observations from land, ocean and upper-air platforms in PREPBUFR format.

ALBEDO ¼ 0:039ρB1 þ 0:504ρB2 � 0:071ρB3

þ 0:105ρB4 þ 0:252ρB5 þ 0:069ρB6

system (Barker et al. ) with 3DVAR techniques (Barker

Other weather data (temperature, wind speed, wind direction and pressure) were taken from meteorological (4)

stations of the Meteonetwork database. Finally, albedo and LAI data derived from satellite

Finally, we generated the 8-day composites of FC, LAI

observations of land cover were used to replace standard

and albedo: every day and for each of the three parameters,

values in WRF simulations for the higher resolution domain.

the map effectively returned as output is composed of pixels whose values are the maximum values that appeared over the last 8 days.

Hydrological modelling

The LST variable was also derived from NRT satellite imagery, for which MOD11_L2 product was used with a

Two distributed hydrological models were used for simulat-

1,000-m spatial resolution.

ing the water balance components: the ﬂash-ﬂood eventbased spatially distributed rainfall–runoff transformation,

Meteorological forecast

including water balance (FEST-WB) (Rabuffetti et al. ()) and the ﬂash-ﬂood event-based spatially distributed

The Weather Research and Forecasting-Advance Research

rainfall–runoff transformation, including energy and water

WRF version 3.61 (WRF-ARW) meteorological model was

balance (FEST-EWB) (Corbari et al. ()). The primary

used to generate daily meteorological forecasts with a forecast

difference between them is in the computation of ET. The

horizon of 9 days and a temporal resolution of 1 hour. These

FEST-WB model derives the actual ET by rescaling the poten-

weather outputs were used to drive the 1-day hydrological simu-

tial ET using a simple empirical approximation, where the

lations. Meteorological ﬁelds from the National Center for

potential ET is computed based only on air temperature

Environmental Prediction (NCEP) Global Forecasting System

measurements (Ravazzani et al. , a). In contrast, the

with 0.25 × 0.25 resolution were used as initial and boundary

FEST-EWB model computes the actual ET by solving the

conditions. For this study, the WRF computation domains com-

system of water mass and energy balance equations

prise the whole of Italy with 18 × 18 km horizontal resolution

(Ravazzani et al. a). The differences in the input par-

(58 × 68 horizontal grid cells) and 28 vertical layers. A second

ameters and meteorological forcings are listed in Table 2.

Page 121

661

Figure 2

G. Ravazzani et al.

Hydrology Research

New technologies for smart farming

48.3

2017

WRF simulation domains.

Both models discretize the computation domain with a mesh of regular square cells (200 × 200 m in this study), in each of which water ﬂuxes are calculated at hourly time step. In particular, SM dynamics, θ, for the generic cell at position i, j, is described by the water balance equation:

was designed to collect vegetation-related parameters (Figure 3(a)). The idea is to let everyone collect useful information, even contributors without a speciﬁc background; therefore, pictures of the most commonly found vegetation species are included as examples, to guide the end-users in deciding

@θi,j 1 ¼ (Pi,j � Ri,j � Di,j � ETi,j ) @t Zi,j

(5)

what species they have just taken a picture of. The mobile app allows including the height of vegetation, directly related to the stage of growth, and if the ﬁeld is ﬂooded or

where P is the precipitation rate, R is runoff ﬂux, D is drainage

not, useful information to know whether the farmer is irri-

ﬂux, ET is evapotranspiration rate and Z is the soil depth. For

gating the ﬁeld.

further details on distributed hydrological models and their

Every collected report, which includes a geocoded and

applications, readers may refer to Boscarello et al. ()

oriented picture and answers to the above-mentioned

and Ravazzani et al. (b, c, ).

group of questions, is automatically uploaded and stored in a remote database (Galeazzo et al. ). To avoid weighing on the contributor’s mobile data quota, an option can be

Crowdsourcing

activated to store reports on the hand-held device and upload them only when a WiFi connection becomes avail-

Based on the idea of volunteers (‘citizen sensors’) providing

able. A webgis interface is used to display data on an

information through their smartphones, a mobile app

OpenStreetMap-based map (Figure 3(b)). Within the

Page 122

662

Table 2

G. Ravazzani et al.

Hydrology Research

New technologies for smart farming

Meteorological forcings and parameters used as input to the FEST-WB and FEST-

48.3

2017

Calibration and validation of the FEST-WB model

EWB models

The 2010–2011 period was used to calibrate and the 2012 to

Input

Unit

FEST-WB

FEST-EWB

Precipitation

Temperature

Solar radiation

W/m2

Wind speed

m/s

Relative humidity

Saturated hydraulic conductivity

m/s

Residual moisture content

–

Saturated moisture content

–

Wilting point

–

Field capacity

–

Pore size index

–

Curve number

–

Soil depth

Vegetation fraction

Calibration and validation of the FEST-EWB model

The FEST-EWB model was calibrated during the period

Crop coefﬁcient

–

LAI

m2/m2

Albedo

–

Minimum stomatal resistance

s/m

Vegetation height

validate the FEST-WB model against SM and ET observations acquired at Cascina Nuova ﬁeld in Livraga. Only values of the parameters of the cell corresponding to the station site could be calibrated as there were no other stations with similar capabilities available in the consortium. Figure 7 shows the comparison between observed and simulated SM and ET, along with rainfall and irrigation amount, during the three growing seasons of 2010, 2011 and 2012. In general, satisfactory results are found in terms both of SM and ET during calibration and validation periods. More details and comments can be found in Ceppi et al. ().

2010–2012 against observed MODIS LST. Hence, soil hydraulic and vegetation parameters were calibrated in each single pixel minimizing the difference between the observed and simulated land surface temperatures, following the procedure developed by Corbari & Mancini ()

server, an algorithm can automatically associate the geo-

and Corbari et al. (). For the entire dataset of 166

localized reports with polygons related to each single ﬁeld

images, statistical parameters between LST from calibrated

using Global Positioning System (GPS) position and com-

FEST-EWB and LST from MODIS were computed: mean

pass direction (Figure 3(c)). Cooperation is in progress

absolute error (MAE) is equal to 0.2 C, root mean square

with the Research Support Service of the European Space

error to 1.8 C, relative error (RE) to 4.2% and the Nash &

Agency to share the collected data for their possible use in

Sutcliffe () index to 0.73. Cities areas were discarded

validation of land cover/use information derived from

from the comparison. In Figure 5, as an example, for 27

Earth observation satellite datasets.

August 2012 at 13:00, MODIS LST and FEST-EWB LST

images before and after the calibration, are shown. The FEST-EWB model was then validated against the

RESULTS AND DISCUSSION

ﬂuxes measured acquired at Cascina Nuova ﬁeld in Livraga. In Figure 4, cumulated ET over the 3 years was

Performance of the hydrological models

reported for observed data and for the calibrated FESTEWB. A RE of 5.6% was found between observations

The FEST-EWB and the FEST-WB models were calibrated

and ET from the calibrated model, while a RE of 44.1%

and validated following different procedures. In fact, the

was obtained if the non-calibrated ET was considered.

FEST-EWB model was calibrated distributed by comparison

SM estimates had a mean RE of 5.9%.

of simulated LST with the observed ones and validated

In general, the hydrological model FEST-EWB, after the

against local SM and ET, whereas the FEST-WB model

calibration procedure, is able to correctly reproduce distrib-

was calibrated locally and SM and ET were measured.

uted LST and local SM and ET during calibration and Page 123

663

Figure 3

G. Ravazzani et al.

New technologies for smart farming

Hydrology Research

48.3

2017

Mobile app graphic interface (a), WebGIS interface (b), and detail of automatic association of a report with the corresponding polygon according to compass direction (c).

Page 124

664

Figure 4

G. Ravazzani et al.

Hydrology Research

New technologies for smart farming

48.3

2017

Comparison between observed and simulated SM and ET at Livraga during 2010–2012.

validation periods (Figure 5). More details of this case study

simulations of both models for the 2015 growing season

can be found in Corbari et al. ().

were performed without a local calibration, but using their own parameters previously calibrated for 2010. Hence,

Comparison between the FEST-WB and the FEST-EWB models

FEST-WB soil parameters were only locally (e.g., Livraga) calibrated, while FEST-EWB ones were calibrated in a distributed way for each pixel of the analysed domain.

SM and ET estimates from FEST-EWB and FEST-WB were

Figure 6 shows the comparison between observed and simu-

then compared at local and basin scales for 2015. The

lated SM and ET, along with rainfall and irrigation, at

Page 125

665

Figure 5

G. Ravazzani et al.

Hydrology Research

New technologies for smart farming

48.3

2017

MODIS LST and RET images reported for the O-SoVeg conﬁguration and after calibration for 27 August 2012 at 13:00.

Secugnago station. SM from FEST-EWB better reproduces

respectively. If all the maps are analysed from 3 June to 30 Sep-

observed SM with a mean RE equal to 0.6% and a Nash &

tember 2015, the mean temporal differences of the spatial mean

Sutcliffe () index equal to 0.78. In contrast, SM from

is equal to 0.08.

FEST-WB has a mean RE of 18.5% with observed data and a

The same comparison was then performed for ET maps.

Nash & Sutcliffe () index equal to �0.13. Hence, SM from

In Figure 7, as an example, the FEST-EWB and FEST-WB

Observed ET at Secugnago site was available concur-

spatial mean and standard deviation are equal to 0.13 mm

rently to model simulations only from Day 154 (3 June) to

and 0.05 mm for FEST-EWB, while for FEST-WB they are

Day 171 (21 June) due to station malfunctioning. Cumulated

equal to 0.037 mm and 0.01 mm.

FEST-EWB and FEST-WB has a relative difference of 21.4%.

maps are reported for 30 September 2015 at 12:00. ET

ET from FEST-EWB and from FEST-WB were then com-

When the entire simulation period is considered, the

pared with observed values until 21 June and a RE equal

mean temporal differences of the spatial mean is equal

to 0.69% and to �7% was obtained, respectively (Figure 6).

to 1.1 mm. These differences are due to different model-

The difference between the two models in computing ET

ling schemes on ET and calibration procedures; in

over the whole growing season was equal to 42.7 mm.

particular, the FEST-WB was calibrated at local scale

FEST-EWB results were also compared at basin scale in each pixel of the domain with the output of the simpliﬁed ver-

only, while the FEST-EWB was calibrated pixel by pixel at basin scale.

sion of FEST-WB in terms of SM and ET. In Figure 7, for 30 September 2015, maps and histograms of simulated SM and

Impact of crowdsourcing data

ET from FEST-EWB and FEST-WB are reported. The SM spatial mean for FEST-EWB is equal to 0.22 with a standard

In order to assess how crowdsourcing data may affect accu-

deviation of 0.09, and for FEST-WB are 0.17 and 0.077,

racy of water balance, SM and ET were simulated with

Page 126

666

Figure 6

G. Ravazzani et al.

Hydrology Research

New technologies for smart farming

48.3

2017

Comparison between observed and simulated SM and ET at Secugnago during 2015.

FEST-EWB at the Secugnago site according to three scenarios: 1. Vegetation weight was changed according to crowdsourcing information acquired with the smartphone application, LAI and albedo were retrieved from remotely sensed images, and crop minimum stomatal resistance (rsmin) was set to 150 s/m. This is the reference scenario whose results were presented in previous sections.

2. We assume the ﬁeld was grass cultivated, and all vegetation parameters were assigned for a grass crop: height ¼ 0.12 m, LAI ¼ 1, rsmin ¼ 70 s/m.

3. We assume the ﬁeld was grass cultivated, height ¼ 0.12, rsmin ¼ 70 s/m, but LAI and albedo were taken from remotely sensed images.

Results are shown in Figure 8. Scenario 2 exhibits a signiﬁcant difference with respect to the reference Page 127

667

Figure 7

G. Ravazzani et al.

Hydrology Research

New technologies for smart farming

48.3

2017

Comparison of maps and histograms between simulated SM and ET from FEST-EWB and FEST-WB for 30 September 2015.

scenario, which means that water balance simulation may lose accuracy if the type of plant that is cultivated and its phenology are not known. The difference between scenario 3 and the reference scenario is lower because information retrieved from remotely sensed images can substantially compensate the lack of information about cultivated plants. As a general comment, the difference is greater when water supply is not enough to sustain ET and this is limited by vegetation parameters.

Veriﬁcation of the weather predictions The WRF meteorological model was daily launched from 3 June 2015 to 30 September 2015 in order to obtain weather forecasts over the two areas of study during the 2015 grow-

Figure 8

Comparison of SM and cumulated ET simulated under three different scenarios as described in the section ‘Comparison between the FEST-WB and the FEST-EWB models’.

ing season. The main meteorological ﬁelds available to feed

Table 3 highlights the performance of the WRF model

the FEST hydrological models were: air temperature and

forecasts in comparison with observed data for the

relative humidity, incoming shortwave solar radiation, pre-

entire forecast horizon of 9 days over the Secugnago

cipitation and wind speed.

site. The forecast of the day þ0, i.e., the forecast of the

Page 128

668

G. Ravazzani et al.

Hydrology Research

New technologies for smart farming

48.3

2017

same day of the model initialization, is here omitted, since

two schemes (ET computed with Hargreaves equation,

it would not be exploitable for irrigation scheduling

FEST-WB; and ET computed by solving the energy

management.

balance, FEST-EWB) for calculating SM at Livraga

Satisfactory results were found over the Secugnago test-

and Secugnago sites. Goodness of forecast was assessed

bed during the 4 months of simulations. In fact, air tempera-

by computing literature ﬁt indexes comparing SM

ture forecasts maintained a bias of about 2 C for the entire

simulated

forecast horizon; in particular, the WRF model tended to

observed meteorological forcings and SM obtained with

underestimate temperatures and even relative humidity fore-

the same hydrological models fed with meteorological

casts were 6–8% below the observed data; in contrast, the

forecasts.

FEST-WB

and

FEST-EWB

fed

with

incoming solar radiation, wind speed and daily precipitation

As shown in Tables 4 and 5, a better correlation (R2) was

were overestimated by the WRF model at about 80–90 W/

found using the energy-balance model in both the two sites,

m , 1.6–1.8 m/s and 2–7 mm, respectively. In general, no

Livraga and Secugnago, respectively; however, the MAE for

outliers were found during the analysed period and no sig-

SM shows fairly good results using both the Hargreaves and

niﬁcant decrement of the WRF performance at increasing

energy-balance equations during the entire forecast horizon,

of lead-time was present.

also due to a good performance of weather forecasts previously described; acceptable values, in fact, were

SM forecast and irrigation scheduling

found between 0.01 and 0.03 from day þ0 to day þ8,

respectively.

The beneﬁt of having a good coupled hydro-meteorolo-

Weather forecasts were afterwards used to drive the FEST Table 3

hydrological |

model

simulations

using

the

gical system many days in advance can be summarized in

MAE for the WRF meteorological model over the Secugnago area from day þ1 to day þ8 as lead time of forecast

MAE

Day þ1

Day þ2

Day þ3

Day þ4

Day þ5

Day þ6

Day þ7

Day þ8

Temperature [ C]

2.43

2.25

2.2

2.17

2.12

2.00

1.94

2.27

Relative humidity [�]

0.06

0.07

0.08

Daily precipitation [mm]

2.23

1.87

3.02

3.15

2.77

3.70

6.70

5.48

Incoming solar radiation [W/m2]

81.04

80.79

80.57

83.58

84.18

90.34

84.75

87.49

Wind speed [m/s]

1.65

1.63

1.71

1.61

1.60

1.58

1.67

1.77

Table 4

Performance for SM forecasts over the Livraga maize ﬁeld using Hargreaves equation (a) and the energy balance (b)

Livraga

dþ0

dþ1

dþ2

dþ3

dþ4

dþ5

dþ6

dþ7

dþ8

(a) SM – Hargreaves R2 [�]

0.88

0.80

0.81

0.80

0.75

0.70

0.68

0.63

0.54

MAE

0.01

0.02

0.03

MRE [%]

2.89%

3.99%

4.11%

4.61%

5.49%

5.81%

6.30%

7.59%

9.23%

(b) SM – EWB R2 [�]

0.94

0.88

0.86

0.82

0.78

0.75

0.71

0.63

MAE

0.00

0.01

MRE [%]

0.32%

0.38%

0.19%

0.08%

0.04%

�0.10%

�0.22%

�0.21%

�0.11%

Page 129

669

Table 5

G. Ravazzani et al.

Hydrology Research

New technologies for smart farming

48.3

2017

Performance for SM forecast over the Secugnago maize ﬁeld using Hargreaves equation (a) and the energy balance (b)

Secugnago

dþ0

dþ1

0.80

0.70

dþ2

dþ3

dþ4

dþ5

dþ6

dþ7

dþ8

(a) SM – Hargreaves R2 [�]

0.71

0.68

0.62

0.56

0.51

0.45

0.36

MAE

0.00

0.01

MRE [%]

0.83%

1.18%

1.27%

1.49%

1.67%

1.75%

2.04%

2.41%

2.90%

0.92

0.85

0.86

0.84

0.78

0.74

0.69

0.62

0.50

(b) SM – EWB R2 [�] MAE

0.01

MRE [%]

0.69%

0.18%

0.01

0.02

0.03

�0.88%

�1.65%

�2.28%

�3.17%

�3.62%

�3.79%

�3.58%

the following picture where a reanalysis for the period 22

Numerical weather predictions were provided by the

June 2015 to 15 July 2015 is shown. In this time span,

WRF-ARW meteorological model with a 10-day lead-time.

according to the MBL consortium regulation, irrigation

Two conﬁgurations of the FEST distributed hydrological

was scheduled on 30 June 2015 and 14 July 2015.

model were tested: the FEST-WB scheme that computes

Figure 9 shows accumulated precipitation and SM forecasts initialized 1 day before (dashed lines) and 8 days before

ET with the Hargreaves equation, and the FEST-EWB that solves the energy balance equation.

(solid lines) the planned irrigation of 30 June. The FEST-

The FEST-WB model was calibrated against SM and

EWB simulation, under the assumption that no irrigation

actual ET measured at Livraga station during the 2010–2012

occurred, is included as well. This demonstrates that irrigation

campaigns. Only parameters of the cells surrounding the Liv-

scheduled on 30 June was necessary in order to maintain SM

raga station could be calibrated as no other measurements

above stress threshold, since no signiﬁcant rainfall was pre-

were available in the MBL area. The FEST-EWB model was

dicted before the next planned irrigation allotment of 14

calibrated during the period 2010–2012 against observed

July, with a consequent high risk of compromising the crop.

MODIS LST. The two models were further validated against SM measured in the 2015 campaign at Secugnago. Comparisons with observations show that, while FEST-EWB was able

SUMMARY AND CONCLUSIONS

to properly simulate SM and ET, FEST-WB, which was not calibrated at Secugnago, showed greater error. Moreover,

This work was part of the SEGUICI project, the aim of which

the comparison of spatial distribution of SM and ET com-

was to develop and integrate smart technologies for water

puted by FEST-WB and FEST-EWB showed signiﬁcant

resources management for civil consumption and irrigation.

differences due to different methods used for their calibration.

The aim of this paper was to assess how mathematical

Calibration using remotely sensed images is an effective

models for weather and hydrological simulations, together

alternative to ground-based observations and provides

with remotely sensed images and crowdsourcing, can help

spatially distributed information impossible to acquire with

in managing irrigation scheduling, by forecasting SM and

conventional technologies.

crop water requirement. The test beds of the project were

Crowdsourcing resulted in a fundamental source of

two maize ﬁelds at Livraga (2010–2012) and Secugnago

information that could increase the accuracy of water bal-

(2015) in the MBL consortium, about 50 km south-east of

ance simulation, with maximum advantage occurring

Milan in northern Italy.

when combined with remotely sensed information.

The SM forecast was accomplished by coupling a

The performances of numerical weather predictions

meteorological model with the FEST hydrological model.

were assessed against air temperature and relative

Page 130

670

Figure 9

G. Ravazzani et al.

Hydrology Research

New technologies for smart farming

48.3

2017

Simulation of the FEST-EWB model without the scheduled irrigation of 30 June; accumulated precipitation and SM forecast by the WRF model initialized 8 days (solid lines) and 1 day (dashed lines) before the planned irrigation.

humidity, incoming shortwave solar radiation, precipi-

possible to get reliable SM forecasts for up to 1 week, and

tation and wind speed observations at Secugnago. Air

this helped farmers to properly decide irrigation scheduling.

temperature forecasts maintained a bias of about 2 C W

for the entire forecast horizon; in particular, the WRF model tended to underestimate temperatures and even relative

humidity

forecasts

were

6–8%

below

ACKNOWLEDGEMENTS

the

observed data. In contrast, the incoming solar radiation,

This work was sponsored by the Lombardy region in

wind speed and daily precipitation were overestimated

the framework of the SEGUICI project. We thank

by the WRF model at about 80–90 W/m , 1.6–1.8 m/s and

ARPA Lombardia (http://www.arpalombardia.it) and the

2–7 mm, respectively.

Meteonetwork Association (http://www.meteonetwork.it)

Weather forecasts were afterwards used to drive the

for providing meteorological observations from automatic

FEST-WB and FEST-EWB models for forecasting SM at Liv-

stations. The editor, Prof. Attilio Castellarin, and three

raga and Secugnago in the 2015 campaign. Goodness of

anonymous reviewers are gratefully acknowledged for

forecast was assessed by computing literature ﬁt indexes

their efforts to improve the quality and contents of this

comparing SM simulated by FEST-WB and FEST-EWB fed

manuscript.

with observed meteorological forcings and SM obtained with the same hydrological models fed with meteorological forecasts. SM forecast was reasonably satisfactory no matter whether the FEST-WB or the FEST-EWB was used. Moreover, results showed how combing meteorological and hydrological model that were correctly calibrated, it was

REFERENCES Amengual, A., Diomede, T., Marsigli, C., Martín, A., Morgillo, A., Romero, R., Papetti, P. & Alonso, S.  A hydrometeorological model intercomparison as a tool to

Page 131

Journal of

Water & Climate Change

ISSN 2040-2244 iwaponline.com/jwcc

Journal of Water and Climate Change covers all aspects of water science, technology, management and innovation in response to climate change, with emphasis on reduction of energy usage. The journal’s scope includes: • • • • • • • • • • • • •

Energy reduction technologies and strategies Hydrology Extreme events (floods, rainstorms, droughts) Energy and nutrient recovery in wastewater treatment Agricultural water use and climate change Water resource management including accounting, water reuse and demand management Technologies for reducing greenhouse emissions for water and wastewater treatment Carbon accounting in the water sector Targets and strategies for carbon emissions reduction Policy and practice of adaptation and mitigation of climate change in the water sector Predictive modelling of water resources Waterborne disease Inland and coastal waters, including both surface and ground waters For more details, visit iwaponline.com/jwcc

Page 134

320

08.2

2017

Assessment of climate change impact on crop yield and irrigation water requirement of two major cereal crops (rice and wheat) in Bhaktapur district, Nepal Lajana Shrestha and Narayan Kumar Shrestha

ABSTRACT Rice and wheat are major cereal crops in Nepal. Climate change impacts are widespread and farmers in developing countries like Nepal are among the most vulnerable. A study was carried out to assess the impact of climate change on yield and irrigation water requirement of these cereal crops in Bhaktapur, Nepal. Laboratory and soil-plant-air-water analysis showed silt-loam being the most dominant soil type in the study area. A yield simulation model, AquaCrop, was able to simulate the crop yield with reasonable accuracy. Future (2030–2060) crop yield simulations, on forcing the Providing Regional Climates for Impacts Studies (PRECIS) based on regional circulation model simulation indicated decreased (based on HadCM3Q0 projection) and increased (based on ECHAM5

Lajana Shrestha Narayan Kumar Shrestha (corresponding author) Center for Post-Graduate Studies, Nepal Engineering College (necCPS), Kathmandu, Pokhara University, Nepal E-mail: shrestha.narayan@hotmail.com Narayan Kumar Shrestha Faculty of Science and Technology, Athabasca University, Edmonton, Alberta, Canada

projection) yield of monsoon rice for A1B scenario, and rather stable yield (for both projection) of winter wheat. Simulation results for management strategies indicated that the crop yield was mainly constrained by water scarcity and fertility stress emphasizing the need for proper water management and fertilizer application. Similarly, a proper deﬁcit irrigation strategy was found to be suitable to stabilize the wheat yield in the dry season. Furthermore, an increase in fertilizer application dose was more effective in fully irrigated conditions than in rainfed conditions. Key words

| AquaCrop, climate change, PRECIS, rice, wheat

INTRODUCTION It is increasingly becoming clear that climate change is a

in order to have better crop yield (Maximay ). Moreover,

real phenomenon and human activities such as burning of

changes in precipitation pattern such as intense rainfall

fossil fuel, deforestation, and so on are primarily to blame.

during a particular month are becoming more frequent

As such, the recent anthropogenic emission of greenhouse

and such events could have a devastating effect on crop pro-

gases is at its highest level (IPCC ). The impacts of

duction especially if they occurred in a sensitive phase of the

increased temperature and elevated CO2 level, intense or

crop, e.g., the ﬂowering stage ( Joshi et al. ).

no rainfall, are widespread on natural systems and on

All these events associated with climate change would

humans (IPCC , ). Water resources are affected,

pose further stress on farmers to produce more and more

and hence the agricultural sector which could have long-

food for an increasingly growing and wealthier population

term effects on food security (Malla ; IFPRI ). It

(FAO ). The case is even more severe in peri-urban

is evident that increase in temperature and carbon dioxide

areas like Bhaktapur district, Nepal, which has experienced

(CO2) have a positive impact on some crops, but the nutrient

rapid urbanization of late (Shrestha et al. b). To cope

levels, soil moisture conditions, water availability for irriga-

with such adverse effects (of climate change), different adap-

tion, and other crop-related conditions should be favorable

tive management practices especially applying proper

doi: 10.2166/wcc.2016.153

Page 135

321

L. Shrestha & N. K. Shrestha

Climate change impact on major cereal crops of Bhaktapur, Nepal

Journal of Water and Climate Change

08.2

2017

fertilizer dose and water application need to be adopted.

(10,240 ha). Only 30% of the agricultural land has round-

Furthermore, adaptation practices such as shifting of crop

the-year irrigation facilities (Poudel et al. ). Because of

plantation date, adjusting cropping area and intensity

the very fertile nature of the land, the district is also known

might need to be considered (Iizumi & Ramankutty ).

as the grain and vegetable store of the valley. Rice, wheat,

There have been some studies which have quantiﬁed the

and maize are the major cereal crops of the district and are

impact of climate change on cereal crop yields at regional

grown in land areas of 4,326, 3,665 and 1,793 ha, respect-

and national scale. For instance, Lal () reported a

ively. The annual production of these cereal crops is 4.5,

decrease of 4–10% in cereal crop yield by 21st century in

2.69 and 2.93 t/ha, respectively (Poudel et al. ).

South Asia. Similarly, Shrestha et al. () reported, based on a study in Myanmar, that future climate condition

Data collection

would increase the paddy rice yield and would thus increase the food security in the region. Acharya & Bhatta () used

Several different techniques were applied for the data collec-

a quantitative modelling approach to calculate the impact of

tion and will be discussed in the next sections.

climate change on agricultural gross domestic product of Nepal, and found positive impact due to increased precipi-

Questionnaire survey, ﬁeld visits and soil samples

tation in future. Karn () reported a decrease of 4.2% in rice yield based on analysis made on 20 major rice growing

Altogether 30 soil samples (see Figure 1) from different

district of Nepal. Palazzoli et al. () carried out a study

locations of the district were taken based on snowball sampling

based on a physically based model in Indrawoti river basin

technique. Moreover, farmers of the sampled land were also

of central Nepal, and found different estimates (�36% to

supplied with questionnaires in order to collect information

þ18% for wheat, and �17% to þ12% for rice) of crop

regarding farming practices, main factors affecting the planta-

yield changes while using different future climate projection

tion of crops, crop yield, variety of crops, crop phonological

data. It is thus evident that very few studies have focused on

stages, and period, time and irrigation practices. To determine

peri-urban regions like Bhaktapur district.

the soil texture class, collected soil samples were submitted to

In this context, this paper analyzes the impact of climate

the Agricultural Technology Center (ATC), Nepal. Although

change on yield response of main cereal crops – rice and

the soil samples were taken from 30 cm depth, uniform soil

wheat in Bhaktapur district, Nepal. It also aims to ﬁnd

proﬁle is considered as suggested by Shrestha ().

ways to stabilize the yield with plausible water and fertilizer application scenarios. An understanding of the impacts of

Climate data

recent climate trends on major cereal crops would help to anticipate impacts of future climate on the agriculture

Daily historical climate data were collected from the Depart-

sector. We believe that the outcome of the study will facili-

ment of Hydrology and Meteorology (DHM), Nepal, for the

tate in formulating suitable adaptation strategies to cope

period 1979–2013 of nearby station named Tribhuvan Inter-

up with the adverse effects of climate change, thereby

national Airport (TIA), Nepal. It should however be noted

increasing food security of the district.

that there exist several meteorological stations in and around the Kathmandu Valley. Considering the fact that the Bhaktapur district is the smallest district of the valley

METHODS

and the TIA station is the nearest to the district, and most of the other stations are lying either on foothills or on the

Study area

hills, use of only one station (the TIA) located on a similar altitude as that of study area can be justiﬁed. The climate

The Bhaktapur district (Figure 1) is the smallest district of

data included daily rainfall, maximum and minimum air

Kathmandu Valley, Nepal. Although peri-urban, the district

temperature, sunshine hours, wind speed and relative

has about 80% of its total area as agricultural land

humidity. Figure 2 shows time series plots of annual rainfall,

Page 136

322

Figure 1

L. Shrestha & N. K. Shrestha

Climate change impact on major cereal crops of Bhaktapur, Nepal

Journal of Water and Climate Change

08.2

2017

The study area (Bhaktapur district) with soil sample points indicated by markers as per soil types: S1 (black triangle), S2 (green pentagon) and S3 (red circle). Also shown, in inset, is the map of Nepal and location of Bhaktapur district. Please refer to the online version of this paper to see this ﬁgure in color: http://dx.doi.org/10.2166/wcc.2016.153.

and minimum and maximum temperature at the station

used to calculate the potential evapotranspiration (ETo). The

during 1979–2013.

tool is based on a theoretical method proposed by Penman and Monteith (Allen et al. ) to calculate the ETo. The

Simulators/models used

tool requires several climatic data such as temperature (maximum and minimum), relative humidity, wind speed, solar

ETo calculator

radiation etc., at user deﬁned time steps. In this study, the tool was run for a daily time scale. Besides calculating the

The ETo calculator, developed by the Land and Water Div-

ETo, temperature data are also produced in a format suitable

ision of the Food and Agriculture Organization (FAO), is

for the AquaCrop model (see below for details on the model). Page 137

323

Figure 2

L. Shrestha & N. K. Shrestha

Climate change impact on major cereal crops of Bhaktapur, Nepal

Journal of Water and Climate Change

08.2

2017

Time series plot of annual rainfall volume (as grey column), and maximum (Tmax, as black diamond) and minimum (as black circle, Tmin) temperature at a meteorological station (Tribhuvan International Airport – TIA) used in the study.

Soil-plant-air-water model

as phonology, crop cover, root depth, biomass production

Soil-plant-air-water (SPAW) is a model which is generally

such as irrigation fertility, and ﬁeld agronomic practices

applied to simulate daily hydrologic water budgets of agri-

(Raes et al. ). The crop ﬁles for Nepal’s general crops

and harvestable yield, and ﬁeld management conditions

cultural landscapes. The embedded hydrologic analysis

were adapted from the study of Shrestha (). The model

involves the evaluation of soil water inﬁltration, conduc-

was built up using the subsequent model outputs of the

tivity, storage, and plant-water relationships (Saxton et al.

ETo calculator, SPAW model, and using information col-

). Soil characteristics such as textural composition,

lected in the questionnaire survey. The model is then

organic matter content, as obtained from the laboratory

calibrated based on the actual yield, also obtained from

tests, were supplied to the model which in turn simulated

the questionnaire survey.

permanent wilting point, ﬁeld capacity (FC), total available water (TAW), and saturated hydraulic conductivity (SAT)

Future climatic projection data

as outputs. These outputs are actually required by the AquaCrop model (see next section for details on the

Future climate projection data were fetched from the Nepal

model).

Climate Data Portal (NCDP ) with spatial resolution of

AquaCrop

mates for Impacts Studies (PRECIS), one of the widely used

25 × 25 km. The data were based on Providing Regional Clidynamical downscaling tools developed at Met Ofﬁce and AquaCrop is a crop-water productivity model which relates

Hadley Center, United Kingdom. The tool uses the atmos-

the soil, crop and atmospheric components. The soil com-

pheric component of the HadCM3 Global Climate Model

ponent

texture

– GCM (Gordon et al. ; Jones et al. , as cited in

composition, and for each textural class, hydraulic charac-

NCDP ). Data of a Regional Climate Model (RCM)

teristics (generally the results of SPAW model) are

run in PRECIS with imposed Lateral Boundary Condition

required. The atmospheric component requires rainfall (gen-

(LBC) as HadCM3Q0 and ECHAM5, both with A1B scen-

erally observed at a meteorological station), temperature

ario (IPCC ), were fetched in the NetCDF format

(generally the result of ETo calculator), evapotranspiration

(NCDP ). The future climate data (2030–2060) would

(generally the result of ETo calculator), and carbon dioxide

then be used for simulating ﬁeld management strategies

concentration (generally taken as default value of the year

for future climate change scenarios. As such, average data

2000 measured at Mauna Loa Observatory, Hawaii). The

of all pixels covering the study area were estimated using

crop component requires information about the crop such

ArcGIS.

Page 138

requires

soil

horizons

different

324

L. Shrestha & N. K. Shrestha

Climate change impact on major cereal crops of Bhaktapur, Nepal

Tested management scenarios

Journal of Water and Climate Change

08.2

2017

D2 are only applicable for winter wheat. Under the fertilizer management scenarios, different fractions of fertilizer as per

Different water and fertilizer scenarios (Table 1) were for-

National Recommended Fertilizer Dose (NRFD) were for-

mulated as plausible adaptation measures to cope with

mulated (Table 1). Finally, all possible permutations of

adverse climate change impact. The calibrated AquaCrop

water and fertilizer application scenarios were tested.

model was run to simulate the crop yield using these scenarios. Under water management scenario, rainfed (RF) and full irrigation (FI) conditions were formulated. These water management scenarios are currently practiced in the study

RESULTS AND DISCUSSION

area. Due to high household demand in the study area because of ever increasing population, and due to the need

Results from social survey

for allowing minimum discharge to maintain the ecological status of surface water, farmers are likely to face water scar-

It was found that 100% of the farmers grow rice during the

city and might not be able to use all the surface water

monsoon season (June to September), and the overwhelm-

sources for irrigation purposes in near future. Hence,

ing majority (>80%) grow wheat during the winter season.

another scenario (deﬁcit irrigation) in which limited water

During winter, the rest (20%) grow maize. The statistics

is applied at the most sensitive growing phase of the crop

indeed justiﬁed our selection of rice and wheat being two

(e.g. before and after ﬂowering), is tested too. Deﬁcit irriga-

major cereal crops of the study area. Besides, 80% of the

tion scheme is a rather promising and tested irrigation

respondents reported that they waited for rainfall in order

technique, especially in rain deﬁcient conditions (Geerts

to sow the crops, and the rest (20%) ﬁrst examined moisture

et al. , ; Geerts & Raes ; Shrestha et al.

content in the soil and then ﬁxed a sowing date.

c). As such, we tested two deﬁcit irrigation scenarios –

Respondents agreed that there had been a decrease in

D1 and D2 (see description in Table 1). While RF and FI

the crop yield, due to several factors including (see Figure 3):

conditions are applicable for monsoon rice, the D1 and

(a) spreading of disease (40%), (b) water scarcity (30%), (c)

Table 1

Water and fertilizer management scenario used in the simulation

Crop

Water Management

Fertilizer Management

Monsoon rice

(a) RF

(a) 150% of NRFD/Nonlimiting (b) 100% of NRFD (c) 50% of NRFD (d) 0% of NRFD

(b) Full Irrigation (FI): (Soil Water Content (SWC) maintained up to 100% FC) Winter wheat

(a) RF (b) Full Irrigation (SWC maintained up to 30% of FC) (c) Deficit Irrigation Strategy D1 (two application of 1/6 Net Irrigation Requirement (Inet) each before and around flowering) (d) Deficit Irrigation Strategy D2 (three applications of 1/6 Inet each, one before, around and after flowering)

(a) 150% of NRFD/Nonlimiting (b) 100% of NRFD (c) 50% of NRFD (d) 0% of NRFD

NRFD: National Recommended Fertilizer Dose in which following chemical content is needed (MoAC ) Component

Fertilizer component required in rice (wheat) expressed in kg/ha

Nitrogen

100 (100)

Phosphorus

30 (50)

Potassium

30 (25)

Zinc sulphate

10 (–)

Page 139

325

L. Shrestha & N. K. Shrestha

Figure 3

Climate change impact on major cereal crops of Bhaktapur, Nepal

Journal of Water and Climate Change

08.2

2017

Major factors affecting the crop growth.

decrease soil fertility (10%), (d) lack of seed varieties (10%), and (e) lack of fertilizer and technology (5% each). Reported crop yield from the farmers (Table 2) of monsoon rice (4.9 t/ha) is comparable to the ﬁndings of MoAC () which stand at 4.5 t/ha. This is also the case for the winter wheat in which responded reported yield (2.5 t/ha) nearly matches with the MoAC () value of 2.69 t/ha. Finally, the phenological periods for monsoon rice and

Figure 4

Textural triangle of collected soil samples.

winter wheat are obtained from the questionnaire survey and it is reported that plantation dates for rice and wheat

increased from 0.5 to 3%, the TAW would double

were July 1 and December 15, respectively.

(Hudson ).

Results from laboratory tests of soil samples

Model results

The extracts of laboratory analysis report have been plotted

ETo-calculator results

on the textural triangle developed by Gerakis & Baer () which indicated that 80% is classed as ‘Silt Loam’,

ETo calculator revealed that the daily ETo have a decreasing

while 16% as ‘Loam’, and remaining 4% as ‘Sandy Loam’

trend for the base period (1979–2013) which is contrary to

(Figure 4). Moreover, organic matter content of the soil

expectation as it is perceived that there would be rise in

samples is found to be below 3%. Increased organic

temperature due to climate change. However, as can be

matter increases water holding capacity and conductivity

seen in Figure 2, the maximum and minimum temperatures

(Saxton & Rawls ). If organic matter content is

are rather stable in the last decade or so. The decreasing ETo trend could then be due to the higher humidity levels in the

Table 2

Mean reported yield from respondents

atmosphere and decreased amount of solar radiation reaching the Earth’s surface. It is well perceived that a small

Rice yield (t/ha) Year (as of 2013)

Before 10 years

Maximum

Minimum

Average

11.9

3.0

8.2

Last year

8.9

2.5

5.8

This year

6.9

1.9

4.9

Year (as of 2013)

Wheat yield (t/ha) Maximum Minimum

Average

Before 10 years

6.7

2.0

3.8

Last year

4.0

1.3

2.5

This year

4.0

1.0

2.5

Page 140

change in solar radiation can bring large amount of change in evapotranspiration (Gad & Gyar ). SPAW model results Based on the soil physical characteristics as determined using Pedo-transfer functions from soil texture using the SPAW model (Saxton & Rawls ), we classiﬁed the soil samples into three classes namely S1, S2 and S3. The classiﬁcation was based on the range of TAW values (refer

326

L. Shrestha & N. K. Shrestha

Climate change impact on major cereal crops of Bhaktapur, Nepal

Journal of Water and Climate Change

08.2

2017

to Table 3). Other characteristics of the soil are presented in

yield) are available, and the simulated yield falls into the

the Table 3. Spatial distribution of soil samples is presented

range of yield as reported (by the respondents) in the ques-

in Figure 1. It is clear that S3 was the most dominant type in

tionnaire survey. However, it should be noted that

the study area.

availability of continuous yearly crop yield data would

AquaCrop model results – model calibration

us to check the accuracy of the model calibration using well

better reflect the accuracy of model calibration. This limited established goodness-of-fit statistics such as bias, coefficient The calibration results of monsoon rice and winter wheat

of correlation, etc. We found that the provision of a 100%

yield, for soil type S1 are shown in Figures 5 and 6, respect-

dose of fertilizer as recommended by NRFD and RF con-

ively. As can be seen, only three data points (of the observed

dition better matched the yield of latest data (the year of 2013) for both crops, which was somehow expected as most of the farmers (80%, details in ‘Results from social

Table 3

Type

Total soil sample

TAW (mean) mm

SAT (mean) vol %

FC (mean) vol %

WP (mean) vol %

100 to 140

150 to 180

190 to 220

Classiﬁcation of collected soil samples

survey’ section) depend on the rainfall occurrence for

Figure 5

AquaCrop model calibration results for monsoon rice (soil type S1).

Figure 6

AquaCrop model calibration results for winter wheat (soil type S1).

sowing and plantation, and in later stages of crop development too. It is clear that monsoon rice yield does not seem to be too sensitive to the total rainfall during crop season. While large variation (600–1,395 mm) in the rainfall is evident,

Page 141

327

L. Shrestha & N. K. Shrestha

Journal of Water and Climate Change

Climate change impact on major cereal crops of Bhaktapur, Nepal

the variation in the monsoon rice yield (2.9–4.2 t/ha) is sup2

08.2

2017

The monsoon rice yield in different fertilizer and water

pressed (Figure 5). The coefﬁcient of correlation (r )

management scenarios is presented in Table 4. As can be

between them is thus very low (0.04). However, winter

seen, in the RF case, increment in fertilizer application

wheat yield is very sensitive to total rainfall occurring

from 0% to 150% of the NRFD resulted in a signiﬁcant

during crop season (Figure 6). Large variation in rainfall

increment in the yield (up to 65%), and the result for the irri-

(25–210 mm) is also reﬂected in large variation in the yield

gated (FI) case is even higher (up to 74%). The net

(0.1–4.4 t/ha) with higher r2 value of 0.37 between them.

contribution of irrigation in the yield increment is below

It therefore implies that RF irrigation can be practiced for

4%. This implies that increment in fertilizer application

monsoon rice while winter wheat needs irrigation infrastruc-

should be practiced while the provision of even full irriga-

ture to ensure timely irrigation and better yield.

tion would barely be beneﬁcial. It might also be due to the

Although the calibration result for soil type S1 is pre-

fact that rainfall is enough during the crop growing period,

sented in Figure 5 (for monsoon rice) and Figure 6 (for

and the development of an irrigation system might not be

winter wheat), the yield scenario for each soil type is

economically viable for rice anyway. These ﬁndings are con-

shown in Figure 7. As can be seen, the yield on type S3 is

sistent with the ﬁndings of Shrestha et al. (c) for the

the highest and has the lowest variation in terms of maxi-

southern plain region (Terai) of Nepal.

mum and minimum yields, which are mainly due to the

The same for the winter wheat (Table 5) illustrated a

higher TAW retaining capacity of S3 (see Table 3). Higher

rather different picture. Winter wheat yield could substan-

TAW means that the soil can hold more moisture in a pro-

tially be increased (up to 110%) by providing optimal

longed no-rain case.

fertilizer dose, and the contribution of irrigation is also

Figure 7

Table 4

AquaCrop simulated monsoon rice (left) and winter wheat (right) yield in the base period (1979–2013) for different soil types (S1, S2 and S3).

Contribution of fertilizer and/or irrigation on monsoon rice yield RF

Full irrigation (FI)

Yield

Increase by fertilizer

Yield

Increase by fertilizer

Increase by irrigation

Fertilizer application dose

t/ha

150% of NRFD

5.23

5.44

100% of NRFD

4.35

4.41

50% of NRFD

3.86

3.89

0.5

0% of NRFD

3.16

–

3.17

–

0.3

AquaCrop simulation results in base period (1979–2013) for soil type S1. NRFD: National Recommended Fertilizer Dose.

Page 142

328

Table 5

L. Shrestha & N. K. Shrestha

Journal of Water and Climate Change

Climate change impact on major cereal crops of Bhaktapur, Nepal

08.2

2017

Contribution of fertilizer and/or irrigation on winter wheat yield increment

RF Increase by Fertilizer application dose

Yield t/ha

Deﬁcit irrigation (D1)

Full irrigation (FI)

fertilizer %

Yield t/ha

Deﬁcit irrigation (D2)

Increase by

fertilizer %

irrigation %

Yield t/ha

fertilizer %

irrigation %

Yield t/ha

fertilizer %

irrigation %

150% of NRFD

2.71

4.62

110

3.67

4.13

100% of NRFD

2.49

3.31

3.65

50% of NRFD

2.34

3.03

2.7

2.88

0% of NRFD

1.94

–

2.2

–

2.07

–

2.15

–

AquaCrop simulation results in base period (1979–2013) for soil type S3. NRFD: National Recommended Fertilizer Dose.

substantial (up to 71%), unlike that observed in the case of

HadCM3Q0 based simulation showed a marked drop

monsoon rice (only up to 4%). Yield response indicated

(�63.78%, see Table 6) in monsoon rice yield in future

that the increment from the increased fertilizer dose

period (2030–2060) as compared to the base period

would further be enhanced in the case of FI rather than

(1979–2013)

RF and deﬁcit irrigation (D1, D2). While it is evident

while

the

ECHAM5

based

simulation

showed þ20.5% increment (see Table 6). Such signiﬁcant

that the yield was highest with full irrigation, the yield at

differences in the yield when using two different climate

deﬁcit irrigation schemes (D1 and D2) would not be too

projection data would surely have implications for policy

low, especially for the D2 case. Hence, proper deﬁcit irriga-

makers. It should however be noted that different results

tion schemes could be the option for the winter wheat

when using different climatic data sets have been widely

along with optimal fertilizer application in order to

reported in the literature. McSweeney & Jones ()

increase the yield.

related such uncertainties to differences in the climate model formulation and the adopted downscaling tech-

AquaCrop – future crop yield in climate change scenario

niques.

Lately, researchers

have

used

Mean

Model

Ensemble (MME) as future climate data in simulation For all the fertilizer and water management scenarios,

models (McSweeney & Jones ) which can be perceived

model simulation showed two contrasting results when

as balancing the extremes of the climate models. Some

different future (2030–2060) climatic data (HadCM3Q0

researchers (e.g. Prudhomme ) thus have warned of

and ECHAM5) were forced. For instance, in calibrated

‘misleading conclusions’ derived from different climate

conditions (fertilizer: 100% of the NRFD, and RF), the

change projection models.

Table 6

Summer rice yield (mean) response in future (2030–2060) relative to baseline period (1979–2014) for both HadCM3Q0 and ECHAM5 forcings RF

Full irrigation (FI) Yield change in future

Yield change in future

Base period yield

HadCM3Q0

ECHAM5

Base period yield

HadCM3Q0

ECHAM5

Fertilizer application dose

t/ha

150% of NRFD

5.23

5.44

20.50

4.41

�36.03

20.04

4.39

�65.97

24.86

100% of NRFD 50% of NRFD

3.89

19.28

3.86

0% of NRFD

3.17

17.98

3.12

�63.78 �62.21 �57.41

�35.60 �35.49 �34.94

19.95 19.69 19.55

Page 143

329

L. Shrestha & N. K. Shrestha

Climate change impact on major cereal crops of Bhaktapur, Nepal

Journal of Water and Climate Change

08.2

2017

The HadCM3Q0 based simulation results (drop in

the ﬁndings of Acharya & Bhatta () who analyzed the

summer rice yield) are in line with the ﬁndings of: (a)

climate change impact in agricultural gross domestic pro-

Karn () who reported about 4% drop in rice yield,

duct of Nepal. Moreover, results showed that there exists

based on the analysis made on 20 districts of Nepal; (b)

greater uncertainty in future crop yield as indicated by

Lal () who reported drop of 4–10% for the South

wider error bars of the box plots as compared to that of cur-

Asia; and (c) Palazzoli et al. () who found a wide

rent case. As expected, the width or variation in box plots

range (�17% to þ12%) when using different climate projec-

are less in the FI case than that in the RF case which is

tions, in Indrawoti river basin of central Nepal. While other

apparently due to lower water stress on the crop. Such a sig-

studies (e.g., Joshi et al. ; Bhatt et al. ; Shrestha et al.

niﬁcant decrease (�36%, for highest fertilizer application

) reported the opposite. As can be seen in Figure 8, the

case �150% NRFD) even for full irrigation (see Table 6)

yield would even drop to near zero level. The yield improved

would mean that temperature stress is the main factor

when full irrigation was introduced but is still lower than

behind such a decrease.

current yield. Unlike that observed in the base period (see

In contrast, the ECHAM5 based simulation showed

Table 4), increasing fertilizer application did not improve

that the crop yield would increase in future. This ﬁnding

the yield in future. These ﬁndings (signiﬁcant increment

is in line with the results of Shrestha et al. () in Myan-

from the provision of full irrigation and negligible contri-

mar; Bhatt et al. () in Koshi river basin, Nepal; and

bution from increasing fertilizer dose) are consistent with

Joshi et al. () across Nepal. For all cases (RF and FI,

Figure 8

AquaCrop simulated monsoon rice yield represented as box plots for (a) rainfed (RF), (b) full irrigation (FI). Mean values are shown as black diamond for base/current (CU), and red triangle for future (FU) based on HadCM3Q0 forcing and blue circle based on ECHAM5 forcing. NL, 100, 50, 0 represents fertilizer application scenario as Non-Limiting, 100%, 50% and 0% as per National Recommended Fertilizer Dose (NRFD), respectively.

Page 144

330

L. Shrestha & N. K. Shrestha

Climate change impact on major cereal crops of Bhaktapur, Nepal

Journal of Water and Climate Change

08.2

2017

and all fertilizer application conditions), the ECHMA5 based simulation showed higher yield than current yield. Furthermore, there exists less variation, as indicated by narrower box plots, in the yield which indicates that there would be sufﬁcient rainfall in the monsoon rice growing season, should the ECHAM5 projection prevail in future. The future simulation results for winter wheat showed different results than that observed for summer rice (Figure 9). The HadCM3Q0 based future simulation showed increment in winter wheat yield in all water management

and

fertilizer

application

scenarios

which

indicates that rainfall and temperature during the winter wheat growing season would be favorable. The ECHAM5 based future simulation however showed increment in certain scenarios and drop in others. In improved water management scenarios (e.g., full irrigation, D1), the future yield is always expected to be higher than current yield (see Figure 9(b) and 9(c), and Table 7). The HadCM3Q0 based simulation results also showed that even deficit irrigation schemes (D1 and D2) would result in better yields (see Figure 9(c) and 9(d), and Table 7). The ECHAM5 based simulation results however showed that the yield would decrease (Table 7) especially in D2 case. While it is not clear if the monsoon rice yield would increase or decrease in future as both future climate data set indicated contrasting result, it is rather clear that the yield of winter wheat can easily be stabilized or even increased adopting proper water management scenarios (FI or D1). Such significant uncertainty in future yield of monsoon rice is indeed a dilemma for policy makes, hence, an effort was made in investigating what caused such a drastic decrease in monsoon rice yield when forcing the HadCM3Q0 projection. It was found that significant temperature stress (consequently higher evapotranspiration and higher the demand of irrigation water) would be the main reason behind the sharp decrease in yield if HadCM3Q0 projection prevail in future (see Figure 10, left). During the base period the temperature stress is very low (almost near zero) and variation of temperature stress is also very low (as indicated by narrower

Figure 9

AquaCrop simulated winter wheat yield represented as box plots for (a) rainfed (RF), (b) full irrigation (FI), (c) deﬁcit irrigation scheme 1(D1), (d) deﬁcit irrigation scheme 2 (D2). Mean values are shown as black diamond for base/current

box plots). In contrast, the HadCM3Q0 based projected

(CU), and red triangle for future (FU) based on HadCM3Q0 forcing and as blue

would lead to rather signiﬁcant temperature stress, ranging

scenario as Non-Limiting, 100%, 50% and 0% as per National Recommended Fertilizer Dose (NRFD), respectively.

from nearly 40% to 10% (Figure 10, left). Extreme

circle based on ECHAM5 forcing. NL, 100, 50, 0 represents fertilizer application

Page 145

13.02

8.33

�3.29

13.49

9.38

�11.14 6.05

08.2

2017

temperature indeed has a negative effect on photosynthesis, primary and secondary metabolism, and stability of various proteins, membranes and cytoskeleton structures, resulting

5.48

HadCM3Q0 %

in low yield. The effect is more pronounced in the reproductive stage. Furthermore, water stress due to low rainfall and high evapotranspiration demand also contributed to the low yield. When a plant does not get a sufﬁcient amount of water

2.15

2.88

4.13

3.65

cantly. According to Shrestha (), crop does not progress well at temperature below 8 C and above 30 C.

7.25

2.59

2.18

2.72

Furthermore, there would be rather unfavorable rainfall

15.46

occurrence and distribution in monsoon rice crop growing 14.44

HadCM3Q0 %

high and low temperatures affect the crop progress signiﬁW

18.26

ECHAM5 %

yield t/ha

for growth, then yield will certainly decrease. Both extreme

14.50

Yield change in future Base period Yield change in future

season (see Figure 10, right). The cumulative rainfall (mean) for the season would be less than 500 mm if

2.07

2.7

3.67

3.31

To further analyze the case, for a purposively selected

22.27

22.77

22.51

22.50

year (2044), it was found that temperature seems to drop below 8 C (even reaching below freezing) for a prolonged W

time of almost 3 months if HadCM3Q0 projection prevail

23.64

23.76

23.81

in future (Figure 11, top right). Similarly, there would be 23.75

yield t/ha HadCM3Q0 %

fall with that occurring in the base period.

very limited rainfall during the rice growing season (Figure 11, top left). Rather, the rainfall peaks seem to be shifted to earlier months with the highest during mid-March. This implies that

2.2 �11.96

3.03

�20.43

4.62

�22.96

�22.09

yield t/ha ECHAM5 %

rice’s plantation date need to be shifted so as to beneﬁt the ample rain. Furthermore, the minimum temperature also seems favorable for shifting of the plantation date. This issue (shifting crop plantation month) is further investigated, results of which have been presented in the next section. If

4.89

there would be rather favorable distribution of rainfall 7.39

28.89

ECHAM5 projection prevails, it is clear from the plots that 26.10

HadCM3Q0 %

Yield change in future

Base period

Full irrigation (FI)

ECHAM5 projection indicates comparable cumulative rain-

ECHAM5 %

Base period

HadCM3Q0 projection prevails in future while the

Yield change in future

Deﬁcit irrigation 1 (D1)

Journal of Water and Climate Change

Climate change impact on major cereal crops of Bhaktapur, Nepal

(Figure 11, bottom left) and temperature barely drops below 8 C (Figure 11, bottom right) meaning that there would be

1.84

2.3 50% of NRFD

0% of NRFD

2.7

2.49

150% of NRFD

100% of NRFD

yield t/ha

minimal water and temperature stress.

Fertilizer application dose

Base period

Winter wheat yield (mean) response in future (2030–2060) relative to baseline period (1979–2014) for both HadCM3Q0 and ECHAM5 forcings

| Table 7

Page 146

ECHAM5 %

L. Shrestha & N. K. Shrestha

Deﬁcit irrigation 2 (D2)

331

AquaCrop – shifting crop plantation season to stabilize crop yield The previous simulation results, in the case of monsoon rice, indicated a marked decrease in yield in all possible management scenarios, should HadCM3Q0 projection prevail in future. Moreover, rainfall and temperature seem to favor

332

L. Shrestha & N. K. Shrestha

Journal of Water and Climate Change

Climate change impact on major cereal crops of Bhaktapur, Nepal

08.2

Figure 10

Temperature stress (left) during the crop growing season for monsoon rice for base and future period. Also shown is the rainfall occurrence during season (right).

Figure 11

Rainfall distribution pattern for the year of 2044 (left) and minimum temperature pattern (right).

2017

an earlier plantation date (mid-March, see Figure 11). As an

and FI conditions under optimal fertilizer application dose

adaptation measure for the climate change impact, and in

(Figure 12). Even under FI conditions, the tradition planta-

order to stabilize the monsoon rice yield, crop plantation

tion date (July) of monsoon rice would give almost the

months were arbitrarily shifted and simulations with both

lowest yield, mainly due to temperature stress as minimum

water management scenarios, RF and FI, are carried out.

temperature in subsequent crop growing months tends to

It has to be noted that the worst future climatic scenario

reach below 8 C. W

(HadCM3Q0 projection) has been considered here, as simulation based on ECHAM5 projection showed increment in

AquaCrop â&#x20AC;&#x201C; net irrigation water requirement

monsoon rice yield. Simulations showed that the March plantation of mon-

The simulation result of net irrigation water requirement

soon rice would result in the maximum yield for both RF

(Inet) also indicates the severe water stress that the monsoon Page 147

333

Figure 12

L. Shrestha & N. K. Shrestha

Climate change impact on major cereal crops of Bhaktapur, Nepal

Journal of Water and Climate Change

08.2

2017

Monsoon rice yield in rainfed (RF) condition (black diamond) and full irrigation (FI) condition (red triangle) in each months of future (HadCM3Q0 based) period under Non-Limiting (NL) fertilizer application. Please refer to the online version of this paper to see this ﬁgure in color: http://dx.doi.org/10.2166/wcc.2016.153.

rice would face (Figure 13, left), should HadCM3Q0 projec-

that less rainfall is expected during winter wheat’s growing

tion prevail in future. The Inet (155 mm) during the base

season which is also evident in Figure 11 (bottom left).

period increased signiﬁcantly to 317 mm which is comparable to Inet value reported by Shrestha et al. (a). Should ECHAM5 projection prevail in future, the favorable

CONCLUSIONS

rainfall distribution during rice growing season has been reﬂected in a very low value (<60 mm) of Inet. With this

An assessment of climate change impact on irrigation water

hindsight, possible adaptation measures might be the pro-

requirement and crop yield of two widely used cereal crops

vision of irrigation facility and shifting of monsoon rice

in Bhaktapur district, Nepal, was made with the help of

plantation date.

social and analytical (using various models) techniques.

On the other hand, the main reasons for a rather stable

Questionnaire survey with 30 farmers, selected using snow-

winter wheat yield (based on HadCM3Q0 projection) are

ball sampling technique, was carried out to gain insights on

due to favorable rainfall distribution and lessened tempera-

the crop, water and fertilizer management practices, and

ture stress (Figure 10, left). However, the wider range of the

harvested yield. Moreover, soil samples from the croplands

error bars of the box plot of Inet, meaning higher variability,

of the selected farmers were taken and later analyzed in a

is of concern to policy makers. Furthermore, the Inet based

laboratory to determine texture composition and organic

on ECHAM5 projection is higher than base period indicating

matter content. SPAW tool was used to determine physical

Figure 13

Net irrigation water requirement in base (black diamond) and in future (HadCM3Q0 – red triangle and ECHAM5 – blue circle) for monsoon rice (left) and winter wheat (right). Please refer to the online version of this paper to see this ﬁgure in color: http://dx.doi.org/10.2166/wcc.2016.153.

Page 148

334

L. Shrestha & N. K. Shrestha

Climate change impact on major cereal crops of Bhaktapur, Nepal

Journal of Water and Climate Change

08.2

2017

characteristics of the samples, and the ETo calculator was

Nepal, to conduct this study. The authors are very grateful

used to estimate daily potential evapotranspiration series.

for Dr Nirman Shrestha and Mr Pabitra Gurung for

To study the crop-yield response on forced climatic, crop,

providing valuable suggestions.

soil, and management data, a yield simulation model namely, the AquaCrop model was calibrated. To realize the possible impacts of climate change 30 years of future cli-

REFERENCES

mate data (2030–2060), as simulated by Providing Regional Climates for Impacts Studies (PRECIS) based on regional circulation

model

simulation

HadCM3Q0

and

ECHAM5 for A1B scenario, were forced to the calibrated AquaCrop model. Finally, some realistic water and fertility management scenarios for the selected crops were formulated and simulated so as to explore ways of stabilizing the crop yield. Laboratory and SPAW analysis conﬁrmed silt loam as being the most dominant soil type in the study area. The AquaCrop model was able to simulate crop yield with reasonable accuracy. Future (2030–2060) crop yield simulations indicated decreased yield (based on HadCM3Q0 projection) and increased yield (based on ECHAM5 projection) of monsoon rice and stable yield of winter wheat as compared to base period (1979–2014) yield. Simulation results for the management strategies indicated that the monsoon rice and winter wheat yields were mainly constrained by water scarcity and fertility stress, and that such a situation can be overcome with improved fertilizer application and proper irrigation management. Moreover, winter wheat yields in the dry season could easily be stabilized and could even be increased with proper deﬁcit irrigation strategies. Results also indicated that an increase in fertilizer application was more effective in fully irrigated conditions than in the RF condition. Analysis relating to shifting of crop plantation date showed a marked increase in the crop yield of the monsoon rice in the future period. This study therefore recommends shifting of crop plantation date, temperature resilient crop genotype (to overcome temperature stress), and proper water and fertilizer management to stabilize the crop yield.

ACKNOWLEDGEMENTS Ms L. Shrestha received ﬁnancial support from Center of Research for Environment, Energy and Water (CREEW),

Acharya, S. K. & Bhatta, G. R.  Impact of Climate Change on Agricultural Growth in Nepal. NRB working paper series, NRB-WP-15. Nepal Rastra Bank, Nepal. Allen, R. G., Pereira, L. S., Raes, D. & Smith, M.  Crop Evapotranspiration: Guidelines for Computing Crop Water Requirements. Food and Agriculture Organization of the United Nations, Rome, Italy. Bhatt, D., Maskey, S., Babel, M., Uhlenbrook, S. & Prasad, K. C.  Climate trends and impacts on crop production in the Koshi River basin of Nepal. Regional Environmental Change 14, 1291–1301. FAO  Climate Change, Water and Food Security. Food and Agriculture Organization of the United Nations, Rome, Italy. Gad, H. E. & Gyar, S. M.  Effect of solar radiation on the crops evapotranspiration in Egypt. In: 14th International Water Technology Conference, Cairo, Egypt, 21–23 March. Geerts, S. & Raes, D.  Deficit irrigation as an on-farm strategy to maximize crop water productivity in dry areas. Agricultural Water Management 96, 1275–1284. Geerts, S., Raes, D., García, M., Vacher, J., Mamani, R., Mendoza, J., Huanca, R., Morales, B., Miranda, R., Cusicanqui, J. & Taboada, C.  Introducing deficit irrigation to stabilize yields of quinoa (Chenopodium quinoa Willd.). European Journal of Agronomy 28, 427–436. Geerts, S., Raes, D. & Garcia, M.  Using AquaCrop to derive deficit irrigation schedules. Agricultural Water Management 98, 213–216. Gerakis, A. & Baer, B.  Computer program for soil textural classification. Soil Science Society of America Journal 63, 807–808. Gordon, C., Cooper, C., Senior, C. A., Banks, H., Gregory, J. M., Johns, T. C., Mitchell, J. F. B. & Wood, R. A.  The simulation of SST, sea ice extents and ocean heat transports in a version of the Hadley Centre coupled model without flux adjustments. Climate Dynamics 16, 147–168. Hudson, B. D.  Soil organic matter and available water capacity. Journal of Soil and Water Conservation 49, 180–194. IFPRI  Climate Change Impact on Agriculture and Costs of Adaptation. Food policy report of International Food Policy Research Institute. International Food Policy Research Institute, Washington, DC. Iizumi, T. & Ramankutty, N.  How do weather and climate influence cropping area and intensity? Global Food Security 4, 46–50. IPCC  IPCC Special Report Emissions Scenarios: Summary for Policymakers. A Special Report of IPCC Working Group III. IPCC, Geneva, Switzerland.

Page 149

335

L. Shrestha & N. K. Shrestha

Climate change impact on major cereal crops of Bhaktapur, Nepal

IPCC  Climate Change 2014: Synthesis Report. Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. IPCC, Geneva, Switzerland. Jones, R. G., Noguer, M., Hassell, D. C., Hudson, D., Wilson, S. S., Jenkins, G. J. & Mitchell, J. F. B.  Generating High Resolution Climate Change Scenarios Using PRECIS. Met. Ofﬁce, Hadley Centre, Exeter. Joshi, N. P., Maharjan, K. L. & Luni, P.  Effect of climate variables on yield of major food-crops in Nepal – A time series analysis. Journal of Contemporary India Studies: Space and Society 1, 19–26. Karn, P. K.  The Impact of Climate Change on Rice Production in Nepal. SANDEE working paper No. 85-14. Lal, M.  Implications of Climate Change on Agricultural Productivity and Food Security in South Asia. Key Vulnerable Regions and Climate Change – Identifying Thresholds for Impacts and Adaptation in Relation to Article 2 of the UNFCCC. Springer, Dordrecht, The Netherlands. Malla, G.  Climate change and its impact on Nepalese agriculture. The Journal of Agriculture and Environment 9, 62–71. Maximay, S.  The Caribbean’s response to climate change impacts. In: Impacts of Climate Change on Food Security in Small Island Developing States (W. Ganpat & W.-A. P. Isaac, eds). pp. 33–66. McSweeney, C. & Jones, R.  Selecting Members of the ‘QUMP’ Perturbed-Physics Ensemble for use with PRECIS. Met. Ofﬁce, Hadley Centre, UK. MoAC  Statistical Information of Nepalese Agriculture 2009/ 10. Ministry of Agriculture and Co-operatives (MoAC), Kathmandu, Nepal. NCDP  Nepal Climate Data Portal User Manual (V 0.6). [Online] Asian Disaster Preparedness Center, Department of Hydrology and Meteorology, Ministry of Environment (Government of Nepal), Asian Development Bank (accessed 29 September 2014). Palazzoli, I., Maskey, S., Uhlenbrook, S., Nana, E. & Bocchiola, D.  Impact of prospective climate change on water resources and crop yields in the Indrawati basin, Nepal. Agricultural Systems 133, 143–157.

Journal of Water and Climate Change

08.2

2017

Poudel, J. P., Tandon, H. & Bhattrai, A.  District Development Profile of Nepal. Mega Publication and Research Center, Kathmandu, Nepal. Prudhomme, C.  GCM and downscaling uncertainty in modelling of current river flow: why is it important for future impacts? In: Climate Variability and Change – Hydrological Impacts. Proceedings of the Fifth FRIEND World Conference, Havana, Cuba, November 2006. Raes, D., Steduto, P., Hsiao, T. C. & Fereres, E.  AquaCrop version 4.0 – reference manual. [Online] Available at: http:// www.fao.org/nr/water/docs/AquaCropV40Chapter2.pdf (accessed 13 December 2014). Saxton, K. E. & Rawls, W. J.  Soil water characteristic estimates by texture and organic matter for hydrologic solutions. Soil Science Society of America Journal 70, 1578–1596. Saxton, K., Willey, P. H. & Rawls, W. J.  Field and pond hydrologic analyses with the SPAW Model. In: An ASABE Meeting Presentation. American Society of Agriculture and Biological Engineer Portland Convention Center, Portland, OR. Shrestha, N.  Improving cereal production in Terai region of Nepal: Assesment of field management strategeis through the model based approach. Dissertation presented in partial fulfillment of the requirements for the degree of Doctor in Bioscience Engineering, KU Leuven, Science, Engineering & Technology, Kathmandu, Nepal. Shrestha, S., Gyawali, B. & Bhattarai, U. a Impacts of climate change on irrigation water requirements for rice-wheat cultivation in Bagmati River basin, Nepal. Journal of Water and Climate Change 4, 422–439. Shrestha, A., Karki, K., Shukla, A. & Sada, R. b Groundwater Extraction: Implications on Local Water Security of periurban, Kathmandu, Nepal. Peri Urban Water Security Discussion Paper Series, Paper No. 7, SaciWATER. Shrestha, N., Raes, D., Vanuytrecht, E. & Sah, S. K. c Cereal yield stabilization in Terai (Nepal) by water and soil fertility management modelling. Agricultural Water Management 122, 53–62. Shrestha, S., Thin, N. M. M. & Deb, P.  Assessment of climate change impacts on irrigation water requirement and rice yield for Ngamoeyeik Irrigation Project in Myanmar. Journal of Water and Climate Change 5, 427–442.

First received 10 December 2015; accepted in revised form 3 September 2016. Available online 27 October 2016

Page 150

08.1

2017

Wavelet analyses of western US streamﬂow with ENSO and PDO Kazi Ali Tamaddun, Ajay Kalra and Sajjad Ahmad

ABSTRACT This study investigated the correlation between western US streamflow and two of the most important oceanic–atmospheric indices having significant effects in this region, namely, El Niño Southern Oscillation (ENSO) and Pacific Decadal Oscillation (PDO). Data from 61 streamflow stations across six different hydrologic regions of the western USA were analyzed, using a study period of 60 years from 1951 to 2010. Continuous wavelet transformation along with cross wavelet transformation and wavelet coherence were used to analyze the interaction between streamflow and climate indices. The results showed that streamflows have changed coincidentally with both ENSO and PDO over the study period at different time-scale bands and at various time intervals. Both ENSO and PDO showed correlation with streamflow change behavior from 1980 to 2005. ENSO showed a strong correlation with streamflow across the entire study period in the 10–12 year band. PDO showed a strong correlation in bands of 8–10 years and bands beyond 16 years. The phase

Kazi Ali Tamaddun Sajjad Ahmad (corresponding author) Department of Civil and Environmental Engineering and Construction, University of Nevada, 4505 S. Maryland Parkway, Las Vegas, NV 89154-4015, USA E-mail: sajjad.ahmad@unlv.edu Ajay Kalra Department of Civil and Environmental Engineering, Southern Illinois University, 1230 Lincoln Drive, Carbondale, IL 62901-6603, USA

relationship showed that both ENSO and PDO preceded streamﬂow change behavior; in some instances, the variables were found to be moving in opposite directions even though they changed simultaneously. The results can be helpful in understanding the relationship between the climate indices and streamﬂow. Key words

| continuous wavelet transformation, cross wavelet transformation, ENSO, PDO, wavelet coherence, western US streamﬂow

INTRODUCTION Understanding the behavior of streamﬂow change can be

& Clark ). Studies have strongly suggested that

considered one of the most important parameters used to

proper documentation and understanding of the hydrolo-

trace changes that have occurred in the hydrologic cycle.

gic variables can be used as effective tools to evaluate

Since streamﬂow measures the ﬂow in natural streams, a

changes occurring in the hydrologic cycle (Clark ;

change in the behavior consequently can threaten the

Birsan et al. ). Hydrologic processes are directly

entire water supply system. The hydrologic cycle, along

related to climate conditions, and changes in hydrologic

with the mass balance mechanism associated with it,

processes can be attributed as a major cause behind the

plays an important role in transporting mass and energy

spatiotemporal patterns of hydrologic events as well as

throughout the hydrosphere (Rice et al. ). Intensiﬁca-

their severity and recurrences (Burn et al. ; Dawadi

tion of parameters in the hydrologic cycle can cause

& Ahmad ; Zhang et al. ). Change in the hydrolo-

extreme events that bring about enormous loss and, sub-

gic cycle has been considered one of the crucial results of

sequently, can endanger the entire water resource

climate warming (Ampitiyawatta & Guo ; Durdu

system (Lins & Slack ; Cayan et al. ; McCabe

).

doi: 10.2166/wcc.2016.162

Page 151

K. A. Tamaddun et al.

ENSO and PDO relationship with western US streamﬂow

Journal of Water and Climate Change

08.1

2017

Many previous studies have determined relationships

oceanic–atmospheric pattern found in the North Paciﬁc

among hydro-climatic parameters (i.e., temperature, precipi-

Ocean, and has a larger area of inﬂuence than ENSO

tation, streamﬂow, etc.) and climate variability (McCabe &

(Hamlet & Lettenmaier ; Miles et al. ; McCabe &

Wolock ; Birsan et al. ; Hamlet & Lettenmaier

Dettinger ; Beebee & Manga ; Trenberth & Fasullo

; Durdu ). Temporal variability of climate change

). Similar to ENSO, PDO has two full phases, i.e., warm

has been found to be related with the change in hydrologic

and cold, and these phases alter with a cycle of around 25 to

variables as well (Burn & Elnur ). Recent works have

50 years (Hamlet & Lettenmaier ; Mantua & Hare ;

studied the relationship between secondary hydrologic par-

Beebee & Manga ). The ﬂuctuations of SST have been

ameters, such as streamﬂow and climate variability (Kalra

found to be a good predictor of hydrologic parameters –

& Ahmad ; Carrier et al. , ; Tamaddun et al.

such as the formation of snowpack, precipitation, soil moist-

). The need to understand the relationship between a

ure, streamﬂow, etc. – since SST affects the air pressure and

change in climate and the consequent change in hydrologic

the wind dynamics above the inﬂuencing zone; this, in turn,

variables (i.e., streamﬂow) is increasing since it is of utmost

affects the hydrology of the surrounding area.

interest to efﬁciently manage sustainable water resources,

In previous studies, ENSO has been identiﬁed as a

especially with the increase in population and with the con-

major factor affecting the atmospheric anomalies (extreme

tinuous and growing demand in the energy sector (Kalra &

conditions) both globally and regionally (Ropelewski &

Ahmad ; Shrestha et al. ; Wu et al. ). Besides list-

Halpert ; Kahya & Dracup ). Studies have found

ing the potential dangers that can occur as a result of climate

PDO to have an inﬂuence on such parameters as snowpack

change (Bates et al. ; IPCC ), studies constantly

formation, precipitation, and streamﬂow in the western

have emphasized the importance of spatio-temporal scales

USA, especially in such regions as the Colorado River

on the change behaviors observed in the hydrologic vari-

Basin (CRB) and California (Dettinger & Cayan ;

ables (Weider & Boutt ).

Hidalgo & Dracup ; Cañón et al. ; Sagarika

Besides understating the relationship between climate

et al. a). Besides understanding the relationship

change and hydrologic variables, studies have focused on

between ENSO and PDO with the various hydrologic par-

ﬁnding correlations among climate indices, which represent

ameters, many studies have focused on understanding the

various oceanic–atmospheric systems, and hydrologic vari-

coupling effect of ENSO and PDO. According to Praskie-

ables; this is because climate indices can be a very

vicz & Chang () on the Willamette Valley of Oregon,

effective tool for forecasting hydrologic cycle behavior. El

La Niña was found to affect the intensity of November pre-

Niño Southern Oscillation (ENSO) and Paciﬁc Decadal

cipitation, while El Niño affected the intensity of April

Oscillation (PDO) are two of the most important oceanic–

precipitation. This study revealed an inverse relationship

atmospheric indices found to have a great inﬂuence on the

between PDO and the intensity of precipitation. A study

climate variability in the western United States (Barnett

on the Upper Colorado River Basin (UCRB) by McCabe

et al. ; Taylor & Hannan ; Beebee & Manga

et al. () found strong correlation between UCRB

). ENSO, an index associated with sea-surface tempera-

streamﬂow and temporal SST ﬂuctuations. Hamlet & Let-

ture (SST) ﬂuctuation, has been identiﬁed as one of the most

tenmaier () observed the effect of lead time of ENSO

dominant oceanic–atmospheric patterns found in the tropics

and PDO on a forecasting model for the Columbia River.

of the Paciﬁc Ocean; in addition, it is considered to be one

Sagarika et al. () studied the shifts (step changes) for

of the prominent factors affecting the western US hydrology

streamﬂow patterns in 240 streamﬂow stations across the

(Barnett et al. ; Cayan et al. ; Taylor & Hannan

continental USA, and observed the coupled effect of the

; Beebee & Manga ). ENSO is a natural cycle

PDO warm and cold phases with the change in ENSO indi-

that occurs on a scale of 2–7 years, which alters between a

ces. Kalra & Ahmad () concluded that climate signals

warm phase (El Niño, positive index) and a cold phase

signiﬁcantly inﬂuenced annual precipitation behavior in

(La Niña, negative index). PDO, an index that represents

the CRB; PDO was found to be more inﬂuential on the

SST ﬂuctuations on a decadal scale, is another important

upper CRB, whereas ENSO was more successful in

Page 152

K. A. Tamaddun et al.

ENSO and PDO relationship with western US streamﬂow

Journal of Water and Climate Change

08.1

2017

predicting precipitation behavior in the lower CRB. Beebee

classiﬁcation of wavelets, and how wavelets work can be

& Manga () studied the relationship between runoff

found in Lau & Weng (), Torrence & Compo (),

generated from snowmelt with ENSO and PDO, and

Torrence & Webster (), Grinsted et al. (), and

suggested some historical time intervals that were found

David & Rajasekaran ().

to be more correlated compared to other intervals. Hoer-

Continuous wavelet transform (CWT), which is best

ling & Kumar () provided an explanation on how

suited for feature extraction, has been used in previous

change in pressure occurs in the Paciﬁc, the subsequent

studies as a useful tool to extract a low signal-to-noise ratio

change in tracks of cyclonic storms, and the effects of

(s/n) in a time series (Grinsted et al. ). In a time

moisture on the western USA. These studies reveal some

series, CWT can analyze intermittent oscillations that are

important insights regarding how ENSO and PDO are

localized; this method performs much better than tra-

changing with respect to each other; however, they do

ditional transformation tools (Foufoula-Georgiou & Kumar

not clarify whether one or both of these indices inﬂuences

; Holschneider ; Grinsted et al. ). As mentioned

certain parameters in the same way.

earlier, the coupling of two time series can provide infor-

Hydrologic and geophysical time series are very com-

mation regarding their changing pattern with respect to

plex to analyze as they are non-stationary in nature and

each other. However, sometimes it becomes important to

they do not follow normally distributed probability functions

understand which of these time series affects a third time

( Jevrejeva et al. ; Önöz & Bayazit ; Grinsted et al.

series more dominantly. The application of cross wavelet

; Milly et al. ; Villarini et al. ; Sagarika et al.

transform (XWT) and wavelet coherency (WTC) analysis

b). As a result, predicting the trend patterns and period-

are useful methods to examine multiple time series that

icities of these time series has drawn much attention in

might be linked in certain ways (Jevrejeva et al. ;

recent times (Grinsted et al. ). The most traditional

Grinsted et al. ; Tang et al. ). XWT, which reveals

mathematical method used to examine periodicities in the

a common power (covariance) and a relative phase relation-

frequency domain is Fourier analysis (Polikar ). The

ship in a wavelet spectrum, is constructed from two separate

underlying drawback of Fourier analysis is it implicitly

CWTs that are supposedly linked in some way (Torrence &

assumes a stationarity in time (Polikar ; David & Raja-

Compo ; Grinsted et al. ). By observing the XWT,

sekaran ); however, this cannot be a useful assumption

the correlation as well as the phase relationship between

for a time series of hydrologic variables, such as streamﬂow.

the parameters can be assessed. To further quantify the cor-

Wavelet transformation has been suggested as a powerful

relation between the parameters, WTC can detect signiﬁcant

tool for analyzing processes that occur over ﬁnite spatio-

coherence even at a lower common power. This technique

temporal domains and are non-stationary in nature, some-

shows how conﬁdence levels can be calculated against red

times containing multiscale resolution (Lau & Weng ).

noise backgrounds (Grinsted et al. ). Through the pro-

Wavelets allow determination of the most signiﬁcant period-

cess of using CWT, XWT, and WTC, a one-dimensional

icities (frequencies) of a time series and can explain how it

time series is transformed into a two-dimensional time–fre-

has changed over time (Kumar & Foufoula-Georgiou ;

quency wavelet spectrum. This spectrum can show the

Percival & Walden ). As a result, wavelet transform-

amplitude of a signal (in this case time series) at different

ation emerged as a better alternative since it could provide

times and frequencies at the same time (Torrence & Webster

information about time and frequency at the same time.

). Studies also suggest the use of wavelets as a better

By altering time and scale variations, wavelet analyses can

alternative compared to other traditional methods for ana-

produce graphs that can show how the frequency changes

lyzing oceanic–climatic ﬂuctuations, since wavelets can

over amplitude with the change in time (Echer et al. ).

follow the gradual changes occurring in a natural frequency

Other studies have suggested using wavelet analyses as a

with better accuracy (Meyers et al. ; Yiou et al. ).

successful statistical tool for analyzing trends and other

Previous literary works motivated this current study to

properties of a time series (Nakken ; Kang & Lin

use CWT as an analysis tool to evaluate the correlation

). A more detailed description of the history of wavelets,

between parameters that other studies have found to be Page 153

K. A. Tamaddun et al.

Journal of Water and Climate Change

ENSO and PDO relationship with western US streamﬂow

08.1

2017

related somehow. Acknowledging some of the limitations of

shows the chosen regions, with spatial distribution of the

previous research, the current study endeavors to address

stations in each region. Geospatial Attributes of Gages for

some of the suggestions that were presented in those

Evaluating Streamﬂow, Version II (Falcone et al. ) pro-

works. With this motivation in mind, this research focused

vides details about the stations having the data as well.

on applying CWT, along with XWT and WTC, on data for

Upper Colorado was excluded from the analyses, since there

61 unimpaired streamﬂow stations (unimpaired stations

were no stations in that region that met the time period of his-

are free from any sort of modiﬁcations in terms of ﬂow

toric data needed in this study (Figure 1).

path and condition) located in the western USA for a

The climate indices datasets used in this study were

period of 60 years (i.e., 1951 to 2010). The primary objective

ENSO and PDO. The data used in this study for ENSO

of the study was to evaluate signiﬁcant periodicities that

and PDO had the same length as the streamﬂow data. For

have simultaneously triggered changing patterns of stream-

both ENSO (http://www.cpc.ncep.noaa.gov) and PDO

ﬂow and climate signals (i.e., ENSO and PDO). Besides

(http://research.jisao.washington.edu), an increase in the

observing simultaneous change patterns, this study quanti-

index value refers to the warm phase and a decrease in

ﬁed the correlations present in the change patterns. Each

index value refers to the cold phase.

station was transformed with CWT to their wavelet spectrum in order to observe their variability (higher power in the wavelet spectrum represented higher variance in data).

METHODOLOGY

A combined streamﬂow continuous wavelet spectrum was constructed using principal component analysis (PCA) of

In the following sections, brief descriptions of CWT, XWT,

the data obtained from each station, and was used to con-

and WTC are provided, based on Torrence & Webster

struct the corresponding XWTs with ENSO and PDO

(), Grinsted et al. (), and Tang et al. (). Inter-

CWTs. The XWTs revealed the common power of stream-

ested readers may refer to Torrence & Compo (),

ﬂow and ENSO/PDO over the study period. Finally, WTC

Jevrejeva et al. (), Souza et al. (), and Beecham &

was performed to quantify the correlation between stream-

Chowdhury () for further details and clariﬁcation.

ﬂow and ENSO/PDO.

The steps followed in the current study are: 1. decomposition of the original time series using CWT;

STUDY AREA AND DATA

2. construction of XWT from two CWTs; 3. WTC analysis between two CWTs.

Out of the 18 hydrologic regions delineated by the United

The following sections describe each step and explain

States Geological Survey (USGS), this study focused on six

how they were used to analyze the relationship between

regions representing the western USA: Rio Grande (13),

two different time series that supposedly are correlated.

Upper Colorado (14), Lower Colorado (15), Great Basin (16), Paciﬁc Northwest (17), and California (18). A detailed

CWT

description of the regions can be found in the hydrologic unit map provided by the USGS (http://water.usgs.gov/GIS/

Wavelets are functions with a zero mean; unlike Fourier

regions.html). Out of the 704 streamﬂow stations listed by

transforms, which are localized only in frequency, wavelets

USGS, published in 2012 as the Hydroclimatic Data Network

have the ability to be stretched and translated in both time

(HCDN) 2009 (Lins ), 61 stations were selected based on

and frequency ( Jevrejeva et al. ). Studies suggest that

the availability of continuous water-year data for 60 years

using CWT is more appropriate for analyzing a time series

from 1951 to 2010. A single station was chosen from each

that has a non-normal distribution (Grinsted et al. ).

stream to remove spatial bias from the data. Additionally, the

Non-normal distributions are frequently found in non-

streamﬂow stations were free from any sort of modiﬁcation

stationary parameters, for example, such hydroclimatic vari-

or alteration in terms of controlling the ﬂow behavior. Figure 1

ables as precipitation and streamﬂow. The advantage of

Page 154

Figure 1

K. A. Tamaddun et al.

Journal of Water and Climate Change

ENSO and PDO relationship with western US streamﬂow

08.1

2017

Map showing the selected regions of the continental USA and the stations within each region. The table at top right shows the number of stations in each region.

using a wavelet transformation is that it allows the analysis

relationship in time–frequency space (Grinsted et al. ).

of non-stationary time series at different frequencies (period-

The cross wavelet spectrum, which shows the covariance

icities) (Foufoula-Georgiou & Kumar ), and can be used

of two time series, occurs from a complex conjugation of

effectively to observe how the frequencies have changed

the two time series. It produces a cross wavelet power spec-

over time. The Morlet wavelet has been suggested in pre-

trum that is used to observe the correlation between the two

vious studies (Torrence & Compo ; Percival &

time series. The phase angle of the cross wavelet power

Walden ) as the most appropriate wavelet function to

shows how the two time series are related in terms of their

be used for analyzing geophysical signals; accordingly, it

phase relationship in time–frequency space (Jevrejeva et al.

was used in this study. A combined streamﬂow CWT was

). The presence of a statistically signiﬁcant covariance

obtained using PCA; the ﬁrst principal component, which

was determined against red noise background (Torrence &

explained 71.21% variance of the data obtained from all

Compo ).

the stations, was used to represent the overall variance in data.

WTC analysis

XWTs and cross wavelet phase angle

The presence of high common power across two different CWTs could be observed by means of the XWT constructed

An XWT was constructed from two CWTs to observe their

from them, as mentioned in the previous section. In order to

high common power (covariance) and relative phase

observe the coherency of two CWTs in the time–frequency Page 155

K. A. Tamaddun et al.

Journal of Water and Climate Change

ENSO and PDO relationship with western US streamﬂow

space, WTC is considered to be more useful (Grinsted et al.

08.1

2017

RESULTS

). WTC analysis shows the common frequency bands and the time intervals of two CWTs that were found to be

In this study, standardized streamﬂow data of 61 stations

correlated (Tang et al. ). The advantage of using WTC

across six western US hydrologic regions were decomposed

is that it quantiﬁes the correlation and shows the presence

using CWT. A combined CWT for the standardized stream-

of signiﬁcant coherence at lower common powers as well.

ﬂow data was obtained using PCA, which represented the

It explains how to calculate conﬁdence levels alongside

entire time series and the amount of variance in the data.

red noise backgrounds (Grinsted et al. ). In this study,

CWTs of standardized ENSO and PDO data were obtained

the Monte Carlo approach (Wallace et al. ) was used

for the chosen study period. Figure 2 shows the standardized

to calculate the signiﬁcance of the wavelet coherence and

time series of the combined streamﬂow, ENSO, and PDO

a 5% signiﬁcance level was chosen against red noise to cal-

along with their CWTs and their respective global wavelet

culate the statistical signiﬁcance.

spectrums. Figures 3 and 4 show the XWT and WTC,

Figure 2

Standardized time series, CWT, and global wavelet spectrum of (a) combined streamflow, (b) ENSO, and (c) PDO. Red and blue represent stronger and weaker powers, respectively. A thick black contour line delineates a 5% significance level against the red noise. The full color version of this figure is available in the online version of this paper, at http://dx.doi.org/10.2166/wcc.2016.162.

Figure 3

Cross wavelet spectrum between a standardized combined streamflow with standardized (a) ENSO and (b) PDO. A thick black contour line delineates a 5% significance level against the red noise (red and blue represent stronger and weaker powers, respectively). The cone of influence (COI), which potentially can distort the picture around the edges, is shown by lighter shades. The arrows represent the relative phase relationship between the two time series. Right (left) pointing arrows show an in-phase (anti-phase) relationship, while vertically upward arrows show that ENSO and PDO leads streamflow by 90 . The full color version of this figure is available in the online version of this paper, at http://dx.doi.org/10.2166/wcc.2016.162. W

Page 156

K. A. Tamaddun et al.

Figure 4

ENSO and PDO relationship with western US streamﬂow

Journal of Water and Climate Change

08.1

2017

Wavelet coherence spectrum between a standardized combined streamflow with standardized (a) ENSO and (b) PDO. A thick black contour line delineates a 5% significance level against the red noise (red and blue represent stronger and weaker powers, respectively). The COI, which potentially can distort the picture around the edges, is shown by lighter shades. The arrows represent the relative phase relationship between the two time series. Right (left) pointing arrows show an in-phase (anti-phase) relationship, while vertically upward arrows show a lag between ENSO and PDO with streamflow. The full color version of this figure is available in the online version of this paper, at http://dx.doi.org/10.2166/wcc.2016.162.

respectively, of the combined streamﬂow with both ENSO

showed the highest peak near 3–5 years’ band. The presence

and PDO.

of higher power was observed near 12–14 years’ band in the global wavelet spectrum as well; however, they were not

CWT

statistically signiﬁcant.

The time series for the standardized combined streamﬂow of

the Paciﬁc Ocean with a time period of 25–50 years. From

all the stations, along with the continuous wavelet power

the wavelet power spectrum of PDO (Figure 2(c)), a substan-

spectrum, is shown in Figure 2(a). Signiﬁcant variabilities

tially high power was found at a 5% signiﬁcance level in 3–7

in the wavelet power spectrum were found in 2–4 years’

years’ band from 1951 to 1962, in 4–6 years’ band from 1986

PDO is another oceanic–atmospheric pattern found in

band from 1970 to 1977, in 6–16 years’ band from 1970 to

to 2001, in 3–4 years from 1982 to 1988, and in 8–12 years’

2010, and in 3–4 years’ band from 1998 to 2002. From

band from 1993 to 2005. From the global wavelet spectrum,

observing the wavelet power spectrum, the highest power

8–12 years’ band was found to have the highest power

(which represents the variance of data) was observed near

among the statistically signiﬁcant regions. Higher powers

the bands of 2–3 years and 12–14 years. The global wavelet

even were observed in 16 years’ band and above; however,

spectrum showed that the highest peak was located near 12–

they were not found to be statistically signiﬁcant.

14 years’ band.

The exact correlation between ENSO/PDO with stream-

ENSO has been identiﬁed as one of the dominant ocea-

ﬂow variations was found to be quite difﬁcult to observe

nic–atmospheric patterns in the tropics of the Paciﬁc Ocean,

from their respective CWTs. However, the comparison of

with a period of 2–7 years. From the wavelet spectrum of

the wavelet power spectra suggested that higher powers

ENSO (Figure 2(b)), from 1976 to 2003, the presence of sig-

(higher variance) near bands of 3–7 years and 8–12 years

niﬁcant high power in 3–7 years’ band was observed. The

were found to be statistically signiﬁcant. Higher powers

presence of signiﬁcant high powers was also observed in

near 3–7 years’ band were found to be present in both the

5–7 years’ band from 1953 to 1962 and in 3–5 years’ band

combined streamﬂow power spectrum and the ENSO

from 1966 to 1975. From the wavelet power spectrum, the

power spectrum. Both the combined streamﬂow power

highest power was observed from 1982 to 1990 near 3–5

spectrum and the PDO power spectrum showed higher

years’ band. In addition, the global wavelet spectrum

powers in 8–12 years’ band. Page 157

K. A. Tamaddun et al.

Journal of Water and Climate Change

ENSO and PDO relationship with western US streamﬂow

XWT

08.1

2017

The XWT analyses of the combined streamﬂow with ENSO and PDO revealed that common powers of ENSO

To understand the correlations between ENSO/PDO with

(coincidence with streamﬂow variation) were found to be

streamﬂow variations, XWT analysis was performed. From

higher compared to PDO. These results were consistent

the XWT of combined streamﬂow and ENSO (Figure 3(a)),

with the CWTs of ENSO and PDO, where ENSO had

it was found that they shared common power in 2–4 years’

more regions of signiﬁcance compared to PDO (Figure 2(b)

band from 1968 to 1976, in 3–4 years’ band from 1981 to

and 2(c)). The time-scale bands with signiﬁcant common

1986, in 3–4 years’ band from 1995 to 2001, in 6–7 years’

powers were in agreement with what was observed in indi-

band from 1992 to 2002, in 8–12 years’ band from 1997 to

vidual CWTs. Even though 12–14 years’ band was not

2006, and in 12–16 years’ band from 1972 to 2005. The

found to be signiﬁcant in ENSO in the CWT (Figure 2(b)),

arrows in the ﬁgure indicate the phase angle relationship

the global wavelet spectrum showed the presence of

between the two time series. In the lower time-scale bands,

higher power in 12–14 years’ band; this justiﬁed the relation-

arrows mostly pointed left, which indicated an anti-phase

ship found from the XWT of combined streamﬂow and

relationship between streamﬂow and ENSO; this meant

ENSO. To be certain that these relationships were not by

they were moving at the same time but in the opposite direc-

mere chance, and to quantify the correlation, WTC analyses

tion. Anti-phase can be interpreted as an increase (decrease)

were performed on combined streamﬂow CWT and ENSO/

in streamﬂow and decrease (increase) in ENSO index, which

PDO CWT.

means a colder (warmer) phase. As the time-scale band increased, arrows were observed to have a greater tendency

WTC analysis

to point straight up, indicating a time lag between ENSO and streamﬂow variation. Arrows indicating straight up indi-

CWT and XWT analyses provided important information

cated that ENSO led streamﬂow by 90 . The phase relation

regarding the correlation between the two time series. How-

can be used to calculate the exact time lag; however, since

ever, to quantify the correlation between the two variables,

it depends on the speciﬁc wavelength of the signal, this step

WTC analysis was performed in this study, in which the

was not performed in this study.

Monte Carlo approach was used to compute the signiﬁcance

XWT analysis of the combined streamﬂow and PDO

of correlation.

(Figure 3(b)) showed common power in 2–3 years’ band

From the WTC of combined streamﬂow and ENSO,

from 1974 to 1981, in 3–4 years’ band from 1972 to 1978,

areas of signiﬁcance were observed in the band of 10–16

in 5–7 years’ band from 1991 to 1998, around 3 year band

years across the entire study period of 60 years, from 1951

during 2000, and in 7–14 years’ band from 1983 to 2008.

to 2010 (Figure 4(a)). The time-scale bandwidths were

Common powers observed at lower time-scale bands were

observed to decrease at both ends of this time period.

lower compared to higher time-scale bands. At lower time

From 1968 to 1995 in the 10–12 years’ band, the correlation

scales (in 2–4 years’ band), arrows indicating phase relation-

coefﬁcient in this area of signiﬁcance varied from 0.8 to

ship were found to point towards both the right and left

approximately 1.0. Arrows indicating phase relationships

during various time intervals across the study period; this

mostly pointed upward; this suggested a lag between

indicated an in-phase and anti-phase relationship, respect-

ENSO and streamﬂow variations, with ENSO leading

ively. In 5–7 years’ band, arrows pointed downward and

streamﬂow by 90 . High correlation values, ranging from

slightly towards both the left and right. In higher time

0.6 to 0.8, were observed in the band of 2–6 years from

scales (in 6–14 years’ band), arrows mostly were found to

1952 to 1978 and from 1987 to 2004. Higher correlation

point straight up, indicating a phase difference of 90 ; this

values were observed as well in the 16 years’ band and

referred to a lag between PDO and streamﬂow variations.

above across the study period; however, they were not

Common powers at time scales higher than 16 years’ band

found to be statistically signiﬁcant.

were observed; however, they were not found to be statistically signiﬁcant. Page 158

The WTC of combined streamﬂow and PDO (Figure 4(b)) showed less areas of signiﬁcance compared to the areas of

K. A. Tamaddun et al.

ENSO and PDO relationship with western US streamﬂow

Journal of Water and Climate Change

08.1

2017

signiﬁcance observed in the WTC of combined streamﬂow

this study. An XWT constructed from two different CWTs

and ENSO. Statistically signiﬁcant areas were found in

showed a common power of the wavelet spectrum, and

10–12 years’ band at the beginning of 1950s, in 8–10 years’

suggested a phase relationship between the time series

band from 2003 to 2010, and above 16 years’ band from

under inspection. By using WTC, which was helpful in quan-

1986 to 2010. Correlation values in these regions were

tifying the correlation, signiﬁcant coherence was found at

found to be in the range of 0.7 to approximately 1.0;

lower common power. The results showed ENSO to have

higher correlation values were found in 10–12 years’ band

a higher correlation than PDO during the study period.

during the 1950s and in the 16 years’ band and above

The most inﬂuential periodicities varied from 8–12 years

from 1986 to 2010. High correlations, in the range of 0.6

for both ENSO and PDO. The interval of 1980 to 2005

to 0.8, were observed from 1975 to 1995 in 12–14 years’

showed the presence of higher correlation with streamﬂow

band and in 8–14 years’ band from 1995 to 2010, although

for both ENSO and PDO. Presence of signiﬁcant regions

they were found to be statistically insigniﬁcant. Presence

in the 16 years’ band and above indicated that more areas

of regions having higher correlation values – in the range

of signiﬁcance (at higher periodicities) could have been

of 0.6 to 0.9 – but not statistically signiﬁcant were observed

explored if a longer study period were chosen.

in some of the other intervals in the study period at lower

CWT analysis of the combined streamﬂow along with

time scales, near the band of 2–5 years from 1968 to 2005

the CWTs of ENSO and PDO indices were formed to

with intervals in between.

observe their individual signiﬁcant variance (high power in

The WTC analyses between combined streamﬂow and

the wavelet spectrum) across the study period. Signiﬁcant

ENSO/PDO showed that ENSO had a much more pro-

high power in streamﬂow wavelet spectrum was found in

nounced correlation with streamﬂow compared to PDO,

bands of 2–4 years, 3–4 years, and 6–16 years at different his-

as ENSO showed the presence of more signiﬁcantly corre-

torical time intervals (Figure 2(a)). The global wavelet

lated areas (high common power). For both ENSO and

spectrum revealed that the highest power for streamﬂow

PDO, the band of 8–16 years was found to be most signiﬁ-

variation occurred in the band of 12–14 years from 1980

cantly correlated. For PDO, regions with high correlation

to 2000. For ENSO, signiﬁcantly high power was observed

were observed in the 16 years’ band and above; however,

in bands of 3–5 years, 3–7 years, and 5–7 years (Figure 2(b)),

due to the limitation of data, the study could not detect

with the highest power in the 3–5 years’ band from 1982 to

the entire band length.

1990. For PDO, signiﬁcantly high power was observed in bands of 3–4 years, 3–7 years, 4–6 years, and 8–12 years (Figure 2(c)). The highest power was observed in the 8–12

DISCUSSION

years’ band from 1993 to 2005. The global wavelet spectrum of PDO also showed the

To understand how streamﬂow in the western USA has

presence of higher power in bands higher than 16 years;

changed with the change in ENSO/PDO, CWT along with

however, they were not found to be statistically signiﬁcant.

XWT and WTC were used in this study. The most signiﬁcant

Observation of individual CWTs revealed information

periodicities that triggered simultaneous variations in the

regarding their changing patterns; however, it was difﬁcult

change patterns were observed to understand the corre-

to formulate any strong correlation between them from

lation

sight only. From observing the individual CWTs, neverthe-

observing high common power in the wavelet spectrum at

less, it could be concluded that both ENSO and PDO had

various time scales through the study period of 60 years

some effect on the variation of streamﬂow, since high

(i.e., 1951–2010), the study investigated the correlation

power bands overlapped in certain regions. Similar to pre-

between ENSO/PDO and streamﬂow variations across the

vious works (Grinsted et al. ; Jevrejeva et al. ),

western USA.

results of the current study reinforced the choice of CWT

between

climate

indices

and

streamﬂow.

In order to analyze two time series at the same time, XWT and WTC between two CWTs were performed in

as a better feature extraction tool, as CWT produced visible high power to represent variance in data. Page 159

K. A. Tamaddun et al.

ENSO and PDO relationship with western US streamﬂow

Journal of Water and Climate Change

08.1

2017

To understand the correlation between the time series

years, and 7–14 years (Figure 3(b)) across different historical

with greater precision, XWTs were constructed from two

periods. The highest common power was observed in 7–14

individual CWTs. These XWTs provided information regard-

years’ band from 1983 to 2008. The arrows indicating a

ing high common power (covariance) with consistent phase

phase relationship in the highest power region mostly

relationships as well as information regarding temporal lags

pointed upward, which indicated a lag between PDO and

between the two time series. The XWT between the com-

streamﬂow (PDO leaded streamﬂow by 90

bined streamﬂow and ENSO (Figure 3(a)) revealed that

where the arrows were pointing exactly upward). Phase

common high power was present in bands of 2–4 years, 3–

relationship at lower time scales were observed to be not

4 years, 6–7 years, 8–12 years, and 12–16 years at different

showing any uniform pattern.

at the points

historical time periods. Highest power was observed in 2–

Similar to ENSO, calculation of exact lag time between

4 years’ band from 1968 to 1973 and in 12–16 years’ band

PDO and streamﬂow variation was not a focus for this cur-

from 1972 to 2002. At lower time scales, in the 2–5 years’

rent study. However, previous studies have investigated the

band, arrows indicating the phase relationship mostly

lag response of PDO and streamﬂow, and found a delay of

pointed to the left, which suggested an anti-phase relation-

several months between oceanic oscillations and streamﬂow

ship between streamﬂow and ENSO, suggesting the

ﬂuctuations. Hanson et al. () studied the relationship

streamﬂow mirrors the behavior of ENSO. In other words,

between different climate variabilities and southwestern US

since they share common power, they both moved at the

discharge ﬂows, and suggested that the lag time between

same time but in opposite directions. At higher time

the PDO index and ﬂow change could vary between 1.5

scales, in the 6–16 years’ band, arrows mostly pointed

and 5 years, depending on the type of ﬂow. Although they

upwards, indicating a lag between ENSO and the variability

were not found to be statistically signiﬁcant, the XWT of

of streamﬂow. Arrows pointing exactly upward suggested

combined streamﬂow and PDO from the current study

that ENSO leads streamﬂow by 90 at those points in time.

revealed the presence of high common power at time

It was possible to calculate exact lag times from the

scales greater than 16 years. The limited data restricted the

phase relationships obtained from XWTs, but they were

conﬁdence for bands beyond 16 years. Since PDO has a

speciﬁc to a certain wavelength. As a result, calculation of

multi-decadal time period (25–50 years), it is probable that

exact lag times was not considered to be within the scope

the presence of more common powers for bands at time

of this study. Previous studies have investigated the lag

scales greater than 16 years were missed.

response of ENSO and streamﬂow, and also observed vari-

WTC assisted in quantifying the correlation between the

able lags between oceanic oscillations and streamﬂow

wavelet spectra and helped to detect signiﬁcant coherence

variations. The overall response time, which can be up to

at low common powers found during the analyses with

several months, is the result of all the lags that occur from

XWTs. From the WTC between combined streamﬂow and

oceanic ﬂuctuations, precipitation events, the time required

ENSO (Figure 4(a)), the continuous presence of common

for snowmelt, and delays in streamﬂow response (Cayan

power was observed in the 10–16 years’ band across the

et al. ; Hanson et al. ). Use of lags and their effects

entire study period. The correlation values in the 10–16

can be found in Trenberth & Hurrell (), Pozo-Vázquez

years’ band were in the range of 0.8 to as high as approxi-

et al. (), and Jevrejeva et al. (). Similar to the results

mately 1.0 around the 10–12 years’ band from 1968 to

of the current study, McCabe & Dettinger () and Beebee

1995. The reason behind such strong common power at

& Manga () found that ENSO had less correlation with

this range of the time scale could be because ENSO itself

mean annual ﬂow from 1920 to 1950, and observed an

has a periodicity of 2–7 years, and the results might have

increased correlation after 1950. In the current study, all

occurred when two ENSO cycles joined together. In

the signiﬁcant correlations observed at 5% signiﬁcance

addition, the presence of high common power was observed

level occurred after 1968 across all time-scale bands.

with a correlation ranging from 0.6 to 0.8 at lower time

High common power between combined streamﬂow

scales in the 2–6 years’ band, though they were not found

and PDO was found in bands of 2–3 years, 3–4 years, 5–7

to be statistically signiﬁcant. The phase relationships found

Page 160

K. A. Tamaddun et al.

Journal of Water and Climate Change

ENSO and PDO relationship with western US streamﬂow

08.1

2017

in the signiﬁcant regions were consistent with what was

as PDO, which has a periodicity (recurrence interval) of

observed in the XWT between ENSO and the streamﬂow

multiple decades, a longer study period would have resulted

of the stations. ENSO was found to lead streamﬂow vari-

in a better understanding of the correlation between the par-

ation by 90 in most of the signiﬁcant regions.

ameters in hand. Analyses of a longer period of data are

The WTC between the combined streamﬂow and PDO

important as well for regions that are currently going

revealed the presence of high correlation in bands of 8–10

through extreme scenarios; for example, the western USA

years, 10–12 years, and beyond 16 years at different intervals

has been experiencing drought for several years. Inclusion

across the study period. The correlation values were found

of a larger number of stations would have provided results

to be as high as approximately 1.0 in 8–10 years’ band

having more reliability, but that would have minimized the

during the 1950s and beyond 16 years’ band from 1986 to

minimum temporal length of the data since many of the

2010. The region found beyond the 16 years’ band suggested

stations do not have longer data records.

that this was likely to continue at even greater time scales. Since the relatively short study period of 60 years could not generate a wavelet spectrum beyond this time scale, it was not possible to investigate beyond this point. PDO has a time period of multiple decades, which explains the presence of common power at higher time scales. A correlation in the range of 0.6 to 0.8 was observed at lower time scales, but was not found to be statistically signiﬁcant. PDO was found to lead streamﬂow by 90

at some

points in time in higher bands. In other regions having a higher common power, PDO and streamflow were mostly found in an anti-phase relationship. Similar anti-phase or inverse relationship was found by Lins () and Dettinger et al. (), which supports the results of the current study. ENSO was found to have a higher correlation with the change in streamflow compared to PDO. Similarly, Beebee & Manga () found a higher correlation between ENSO and the mean annual discharge compared to PDO while studying snowmelt and consequential runoff in Oregon. They found mean annual discharge to be more correlated than temperature and precipitation, and concluded that the underlying reason might be because the discharge represents the spatial average of a much smaller area compared to broader climatic zones of temperature and precipitation (Beebee & Manga ). This phenomenon, that flow behavior can represent a change occurring in a localized area with better accuracy, influenced the current study to work with streamflows of a particular region – in

CONCLUSION In this study, data from 61 unimpaired streamflow stations with 60 years of continuous data (i.e., 1951–2010) were obtained across six hydrologic regions in the western USA to evaluate the correlation between streamflow and two major oceanic–atmospheric patterns, also known as climate signals, of the Pacific Ocean, namely, ENSO and PDO. To understand these relationships, CWT along with XWT and WTC were applied. The study investigated the correlation between the parameters and also provided some insights regarding the significant frequencies (periodicities) of the multiple time series that were analyzed. The results of this study indicated the presence of multiple significant time scales (bandwidths), which are important in understanding the relationships between streamflow and the oceanic–atmospheric patterns (i.e., ENSO and PDO). The results indicated that both ENSO and PDO had significant correlation with the streamflow variation in the 8–16 years’ band during the study period. In addition, ENSO showed the presence of significant correlation at lower time scales, i.e., 2–5 years’ band. The presence of high correlation was found with PDO in bands of 16 years and above. Limitations due to the length of data prevented the current study analyzing results beyond the bands of 16 years. The major contributions of this study are as follows:

this case, the western region – rather than working with the entire United States. A longer study period would have allowed the current

•

A continuous wavelet-based analysis for unimpaired streamﬂow stations across the entire western USA to

study to investigate the wavelet spectrum at time scales

evaluate the coupled effect of streamﬂow change with

beyond 16 years. For oceanic–atmospheric patterns, such

oceanic–atmospheric patterns (ENSO and PDO). Page 161

•

K. A. Tamaddun et al.

ENSO and PDO relationship with western US streamﬂow

Application of cross wavelet and WTC analyses to understand the relationship between the parameters chosen (streamﬂow, ENSO, and PDO variations).

•

Evaluation of the most signiﬁcant periodicities (frequencies) that affect the streamﬂow change patterns.

•

Quantiﬁcation of the correlations observed between the parameters.

•

Conforming to the results of previous works using a comparatively recent approach. The scope of the current study can be extended by ana-

lyzing data of greater lengths. A greater length of data could be obtained by using various reconstruction methods that have been found effective in extrapolating data in previous studies. Reconstruction could be helpful in interpolating missing data (data dropout) or in cases of data irregularities. As for the record, similar methods could be applied to climate signals’ data as well to obtain data of greater length. Incorporation of reconstructed (interpolated) data in wavelet analysis has not been well explored in the field of hydrologic time series analyses. Potentially, this can be an opportunity for further research since there has been some work in signal processing dealing with similar techniques. Analyzing other oceanic–atmospheric indices could be possible as well by applying the methods used in this study. Another plausible extension of this work could be the calculation of precise lag times at specific wavelengths. The results of this study provided insights regarding the coupled behavior of streamflow in the western USA with the changes in ENSO and PDO indices. The study focused on formulating a correlation between the parameters in hand. The results provided information about the periodicities of the fluctuation patterns and presented insight regarding their effects over the historical time series of streamflow. These findings can be helpful to water managers to get a better understanding of the relationships between oceanic– atmospheric patterns and streamflow.

REFERENCES Ampitiyawatta, A. D. & Guo, S.  Precipitation trends in the Kalu Ganga basin in Sri Lanka. Journal of Agricultural Science 4 (1), 10–18.

Page 162

Journal of Water and Climate Change

08.1

2017

Barnett, T. P., Pierce, D. W., Saravanan, R. & Planck, M.  Origins of the midlatitude pacific decadal variability. Geophysical Research Letters 26 (10), 1453–1456. Bates, B., Kundzewicz, Z. W., Wu, S. & Palutikof, J.  Climate Change and Water. Technical Paper VI. Intergovernmental Panel on Climate Change (IPCC), Geneva, Switzerland. Beebee, R. A. & Manga, M.  Variation in the relationship between snowmelt runoff in Oregon and ENSO and PDO. Journal of the American Water Resources Association 40 (02099), 1011–1024. Beecham, S. & Chowdhury, R. K.  Temporal characteristics and variability of point rainfall: a statistical and wavelet analysis. International Journal of Climatology 30, 458–473. Birsan, M. V., Molnar, P., Burlando, P. & Pfaundler, M.  Streamflow trends in Switzerland. Journal of Hydrology 314 (1–4), 312–329. Birsan, M. V., Zaharia, L., Chendes, V. & Branescu, E.  Recent trends in streamflow in Romania (1976–2005). Romanian Reports in Physics 64 (1), 275–280. Burn, D. H. & Elnur, M. A. H.  Detection of hydrologic trends and variability. Journal of Hydrology 255 (1), 107–122. Burn, D. H., Sharif, M. & Zhang, K.  Detection of trends in hydrological extremes for Canadian watersheds. Hydrological Processes 24 (13), 1781–1790. Cañón, J., González, J. & Valdés, J.  Precipitation in the Colorado River Basin and its low frequency associations with PDO and ENSO signals. Journal of Hydrology 333 (2–4), 252–264. Carrier, C., Kalra, A. & Ahmad, S.  Using Paleo reconstructions to improve streamflow forecast lead time in the western United States. Journal of the American Water Resources Association 49 (6), 1351–1366. Carrier, C., Kalra, A. & Ahmad, S.  Long-range precipitation forecast using paleoclimate reconstructions in the western United States. Journal of Mountain Science 13 (4), 614–632. Cayan, D. R., Redmond, K. T. & Riddle, L. G.  ENSO and hydrological extreme in the western United States. Journal of Climate 12, 2881–2893. Cayan, D. R., Kammerdiener, S. A., Dettinger, M. D., Caprio, J. M. & Peterson, D. H.  Changes in the onset of spring in the western United States. Bulletin of the American Meteorological Society 82 (3), 399–415. Clark, G. M.  Changes in patterns of streamflow from unregulated watersheds in Idaho, western Wyoming, and Northern Nevada. Journal of the American Water Resources Association 46 (3), 486–497. David, V. S. & Rajasekaran, S.  Pattern Recognition Using Neural and Functional Networks. Springer-Verlag, Berlin, Germany. Dawadi, S. & Ahmad, S.  Changing climatic conditions in the Colorado River Basin: implications for water resources management. Journal of Hydrology 430–431, 127–141. Dettinger, M. D. & Cayan, D. R.  Large-scale atmospheric forcing of recent trends toward early snowmelt runoff in California. Journal of Climate 8, 606–623.

K. A. Tamaddun et al.

ENSO and PDO relationship with western US streamﬂow

Dettinger, M. D., Battisti, D. S., Garreaud, R. D., McCabe, G. J. & Bitz, C. M.  Interhemispheric effects of interannual and decadal ENSO-like climate variations on the Americas. Interhemispheric Climate Linkages. Academic Press, Elsevier, Amsterdam, The Netherlands, pp. 1–16. Durdu, Ö. F.  Effects of climate change on water resources of the Büyük Menderes River basin, western Turkey. Turkish Journal of Agriculture and Forestry 34 (4), 319–332. Echer, M. P. S., Echer, E., Nordemann, D. J., Rigozo, N. R. & Prestes, A.  Wavelet analysis of a centennial (1895–1994) southern Brazil rainfall series. Climatic Change 87, 489–497. Falcone, J. A., Carlisle, D. M., Wolock, D. M. & Meador, M. R.  GAGES: a stream gage database for evaluating natural and altered flow conditions in the conterminous United States. Ecology 91 (2), 621. Foufoula-Georgiou, E. & Kumar, P.  Wavelets in Geophysics. Academic Press, San Diego, CA, USA, p. 373. Grinsted, A., Moore, J. C. & Jevrejeva, S.  Application of the cross wavelet transform and wavelet coherence to geophysical time series. Nonlinear Processes in Geophysics 11, 561–566. Hamlet, A. F. & Lettenmaier, D. P.  Columbia river streamflow forecasting based on ENSO and PDO climate signals. Journal of Water Resources Planning and Management 125 (6), 333–341. Hamlet, A. F. & Lettenmaier, D. P.  Effects of 20th century warming and climate variability on flood risk in the western US. Water Resources Research 43 (6), W06427. Hanson, R. T., Newhouse, M. W. & Dettinger, M. D.  A methodology to assess relations between climatic variability and variations in hydrologic time series in the southwestern United States. Journal of Hydrology 287 (1), 252–269. Hidalgo, H. G. & Dracup, J. A.  ENSO and PDO effects on hydroclimatic variations of the Upper Colorado River Basin. Journal of Hydrometeorology 4, 5–23. Hoerling, M. P. & Kumar, A.  Understanding and predicting extratropical teleconnections related to ENSO. In: El Niño and the Southern Oscillation: Multi-scale Variations and Global and Regional Impacts (H. F. Diaz & V. Markgraf, eds). Cambridge University Press, Cambridge, UK, pp. 57–88. Holschneider, M.  Wavelets: An Analysis Tool. Oxford University Press, New York, USA, p. 455. Intergovernmental Panel on Climate Change, IPCC  AR5 Working Group 3, Mitigation of Climate Change – Contribution of Working Group III. Available at: http:// report.mitigation2014.org/spm/ipcc_wg3_ar5_summary-forpolicymakers_approved.pdf (accessed 15 July 2015). Jevrejeva, S., Moore, J. C. & Grinsted, A.  Influence of the arctic oscillation and El Niño-Southern Oscillation (ENSO) on ice conditions in the Baltic Sea: the wavelet approach. Journal of Geophysical Research 108 (D21), 4677. Kahya, E. & Dracup, J. A.  U.S. streamflow patterns in relation to the El Niño/ Southern Oscillation. Water Resources Research 29 (8), 2491–2503.

Journal of Water and Climate Change

08.1

2017

Kalra, A. & Ahmad, S.  Evaluating changes and estimating seasonal precipitation for the Colorado River Basin using a stochastic nonparametric disaggregation technique. Water Resources Research 47, W05555. Kalra, A. & Ahmad, S.  Estimating annual precipitation for the Colorado River Basin using oceanic-atmospheric oscillations. Water Resources Research 48, W06527. Kang, S. & Lin, H.  Wavelet analysis of hydrological and water quality signals in an agricultural watershed. Journal of Hydrology 338, 1–14. Kumar, P. & Foufoula-Georgiou, E.  Wavelet analysis for geophysical applications. Reviews of Geophysics 35, 385–412. Lau, K. M. & Weng, H.  Climate signal detection using wavelet transform: how to make a time series sing. Bulletin of the American Meteorological Society 76, 2391–2402. Lins, H. F.  Regional streamflow regimes and hydroclimatology of the United States. Water Resources Research 33, 1655–1668. Lins, H. F.  USGS Hydro-Climatic Data Network 2009 (HCDN–2009), US Geological Survey Fact Sheet 2012–3047, p. 4. http://pubs.usgs.gov/fs/2012/ 3047/. Lins, H. F. & Slack, J.  Streamflow trends in the United States. Geophysical Research Letters 26, 227–230. Mantua, N. J. & Hare, S. R.  The pacific decadal oscillation. Journal of Oceanography 58, 35–44. McCabe, G. J. & Clark, M. P.  Trends and variability in snowmelt runoff in the western United States. Journal of Hydrometeorology 6 (4), 476–482. McCabe, G. J. & Dettinger, M. D.  Decadal variations in the strength of ENSO teleconnections with precipitation in the western United States. International Journal of Climatology 19 (13), 1399–1410. McCabe, G. J. & Dettinger, M. D.  Primary modes and predictability of year-to-year snowpack variations in the western United States from teleconnections with Pacific Ocean climate. Journal of Hydrometeorology 3 (1), 13–25. McCabe, G. J. & Wolock, D. M.  A step increase in streamflow in the conterminous United States. Geophysical Research Letters 29 (24), 2185. McCabe, G. J., Betancourt, J. L. & Hidalgo, H. G.  Associations of decadal to multidecadal sea-surface temperature variability with Upper Colorado river flow. Journal of the American Water Resources Association 43 (1), 183–192. Meyers, S. D., Kelly, B. G. & O’Brian, J. J.  An introduction to wavelet analysis in oceanography and meteorology: with application to the dispersion of Yanai waves. Monthly Weather Review 121, 2858–2866. Miles, E. L., Snover, A. K., Hamlet, A. F., Callahan, B. & Fluharty, D.  Pacific northwest regional assessment: the impacts of climate variability and climate change on the water resources of the Columbia River Basin. Journal of the American Water Resources Association 36 (2), 399–420. Milly, P., Julio, B., Malin, F., Robert, M., Zbigniew, W., Dennis, P. & Ronald, J.  Stationarity is dead. Ground Water News Views 4 (1), 6–8.

Page 163

K. A. Tamaddun et al.

ENSO and PDO relationship with western US streamﬂow

Nakken, M.  Wavelet analysis of rainfall-runoff variability isolating climatic from anthropogenic patterns. Environmental Modelling and Software 14 (4), 283–295. Önöz, B. & Bayazit, M.  The power of statistical tests for trend detection. Turkish Journal of Engineering and Environmental Science 27 (4), 247–251. Percival, D. B. & Walden, A. T.  Wavelet Methods for Time Series Analysis. Cambridge University Press, Cambridge, p. 594. Polikar, R.  The wavelet tutorial. Available at: http://users. rowan.edu/∼polikar/wavelets/wttutorial.html. Pozo-Vázquez, D., Esteban-Parra, M. J., Radrigo, F. S. & CastroDiez, Y.  The association between ENSO and winter atmospheric circulation and temperature in the North Atlantic region. Journal of Climate 14, 3408–3420. Praskievicz, S. & Chang, H.  Winter precipitation intensity and ENSO/PDO variability in the Willamette Valley of Oregon. International Journal of Climatology 29 (13), 2033–2039. Rice, J. S., Emanuel, R. E., Vose, J. M. & Nelson, S. A. C.  Continental U.S. streamflow trends from 1940 to 2009 and their relationships with watershed spatial characteristics. Water Resources Research 51, 6262–6275. Ropelewski, C. F. & Halpert, M. S.  North American precipitation and temperature patterns associated with ElNino-Southern Oscillation (ENSO). Monthly Weather Review 114, 2165–2352. Sagarika, S., Kalra, A. & Ahmad, S.  Evaluating the effect of persistence on long-term trends and analyzing shifts in streamflows of the continental United States. Journal of Hydrology 517, 36–53. Sagarika, S., Kalra, A. & Ahmad, S. a Interconnection between oceanic-atmospheric indices and variability in the US streamflow. Journal of Hydrology 525, 724–736. Sagarika, S., Kalra, A. & Ahmad, S. b Pacific Ocean SST and Z500 climate variability and western U.S. seasonal streamflow. International Journal of Climatology 36, 1515– 1533. Shrestha, E., Ahmad, S., Johnson, W. & Batista, J. R.  The carbon footprint of water management policy options. Energy Policy 42, 201–212. Souza, E. M. P., Echer, E., Nordemann, D. J., Rigozo, N. R. & Prestes, A.  Wavelet analysis of a centennial (1895–1994) southern Brazil rainfall series (Pelotas, 31 460 19″S 52 200 33″W). Climatic Change 87 (3–4), 489–497. W

Journal of Water and Climate Change

08.1

2017

Tamaddun, K., Kalra, A. & Ahmad, S.  Identification of streamflow changes across the Continental United States using variable record lengths. Hydrology 3 (2), 24. Tang, C., Chen, D., Crosby, B. T., Piechota, T. C. & Wheaton, J. M.  Is the PDO or AMO the climate driver of soil moisture in the Salmon River Basin, Idaho? Global and Planetary Change 120, 16–23. Taylor, G. H. & Hannan, C.  The Climate of Oregon. Oregon State University Press, Corvallis, OR, USA. Torrence, C. & Compo, G. P.  A practical guide to wavelet analysis. Bulletin of the American Meteorological Society 79 (1), 61–78. Torrence, C. & Webster, P.  Interdecadal changes in the ENSO–monsoon system. Journal of Climate 2679–2690. Available at: http://journals.ametsoc.org/doi/abs/10.1175/ 1520-0442(1999)012%3C2679:ICITEM%3E2.0.CO;2 (accessed 29 March 2015). Trenberth, K. E. & Fasullo, J.  Water and energy budgets of hurricanes and implications for climate change. Journal of Geophysical Research Atmospheres 112 (D23), 107. Trenberth, K. E. & Hurrell, J. W.  Decadal atmosphere-ocean variations in the Pacific. Climate Dynamics 9, 303–319. Villarini, G., Serinaldi, F., Smith, J. A. & Krajewski, W. F.  On the stationarity of annual flood peaks in the continental United States during the 20th century. Water Resources Research 45 (8). Wallace, J. M., Zhang, Y. & Lau, K. H.  Structure and seasonality of interannual and interdecadal variability of the geopotential height and temperature-fields in the northern-hemisphere troposphere. Journal of Climate 6 (11), 2063–2082. Weider, K. & Boutt, D. F.  Heterogeneous water table response to climate revealed by 60 years of ground water data. Geophysical Research Letters 37 (24), 10–15. Wu, G., Li, L., Ahmad, S., Chen, X. & Pan, X.  A dynamic model for vulnerability assessment of regional water resources in arid areas: a case study of Bayingolin, China. Water Resources Management 27 (8), 3085–3101. Yiou, P., Sornette, D. & Ghil, M.  Data-adaptive wavelets and multi-scale single spectra analysis. Physica D 142, 254–290. Zhang, F., Li, L., Ahmad, S. & Li, X.  Using path analysis to identify the influence of climatic factors on spring peak flow dominated by snowmelt in an Alpine watershed. Journal of Mountain Science 11 (4), 990–1000.

First received 15 December 2015; accepted in revised form 15 July 2016. Available online 17 August 2016

Page 164

Journal of

Water & Health

ISSN 1477-8920 iwaponline.com/jwh

Journal of Water and Health covers the dissemination of information on the health implications and control of waterborne microorganisms and chemical substances in the broadest sense for developing and developed countries worldwide. This includes microbial toxins, chemical quality and the aesthetic qualities of water. The journal’s scope includes: • • • • • • • • • • • • • • • • • • •

Epidemiology Risk assessment Detection and ecology of pathogens in the environment Water and wastewater treatment Disinfection, and disinfection by-products Indicators of water and waste quality Regulatory issues and standard development Water quality surveys and assessments Monitoring Microbial toxins (including cyanobacteria) Chemical and physical quality of water as it affects health Endocrine disruptors Taste and odour Impacts of water quality on food quality Impact of climate change on water quality Water based diseases Water-based insect vectors of disease Water policy and health Health effectiveness of water management For more details, visit iwaponline.com/jwh

Page 168

175

15.2

2017

Towards a research agenda for water, sanitation and antimicrobial resistance Susanne Wuijts, Harold H. J. L. van den Berg, Jennifer Miller, Lydia Abebe, Mark Sobsey, Antoine Andremont, Kate O. Medlicott, Mark W. J. van Passel and Ana Maria de Roda Husman

ABSTRACT Clinically relevant antimicrobial resistant bacteria, genetic resistance elements, and antibiotic residues (so-called AMR) from human and animal waste are abundantly present in environmental samples. This presence could lead to human exposure to AMR. In 2015, the World Health Organization (WHO) developed a Global Action Plan for Antimicrobial Resistance with one of its strategic objectives being to strengthen knowledge through surveillance and research. With respect to a strategic research agenda on water, sanitation and hygiene and AMR, WHO organized a workshop to solicit input by scientists and other stakeholders. The workshop resulted in three main conclusions. The ﬁrst conclusion was that guidance is needed on how to reduce the spread of AMR to humans via the environment and to introduce effective intervention measures. Second, human exposure to AMR via water and its health impact should be investigated and quantiﬁed, in order to compare with other human exposure routes, such as direct transmission or via food consumption. Finally, a uniform and global surveillance strategy that complements existing strategies and includes analytical methods that can be used in low-income countries too, is needed to monitor the magnitude and dissemination of AMR. Key words

| antibiotics, antimicrobial resistance (AMR), risk assessment, risk management, sanitation, water

Susanne Wuijts Harold H. J. L. van den Berg Mark W. J. van Passel Ana Maria de Roda Husman (corresponding author) National Institute for Public Health and the Environment (RIVM), P.O. Box 1, 3720 BA Bilthoven, The Netherlands E-mail: ana.maria.de.roda.husman@rivm.nl Jennifer Miller Virginia Polytechnic Institute and State University, Blacksburg, VA, USA Lydia Abebe Mark Sobsey University of North Carolina at Chapel Hill, Chapel Hill, NC, USA Antoine Andremont Diderot Medical School, University of Paris, Paris, France and Bichat Hospital Bacteriology Laboratory, Paris, France Kate O. Medlicott World Health Organization (WHO), Geneva, Switzerland Ana Maria de Roda Husman Institute for Risk Assessment Sciences (IRAS) of Utrecht University, Utrecht, The Netherlands

INTRODUCTION ‘Without urgent, coordinated action, the world is headed for

In May 2015 the World Health Assembly of the WHO

a post-antibiotic era, in which common infections which

approved the Global Action Plan on Antimicrobial Resist-

have been treatable for decades can once again kill.’ Dr

ance (GAP on AMR) (WHO a). AMR elements,

Keiji Fukuda, World Health Organization (WHO) Assistant

including resistant bacteria and antibiotic resistance genes

Director-General for Health Security

(ARGs or AMR genes), as well as antibiotic residues, are common in water, wastewater, and feces. Therefore, under-

This is an Open Access article distributed under the terms of the Creative

standing and addressing the role of water, sanitation and

Commons Attribution Licence (CC BY-NC-ND 4.0), which permits copying

hygiene (WaSH) in combatting AMR, including antibiotic

and redistribution for non-commercial purposes with no derivatives, provided the original work is properly cited (http://creativecommons.org/

resistance, is a critical element of the GAP on AMR.

licenses/by-nc-nd/4.0/).

Box 1 summarizes the strategic objectives of the GAP on

doi: 10.2166/wh.2017.124

Page 169

176

S. Wuijts et al.

Journal of Water and Health

Towards a research agenda for water, sanitation and AMR

15.2

2017

environment in the spread of AMR, mostly antibiotic resistBox 1

Strategic objectives of WHO’s Global Action Plan on Antimicrobial Resist-

ance (GAP on AMR)

ance,

and

possible

adverse

health

outcomes

environmental exposures in order to identify knowledge

1. To improve awareness and understanding of AMR

gaps and to develop a research agenda for WaSH aspects

2. To strengthen knowledge through surveillance and

of AMR. Key issues in risk assessment, risk management,

research

and monitoring and surveillance were discussed. This

3. To reduce the incidence of infection

research agenda aims to identify knowledge gaps that need

4. To optimize the use of antimicrobial agents

to be addressed in order to achieve the WaSH-related objec-

5. To develop the economic case for sustainable invest-

tives (objectives 1–3, see Box 1) of the GAP on AMR and

ment that takes account of the needs of all

can thus be a foundation for future global guidance and

countries, and increase investment in new medi-

action on WaSH and AMR.

cines,

diagnostic

tools,

vaccines

and

other

interventions.

The program started with several presentations on the latest insights regarding AMR and WaSH from both health and environmental perspectives. During breakout sessions, the participants were divided into three groups to consider

AMR. The role of WaSH in combatting AMR focuses on improved awareness and understanding through surveillance and research, in order to reduce the incidence of AMR infection. WaSH thus contributes to objectives 1–3

questions on the following topics:

• • •

risk assessment risk management monitoring and surveillance.

of the GAP on AMR (Box 1). Current

WHO

Guidelines

for

Drinking-Water,

Recreational Water, and Safe Use of Wastewater do not yet contain information on antibiotics and other antimicrobial agents, their metabolites, AMR bacteria, or AMR genes. Occurrence and trend data for these elements are needed for risk assessment and risk management strategies for health-related AMR in the environment to be developed and implemented. A WHO workshop was organized as a side event of the IWA Health Related Water Microbiology Symposium on September 18, 2015, in Lisbon, to identify knowledge gaps and to develop a research agenda for WaSH aspects of AMR. This paper describes the presentations, input, discussions, discussion and results of this workshop as building blocks for a research agenda.

RESULTS Prof. Dr Ana Maria de Roda Husman of the National Institute of Public Health and the Environment (RIVM) of the Netherlands opened the discussion on the importance of AMR in the environment. The complex interaction of the natural environment, i.e., water, soil, and air (Huijbers et al. ), and the interplay of AMR bacteria, AMR genes, and antibiotic residues in the environment were highlighted. AMR originates from humans and animals exposed to antibiotics, and from the environment itself; thus, AMR should be approached as a ‘One Health’ problem. Sources for AMR include, for example, wastewater and manure, as was shown during her presentation. There needs to be recognition that there are multiple uses of water, such as washing, irrigation, recreation, and drinking, that contribute to the increasing risk of exposure to

METHODS

AMR. Existing safety plans, such as Water Safety Plans and

During the workshop ‘Developing a Research Agenda for

and De Roda Husman suggested that greater attention

Water, Sanitation and Hygiene (WaSH) and Antimicrobial

should be afforded on AMR in such Safety Plans.

Sanitary Safety Plans, do not speciﬁcally address AMR yet

resistance (AMR)’, input from the scientiﬁc community of

Kate Medlicott of the WHO went on to summarize the

water professionals was solicited on the role of the

WHO GAP on AMR. The World Health Assembly at its

Page 170

177

S. Wuijts et al.

Journal of Water and Health

Towards a research agenda for water, sanitation and AMR

15.2

2017

67th session adopted resolution WHA 67.25 on combatting

NDM-1 from India to the UK. Andremont stressed the

antimicrobial resistance (WHO a). Through this resol-

importance of sanitation in his presentation. Poor sanitation

ution, the Health Assembly requested the development of

or the lack of sanitation is an important pressure on AMR

a draft GAP to combat AMR, including antibiotic resistance.

(Andremont & Walsh ). The One Health concept links

The WHO has led the development of a GAP that reﬂects

the environment, agriculture/food, and sanitation/commu-

the commitment, perspectives, and roles of all relevant sta-

nity health in an integrated risk-based approach.

keholders, and in which everyone has clear and shared

The next speaker, Lydia Abebe of the University of

ownership and responsibilities. The action plan is built on

North Carolina (USA), presented the preliminary ﬁndings

six guiding principles: public and stakeholder engagement,

of an ongoing systematic literature review of WaSH and

actions based on best available knowledge and evidence,

AMR that focused on the current status and gaps in knowl-

‘prevention ﬁrst’, ‘access not excess’, sustainability, and

edge. The literature review focused on AMR bacteria in the

incremental targets for implementation.

environment and associated human health implications.

Furthermore, a brieﬁng note on AMR in the environment was prepared (WHO b).

Abebe discussed the purpose of the review, which is to evaluate the role of environmental exposure to AMR bac-

Prof. Dr Antoine Andremont, of Diderot Medical School

teria and human health outcomes through evaluating the

and Bichat Hospital Bacteriology Laboratory (France),

methods used to create the linkages. Expected outcomes

explained the role of sanitation in the development and

from the review will be an assessment of methods used to

spread of AMR. In his presentation, AMR was addressed

create environmental linkages between transmission of

from a medical perspective and he indicated the need to con-

AMR bacteria in environmental and human reservoirs to

sider the roles of the environment and agriculture in addition

human health outcomes to identify gaps, and thereby

to clinical contributions to the development of resistance

make recommendations for establishing stronger evidence

(Allen et al. ; Graham et al. ; Zhang et al. ). He

for links between environmental exposure to AMR bacteria

demonstrated that, in France, single-antibiotic resistance

and adverse human health outcomes. This work will lead to

was predominantly of hospital origin, but this has evolved

a literature review that focuses on AMR and WaSH from an

to community-borne (food and environment) transmission

integrated One Health perspective, and it is envisaged that

whereby multi-drug resistant bacteria return from the

this will serve to stimulate a research agenda on AMR and

environment back into the hospital.

WaSH.

Two examples of major AMR genes impacting human

Prof. Dr Mark Sobsey of the University of North Caro-

health and coming from environmental sources were

lina (USA) continued with an overview of the research

presented:

topics from the Joint Programming Initiative on Antimicro-

•

bial Resistance (JPI-AMR) agenda for the European Union

ESBL genes confer antibiotic resistance to all beta-lactams

except

carbapenems

(plus

multi-resistance)

(Humeniuk et al. );

•

NDM-1 genes confer the same phenotype plus resistance to carbapenems (Kumarasamy et al. ).

(JPI-AMR ). The JPI-AMR programme seeks to harmonize AMR priorities and research initiatives to address research gaps. There are six priority topics that will be targeted

reduce

AMR:

therapeutics

(alternatives to

antibiotics), diagnostics (treatment and prevention of infec-

These examples show that there is a feedback loop,

tion), surveillance (monitoring, including of environmental

whereby more infection results in more antibiotic use,

reservoirs), transmission, environment (including sources,

which results in more antibiotic resistance. More ESBL

selection and dissemination mechanisms), and interventions

infection leads to more use of carbapenems, which leads

(for example, treatment technologies). One of the priority

to the rise in carbapenem (NDM-1) resistance (Rossolini

topics, selection and dissemination mechanisms in the

et al. ).

environment, emphasizes the assessment of the contribution

The contribution of ‘medical tourism’ to AMR was

of pollution of the environment with antibiotics, antibiotic

shown by Kumarasamy et al. (), with the spread of

residues, and AMR bacteria on the sources, occurrence, Page 171

178

S. Wuijts et al.

Journal of Water and Health

Towards a research agenda for water, sanitation and AMR

15.2

2017

and spread of AMR and the development of strategies to

excreta and manure, but it is more difﬁcult to identify and

minimize environmental contamination by antibiotics and

quantify these conditions in the other domains of the

AMR bacteria.

water cycle. Research is needed on the selection of bacteria,

Finally, Prof. Dr Mark Sobsey discussed the urgency of

AMR properties, locations, sample matrices, and standar-

new research into AMR surveillance. Many studies have

dized analytical methods for monitoring. The HACCP

been carried out to detect AMR in environmental samples.

approach (Hazard Analysis and Critical Control Points)

Nevertheless, there is no organized, harmonized and func-

could be a useful method to identify the ‘critical control

tional system for AMR surveillance in the environment.

points’, bacterial analytical methods, and matrices. Water

Globally used monitoring methods for environmental

Safety Plans and Sanitary Safety Plans are primarily based

microbial surveillance are the detection of Escherichia coli

on this approach. It is important to ﬁnd the most important

and intestinal enterococci. These methods are based on

pathways, whether these be drinking water or other environ-

the detection of fecal contamination, but no global or

mental exposure pathways and media, e.g., irrigation with

national surveillance systems are in place for the detection

wastewater. This knowledge would help to inform the

of environmental AMR. Dr Sobsey presented the need for

public and other stakeholders on effective measures.

the establishment of an international, standardized surveil-

There is discussion on the importance of water as a path-

lance programme for AMR and antibiotic use in human

way for AMR compared to other exposure routes, such as

and agriculture settings that includes targeted environ-

food or from person to person. However, this does not

mental monitoring relevant to human exposures. Potential

mean that inadequate sanitation and fecally contaminated

approaches to environmental surveillance for AMR bacteria

water could not be important routes for AMR transmission.

were discussed.

Little is known about exposure to AMR through WaSH routes and the resulting effects on public health. Therefore,

Risk assessment During this breakout session, facilitated by Prof. Dr Antoine Andremont, the following questions were asked of the

research is needed in this ﬁeld. So far, experiments to transfer genes in laboratory simulations have not succeeded. The potential risk of infection by AMR bacteria through the consumption of drinking water gives rise to the public’s

participants:

questions and concern. This is especially a concern in areas

•

implemented. Governments and water companies need to

What are the needs to identify and quantify the sources, occurrence, and transport of AMR bacteria and their genes?

•

What are the needs to estimate risk of AMR bacteria to human health? Multiple studies (references were made by participants

to studies conducted in Germany, New Zealand, Australia,

in which water reuse projects are being developed and address these questions supported by scientiﬁc data that are based on actual evidence of exposure and observed health risks. It is important that these data are collected in a transparent way with good study design and methods, allowing for easy comparison with studies in other countries or regions.

Denmark, the UK, the Netherlands, Thailand, and South

Following this inventory, the discussion shifted towards

Africa) have been carried out on the identiﬁcation of anti-

the needs of resource-limited countries. One of the participants

biotic (AB) residues, and AMR bacteria and AMR genes

sketched the situation in India, where generic antibiotics are

throughout the water cycle as well as certain ‘hotspots’

very inexpensive and readily available for a large population,

such as hospital wastewater systems, biogas plants, waste-

and there is often poor sanitation. A worldwide analysis

water treatment plants, and livestock such as poultry (New

(Woerther et al. ) demonstrated that Asia is one of the con-

Zealand). The monitoring data now available are insufﬁ-

tinents with the highest incidence of ESBL-enterobacteria

cient to identify occurrence status and trends in the

fecal carriage. From the African region, the aspect of the sus-

appearance of AMR. It is clear, however, that AMR is

ceptibility of a resource-limited population to infections from

found throughout the pathways from human and animal to

WaSH exposures was mentioned in relation to the increased

Page 172

179

S. Wuijts et al.

Journal of Water and Health

Towards a research agenda for water, sanitation and AMR

15.2

2017

risk of AMR infections. The limited resources in low resource

of AMR bacteria, AMR genes, and antibiotic residues in raw

settings call for a pragmatic approach with a baseline surveil-

sewage, although, not using a standardized methodology.

lance strategy supported by appropriate monitoring and

Moreover, quantitative information is needed on potential

further strengthening of water and sanitation conditions.

environmental concentrations; loading mass/volume; and

Research is needed to develop these approaches.

total load in animals, agriculture, wildlife, household, and hospital settings. Loading and concentration data should include those related to antibiotic metabolites (i.e., excreted

Risk management

forms and potential environmental transformation pro-

During this breakout session, facilitated by Prof. Dr Ana Maria De Roda Husman, the following questions were asked of the participants:

•

With respect to treatment, there are serious knowledge gaps around fate (persistence and survival), interactions,

ment systems to identify policies, practices, and tools to

and treatment efﬁciency (removal, log reductions) of anti-

minimize human exposure?

biotic compounds, metabolites, AMR bacteria, and AMR

agement

systems

should

focus

prevention

and

treatment. Educational materials to increase awareness of appropriate use of antibiotics and proper antibiotic disposal is required in order to reduce the release of antibiotics to wastewater and the environment. The group suggested that a literature review of the drug/pharmaceutical management practices of various countries would generate ideas of policy and waste management systems for the safe and environmentally protective disposal of antibiotics. The literature review should include veterinary (animal) and agriculture, household, and hospital/healthcare practices for antibiotic use and the disposal of unused antibiotics and the governance structure in place for this. In terms of water and wastewater technology applications, three dominant questions centered on identifying basic mass balance inputs: What levels of AMR bacteria, genetic material, or antibiotic residues are entering the treatment system?

•

What levels are removed or could potentially be removed?

•

between samples.

What are the needs with regard to practical risk manage-

Research in relation to the establishment of risk man-

•

toilet ﬂushing). These research questions may be challenging and efforts will be required to ensure that analytical methods and detection limits are adequate and standardized

What are the needs with regard to water and wastewater treatment technologies?

•

ducts) as well as parent compounds (direct disposal via

What levels in efﬂuent and biosolids are necessary to protect the receiving environment and ultimately beneﬁt the clinical settings?

genes in water and wastewater technologies. Treatment studies should consider both AMR bacteria as well as AMR genes because DNA may persist despite death of the cell or biological entity. Alternatively, there may be other treatment markers or indicators of AMR bacteria or genes present as indicators of removal or reduction. ‘Critical control points’ should be identiﬁed. Research into treatment technologies should also consider low-cost technologies, appropriate developing world technologies, conventional treatment (water and wastewater), non-standard techniques such as solar/sunlight, and septic systems with/without reticulated water supply. There is a need for criteria and guidelines to assess technologies in order to assist policymakers and utility managers in identifying appropriate technologies and their performance capabilities. To address the third dominant question, the group noted that treatment requires a goal. How much removal is necessary to make a difference in controlling ARM impacts on the environment and human health? What levels in the environment are acceptable with regard to public health protection? How can we work to quantify health impacts associated with reductions in drug usage or drug concentrations (will reducing drug usage have negative health impacts)? What are the health impacts associated with AMR bacteria or

With regard to answering these questions, the group

genes and loading to the environment, that is, will reducing

noted that studies have been undertaken on the occurrence

AMR bacteria or genes loading to the environment improve Page 173

180

S. Wuijts et al.

Journal of Water and Health

Towards a research agenda for water, sanitation and AMR

15.2

2017

clinical health outcomes? A potential method of analysis is

Studies on environmental surveillance of AMR do not

a Quantitative Microbial Risk Assessment (QMRA) as

always link to clinical relevance, e.g., studies do not focus

suggested by Ashbolt et al. (). An example QMRA

on obtaining data relevant to human exposures from

study for exposure to ESBL in recreational waters was pub-

environmental pathways, AMR bacteria of human health/

lished recently (Schijven et al. ), but further studies on

clinical concern, or human exposure media; studies do not

this subject are necessary.

focus on known major sources of AMR release to the environment. To improve the linkage to clinical relevance,

Monitoring and surveillance

the group discussed the importance of combining data

During this breakout session, Prof. Dr Mark Sobsey initiated

Health approach; communication from the environmental

the discussion with the following questions to the participants:

• • •

What are the needs with regard to monitoring? What are the needs with regard to surveillance? What are the needs with regard to regulatory activities and agents?

from humans, animals, and the environment; the One domain to clinical and veterinary domains; and, linking human surveillance with environmental surveillance. To better address the links between human, animal, and environmental data, designing surveillance strategies with a harmonized and tiered approach was recommended. As a result of monitoring and surveillance, data on AMR in the environment will be collected, which can form the

Numerous studies have detected AMR in environmental

evidence base to take actions to minimize exposure and

samples through a variety of culture and molecular methods

human health risks. Therefore, a threshold (regarding risk

(Huijbers et al. ). Nevertheless, there is no organized,

level and safety) of AMR in the environment within a regu-

harmonized, and functional system for AMR surveillance

latory framework is needed. The threshold should answer

in the environment.

the following questions: What is an acceptable level of

There are different reasons to perform environmental

AMR risk? When should management take action to further

surveillance of AMR, such as identifying emerging genes

minimize risk? Furthermore, the gathered data should be

and a potential genetic relationship; characterization of

communicated to other relevant ﬁelds and stakeholders,

pharmaceutical waste; identifying AMR bacteria, AMR

such as, for example, healthcare professionals, policy-

genes, and antibiotic residues entering the environment;

makers, and water and sanitation experts, and provide

and, identifying other hot spots compared to sewage

advice on reduction or removal of AMR, including rec-

(source tracking). Appropriate methods should be estab-

ommendations on cost-effectiveness.

lished depending on the purpose of AMR surveillance in the environment. During the discussion, different methods were mentioned to identify and quantify AMR bacteria,

DISCUSSION

AMR genes, and antibiotic residues in the environment, such as culture methods and molecular detection methods.

The WHO workshop organized on September 18, 2015, in

Initially, selection of an index parameter needs to be set.

Lisbon, provided the opportunity for participants to contrib-

Different index parameters were discussed: antibiotics,

ute to a research agenda on WaSH and AMR. For each of

detection of AMR in fecal indicators (such as E. coli and

the three key topics discussed, namely risk assessment,

intestinal enterococci), other clinically relevant microorgan-

risk management, and monitoring and surveillance, it was

isms (such as Clostridium difﬁcile, Staphylococcus aureus,

evidently demonstrated that there are more open questions

and bacteriophages), horizontal gene transfer and metage-

than answers at this time. Box 2 links research questions

nomics. Then, standard methods should be prescribed to

to GAP on AMR objectives. A research agenda should be

detect these index parameters. A tiered approach was rec-

consistent with and support the process of focusing on rel-

ommended because of different resource settings, and

evant questions and sharing best practices. Box 2

speciﬁc guidance on this approach is greatly needed.

summarizes the output of the workshop discussion,

Page 174

15.1

2017

Safe drinking water and waterborne outbreaks N. A. Moreira and M. Bondelind

ABSTRACT The present work compiles a review on drinking waterborne outbreaks, with the perspective of production and distribution of microbiologically safe water, during 2000–2014. The outbreaks are categorised in raw water contamination, treatment deﬁciencies and distribution network failure. The main causes for contamination were: for groundwater, intrusion of animal faeces or wastewater due to heavy rain; in surface water, discharge of wastewater into the water source and increased turbidity and colour; at treatment plants, malfunctioning of the disinfection equipment; and for distribution systems, cross-connections, pipe breaks and wastewater intrusion into the network. Pathogens causing the largest number of affected consumers were Cryptosporidium, norovirus, Giardia, Campylobacter, and rotavirus. The largest number of different pathogens was found for the

N. A. Moreira Cranﬁeld Water Science Institute, Cranﬁeld University, Bedfordshire MK43 0AL, UK N. A. Moreira M. Bondelind (corresponding author) Department of Civil and Environmental Engineering, Chalmers, Sven Hultins gata 8, Göteborg 412 96, Sweden E-mail: mia.bondelind@chalmers.se

treatment works and the distribution network. The largest number of affected consumers with gastrointestinal illness was for contamination events from a surface water source, while the largest number of individual events occurred for the distribution network. Key words

| distribution network, drinking water, pathogens, waterborne outbreak, water safety plan, water treatment

INTRODUCTION Drinking water safety plays a signiﬁcant role in establishing

temperature, increases in pH and larger alkalinity generation

the quality of human life in modern societies. In that per-

in the lakes themselves. Additionally, sewage discharge from

spective, problems with microbial pathogens within the

combined sewage systems caused by heavy rainfall has been

production and distribution of drinking water can have an

demonstrated to spread waterborne pathogens within the sur-

important impact on public health. The occurrence of a

face waters. Furthermore, increased temperatures may

waterborne disease outbreak (WBO) may also have the

increase disinfection by-product formation rates in surface

effect of lowering trust, increase perceived risk and decrease

waters at natural temperatures, between 5 and 30 C

acceptance for the drinking water (Bratanova et al. ).

(Delpha et al. ). Consequently, environmental contami-

Waterborne outbreaks are caused by drinking water con-

nation, intensive livestock rearing, surface water and

tamination worldwide (Karanis et al. ). One of the most

discharge of wastewater into drinking water sources are

challenging issues facing the drinking water treatment

risk factors that need to be addressed (Chalmers ).

plants (WTP) are the uncertainties related to climate

In the production of safe and aesthetically suitable water

change and the effect it will have on the surface water quality.

for human consumption, the analysis and evaluation of risks

Increase of extreme hydrological events in addition to

to the complete drinking water system, from the catchment

changes in air temperature may increase the risk of WBOs.

until it reaches the consumer, is considered of paramount

The most vulnerable water bodies to future climate changes

importance by the World Health Organization (WHO). To

are likely to be shallow lakes, where the chemical processes

achieve that aim, a framework for safe drinking water was

will be altered by the impact of an increase in water

developed by the WHO throughout the application of

doi: 10.2166/wh.2016.103

Page 175

N. A. Moreira & M. Bondelind

Safe drinking water and waterborne outbreaks

Journal of Water and Health

15.1

2017

guidelines designated as water safety plans (WSP) (WHO

may not have been mentioned in scientiﬁc publications. In

). Through the WSP, hazards and hazardous events that

total 66 reviewed articles were found to be eligible accordingly

can affect the safety of the production of drinking water

to the criteria: (i) data in the timeframe 2000–2014; (ii) drink-

from the catchment to consumer are identiﬁed. The risks

ing water outbreak conﬁned geographically to Europe, North

associated with the events are assessed and control points

America and New Zealand; (iii) surveillance of potential fac-

and barriers are implemented if needed. The WSP should

tors of interest to the drinking water industry affecting the

be reviewed regularly and continuously updated (Bartram

occurrence of parasite transmission hazards.

et al. ). To quantify the barrier effect and the treatment

The time frame for this study is 2000–2014. Regulations

required, the Microbial Barrier Analysis model (MBA) can

are continuously being updated and implemented for

be used (Ødegaard & Østerhus ). The raw water quality

improved safety of drinking water. Therefore, only recent

is evaluated and according to its quality the necessary treat-

events that may be of interest for the water industry today

ment efﬁciency is determined. Thereafter the removal and

are included in this review. For example, the United King-

inactivation efﬁciency of the barriers installed at the WTP

dom alone was responsible for 73.6% of the waterborne

are calculated. The difference between the required and the

outbreaks in Europe until 2003 (Karanis et al. ). The

calculated barrier efﬁciency shows if supplementary surveil-

implementation of a new set of regulations in the year

lance or additional treatment is required.

2000, concerning drinking water production, that took

In spite of the generalised use of risk ranking in WSP, the evaluation and comparison of water safety measures does

place in the UK led to reductions in cryptosporidiosis that were considered statistically relevant (Lake et al. ).

not have a common and structured approach (Lindhe et al.

In this review drinking water outbreaks conﬁned geo-

). As a result, the primary safety procedures against

graphically to Europe, North America and New Zealand

microbiological hazards are still capable sanitation and

have been reviewed. Here public national systems to register

drinking water infrastructures (Baldursson & Karanis ).

the occurrence of waterborne outbreaks are available. In

Thus, reviewing WBOs associated with drinking water pro-

developing countries the information related with WBOs is

duction can help to shed light on the most problematic

less available or even absent and therefore these countries

issues faced by the water industry. The aim of the present

have not been included in this review (Baldursson & Karanis

work is to review causes for drinking water disease out-

). Thus the available reports of incidents, according to the

breaks, assessing possible patterns and accountability issues

stipulated eligibility criteria, resulted in the inclusion of 15

for those events in order to improve drinking water safety.

countries: Canada, Denmark, Finland, France, Greece, Ireland, Italy, Netherlands, New Zealand, Norway, Spain, Sweden, Switzerland, the UK and the USA. The creation of

METHOD

public national systems to register the frequency and prevalence of waterborne outbreaks or protozoan infections may

This study of causes for drinking water disease outbreaks is

vary among the countries. The surveillance of potential fac-

based on information and literature collected from sources

tors of interest to the drinking water industry affecting the

including Scopus, Eurosurveillance, PubMed, New Zealand’s

occurrence of parasite transmission hazards has to be

Institute of Environmental Science and Research (ESR),

known for the event to be included in this review.

Canada Communicable Disease Report (CCDR) and Morbid-

The results of this review are summarised in Tables 1–4

ity and Mortality Weekly Report from the USA CDC (Centers

that present the year of the event; country and speciﬁc

for Disease Control and Prevention). Keywords used in the

location (when available); estimated number of infections;

search comprised: waterborne, water treatment, outbreak,

population served by the water works or distribution

Cryptosporidium, Campylobacter, Giardia, norovirus, rota-

system; causative agent; probable cause for the outbreak to

virus, and adenovirus. The number of identiﬁed outbreaks

occur; and key reference. The medium value was used

may be misrepresentative because of the voluntary nature of

when the number of estimated cases was presented in the

reporting processes (Brunkard et al. ) or that the events

form of an interval in the reviewed articles.

Page 176

Table 1

N. A. Moreira & M. Bondelind

Journal of Water and Health

Safe drinking water and waterborne outbreaks

15.1

2017

List of outbreaks originated from raw water contamination (groundwater)

Year

Location, country

Est. cases

Pop. served

Causative agent

Probable causes for outbreak occurring

Reference

2000

Walkerton, Canada

2,300

4,800

Campylobacter and E. coli

Contamination from livestock faecal residue following heavy rainfall

Hrudey et al. ()

2000

Clitheroe, UK

17,252

Cryptosporidium

Contamination with animal faeces following abnormally heavy rain

Howe et al. ()

2001

Southern Finland

1,000

18,000

Campylobacter

Floodwater from a dike contaminated by runoff (probably from animal sources)

Hänninen et al. ()

2002

Isère, France

2,000

5,600

Norovirus

Heavy rains lead to overﬂow in the sewage treatment works upstream and the ﬂooding of raw water borehole

Tillaut et al. ()

2002

Transtrand, Sweden

500

772

Norovirus

Crack in sewage pipe 10 m from one of the supplying wells

Carrique-Mas et al. ()

2004

Ohio, USA

1,450

Unknown

Campylobacter and norovirus

Multiple contamination of aquifer from onsite septic systems, land application of sludge and inﬁltration of run-off

O’Reilly et al. ()

2005

Xanthi, Greece

709

13,956

Norovirus

Contamination of well following a heavy rain event

Papadopoulos et al. ()

2006

Xanthi, Greece

1,640

100,882

Norovirus

Groundwater contamination following a heavy rain event

Vantarakis et al. ()

2006

Portlaw, Ireland

Unknown

Cryptosporidium

Moderate risk of groundwater contamination previously identiﬁed; UV treatment unit was commissioned

HPSC ()

2009

Evertsberg, Sweden

200

400

Norovirus

Well contaminated by snowmelt

Riera-Montes et al. ()

2011

Agrigento, Italy

156

4,965

Norovirus

Inﬁltration of contaminated surﬁcial waters following heavy rain

Giammanco et al. ()

RESULTS

origin of the drinking water supply: groundwater-related WBOs in Table 1, and surface water-related WBOs in Table 2.

Three areas of the WBOs origins in the drinking water sys-

Eleven drinking water-related outbreaks were associated

tems are analysed in this paper: raw water contamination;

with groundwater contamination, which instigated gastroin-

treatment deﬁciencies at the waterworks; and distribution

testinal illness amongst an estimated total of 10,021

systems failure.

consumers (Table 1, Figure 1). Even though the large majority (82%) of reported outbreaks originated by groundwater contamination occurred before 2007, no time-related pattern can

WBOs caused by raw water contamination

be inferred due to the signiﬁcant delay between incidents and dates of reporting.

The probable causes for outbreaks correlated with the con-

The aetiological agents for the events with groundwater

tamination of raw water in the catchment areas are shown

contamination were norovirus in six outbreaks, Cryptospori-

in Tables 1 and 2 and Figures 1–3. The enteric disease out-

dium in two events, one event with Campylobacter, one

breaks have been divided into two categories, specifying the

with two bacterial pathogens (Escherichia coli and Page 177

N. A. Moreira & M. Bondelind

Table 2

Journal of Water and Health

Safe drinking water and waterborne outbreaks

15.1

2017

List of outbreaks originated from raw water contamination (surface water)

Year

Location, country

Est. cases

Pop. served

Causative agent

Probable causes for outbreak occurring

Reference

2002

Midlands, Ireland

>31

25,000

Cryptosporidium

Contamination with farmyard slurry and manure following very heavy rains

Jennings & Rhatigan ()

2002

St. Maria de Palautordera, Spain

756

6,343

Shigella

Heavy rain led mud and organic material into the WTP

Arias et al. ()

2004

Bergen, Norway

6,000

48,000

Giardia

Leaking sewage pipes with drainage to the raw water source

Nygård et al. (), Røstum et al. ()

2005

Gwynedd and Anglesey, UK

231

60,000

Cryptosporidium

Natural (wildlife) contamination, septic tanks and sewage treatment works; streaming and stratiﬁcation in raw water (lake); UV system subsequently installed

Mason et al. (), Chalmers et al. ()

2005

South East England, UK

140

Unknown

Cryptosporidium

Low water levels in the river may have reduced dilution from sewage discharge

Nichols et al. ()

2005

Oregon, USA

Unknown

Campylobacter and E. coli

Inadequate treatment after heavy rainfall conditions

Yoder et al. ()

2006

Cardrona, New Zealand

218

3,800

Norovirus

Contamination from sewage overﬂow

Hewitt et al. ()

2007

Galway, Ireland

304

Unknown

Cryptosporidium

Very wet winter contributed to contamination of lake probably due to run-off from land following slurry spreading

Pelly et al. (), HPSC ()

2008

Lilla Edet, Sweden

2,400

7,500

Norovirus

Contaminated raw water from point source pollution caused by wastewater

Larsson et al. ()

2009

San Felice del Benaco, Italy

299

3,360

Rotavirus and norovirus

Contamination of lake due to overcapacity of the sewage system and/or illegal discharge

Scarcella et al. ()

2010

Östersund, Sweden

27,000

51,000

Cryptosporidium

Faecal contamination of raw water

Widerström et al. ()

2011

Skellefteå, Sweden

20,000

71,580

Cryptosporidium

Contamination from wastewater

Andersson et al. ()

2012

Elassona, Greece

3,620

37,264

Rotavirus

Heavy rain lead to increased coloured water

Mellou et al. ()

Campylobacter), and also one with both norovirus and

raw water quality, sewage contamination, and snowmelt

Campylobacter. Taking into account the information dis-

were associated with one event each; ﬁnally, multiple con-

played in Table 1 and Figures 2 and 3, norovirus is the

tamination causes were responsible for one outbreak.

prevailing pathogen being present in seven of the WBOs,

Surﬁcial run-off seems to be the suspected cause for the

even though on one occasion as part of a multi-agent out-

large majority (73%) of raw water contamination occur-

break. Campylobacter, on the other hand, was present in

rences, since the events are mostly caused by inﬁltration of

three outbreaks, but only on one occasion was it the single

polluted water subsequent to heavy rainfall circumstances.

detected aetiological agent.

In three outbreaks, animal faecal residues were the probable

Several causes of the WBOs for the events with ground-

origin for the microbiological contamination.

water contamination are presented, where heavy rain was

The outbreaks for the events with groundwater contami-

linked to six outbreaks; contaminated runoff, decreased

nation show that ﬁve countries endured more than a 1,000

Page 178

Table 3

N. A. Moreira & M. Bondelind

Journal of Water and Health

Safe drinking water and waterborne outbreaks

15.1

2017

List of outbreaks originated from treatment deﬁciencies at the WTPs

Year

Location, country

Est. cases

Pop. served

Causative agent

Probable causes for outbreak occurring

Reference

2000

Gourdon, France

2,600

7,088

Campylobacter, rotavirus and norovirus

Failure in the chlorination system (and possible contamination of groundwater from agricultural run-off)

Gallay et al. ()

2000

Colorado, USA

Unknown

Giardia

Multiple failures in the pumping mechanism and ﬁltration system; inadequate time for chlorination due to increased demand

Lee et al. ()

2001

Saskatchewan, Canada

6,450

18,000

Cryptosporidium

Treatment deﬁciencies after maintenance work because of increased turbidity

Stirling et al. ()

2001

Hawkes Bay, New Zealand

186

295

Campylobacter

Malfunction in the UV system and delayed installation of replacement components

Thornley et al. ()

2001

Torres de Segre, Spain

344

1,880

Campylobacter

Failure in chlorination system

Godoy et al. ()

2001

Switzerland

650

Unknown

Norovirus

Treatment failure following deﬁciencies in chlorine and/or ozone application

Fretz et al. ()

2001

Pennsylvania, USA

Unknown

Unspeciﬁed treatment deﬁciency; no chlorine residual in the drinking water

Blackburn et al. ()

2001

Wyoming, USA

Unknown

Norovirus

Failure of pellet chlorinator and septic tank contamination

Blackburn et al. ()

2004

Ireland

25,000

Cryptosporidium

High demand and turbidity issues lead to unﬁltered water mixed with ﬁltered water

O’Toole et al. ()

2004

New Zealand

Unknown

Shigella

Treatment failure and inadequate raw water source

ESR ()

2004

Montana, USA

Unknown

Salmonella

UV disinfection unit found to be out of service

Liang et al. ()

2005

Carlow, Ireland

25,000

Cryptosporidium and Giardia

Aging plant with turbidity problems in highly agricultural basin; sewage treatment plants upstream; rainfall peak

Roch et al. ()

2006

Apulia, Italy

2,860

Unknown

Rotavirus and norovirus

Technical problems with chlorination

Martinelli et al. ()

2006

Valencia d’Aneu, Spain

180

Shigella

Chlorinator froze and stopped working; possible illegal discharge of wastewater near raw water source

Godoy et al. ()

2006

Indiana, USA

Unknown

Campylobacter

Inadequate chlorination of the water supply; cross-contamination also possible when testing a new water main

Yoder et al. ()

2007

Florida, USA

1,663

Unknown

Operation and maintenance deﬁciencies in water treatment

Brunkard et al. ()

2010

Åhus, Sweden

Unknown

Enterococci and E. coli

Salt used in the water softening process was contaminated; rapid intervention of the municipality may have prevented an outbreak

Norberg ()

2012

Darﬁeld, New Zealand

138

3,280

Campylobacter

Pump failure lead to exclusive use of river raw water; heavy rains resulted in increased turbidity, no multi-barrier approach

Bartholomew et al. ()

Page 179

Table 4

N. A. Moreira & M. Bondelind

Journal of Water and Health

Safe drinking water and waterborne outbreaks

15.1

2017

List of outbreaks originated from distribution systems failure

Year

Location, country

Est. cases

Pop. served

Causative agent

Probable causes for outbreak occurring

Reference

2000

Strasbourg, France

60,000

Unknown

Main repair in the network

Deshayes & Schmitt ()

2000

Bari, Italy

344

1,000

Norovirus

Break in pipeline public supply connecting to resort tank

Boccia et al. ()

2000

Belfast, UK

117

Unknown

Cryptosporidium

Seepage of raw sewage from a septic tank into the water distribution system

Glaberman et al. ()

2000

South Wales, UK

281

Unknown

Campylobacter

Seepage of surface water contaminated by agricultural waste following heavy rainfall into drinking water reservoir

Richardson et al. ()

2000

Ohio, USA

Unknown

E. coli

Possible back-siphonage from animal barn

Lee et al. ()

2001

Darcy le Fort, France

563

1,100

Cryptosporidium, rotavirus, Campylobacter and E. coli

Sewage contamination occurred in the distribution network upstream to the city

Dalle et al. ()

2001

Lleida, Spain

293

Norovirus

Contamination of reservoir due to lack of maintenance and structural deﬁciencies

Godoy et al. ()

2001

Utrecht, The Netherlands

1,866

Norovirus

Drinking water system connected to grey water system in maintenance work; cross-connection not removed

Fernandes et al. ()

2001

Belfast, UK

230

Unknown

Cryptosporidium

Wastewater into the drinking water supply due to a blocked drain

Glaberman et al. ()

2002

Vicenza, Italy

670

3,006

Unknown

Broken sewage pipe allowed untreated water from the river to enter the city aqueduct

Tramarin et al. ()

2002

Switzerland

125

Unknown

Norovirus

Faeces related contamination from a sewage leakage

Fretz et al. ()

2004

Ohio, USA

1,450

Unknown

Campylobacter, norovirus and Giardia

Unspeciﬁed distribution system deﬁciency related with untreated groundwater

Liang et al. ()

2007

Køge, Denmark

140

5,802

Campylobacter, E. coli and norovirus

Technical and human error at sewage treatment work allowed partially ﬁltered wastewater to enter the drinking water system

Vestergaard et al. ()

2007

Nokia, Finland

8,453

30,016

Norovirus, Campylobacter and Giardia

Drinking water network contaminated by treated sewage efﬂuent

Laine et al. ()

2007

Västerås, Sweden

400

Unknown

Leaked sewage into drinking water network during maintenance work on a pipeline

Nilsson ()

2008

Zurich, Switzerland

126

2,000

Campylobacter and norovirus

Input of highly pressurised washwater from sewage plant into the drinking water system

Breitenmoser et al. ()

2008

Northampton, UK

>422

250,000

Cryptosporidium

Dead rabbit found in a tank containing drinking water at the treatment works

Smith et al. (), Chalmers () (continued)

Page 180

Table 4

N. A. Moreira & M. Bondelind

Journal of Water and Health

Safe drinking water and waterborne outbreaks

15.1

2017

continued

Year

Location, country

Est. cases

Pop. served

Causative agent

Probable causes for outbreak occurring

Reference

2008

Colorado, USA

1,300

Unknown

Salmonella

Likely animal contamination of a storage tank

Brunkard et al. ()

2009

Utah, USA

Unknown

Giardia

Cross-connection between potable and non-potable water sources resulting in backﬂow

Hilborn et al. ()

2010

Køge, Denmark

409

20,000

Campylobacter

Contamination of central water supply system by unknown mechanism

Gubbels et al. ()

2010

Öland, Sweden

200

Unknown

Norovirus

Untreated water from well in the drinking water network

Hallin ()

2010

Utah, USA

628

Unknown

Campylobacter

Cross-connection between potable and non-potable water sources resulting in backﬂow

Hilborn et al. ()

2012

Kilkis, Greece

1,538

Norovirus

Heavy snowfall and runoff, low temperatures and 15 days without use of school’s public water supply increased microbial load

Mellou et al. ()

2012

Kalundborg, Denmark

187

Unknown

Norovirus

Contamination from sewage pipe due to fall in pressure, throughout water supply system repairs

van Alphen et al. ()

2012

Vuorela, Finland

800

2,931

Sapovirus and E. coli

Main pipe accidently broken during road construction; ﬂushing after breakage repair proved insufﬁcient and storage reservoir was contaminated

Jalava et al. ()

2013

Guipuzko, Spain

238

650

Norovirus and rotavirus

Cross-connection between drinking water supplies and industrial water taken from a river

Altzibar et al. ()

Figure 1

The number of events of WBOs and the number of cases of illnesses among the consumers.

Page 181

N. A. Moreira & M. Bondelind

Journal of Water and Health

Safe drinking water and waterborne outbreaks

15.1

2017

Figure 2

The total number of affected consumers for each pathogen. If several pathogens were present during one outbreak, the number of affected consumers have been divided with the number of present pathogens.

Figure 3

The number of cases of WBOs where each pathogen was present. If several pathogens were present, each occasion has been divided into fractions for each pathogen.

cases of infectious gastrointestinal illness, in one single

causative pathogen in one outbreak each and multiple

event: Canada, Finland, France, Greece and the USA.

aetiologies were responsible in two outbreaks.

Thirteen waterborne outbreaks caused by contaminated

For surface water contamination events the causes of

surface water have been identiďŹ ed (Table 2, Figure 1). A

the WBOs were heavy rainfall, sewage contamination,

time-related pattern could be suggested for the outbreaks ori-

animal or farming activities and increased organic matter.

ginated by surface water contamination where a majority of

The majority of the infections in the identiďŹ ed events were

the cases of illness (87%) occurred after 2007, but that may

related to wastewater contamination.

be due to selection bias. The aetiological agents for the events with surface water

The highest number of estimated cases caused by surface

water

contamination

was

concentrated

contamination were the protozoan pathogen Cryptospori-

only one country (Sweden), responsible for 49,400

dium in six events while norovirus was present in two

infected drinking water consumers, mainly due to two

outbreaks. Shigella, Giardia, and rotovirus were the

especially large outbreaks in 2010 and 2011. The

Page 182

N. A. Moreira & M. Bondelind

Journal of Water and Health

Safe drinking water and waterborne outbreaks

15.1

2017

second largest number of affected consumers was

malfunction. Multiple aetiologies were present in seven out-

located in Norway.

breaks, and in many of them bacterial, viral and protozoan

WBOs caused by treatment failure

unidentiﬁed aetiologies. In the remaining outbreaks one

pathogens were simultaneously identiﬁed. Three WBOs had single aetiological agent was detected: norovirus was responCryptosporidium

Analysing the 18 reviewed incidents originated by treatment

sible

deﬁciencies in the drinking water production, which are dis-

Campylobacter were causative of three outbreaks each, E.

played in Table 3 and Figures 1–3, it can be observed that

coli, Giardia and Salmonella were the single agent in one out-

several causative agents are present and no obvious one is pre-

break each.

for

seven

outbreaks,

and

dominant. Nevertheless, Campylobacter was the most frequent

The available information regarding the causes of distri-

aetiology, present in almost one-third of the outbreaks

bution systems failures show that cross-connections are the

although not exclusively in one of those events. Norovirus

main cause for outbreaks in the distribution system. Other

was present in two out of four outbreaks as part of a multiple

identiﬁed causes were maintenance or repair works in the

pathogen occurrence. Cryptosporidium was responsible for

water mains, intrusion of sewage due to leakage, distribution

three outbreaks but in one of those as part of a mixed-agent out-

system reservoir contamination and regrowth in the distri-

break. Both rotavirus WBOs and one of the Giardia outbreaks

bution network due to low demand. The cause that

were part of events with multiple aetiologies. Shigella, Salmo-

affected the highest number of consumers was intrusion of

nella, Enterococci and E. coli were also present in occurrences

water into the distribution network.

leading to the contamination of the drinking water.

More than half of the estimated cases of illnesses caused

The technical reasons that ultimately led to the outbreaks

by waterborne outbreaks originating from distribution sys-

can be divided into two main groups. The ﬁrst group has 11

tems failure were located in Finland and together with the

outbreaks caused by disinfection-related problems and the

USA almost three quarters of the affected consumers are

second group has four WBOs related to difﬁculties with

accounted for. In the USA ﬁve outbreaks occurred while

increased turbidity in the inﬂow of raw water. The treatment

in Finland only two outbreaks were identiﬁed. Among the

deﬁciencies were sometimes loosely associated with main-

remaining countries the UK and Denmark have four and

tenance work or strain within the treatment process train in

three identiﬁed outbreaks, respectively, while the remaining

coping with increased demand. An event in Sweden demon-

countries have fewer identiﬁed outbreaks.

strates that chemicals used in the production of water can be contaminated. In this event salt used in the water softening process was contaminated with Enterococci and E. coli.

DISCUSSION

The location of seven of the reported illnesses caused by waterborne outbreaks originated from treatment deﬁciencies

In this paper the causes of WBOs have been investigated. The

in North America, where Canada had one outbreak and the

main causes for contamination of groundwater sources ident-

USA six occurrences with signiﬁcant impact. Within Europe

iﬁed in this paper were the intrusion of animal faeces or

a total number of eight outbreaks occurred which corresponds

wastewater due to heavy rains. Even if the large majority of

to 43% of estimated cases. In Italy and France the outbreaks

the reported events occurred before 2007, a time-related pat-

were larger and caused more than 2,500 cases of gastrointes-

tern cannot be inferred and further measures to reduce the

tinal illnesses. Finally, in New Zealand the three reported

contamination risks to the raw water and the catchment

WBOs only affected a smaller number of consumers.

areas should be thoroughly implemented, with the establish-

WBOs caused by distribution systems failure

contamination sources, for instance. The outbreaks origi-

ment of protection areas and identiﬁcation of potential nated by surface water contamination did on the other The 26 incidents that were reviewed for this section, Table 4

hand occur after 2007 for the majority of the cases of illness,

and Figures 1–3, were the consequence of network

but this does not sanction any assumption regarding the Page 183

N. A. Moreira & M. Bondelind

Safe drinking water and waterborne outbreaks

Journal of Water and Health

15.1

2017

protection of raw water sources. The main causes for con-

The distribution network had the highest number of indi-

tamination of surface waters, identiﬁed in this study, are the

vidual events of WBOs. However, the number of affected

discharge of wastewater into the water source and increased

consumers was low for each event, and therefore the total

turbidity and colour of the water. These events may occur

number of affected consumers is not very high. The causes

during heavy rains but also at low water levels. This indicates

identiﬁed in this study for WBOs at the distribution network

that further measures to reduce the contamination risks to the

were cross-connections, pipe breaks and wastewater intru-

raw water and the catchment areas still need to be

sion into the network. Also, cases of contamination of

implemented for surface water sources. Measures that

distribution system reservoirs are reported. One event in

could be applied are the establishment of protection areas,

Greece highlights the magnitude of the challenge posed by

the identiﬁcation of potential contamination sources and

norovirus because of its persistence in water. Previous work

increased monitoring of raw water quality parameters.

has demonstrated a persistence that can be higher than 15

Cryptosporidium, norovirus, Giardia, Campylobacter and rotavirus were the main pathogens causing the highest

days (Seitz et al. ), and that it is resistant at low levels of chlorine disinfection (Kambhampati et al. ).

amount of affected consumers (Figure 2), however, the

In this study causes and pathogens of WBOs have been

choice of keywords in the literature search may have intro-

critically evaluated. Limitations in this study are that out-

duced a bias which downplayed the role of other causative

breaks have only been evaluated if the cause of the event

agents. The identiﬁed pathogens have in common a moder-

was indicated in the reference and if the event was present

ately to long persistence in water supplies and are

in the chosen databases. In a recent review the responsible

moderately to highly infective (Åström ). Both Cryptospor-

authorities and the water industry were directly contacted

idium and Giardia are highly resistant to chlorine disinfection,

about recent WBOs in the Nordic countries (Guzman Her-

and turbidity control (e.g. chemical coagulation followed by ﬁl-

rador et al. ). In total, 175 outbreaks were identiﬁed

tration) is essential for adequate treatment of the water. The

which exceeds the number of outbreaks identiﬁed in our

highest number of different pathogens has been identiﬁed for

study. However, the number of cases of illnesses is of the

the WTP and the distribution network. Although the number

same order of magnitude for Sweden, Finland and Den-

of identiﬁed events was larger for the distribution system in

mark, if adjusted for the year 1998–1999 (Miettinen et al.

comparison to the number of surface water outbreaks, the

). Consequently, this indicates that the identiﬁed

number of consumers with gastrointestinal illness was highest

causes for outbreaks in this review may not cover minor

for contamination events related with a surface water source,

events that have only affected a small number of consumers.

around six times higher than for groundwater contamination

This work has not addressed the differences between

(Figure 1). However, to prevent the outbreaks in these

small and large WTPs. The tendency is that medium and

occasions the WTPs would have had to adequately treat the

large waterworks receive more attention than small ones

contaminated water and, thus, the failure has not only

in these systematic approaches (Coulibaldy & Rodriguez

occurred in the source water but also at the WTPs.

). In a study published in 2011 that analysed small

The main failure at WTPs causing a WBO has been ident-

WTPs in Finland, it was indicated that nonconformity in

iﬁed to be the malfunctioning of the UV treatment step or the

the production of microbiological safe drinking water is

chlorination equipment. Thereafter comes increased turbidity,

more probable in small rather than large waterworks that

maintenance work, high or low demand of water and mal-

were distributing water to a minimum of a 1,000 consumers

functioning equipment (e.g. pumps). For many of the events,

(Zacheus & Miettinen ). Previous reviews have high-

several failures have occurred simultaneously. To reduce the

lighted that the number of small waterborne outbreaks

risk of a WBO, a risk assessment tool for the disinfection

that are not reported or that are merely poorly documented

step has been developed in Norway. The tool can be used to

is not negligible (Hrudey & Hrudey ). In countries like

identify risks within the disinfection processes of chlorination,

Finland where the number of affected consumers is below

UV and ozonation, and thus enabling the prevention of

0.01% (the US EPA guideline), it is considered that the pro-

WBOs (Ødegaard et al. ).

duction of safe drinking water in all types of settings and/or

Page 184

N. A. Moreira & M. Bondelind

Journal of Water and Health

Safe drinking water and waterborne outbreaks

15.1

2017

limitations is not guaranteed and more measures need to be

survey bias had an impact on these results. The highest

implemented (Zacheus & Miettinen ).

number of different pathogens has been identiﬁed for the

The main objective for the water treatment systems is to

WTP and the distribution network. The highest number of

deliver drinking water to consumers that is both aesthetically

affected consumers with gastrointestinal illness was for con-

suitable and safe (Zhang et al. ). With continuously chan-

tamination events with a surface water source, while the

ging raw water quality, variations in water demand and

highest number of events of WBOs occurred for the distri-

operational challenges at the WTP, risk assessment of the

bution network.

water treatment systems have become increasingly important. This has also been stressed by the World Health Organization. Many tools are available for risk assessment of the water treat-

REFERENCES

ment systems. However, identifying possible risk scenarios proves challenging. We expect that this critical evaluation of the causes of WBOs will help the water industry in their work with WSP to identify risks that may lead to waterborne outbreaks. This paper clearly demonstrates the need for further research to reduce the risks of WBOs and the need for wellfounded guidelines for identiﬁcation of risks in the production of drinking water. Additionally, it is suggested that experiences on WBOs are shared within and between water companies and researchers to improve risk analysis tools and risk reduction measures in order to provide safe drinking water.

CONCLUSIONS The importance of identifying and addressing the potential risks in the drinking water systems is of the foremost signiﬁcance to prevent outbreaks and assure the deliverance of safe water to consumers. The main causes of contamination identiﬁed in this review are as follows:

•

Groundwater sources: intrusion of animal faeces or wastewater due to heavy rains.

•

Surface water sources: discharge of wastewater into the water source and increased turbidity and colour of the water.

•

WTP: malfunctioning of the disinfection, increased turbidity, maintenance work, high or low demand of water and malfunctioning equipment (e.g. pumps).

•

Distribution network: cross-connections, pipe breaks, wastewater intrusion into the pipe network, and contamination of reservoirs. The main pathogens causing the highest amount of

affected consumers are Cryptosporidium, norovirus, Giardia, Campylobacter and rotavirus, but it is possible that

Altzibar, J. M., Zigorraga, C., Rodriguez, R., Leturia, N., Garmendia, A., Rodriguez, A., Alkorta, M. & Arriola, L.  Outbreak of acute gastroenteritis caused by contamination of drinking water in a factory, the Basque Country. J. Water Health 13 (1), 168–173. Andersson, T., Bjelkmar, P., Hulth, A., Lindh, J., Stenmark, S. & Widerström, M.  Syndromic surveillance for local outbreak detection and awareness: evaluating outbreak signals of acute gastroenteritis in telephone triage, web-based queries and over-the-counter pharmacy sales. Epidemiol. Infect. 142, 303–313. Arias, C., Sala, M. R., Dominguez, A., Bartolomé, R., Benavente, A., Veciana, P., Pedrol, A. & Hoyo, G.  Waterborne epidemic outbreak of Shigella sonnei gastroenteritis in Santa Maria de Palautordera, Catalonia, Spain. Epidemiol. Infect. 134 (3), 598–604. Åström, J.  Microbial Risks in Surface Water Sources. PhD Thesis. Chalmers University of Technology, Gothenburg. Baldursson, S. & Karanis, P.  Waterborne transmission of protozoan parasites: review of worldwide breaks – an update 2004–2010. Water Res. 45, 6603–6614. Bartholomew, N., Brunton, C., Mitchell, P., Williamson, J. & Gilpin, B.  A waterborne outbreak of campylobacteriosis in the South Island of New Zealand due to a failure to implement a multi-barrier approach. J. Water Health 12 (3), 555–563. Bartram, J., Corrales, L., Davison, A., Deere, D., Drury, D., Gordon, B., Howard, G., Rinehold, A. & Stevens, M.  Water Safety Plan Manual: Step-by-Step Risk Management for Drinking Water Suppliers. WHO, Geneva. Blackburn, B. G., Craun, G., Yoder, J., Hill, V. & Calderon, R.  Surveillance for Waterborne-Disease Outbreaks Associated with Drinking Water – United States, 2001–2002. Centers for Disease Control and Prevention, Atlanta. Boccia, D., Eugenio Tozzi, A., Cotter, B., Rizzo, C., Russo, T., Buttinelli, G., Caprioli, A., Marziano, M. L. & Ruggeri, F. M.  Waterborne outbreak of Norwalk-like virus gastroenteritis at a tourist resort, Italy. Emerg. Infect. Dis. 8 (6), 563–568. Bratanova, B., Morrison, G., Fife-Schaw, C., Chenoweth, J. & Mangold, M.  Restoring drinking water acceptance following a waterborne disease outbreak: the role of trust,

Page 185

N. A. Moreira & M. Bondelind

Safe drinking water and waterborne outbreaks

risk perception, and communication. J. Appl. Soc. Psychol. 43, 1761–1770. Breitenmoser, A., Fretz, R., Schmid, J., Besl, A. & Etter, R.  Outbreak of acute gastroenteritis due to a washwatercontaminated water supply, Switzerland, 2008. J. Water Health 9 (3), 569–576. Brunkard, J. M., Ailes, E., Roberts, V., Hill, V. & Hilborn, E.  Surveillance for waterborne disease outbreaks associated with drinking water – United States, 2007–2008. MMWR Morbid. Mortal. Wkly Rep. 60 (12), 38–75. Carrique-Mas, J., Andersson, Y., Petersén, B., Hedlund, K.-O., Sjögren, N. & Giesecke, J.  A Norwalk-like virus waterborne community outbreak in a Swedish village during peak holiday season. Epidemiol. Infect. 131, 737–744. Chalmers, R. M.  Waterborne outbreaks of cryptosporidiosis. Ann. Ist. Super Sanità 48 (4), 429–446. Chalmers, R. M., Robinson, G., Elwin, K., Hadﬁeld, S., Thomas, E., Watkins, J., Casemore, D. & Kay, D.  Detection of Cryptosporidium species and sources of contamination with Cryptosporidium hominis during a waterborne outbreak in northwest Wales. J. Water Health 8 (2), 311–325. Coulibaldy, H. D. & Rodriguez, M.  Development of performance indicators for small Quebec drinking water utilities. J. Environ. Manage. 74, 243–255. Dalle, F., Roz, P., Dautin, G., Di-Palma, M., Kohli, E., Sire-Bidault, C., Fleischmann, M. G., Gallay, A., Carbonel, S., Bon, F., Tillier, C., Beaudeau, P. & Bonnin, A.  Molecular characterization of isolates of waterborne Cryptosporidium spp. collected during an outbreak of gastroenteritis in South Burgundy, France. J. Clin. Microbiol. 41 (6), 2690–2693. Delpha, I., Jung, A.-V., Baures, E., Clement, M. & Thomas, O.  Impacts of climate change on surface water quality in relation to drinking water production. Environ. Int. 35, 1225–1233. Deshayes, F. & Schmitt, M.  Pollution du réseau d’eau potable à Strasbourg et survenue concomitante de gastroentérites. BEH 2, 5–7. Available from: www.invs.sante.fr/beh/2001/ 02/beh_02_2001.pdf. ESR  Annual Summary of Outbreaks in New Zealand 2004. Institute of Environmental Science and Research. Available from: https://surv.esr.cri.nz/PDF_surveillance/AnnualRpt/ AnnualOutbreak/2004/2004OutbreakRpt.pdf. Fernandes, T., Schout, C., Husman, A., Eilander, A., Vennema, H. & Van Duynhoven, Y.  Gastroenteritis associated with accidental contamination of drinking water with partially treated water. Epidemiol. Infect. 135, 818–826. Fretz, R., Svoboda, P., Lüthi, T. M., Tanner, M. & Baumgartner, A.  Outbreaks of gastroenteritis due to infections with Norovirus in Switzerland, 2001–2003. Epidemiol. Infect. 133 (3), 429–437. Gallay, A., De Valk, H., Cournot, M., Ladeuil, B., Hemery, C., Castor, C., Bon, F., Mégraud, F., Le Cann, P., Desenclos, J. C. & Outbreak Investigation Team  A large multi-pathogen waterborne community outbreak linked to faecal contamination of a groundwater system, France, 2000. Clin. Microbiol. Infect. 12, 561–570.

Page 186

Journal of Water and Health

15.1

2017

Giammanco, G., Di Bartolo, I., Purpari, G., Costantino, C., Rotolo, V., Spoto, V., Geraci, G., Bosco, G., Petralia, A., Guercio, A., Macaluso, G., Calamusa, G., Dr Grazia, S., Ruggeri, F. M., Vitale, F., Maida, C. M. & Mammina, C.  Investigation and control of a Norovirus outbreak of probable waterborne transmission through a municipal groundwater system. J. Water Health 12 (3), 452–464. Glaberman, S., Moore, J., Lowery, C., Chalmers, R., Sulaiman, I., Elwin, K., Rooney, P. J., Millar, B. C., Dooley, J. S., Lal, A. A. & Xiao, L.  Three drinking water-associated Cryptosporidiosis outbreaks, Northern Ireland. Emerg. Infect. Dis. 8 (6), 631–633. Godoy, P., Artigues, A., Nuín, C., Aramburu, J., Pérez, M., Domínguez, A. & Salleras, L.  Brote comunitario de gastroenteritis por Campylobacter jejuni originado por el consumo de agua del suministro público. Med. Clin. 119 (18), 696–698. Godoy, P., Nuín, C., Alsedà, M., Llovet, T., Mazan, R. & Domínguez, Á.  Brote de gastroenteritis por Norovirus causado por el consumo de agua de suministro público. Rev. Clin. Esp. 206 (9), 435–437. Godoy, P., Bartolomé, R., Torres, J., Espinet, L., Escobar, A., Nuin, C. & Domínguez, A.  Brote de gastroenteritis por el consumo de agua de suministro público causado por Shigella sonnei. Gac. Sanit. 25 (5), 363–367. Gubbels, S. M., Kuhn, K. G., Larsson, J. T., Adelhardt, M., Engberg, J., Ingildsen, P., Hollesen, L. W., Muchitsch, S., Molbak, K. & Ethelberg, S.  A waterborne outbreak with a single clone of Campylobacter jejuni in the Danish town of Køge in May 2010. Scand. J. Infect. Dis. 44, 586–594. Guzman Herrador, B. R., Carlander, A., Ethelberg, S., De Blasio, B. F., Kuusi, M., Lund, V., Löfdahl, M., MacDonald, E., Nichols, G., Schönning, C., Sudre, B., Trönnberg, L., Vold, L., Semenza, J. C. & Nygård, K.  Waterborne outbreaks in the Nordic countries, 1998 to 2012. Eurosurveillance 20 (24), 1–10. Hallin, E.  Norovirus i vatten – en litteraturstudie (Norovirus in rater – a literature review). Svenskt Vatten Utveckling, Stockholm. Hänninen, M., Haajanen, H., Pummi, T., Wermundsen, K., Katila, M.-L., Sarkkinen, H., Miettinen, I. & Rautelin, H.  Detection and typing of Campylobacter jejuni and Campylobacter coli and analysis of indicator organisms in three waterborne outbreaks in Finland. Appl. Environ. Microbiol. 69 (3), 1391–1396. Hewitt, J., Bell, D., Simmons, G., Rivera-Aban, M., Wolf, S. & Greening, G.  Gastroenteritis outbreak caused by waterborne Norovirus at a New Zealand Ski Resort. Appl. Environ. Microbiol. 73 (24), 7853–7857. Hilborn, E. D., Wade, T., Hicks, L., Garrison, L. & Gargano, J.  Surveillance for Waterborne Disease Outbreaks Associated with Drinking Water and Other Nonrecreational Water – United States, 2009–2010. Centers for Disease Control and Prevention, Atlanta. Howe, A. D., Forster, S., Morton, S., Marshall, R., Osborn, K., Wright, P. & Hunter, P.  Cryptosporidium oocysts in a

N. A. Moreira & M. Bondelind

Safe drinking water and waterborne outbreaks

water supply associated with a Cryptosporidiosis outbreak. Emerg. Infect. Dis. 8 (6), 619–624. HPSC  Annual Report 2006. Health Protection Surveillance Centre, Dublin. HPSC  Annual Report 2007. Health Protection Surveillance Centre, Dublin. Hrudey, S. E. & Hrudey, E.  Published case studies of waterborne disease outbreaks – evidence of a recurrent threat. Water Environ. Res. 79 (3), 233–245. Hrudey, S. E., Payment, P., Huck, P. M., Gillham, R. W. & Hrudey, E. J.  A fatal waterborne disease epidemic in Walkerton, Ontario: comparison with other waterborne outbreaks in the developed world. Water Sci. Technol. 47 (3), 7–14. Jalava, K., Rintala, H., Ollgren, J., Maunula, L., Gomez-Alvarez, V., Revez, J., Palander, M., Antikainen, J., Kauppinen, A., Räsänen, P., Siponen, S., Nyholm, O., Kyyhkynen, A., Hakkarainen, S., Merentie, J., Pärnänen, M., Loginov, R., Ryu, H., Kuusi, M., Siitonen, A., Miettinen, I., Santo Domingo, J. W., Hänninen, M.-L. & Pitkänen, T.  Novel microbiological and spatial statistical methods to improve strength of epidemiological evidence in a community-wide waterborne outbreak. PLoS ONE 9 (8), 104713. Jennings, P. & Rhatigan, A.  Cryptosporidiosis outbreak in Ireland linked to public water supply. Eurosurveillance 6 (22). Available from: www.eurosurveillance.org/ViewArticle. aspx?ArticleId=2089. Kambhampati, A., Koopmans, M. & Lopman, B. A.  Burden of norovirus in healthcare facilities and strategies for outbreak control. J. Hosp. Infect. 89, 296–301. Karanis, P., Kourenti, C. & Smith, H.  Waterborne transmission of protozoan parasites: a worldwide review of outbreaks and lessons learnt. J. Water Health 5 (1), 1–38. Laine, J., Huovinen, E., Virtanen, M. J., Snellman, M., Lumio, J., Ruutu, P., Kujansuu, E., Vuento, R., Pitkänen, T., Miettinen, I., Herrala, J., Lepistö, O., Antonen, J., Helenius, J., Hänninen, M. L., Maunula, L., Mustonen, J., Kuusi, M. & Pirkanmaa, Waterborne Outbreak Study Group  An extensive gastroenteritis outbreak after drinking-water contamination by sewage efﬂuent, Finland. Epidemiol. Infect. 139 (7), 1105–1113. Lake, I. R., Nichols, G., Bentham, G., Harrison, F., Hunter, P. & Sari Kovats, R.  Cryptosporidiosis decline after regulation, England and Wales, 1989–2005. Emerg. Infect. Dis. 13 (4), 623–625. Larsson, C., Andersson, Y., Allestam, G., Lindqvist, A., Nenonen, N. & Bergstedt, O.  Epidemiology and estimated costs of a large waterborne outbreak of norovirus infection in Sweden. Epidemiol. Infect. 142, 592–600. Lee, S. H., Levy, D., Craun, G., Beach, M. & Calderon, R.  Surveillance for Waterborne-Disease Outbreaks – United States, 1999–2000. Centers for Disease Control and Prevention, Atlanta. Liang, J. L., Dziuban, E. J., Craun, G., Hill, V. & Moore, M. R.  Surveillance for Waterborne Disease and Outbreaks Associated with Drinking Water and Water not Intended for

Journal of Water and Health

15.1

2017

Drinking – United States, 2003–2004. Centers for Disease Control and Prevention, Atlanta. Lindhe, A., Rósen, L., Norberg, T., Røstum, J. & Petersson, T.  Uncertainty modelling in multi-criteria analysis of water safety measures. Environ. Sys. Decis. 33, 195–208. Martinelli, D., Prato, R., Chironna, M., Sallustio, A., Caputi, G., Conversano, M., Cioﬁ Degli Atti, M., D’Ancona, F. P., Germinario, C. A. & Quarto, M.  Large outbreak of viral gastroenteritis caused by contaminated drinking water in Apulia, Italy, May–October 2006. Eurosurveillance 12 (16), E070419.1 Mason, B. W., Chalmers, R. M., Carnicer-Pont, D. & Casemore, D. P.  A Cryptosporidium hominis outbreak in NorthWest Wales associated with low oocyst counts in treated drinking water. J. Water Health 8 (2), 299–310. Mellou, K., Sideroglou, T., Potamiti-Komi, M., Kokkinos, P., Ziros, P., Georgakopoulou, T. & Vantarakis, A.  Epidemiological investigation of two parallel gastroenteritis outbreaks in school settings. BMC Public Health 13 (241), 1–7. Mellou, K., Katsioulis, A., Potamiti-Komi, M., Pournaras, S., Kyritsi, M., Katsiaﬂaka, A., Kallimani, A., Kokkinos, P., Petinaki, E., Sideroglou, T., Georgakopoulou, T., Vantarakis, A. & Hadjichristodoulou, C.  A large waterborne gastroenteritis outbreak in central Greece, March 2012: challenges for the investigation and management. Epidemiol. Infect. 142, 40–50. Miettinen, I. T., Zacheus, O., Von Bonsdorff, C. H. & Vartiainen, T.  Waterborne epidemics in Finland in 1998–1999. Water Sci. Technol. 43, 67–71. Nichols, G., Chalmers, R., Lake, I., Sopwith, W., Regan, M., Hunter, P., Grenfell, P., Harrison, F. & Lane, C.  Cryptosporidiosis: A Report on the Surveillance and Epidemiology of Cryptosporidium Infection in England and Wales. Drinking Water Directorate Contract Number DWI 70/2/201, 129. Nilsson, L.  Cirkulation: Många sårbara punkter i dricksvattenkedjan. Available from: www.cirkulation.com/ 2008/02/manga-sarbara-punkter-i-dricksvattenkedjan/. (accessed 13 October 2014). Norberg, P.  Orsaksutredning bakteriekontamination av Åhus dricksvatten 2010. C4 Teknik, Kristianstads Kommun, Kristianstad. Nygård, K., Schimmer, B., Søbstad, Ø., Walde, A., Tveit, I., Langeland, N., Hausken, T. & Aavitsland, P.  A large community outbreak of waterborne giardiasis-delayed detection in a non-endemic urban area. BMC Public Health 6 (141), doi:10.1186/1471-2458-6-141. Ødegaard, H. & Østerhus, S.  Microbial Barrier Analysis (MBA) – A Guideline. Norsk Vann, Hamar. Ødegaard, H., Fiksdal, L. & Østerhus, S.  Optimal desinfeksjonspraksis for drikkevann fase 1, Report 147. Norsk vann og avløp BA, Norvar, Hamar. O’Toole, C. E., Jennings, P., Meagher, G. & Kelly, I.  Cryptosporidium outbreak in a continuously tested public

Page 187

N. A. Moreira & M. Bondelind

Safe drinking water and waterborne outbreaks

water supply. Epi-Insight (National Disease Surveillance Centre) 5 (10), 1. O’Reilly, C. E., Bowen, A., Perez, N., Sarisky, J., Shepherd, C. A., Miller, M. D., Hubbard, B. C., Herring, M., Buchanan, S. D., Fitzgerald, C. C., Hill, V., Arrowood, M. J., Xiao, L. X., Hoekstra, R. M., Mintz, E. D., Lynch, M. F. & Outbreak Working Group  A waterborne outbreak of gastroenteritis with multiple etiologies among Resort Island Visitors and Residents: Ohio, 2004. Clin. Infect. Dis. 44 (4), 506–512. Papadopoulos, V., Vlachos, O., Isidoridou, E., Kasmeridis, C., Pappa, Z., Goutzouvelidis, A. & Filippou, F.  A gastroenteritis outbreak due to Norovirus infection in Xanthi, Northern Greece: management and public health consequences. J. Gastrointest. Liver Dis. 15 (1), 27–30. Pelly, H., Cormican, M., O’Donovan, D., Chalmers, R. M., Hanahoe, B., Cloughley, R., McKeown, P. & Corbett-Feeney, G.  A large outbreak of cryptosporidiosis in western Ireland linked to public water supply: a preliminary report. Eurosurveillance 12 (18). Available from: www. eurosurveillance.org/ViewArticle.aspx?ArticleId=3187. Richardson, G., Thomas, D., Smith, R. M. & Nehaul, L.  A community outbreak of Campylobacter jejuni infection from a chlorinated public water supply. Epidemiol. Infect. 135 (7), 1151–1158. Riera-Montes, M., Brus Sjölander, K., Allestam, G., Hallin, E., Hedlund, K.-O. & Löfdahl, M.  Waterborne norovirus outbreak in a municipal drinking-water supply in Sweden. Epidemiol. Infect. 139 (12), 1928–1935. Roch, B. A., O’Byrne, A. M., Leane, G., O’Hare, C. & Menton, F.  Cryptosporidiosis Outbreak in Carlow Town and Environs 2005. Health Service Executive – South Eastern Area. Department of Public Health, Kilkenny. Røstum, J., Aasen, A. & Eikebrokk, B.  Risk and vulnerability assessment (“Ros-Analysis”) of the Bergen water supply system – a source to tap approach. In: Risk Management of Water Supply and Sanitation Systems (P. Hlavinek, C. Popovska, J. Marsalek, I. Mahrikova & T. Kukharchyk, eds). Springer, The Netherlands, pp. 73–83. Scarcella, C., Carasi, S., Cadoria, F., Macchi, L., Pavan, A., Salamana, M., Alborali, G. L., Losio, M. N., Boni, P., Lavazza, A. & Seyler, T.  An outbreak of viral gastroenteritis linked to municipal water supply, Lombardy, Italy, June 2009. Eurosurveillance 14 (29), 1–3. Seitz, S. R., Leon, J. S., Schwab, K. J., Lyon, G. M., Dowd, M., McDaniels, M., Abdulhaﬁd, G., Fernandez, M. L., Lindesmith, L. C., Baric, R. S. & Moe, C. L.  Norovirus infectivity in humans and persistence in water. Appl. Environ. Microbiol. 77 (19), 6884–6888. Smith, S., Elliot, A. J., Mallaghan, C., Modha, D., Hippisley-Cox, J., Large, S., Regan, M. & Smith, G. E.  Value of syndromic surveillance in monitoring a focal waterborne outbreak due to an unusual Cryptosporidium genotype in

Journal of Water and Health

15.1

2017

Northamptonshire, United Kingdom, June–July 2008. Eurosurveillance 15 (33), 1–9. Stirling, R., Aramini, J., Ellis, A., Lim, G., Meyers, R., Fleury, M. & Werker, D.  Waterborne cryptosporidiosis outbreak, North Battleford, Saskatchewan, Spring 2001. Can. Commun. Dis. Rep. 27 (22), 185–192. Thornley, C., McDowell, R., Lopez, L. & Baker, M.  Annual Summary of Outbreaks in New Zealand 2001 ESR. Available from: https://surv.esr.cri.nz/PDF_surveillance/AnnualRpt/ AnnualOutbreak/2001/2001OutbreakRpt.pdf. Tillaut, H., Encrenaz, N., Checlair, E., Alexandre-Bird, A., Santo, E. & Beaudeau, P.  Epidémie de gastro-entérite, Isère, novembre 2002. BEH 12 (3–4), 47–48. Available from: http:// opac.invs.sante.fr/doc_num.php?explnum_id=5801. Tramarin, A., Fabris, P., Bishai, D., Selle, V. & De Lalla, F.  Waterborne infections in the era of bioterrorism. Lancet 360 (9346), 1699. van Alphen, L. B., Dorléans, F., Schultz, A., Fonager, J., Ethelberg, S., Dalgaard, C., Adelhardt, M., Engberg, J. H., Fischer, T. K. & Lassen, S. G.  The application of new molecular methods in the investigation of a waterborne outbreak of Norovirus in Denmark, 2012. PLoS ONE 9 (9), e105053. Vantarakis, A., Mellou, K., Spala, G., Kokkinos, P. & Alamanos, Y.  A gastroenteritis outbreak caused by Noroviruses in Greece. Int. J. Environ. Res. Public Health 8, 3468–3478. Vestergaard, L., Olsen, K., Stensvold, C., Böttiger, B., Adelhardt, M., Lisby, M., Mørk, L. & Mølbak, K.  Outbreak of severe gastroenteritis with multiple aetiologies caused by contaminated drinking water in Denmark, January 2007. Eurosurveillance 12 (13), 3164. WHO  Guidelines for Drinking-Water Quality. 4th edn. World Health Organization, Geneva. Widerström, M., Schönning, C., Lilja, M., Lebbad, M., Ljung, T., Allestam, G., Ferm, M., Björkholm, B., Hansen, A., Hiltula, J., Långmark, J., Löfdahl, M., Omberg, M., Reuterwall, C., Samuelsson, E., Widgren, K., Wallensten, A. & Lindh, J.  Large outbreak of Cryptosporidium hominis infection transmitted through the public water supply, Sweden. Emerg. Infect. Dis. 20 (4), 581–589. Yoder, J., Hlavsa, M., Craun, G., Hill, V., Roberts, V., Yu, P., Hicks, L. A., Alexander, N. T., Calderon, R. L., Roy, S. L. & Beach, M. J.  Surveillance for waterborne disease and outbreaks associated with drinking water and water not intended for drinking – United States, 2005–2006. MMWR Morb. Mortal. Wkly Rep. 57 (SS-9), 39–69. Zacheus, O. & Miettinen, I.  Increased information on waterborne outbreaks through efﬁcient notiﬁcation system enforces actions towards safe drinking water. J. Water Health 9 (4), 763–772. Zhang, K., Achari, G., Sadiq, R., Langford, C. & Dore, M.  An integrated performance assessment framework for water treatment plants. Water Res. 46, 1673–1683.

First received 26 April 2016; accepted in revised form 2 September 2016. Available online 25 October 2016

Page 188

Journal of

Water Reuse & Desalination

ISSN 2220-1319 iwaponline.com/jwrd

Journal of Water Reuse and Desalination is an international open access journal focused on the science and technology, policy, regulation, social and economic aspects and applications of sustainable sources of water to cope with water scarcity, including new sources of non-conventional water. The journal’s scope includes: • Water reuse and desalination management, including public health; environmental health; regulatory policies; public perception/education and outreach; and economic aspects of water reuse and desalination. • Water reuse and desalination sustainability, including environmental impacts and mitigation; food-energy-water nexus; integrated resources planning/water supply resiliency; and reuse in developing countries. • Water reuse and desalination applications, including agricultural reuse; environmental restoration, greywater, stormwater treatment and reuse; municipal, industrial, and environmental non-potable reuse applications; and potable water reuse applications. For more details, visit iwaponline.com/jwrd

Page 192

387

07.4

2017

Heavy metal removal from wastewater using various adsorbents: a review Renu, Madhu Agarwal and K. Singh

ABSTRACT Heavy metals are discharged into water from various industries. They can be toxic or carcinogenic in nature and can cause severe problems for humans and aquatic ecosystems. Thus, the removal of heavy metals from wastewater is a serious problem. The adsorption process is widely used for the removal of heavy metals from wastewater because of its low cost, availability and eco-friendly nature. Both commercial adsorbents and bioadsorbents are used for the removal of heavy metals from wastewater,

Renu Madhu Agarwal (corresponding author) K. Singh Department of Chemical Engineering, Malaviya National Institute of Technology, JLN Marg, Jaipur 302017, India E-mail: madhunaresh@gmail.com

with high removal capacity. This review article aims to compile scattered information on the different adsorbents that are used for heavy metal removal and to provide information on the commercially available and natural bioadsorbents used for removal of chromium, cadmium and copper, in particular. Key words

| adsorbents, adsorption capacity, heavy metal, wastewater

INTRODUCTION Discharge from industry contains various organic and inor-

nickel, zinc, lead, mercury and arsenic, respectively (Gopa-

ganic pollutants. Among these pollutants are heavy metals

lakrishnan et al. ). Various treatment technologies

which can be toxic and/or carcinogenic and which are

employed for the removal of heavy metals include chemical

harmful to humans and other living species (MacCarthy

precipitation, ion exchange, chemical oxidation, reduction,

et al. ; Clement et al. ; Renge et al. ). The

reverse osmosis, ultraﬁltration, electrodialysis and adsorp-

heavy metals of most concern from various industries

tion (Fu & Wang ). Among these methods, adsorption

include lead (Pb), zinc (Zn), copper (Cu), arsenic (As), cad-

is the most efﬁcient as the other techniques have inherent

mium (Cd), chromium (Cr), nickel (Ni) and mercury (Hg)

limitations such as the generation of a large amount of

(Mehdipour et al. ). They originate from sources such

sludge, low efﬁciency, sensitive operating conditions and

as metal complex dyes, pesticides, fertilisers, ﬁxing agents

costly disposal. The adsorption method is a relatively new

(which are added to dyes to improve dye adsorption onto

process and is emerging as a potentially preferred alternative

the ﬁbres), mordants, pigments and bleaching agents (Rao

for the removal of heavy metals because it provides ﬂexi-

et al. ). In developed countries, legislation is becoming

bility in design, high-quality treated efﬂuent and is

increasingly stringent for heavy metal limits in wastewater.

reversible and the adsorbent can be regenerated (Fu &

In India, the current maximum contaminant level (ppm–

Wang ). The speciﬁc sources of chromium are leather

mg/mL) for heavy metals is 0.05, 0.01, 0.25, 0.20, 0.80,

tanning, electroplating, nuclear power plants and textile

0.006, 0.00003, 0.050 for chromium, cadmium, copper,

industries. Chromium(VI) is an oxidising agent, is carcinogenic in nature and is also harmful to plants and animals

This is an Open Access article distributed under the terms of the Creative

(Barnhart ). Exposure to chromium(VI) can cause

Commons Attribution Licence (CC BY-NC-ND 4.0), which permits copying

cancer in the digestive tract and lungs, epigastric pain,

and redistribution for non-commercial purposes with no derivatives, provided the original work is properly cited (http://creativecommons.org/

nausea, severe diarrhoea, vomiting and haemorrhage

licenses/by-nc-nd/4.0/).

(Mohanty et al. ). Although chromium can access

doi: 10.2166/wrd.2016.104

Page 193

388

Renu et al.

Journal of Water Reuse and Desalination

Heavy metal removal from wastewater using various adsorbents

07.4

2017

many oxidation states, chromium(VI) and chromium(III)

ions from wastewater. Commercial adsorbents are those

are the species that are mainly found in industrial efﬂuents

adsorbents which are produced commercially on a large

(Mohan & Pittman ). Chromium(VI) is more toxic

scale, such as activated carbon, silica gel, alumina, etc., how-

than chromium(III) and is of more concern (Al-Othman

ever they are costly. Natural bioadsorbents are those

et al. ). The United States Environmental Protection

obtained from biological material and are comparatively

Agency (USEPA) has set the maximum chromium levels

cheap. However, cost analysis is an important criterion for

in drinking water at 0.1 ppm. The USEPA has classiﬁed cad-

selection of an adsorbent for heavy metal removal from

mium as a human carcinogen and it is known to cause

wastewater. The cost of the adsorption process depends on

deleterious effects to health and bone demineralisation

the cost of the adsorbent. For instance, the cost of commer-

either through direct bone damage or as a result of renal dys-

cial activated carbon is Rs. 500/kg; however, the cost of

function (Fu & Wang ). The major sources of cadmium

bioadsorbents is in the range of Rs. 4.4–36.89/kg, which is

include metal reﬁneries, smelting, mining and the photo-

much less as compared to the commercial adsorbents

graphic industry and it is listed as a Category-I carcinogen

(Gupta & Babu ). A comprehensive approach has

by the International Agency for Research on Cancer

been followed to cover all signiﬁcant work done in this

(IARC) and a group B-I carcinogen by the USEPA (Friberg

ﬁeld to date, and a ﬁnal evaluation has been made on the

et al. ). Copper is an essential element and is required

most efﬁcient adsorbent(s) to date.

for enzyme synthesis as well as tissue and bone development. Copper(II) is toxic and carcinogenic when it is ingested in large amounts and causes headache, vomiting, nausea, liver and kidney failure, respiratory problems and

ADSORBENTS USED FOR REMOVAL OF HEAVY METALS FROM WASTEWATER

abdominal pain (Ren et al. ; Hu et al. ; Lan et al. ). The USEPA has set the copper limit at 1.3 ppm in

There are a number of types of adsorbent that are used for

industrial efﬂuents. Industrial sources of copper include

the efﬁcient removal of heavy metal removal from waste-

smelting, mining, electroplating, surface ﬁnishing, electric

water that are both commercial and/or bioadsorbents.

appliances, electrolysis and electrical components (Yin et al.

These are described as follows.

; Bilal et al. ; Lan et al. ). Nickel is a human carcinogen in nature and causes kidney and lung problems,

Commercially available adsorbents for chromium

gastrointestinal distress, skin dermatitis and pulmonary ﬁbro-

removal

sis (Borba et al. ). Zinc is essential for human health but large quantities of zinc can cause skin irritation, stomach

Graphene

cramps, vomiting and anaemia (Oyaro et al. ). Similarly, lead is harmful to human health and can damage kidney,

Nanomaterials are efﬁcient adsorbents for the removal of

liver, reproductive system and brain functions (Naseem &

heavy metals from wastewater because of their high surface

Tahir ). Mercury is also harmful and it is a neurotoxin

area, enhanced active sites and the functional groups that

that can affect the central nervous system. If it is exceeded

are present on their surface (Gopalakrishnan et al. ).

in concentration it can cause pulmonary, chest pain and dys-

Graphene is a carbon-based nanomaterial with a two-dimen-

pnoea (Namasivayam & Kadirvelu ). Arsenic can cause

sional structure, high speciﬁc surface area and good

skin, lung, bladder and kidney cancer, muscular weakness,

chemical stability. It is available in various forms such as

loss of appetite, and nausea (Mohan & Pittman ).

pristine graphene, graphene oxide and reduced graphene

Due to stringent regulations for heavy metals, their

oxide. Graphene may be oxidised to add hydrophilic

removal has become a serious environmental problem.

groups for heavy metal removal (Thangavel & Venugopal

This review surveys the various commercially available

). Yang et al. (a) adsorbed chromium onto the sur-

adsorbents and natural biosorbents used over the past dec-

face of graphene oxide and the maximum adsorption

ades for the removal of chromium, cadmium and copper

capacity found was around 92.65 mg/g at an optimum pH

Page 194

389

Renu et al.

Journal of Water Reuse and Desalination

Heavy metal removal from wastewater using various adsorbents

07.4

2017

of 5. This adsorption of chromium on graphene oxide was

et al. ), i.e., 7.78 mg/g, 15.4 mg/g, 8.8 mg/g, respectively.

found to be endothermic and spontaneous. Gopalakrishnan

Table 1 summarises the graphene-related work that has been

et al. () have also oxidised graphene for the addition of

reported in this area.

�COOH, �C¼O and �OH functional groups onto the sur-

face using a modiﬁed Hummer’s method (Hummers &

Activated carbon

Offeman ). The novelty of their work is that only 70 mg of graphene oxide has been utilised for 100% removal

Modern industries began production of active carbon in

of chromium from wastewater effectively at an optimum pH

1900–1901 to replace bone char in the sugar reﬁning indus-

of 8. Graphene composite materials have been developed by

try (Bansal et al. ) and powdered activated carbon was

a number of authors for the removal of heavy metals.

ﬁrst produced commercially in Europe in the early 19th cen-

Li et al. () functionalised graphene oxide with magnetic

tury, using wood as a raw material (Mantell ). Activated

cyclodextrin chitosan for the removal of chromium since

carbon can be obtained from any material which has high

magnetic cyclodextrin chitosan has favourable properties

carbon content. Activated carbon is a good adsorbent for

such as high adsorption capacity and magnetic property

chromium removal because it has a well-developed porous

which assists in the separation process. Guo et al. ()

structure and a high internal surface area for adsorption

functionalised graphene with a ferro/ferric oxide composite

(Anirudhan & Sreekumari ). However, because coal-

for chromium removal with a maximum adsorption capacity

based activated carbon is expensive, its use has been

of 17.29 mg/g which is higher as compared to the adsorp-

restricted and further efforts have been made to convert

tion capacity of other magnetic adsorbents, such as

cheap and abundant agricultural waste into activated

Fe@Fe2O3 core-shell nanowires (Ai et al. ), chitosan-

carbon (Anirudhan & Sreekumari ). Activated carbon

coated MnFe2O4 nanoparticles (Xiao et al. ) and

is now prepared from various agricultural wastes such as

Fe3O4-polyethyleneimine (PEI)-montmorillonite (Larraza

rubber wood sawdust (Karthikeyan et al. ), moso and

Table 1

Chromium removal using graphene, graphine oxide and modiﬁed graphine as an adsorbent Metal concentration

Optimum

Best model

Contact time

Adsorbent

Adsorbent capacity

Removal per cent

Adsorbent

(ppm-mg/L)

ﬁt

(min)

dose (g/L)

(mg/g)

(%)

References

Graphene oxide based inverse spinel nickel ferrite composite

1,000

Langmuir

120

0.125–2.5

–

Lingamdinne et al. ()

Zero-valent iron assembled on magnetic Fe3O4/graphene nanocomposites

40–100

Langmuir

120

–

101

83.8%

Lv et al. ()

Zero-valent iron decorated on graphene nanosheets

15–35

Langmuir

1.0

–

70%

Li et al. ()

Copolymer of dimethylaminoethyl methacrylate with graphene oxide

–

1.1

–

82.4

93%

Ma et al. ()

Graphene sand composite (GSC)

8–20

1.5

Langmuir

2859.38

93%

Dubey et al. ()

Graphene oxide

Langmuir

–

43.72

92.65%

Yang et al. (a)

Modiﬁed graphene (GN) with cetyltrimethylammonium bromide

50, 100

Langmuir

400

21.57

98.2%

Wu et al. ()

Page 195

390

Renu et al.

Journal of Water Reuse and Desalination

Heavy metal removal from wastewater using various adsorbents

07.4

2017

ma bamboo (Lo et al. ), viticulture industry wastes, grape

once activated maa bamboo and 91.7% removal using

stalk, lex, pomace (Sardella et al. ), hazelnut shell acti-

twice activated maa bamboo. Removal efﬁciency decreases

vated carbon (Kobya ), coconut tree sawdust (Selvi

for once activated moso bamboo and twice activated moso

et al. ), coconut shell carbon (Babel & Kurniawan

bamboo by 20–77% because their average pore diameter is

), sugarcane bagasse (Sharma & Forster ), treated

less than 2 nm and major pores were mesopores. Kobya

sawdust of Indian rose wood (Garg et al. ), wood acti-

() prepared activated carbon using hazelnut shell and

vated carbon (Selomulya et al. ), tyre activated carbon

maximum adsorption capacity of 170 mg/g was obtained

(Hamadi et al. ), coconut shell activated carbon (Selo-

at an optimum pH 1 which is higher than adsorption

mulya et al. ) and palm shell (Saifuddin & Kumaran

capacity of other adsorbents such as wood activated

; Owlad et al. ; Kundu et al. ; Nizamuddin

carbon (Selomulya et al. ), tyre activated carbon

et al. , ; Sabzoi et al. ; Thangalazhy-Gopakumar

(Hamadi et al. ) and coconut shell activated carbon

et al. ), etc.

(Selomulya et al. ) which is only 87.6 mg/g, 58.5 mg/g

Karthikeyan et al. () removed chromium from

and 107.1 mg/g, respectively. Table 2 summarises the

wastewater using activated carbon derived from rubber

reported use of activated carbon for chromium removal

wood sawdust and 44 mg/g maximum adsorption capacity

from wastewater.

was obtained at an optimum pH 2. Maximum adsorption capacity obtained in their work was higher as compared to

Carbon nanotubes

other adsorbents such as coconut tree sawdust (Selvi et al. ), coconut shell carbon (Babel & Kurniawan ),

Carbon nanotubes are efﬁcient adsorbents for heavy metal

sugarcane bagasse (Sharma & Forster ) and treated saw-

removal because they possess chemical stability, large surface

dust of Indian rose wood (Garg et al. ), which were only

area, excellent mechanical and electrical properties, adsorp-

3.60 mg/g, 10.88 mg/g, 13.40 mg/g and 10 mg/g, respect-

tion property and well-developed mesopores (Gupta et al.

ively. Lo et al. () derived activated carbon from moso

; Mubarak et al. a; Al-Khaldi et al. ). They can

and ma bamboo, and 100% removal was obtained using

also be further modiﬁed by chemical treatment to increase

Table 2

Chromium removal using activated carbon as an adsorbent Metal concentration

Optimum

Contact time

Adsorbent

Removal per cent

Adsorbent derived from

(mg/L)

Best model ﬁt

(min)

dose (g/L)

capacity (mg/g)

(%)

References

Acrylonitriledivinylbenzene copolymer

Freundlich

420

0.6

101.2

80%

Duranoğlu et al. ()

Syzygium jambolanum nut carbon

20–100

Langmuir

240

–

100%

Muthukumaran & Beulah ()

Green alga Ulva lactuca

5–50, 5–250

Langmuir

10.61 112.36

98%

El-Sikaily et al. ()

Jatropha wood

30–100

2–10

Langmuir

360

0.6–2

106.4–140.8

–

Gueye et al. ()

Tamarind wood

10–50

6.5

Langmuir Freundlich

–

28%

Acharya et al. ()

Pterocladia capillacea

5–100

Langmuir

120

3–10

100%

El Nemr et al. ()

Zizania caduciﬂora

10–50

2–3

Freundlich

0.8

2.7

84.8%

Liu et al. ()

Prawn shell

25–125

–

Langmuir Freundlich

31.4

–

100

98%

Arulkumar et al. ()

Page 196

391

Renu et al.

Journal of Water Reuse and Desalination

Heavy metal removal from wastewater using various adsorbents

07.4

2017

adsorption capacity (Chen et al. ; Mubarak et al. ,

et al. ). This material is not soluble in water and possesses

a, b, c, a, c; Ruthiraan et al. b). Hu

a granular structure, chemical stability and good mechanical

et al. () removed chromium using oxidised multi-walled

strength (Chuah et al. ). Silica is derived from rice husk

carbon nanotubes and 100% maximum removal was

using sol gel technique and has an afﬁnity for chromium

achieved at an optimum pH of 2.88. Gupta et al. () com-

(Adam et al. ). Thus, Oladoja et al. () incorporated

bined the adsorptive property of multi-walled carbon

iron oxide into silica derived from rice husk, calling it modi-

nanotubes with the magnetic property of iron oxide. The

ﬁed rice husk derived silica. This modiﬁed rice husk

advantages of this composite are high surface area, can be

derived silica has higher adsorption (63.69 mg/g) as com-

used for contaminant removal and can be controlled and

pared to the silica derived from raw rice husk. Rice husk in

removed from the medium using a simple magnetic process.

its natural form and in modiﬁed form (activated carbon modi-

A maximum removal of 88% at pH 6 was obtained. Luo

ﬁed using ozone) was used for the removal of chromium(VI)

et al. () prepared manganese dioxide/iron oxide/acid oxi-

and results compared (Bishnoi et al. ; Sugashini &

dised multi-walled carbon nanotube nanocomposites for

Begum ). It was found that ozone modiﬁed rice husk

chromium removal. Manganese dioxide is a scavenger of aqu-

shows a higher removal capacity than raw rice husk. Suga-

eous trace metals because of its high adsorption capacity but

shini & Begum () modiﬁed rice husk by treating it with

the use of pure manganese dioxide is not favoured because of

ozone to produce activated carbon for chromium removal

the high cost and its unfavourable physical and chemical

with 86% removal being reported. Ozone was used for acti-

properties. The maximum adsorption capacity of the above

vation because it is a strong oxidising agent, stable and can

nanocomposite was 186.9 mg/g with a maximum removal of

be regenerated. Rice husk can also be modiﬁed by prep-

85% at an optimum pH of 2. Mubarak et al. (b) functiona-

aration of biochar. Biochar is a carbon-rich solid by-product

lised carbon nanotubes for chromium removal using nitric

resulting from the pyrolysis of rice husk under oxygen-free

acid and potassium permagnate in 3:1 volume ratio and com-

and low temperature conditions (Lehmann ; Woolf

pared the removal capacity with non-functionalised carbon

et al. ; Mubarak et al. , c; Agrafioti et al. ;

nanotubes. They found that maximum adsorption capacity

Ruthiraan et al. a, b). Biochar has the ability to

for functionalised carbon nanotubes was 2.517 mg/g while

adsorb heavy metals because of electrostatic interactions

for non-functionalised carbon nanotubes it was 2.49 mg/g,

between the negative surface charge and the metal cations,

and removal capacity for functionalised carbon nanotubes

as well as ion exchange between biochar surface protons

(87.6%) was higher compared to non-functionalised carbon

and metal cations (Machida et al. ; Lehmann ;

nanotubes (83%). Mubarak et al. (b) produced carbon

Woolf et al. ; Xu et al. ; Thines et al. , ). Agra-

nanotubes using microwave heating for comparative study

fioti et al. () modiﬁed rice husk by pyrolysis for chromium

of the removal of chromium with another heavy metal (i.e.,

removal with 95% removal reported. Table 4 summarises the

lead). Microwave heating provides a fast and uniform heating

reported use of rice husk for chromium removal from

rate and it accelerates reaction and gives a higher yield. The

wastewater.

maximum adsorption capacity obtained for chromium was 24.45 mg/g and removal efﬁciency obtained was 95% at an

Surfactant modiﬁed waste

optimum pH 8. Table 3 summarises the reported use of carbon nanotubes for chromium removal from wastewater.

Various agricultural wastes have been modiﬁed using surfac-

Bio-adsorbents for chromium removal from wastewater

; Nadeem et al. ; Jing et al. ; Min et al. ).

tants (Bingol et al. ; Namasivayam & Sureshkumar Surfactants are amphipathic substances with both lyophobic Rice husk

and lyophilic groups with the capability of forming selfassociated clusters. Depending upon the nature of their

Rice husk consists of cellulose (32.24%), lignin (21.44%),

hydrophilic group they can be cationic (positive charge),

hemicellulose (21.34%) and mineral ash (15.05%) (Rahman

anionic (negative charge), non-ionic (no apparent charge) Page 197

392

Table 3

Renu et al.

Journal of Water Reuse and Desalination

Heavy metal removal from wastewater using various adsorbents

07.4

2017

Chromium removal using carbon nanotubes as an adsorbent Metal concentration

Optimum

Contact time

Adsorbent

Adsorbent capacity

Removal per cent

Adsorbent

(mg/L)

Best model ﬁt

(min)

dose (g/L)

(mg/g)

(%)

References

Nitric acid oxidised carbon nanotube

–

150

0.5

18%

Atieh et al. ()

Composite of carbon nanotubes and activated alumina

100

Langmuir Freundlich

240

2.5

264.5

>95%

Sankararamakrishnan et al. ()

Nitrogen-doped magnetic CNTs

12.82

Langmuir

720

0.2

638.56

>97%

Shin et al. ()

CNT supported by activated carbon

0.5

Langmuir

0.04

72%

Atieh ()

Cigarette ﬁlter with MWCNT and graphene

–

63–79%

Yu et al. ()

Oxidised multi-walled carbon nanotubes

2.88

Langmuir adsorption isotherm

9,900

75–1.25

4.2615

100%

Hu et al. ()

Composite of multiwalled carbon nanotubes and iron oxide

–

10–60

0.1–2

–

88%

Gupta et al. ()

Manganese dioxide/ iron oxide/acid oxidised multiwalled carbon nanotube nanocomposites

50–300

Langmuir

150

186.9

85%

Luo et al. ()

Carbon nanotubes functionalised using nitric acid and potassium permagnate

Langmuir and Freundlich

120

0.1

2.47, 2.48

87.6%

Mubarak et al. (b)

Carbon nanotube produced using microwave heating

Langmuir and Freundlich

24.45

95%

Mubarak et al. (b)

Table 4

Chromium removal using rice husk as an adsorbent Metal

Adsorbent

concentration (mg/L)

Optimum pH

Best model ﬁt

Contact time (min)

Adsorbent dose (g/L)

capacity (mg/g)

Removal per cent (%)

Iron oxide incorporated into silica derived from rice husk

50–300

Langmuir

120

2.0

63.69

71%

Oladoja et al. ()

Ozone-treated rice husk

50, 100

Freundlich

150

4.0

8.7–13.1

86%

Sugashini & Begum ()

Modiﬁed rice husk

190, 850

6.8

Freundlich

5,760

1–16

–

95%

Agrafioti et al. ()

Adsorbent

Page 198

References

393

Renu et al.

Journal of Water Reuse and Desalination

Heavy metal removal from wastewater using various adsorbents

07.4

2017

and zwitterionic (both charges are present); because of these

Ahmad et al. () reported chromium removal using

characteristics surfactant modiﬁed adsorbents are superior

chromium-resistant reducing bacteria Acinetobacter haemo-

in removal efﬁciency and promote selective adsorption

lyticus inside sugarcane bagasse; this bacteria converts

(Nadeem et al. ; Rosen & Kunjappu ). These

Cr(VI) into Cr(III) which is less toxic and less soluble as com-

researchers modiﬁed carbon powder obtained from the

pared to Cr(VI), and a removal of more than 90% was

husks and pods of Moringa oleifera using the surfactant

obtained. Chemicals used for modiﬁcation of sugarcane

cetyltrimethyl ammonium bromide. This process improved

bagasse

are

succinic

anhydride,

EDTA

dianhydride

the removal efﬁciency of the carbon powder with an adsorp-

(EDTAD), xanthate, pyromellitic anhydride, sulphuric acid,

tion capacity of 27 mg/g being reported at an optimum pH

citric acid, sodium bicarbonate, ethylenediamine, etc. These

of 8. Similarly, Namasivayam & Sureshkumar () modi-

acids work as good chelating agents, so they become polymer-

hexadecyltrimethyl

ised with sugarcane bagasse because it increases the number

ammonium bromide surfactant to increase the removal efﬁ-

of chelating sites and helps in heavy metal removal from

ciency of chromium. They reported a maximum adsorption

wastewater. Garg et al. () used succinic acid for modiﬁ-

capacity of 76.3 mg/g at an optimum pH of 2. Table 5 sum-

cation of sugarcane bagasse and reported 92% removal

marises the reported use of surfactant modiﬁed waste as an

obtained at an optimum pH of 2. Cronje et al. () removed

adsorbent for chromium removal.

chromium by activating sugarcane bagasse with zinc chlor-

ﬁed

coconut

coir

pith

using

ide, and >87% chromium was reported at an optimum pH of 8.58. Table 6 summarises the reported use of sugarcane

Modiﬁed sugarcane bagasse

bagasse as an adsorbent for chromium removal.

Sugarcane bagasse is a by-product of agricultural wastes that consists of cellulose (50%), polyoses (27%) and lignin (23%).

Modiﬁed wheat bran

Due to these biological component polymers, sugarcane bagasse is rich in hydroxyl and phenolic groups and these

Wheat bran is an agricultural by-product which can be used

groups can be chemically modiﬁed to improve adsorption

for the removal of heavy metals and is obtained from the

capacity (Ngah & Hanafiah ). Sugarcane bagasse is

shell of ﬂour mill wheat seeds. It is economically viable, bio-

obtained from the ﬁbrous material left after cane stalk crush-

degradable and consists of many nutrients such as protein,

ing and juice extraction. Sugarcane bagasse originates from

minerals, fatty acids and dietary ﬁbres (Kaya et al. ). It

the outer rind and inner pith (Ullah et al. ) and has

has various organic functional groups with a surface area

been used in the natural form as well as in a modiﬁed form.

of 441 m2/g and a ﬁxed carbon content of 31.78% (Singh

Table 5

Chromium removal using surfactant modiﬁed waste as an adsorbent

Contact

Metal

Adsorbent

Removal

time

Adsorbent

capacity

per cent

Best model ﬁt

(min)

dose (g/L)

(mg/g)

(%)

References

Langmuir

120

29.96

98%

Nadeem et al. ()

Langmuir, Freundlich, Dubinin– Radushkevich

0.5–6.0

76.3

96%

Namasivayam & Sureshkumar ()

concentration

Optimum

Adsorbent

(mg/L)

Coconut coir pith modiﬁed by using surfactant cetyltrimethyl ammonium bromide

Coconut coir pith modiﬁed by using hexadecyltrimethyl ammonium bromide surfactant

20–60

Page 199

394

Table 6

Renu et al.

Journal of Water Reuse and Desalination

Heavy metal removal from wastewater using various adsorbents

07.4

2017

Chromium removal using modiﬁed sugarcane bagasse as an adsorbent Metal concentration

Optimum

Best model

Contact

Adsorbent

Removal per

Adsorbent

(mg/L)

ﬁt

time (min)

dose (g/L)

capacity (mg/g)

cent (%)

References

Acinetobacter haemolyticus bacteria inside sugarcane bagasse

10–100

–

2,880

–

>90%

Ahmad et al. ()

Succinic acid modiﬁed sugarcane bagasse

–

92%

Garg et al. ()

Sugarcane bagasse activated with zinc chloride

77.5

8.58

–

6.85

–

>87%

Cronje et al. ()

et al. ). It has various functional groups, such as meth-

Modiﬁed coconut waste

oxy, phenolic hydroxyl and carbonyl, that have the ability to bind heavy metals (Ravat et al. ). Farajzadeh &

Coconut waste is also used as an adsorbent for chromium

Monji () demonstrated the removal of chromium

removal. Its sorption properties are due to the presence of

using wheat bran with a maximum adsorption capacity of

coordinating functional groups such as hydroxyl and car-

93 mg/g and a maximum removal of 89%. Wheat bran can

boxyl (Tan et al. ). Coconut coir pith and coconut

be modiﬁed by using different acids to increase removal

shell are coconut wastes suitable for heavy metal removal.

capacity (Al-Khaldi et al. ). The thermo-chemical inter-

Coir pith is a light ﬂuffy biomaterial and is generated

action between wheat bran and acids increases with

during the separation process of ﬁbre from coconut husk

temperature. Thus, Özer & Özer () modiﬁed wheat

(Namasivayam & Sureshkumar ). Notably, 7.5 million

bran using sulphuric acid and demonstrated chromium

tons per year of coconut is produced in India (Chadha

removal with an adsorption capacity of up to 133 mg/g at

). Raw coir pith consists of 35% cellulose, 1.8% fats,

an optimum pH of 1.5. Kaya et al. () used tartaric acid

25.2% lignin and resin, 7.5% pentosans, 8.7% ash content,

for modiﬁcation of wheat bran and reported a 51% removal

11.9% moisture content and 10.6% other substances (Dan

without modiﬁcation, while after modiﬁcation, removal was

). Namasivayam & Sureshkumar () modiﬁed coir

up to 90% at pH 2 and the maximum adsorption capacity

pith using the surfactant hexadecyltrimethylammonium bro-

was reported to be 4.53 mg of Cr(VI)/g and 5.28 mg of

mide for chromium removal. The maximum removal

Cr(VI)/g at pH 2.2, without and with modiﬁcation, respect-

obtained with this material was reported as being higher

ively. Table 7 summarises the reported use of modiﬁed

than 90% at an optimum pH of 2 and the maximum adsorp-

wheat bran as an adsorbent for chromium removal.

tion capacity was 76.3 mg/g. This was higher than the

Table 7

Chromium removal using modiﬁed wheat bran as an adsorbent Metal

Adsorbent

concentration (mg/L)

Optimum pH

Wheat bran

Wheat bran modiﬁed using sulphuric acid

50, 100

Wheat bran modiﬁed using tartaric acid

Page 200

Best model ﬁt

Contact time (min)

Adsorbent dose (g/L)

Adsorbent capacity (mg/g)

Removal per cent (%)

89%

Farajzadeh & Monji ()

1.5

Langmuir

300

2.0

133

99.9%

Özer & Özer ()

2, 2.2

Freundlich

15–1,440

5.28

90%

Kaya et al. ()

References

395

Renu et al.

Journal of Water Reuse and Desalination

Heavy metal removal from wastewater using various adsorbents

07.4

2017

maximum adsorption capacity obtained using raw coir pith

interfered with by the presence of iron as more than one

which was only 1.24 mg/g (Sumathi et al. ). This

heavy metal in the mixture may increase, decrease or may

demonstrates that the adsorption capacity obtained after

not affect removal performance of the adsorbent. The

modiﬁcation was much higher. Similarly, Shen et al. ()

removal per cent and adsorption capacity obtained in

removed chromium using coconut coir and derived char

single phase (presence of chromium only) was 51% and

and reported a maximum removal of 70%. Table 8 summar-

4.79 mg/g while for the binary system (presence of chro-

ises the reported use of modiﬁed coconut waste as an

mium along with iron) it was 79% and 7.60 mg/g. López-

adsorbent for chromium removal.

Téllez et al. () removed chromium by preparing a composite that incorporates iron nanoparticles into orange peel pith. It was found that for this composite the percentage

Modiﬁed orange peel waste

removal and adsorption capacity were 71% and 5.37 mg/g, Orange peel is used as an adsorbent for the removal of chro-

respectively, as compared to raw orange peel, i.e., 34%

mium from wastewater because it contains cellulose,

and 1.90 mg/g, respectively. Table 9 summarises the

hemicelluloses, pectin (galacturonic acid) and lignin (Feng

reported use of modiﬁed orange peel waste as an adsorbent

et al. ). These components also have various coordinat-

for chromium removal from wastewater.

ing functional groups including carboxylic and phenolic acid groups which can bind heavy metals. Orange peel is

Modiﬁed sawdust

an attractive adsorbent because of its availability and low cost (Feng et al. ). Marín et al. () studied the role

As a solid waste, sawdust is produced in large quantities at

of three major functional groups (amine, carboxyl and

sawmills. It contains primarily lignin and cellulose. Sawdust

hydroxyl) on chromium removal where the bioadsorbent

has been used as an adsorbent for heavy metal removal and

(orange peel) was chemically modiﬁed by esteriﬁcation,

shows good removal (Shukla et al. ). Sawdust is

acetylation and methylation in order to selectively block

obtained by cutting, grinding, drilling, sanding or by pulver-

the functional groups. Thus, esteriﬁcation decreased

ising wood with a saw or other tool producing ﬁne wood

removal capacity, which indicates that the carboxylic

particles. Argun et al. () used hydrochloric acid modi-

groups present in the adsorbent are important for chromium

ﬁed oak sawdust (Quercus coccifera) for the removal of

removal and that the amine and hydroxyl groups have a neg-

chromium. This treatment increases the proportion of

ligible effect. The maximum adsorption capacity reported by

active surfaces and prevents the elution of tannin com-

these researchers was 40.56 mg/g. Lugo-Lugo et al. ()

pounds that would stain treated water. The maximum

biosorbed chromium on pre-treated orange peel in both

removal efﬁciency reported was 84% for Cr(VI) at pH 3

single (presence of chromium only) and binary mixtures

and the maximum adsorption capacity was 1.70 mg/g for

(presence of chromium along with iron). It was observed

Cr(VI) at pH 3. Politi & Sidiras () used pine sawdust

that in the binary mixture, removal of chromium was

modiﬁed with 0.11–3.6 N sulphuric acid for the removal of

Table 8

Chromium removal using modiﬁed coconut waste as an adsorbent Metal concentration

Optimum

Contact

Adsorbent

Adsorbent capacity

Removal per cent

Adsorbent

(mg/L)

Best model ﬁt

time (min)

dose (g/L)

(mg/g)

(%)

References

Modiﬁed coir pith using the surfactant hexadecyltrimethylammonium bromide

20–100

Langmuir, 30–90 Freundlich and Dubinin– Raduskevich

76.3, 1.24

>90%

Namasivayam & Sureshkumar ()

Coconut coir and derived char

10–500

–

7,200

1.0

70.4

70%

Shen et al. ()

Page 201

396

Table 9

Renu et al.

Journal of Water Reuse and Desalination

Heavy metal removal from wastewater using various adsorbents

07.4

2017

Chromium removal using modiﬁed orange peel waste as an adsorbent Metal concentration

Optimum

Contact

Adsorbent

Adsorbent capacity

Adsorbent

(mg/L)

Removal

Best model ﬁt

time (min)

dose (g/L)

(mg/g)

per cent (%)

References

Modiﬁed orange peel

0–500

Langmuir

4,320

4.0

40.56

82%

Marín et al. ()

Pre-treated orange peel

Langmuir model

260

10.0

4.79, 7.60

51%, 79%

Lugo-Lugo et al. ()

Composite of iron nanoparticles into orange peel pith

10–50

Langmuir

5.0

1.90, 5.37

34%, 71%

López-Téllez et al. ()

chromium and reported a maximum adsorption capacity of

to the adsorbent) has an afﬁnity for chromium. Egg shells

20.3 mg/g and 86% removal at pH 2. Table 10 summarises

have been used for the removal of chromium from water

the reported use of modiﬁed sawdust as an adsorbent for

in both modified and non-modified forms. Modification is

chromium removal from wastewater.

carried out by calcinating at high temperatures. After calcination the structure changes due to the development of pores via the emission of carbondioxide gas (Rohim et al.

Modiﬁed egg shell

). Daraei et al. () used egg shell for chromium Although chicken eggs are a worldwide daily food they also

removal and reported 93% removal at an optimum pH 5

pose environmental problems. For example, in the United

and 1.45 mg/g of maximum adsorption capacity. Liu &

States, about 150,000 tons of this material is disposed of in

Huang () modiﬁed egg shell using PEI. The PEI functio-

landﬁlls every year (Toro et al. ). Egg shell has an out-

nalises the eggshell membrane (ESM) via cross-linking

standing mechanical performance, such as an excellent

reactions between various functional groups. The prepared

combination of stiffness, strength, impact resistance and

bioadsorbent is reported as interacting strongly with chro-

toughness. The composition is about 95% calcium carbon-

mium(VI), and the uptake capacity of the PEI–ESM was

ate (which occurs in two crystal forms: hexagonal calcite

increased by 105% compared with the unmodiﬁed egg

and rhombohedral aragonite) and 5% organic materials.

shell with a maximum removal of 90% and a maximum

The amine and amide groups of the proteins on the surface

adsorption capacity of up to 160 mg/g at an optimum pH

of particulate egg shell are a potential source of hardening

of 3. Table 11 summarises the reported use of modiﬁed

agent and help in chromium removal via chelation (Guru

egg shell as an adsorbent for chromium removal from

& Dash ) and this hardening agent (providing strength

wastewater.

Table 10

Chromium removal using modiﬁed sawdust as an adsorbent Adsorbent

Metal Contact

Adsorbent

capacity

Removal

Best model ﬁt

time (min)

dose (g/L)

(mg/g)

per cent (%)

References

Langmuir, D–R isotherms

0–720

1.70

84%

Argun et al. ()

Freundlich

240

20.3

–

Politi & Sidiras ()

concentration

Optimum

Adsorbent

(mg/L)

Hydrochloric acid modiﬁed oak sawdust (Quercus coccifera)

0.1–100

Sulphuric acid modiﬁed pine sawdust

15–75

Page 202

397

Table 11

Renu et al.

Journal of Water Reuse and Desalination

Heavy metal removal from wastewater using various adsorbents

07.4

2017

Chromium removal using modiﬁed egg shell as an adsorbent Metal concentration

Optimum

Best model

Contact

Adsorbent

Removal per

Adsorbent

(mg/L)

ﬁt

time (min)

dose (g/L)

capacity (mg/g)

cent (%)

References

Egg shell

5–30

Freundlich

3.5

1.45

93%

Daraei et al. ()

Egg shell modiﬁed using PEI

100

Langmuir

10–1,440

10–40

160

90%

Liu & Huang ()

Commercially available adsorbents for cadmium

regular two-dimensional hexagonal array of channels with a

removal from wastewater

pore diameter of the order of 7–10 nm. The reported removal was 98% at pH > 4.5. Similarly, Burke et al. () also used

Mesoporous silica

aminopropyl and mercaptopropyl, functionalised and bi-functionalised, large pore mesoporous silica spheres for the

Mesoporous silica is a highly ordered material which possesses

removal of chromium from wastewater. These researchers

a regular two-dimensional hexagonal array of channels. Meso-

reported a maximum sorption capacity of 43.16 mg/g for Cr.

porous silica is efﬁcient for cadmium removal because of its

Pérez-Quintanilla et al. () modiﬁed silica and amorphous

high surface area and 2–10 nm pore size (Bhattacharyya

silica using 2-mercaptopyridine and reported maximum

et al. ). Mesoporous silica may be chemically modiﬁed

adsorption capacities of 205 mg/g and 97 mg/g, respectively.

via the attachment of groups including carboxylic acid, sulfo-

Table 12 documents the available data for mesoporous silica

nic acid and amino-carbonyl. Javadian et al. () prepared

for cadmium removal from wastewater.

polyaniline/polypyrrole/hexagonal type mesoporous silica for cadmium removal and reported a removal of 99.2% cad-

Chitosan

mium at an optimum pH of 8. Hajiaghababaei et al. () modiﬁed SBA-15 nanoporous silica by functionalising it with

Chitosan is a derivative of the N-deacetylation of chitin which

ethylenediamine. SBA-15 is a highly ordered material with a

is a naturally occurring polysaccharide obtained from

Table 12

Cadmium removal using mesoporous silica as an adsorbent Adsorbent

Metal Adsorbent functionalised

concentration

Optimum

Best model

Contact

Adsorbent

capacity

Removal

with

(mg/L)

ﬁt

time (min)

dose (g/L)

(mg/g)

per cent (%)

References

Silica functionalised with mono amino and mercapto groups

Langmuir

1,440– 2,880

12.36, 14.61, 28.10

80%

Machida et al. ()

Amino functionalised silica

Langmuir

120

18.25

90%

Heidari et al. ()

Amino functionalised mesoporous silica

5–300

–

Langmuir

1,440

1.11

93.30

100%

Aguado et al. ()

Iminodiacetic acidmodiﬁed mesoporous SBA-15

50–1,000

5.6

Langmuir

1,440

4.0

–

99.8%

Gao et al. ()

Polyamine-functionalised

100

5.5–7

–

2,880

–

70%

Alothman & Apblett ()

Page 203

398

Renu et al.

Journal of Water Reuse and Desalination

Heavy metal removal from wastewater using various adsorbents

07.4

2017

crustaceans. Chitosan is an efﬁcient adsorbent for the removal

). Similarly, Hydari et al. () modiﬁed chitosan by coat-

of heavy metals (Ren et al. ). Chitosan is cheap, hydrophilic

ing with activated carbon and reported an adsorption capacity

and biodegradable and it also offers ease of derivatisation. It

of 52.63 mg/g adsorption capacity at an optimum pH of 6 with

contains amino and hydroxyl groups that may form chelates

100% removal. Table 13 presents cadmium removal data for

with heavy metals (Huo et al. ; Hu et al. ). Chitosan

chitosan as an adsorbent from wastewater.

has the advantage of being cheap yet effective, but has the disadvantages of being mechanically weak, soluble under acidic

Zeolite

conditions and may leach carbohydrate when used in raw form (Ren et al. ; Huo et al. ). Various efforts have

Zeolites are among the best adsorbents for the removal of cad-

been made to stabilise chitosan using cross-linking agents,

mium because they are composed of hydrated aluminosilicate

but this results in a decrease in adsorption capacity (Wang

minerals made from the interlinked tetrahedra of alumina

et al. ). Thus, Chen et al. () have developed ‘ion imprint

(AlO4) and silica (SiO4) moieties (Choi et al. ). Zeolite

technology’ for achieving higher adsorption capacity and stab-

has good ion exchange properties, a high surface area and a

ility. This involves the development of a novel adsorbent that is

hydrophilic character which makes them suitable for seques-

a thiourea-modiﬁed magnetic ion imprinted chitosan/TiO2

tration of cadmium. Modiﬁed zeolite provides a higher

composite for the removal of cadmium. The maximum adsorp-

adsorption capacity compared to natural zeolite. There are

tion capacity obtained for this material was reported to be

different methods for zeolite modiﬁcation. For example, nano-

256.41 mg/g at an optimum pH of 7. Chitosan has also been

sized zeolite has more accessible pores which make it more

modiﬁed by a coating process involving ceramic alumina. Coat-

suitable for heavy metal removal. Among nanosized zeolite

ing helps increase accessibility of binding sites and improves

adsorbents, NaX nanozeolite (Ansari et al. ) (in molar

mechanical stability. Maximum adsorption capacity obtained

ratio of 5.5 Na2O:1.0 Al2O3:4.0 SiO2:190 H2O) is widely

was reported to be 108.7 mg/g at an optimum pH of 6 and

used for cadmium removal from wastewater (Erdem et al.

the maximum removal reported was 93.76% (Wan et al.

; Jha et al. ; Ibrahim et al. ; Aliabadi et al. ;

Table 13

Cadmium removal using chitosan as an adsorbent Metal concentration

Optimum

Contact time

Adsorbent

Adsorption capacity

Adsorbent

(mg/L)

Removal

Best model ﬁt

(min)

dose (g/L)

(mg/g)

(%)

References

α-Ketoglutaricacidmodiﬁed magnetic chitosan

100–500

Langmuir

0.04

201.2

93%

Yang et al. (b)

Electrospun nanoﬁbre membrane of PEO/ chitosan

50–1,000

Freundlich, Langmuir and Dubinin– Radushkevich

120

–

248.1

72%

Aliabadi et al. ()

Nano-hydroxyapatite/ chitosan composite

100–500

5.6

Freundlich and Langmuir

5.0

92, 122

92%

Salah et al. ()

Polyaniline grafted cross-linked chitosan beads

40–220

Langmuir

120

4.5

145

99.6%

Igberase & Osifo ()

O-carboxymethyl functionalisation of chitosan

675

–

1,440

–

95%

Borsagli et al. ()

Multi-walled carbon nanotubes modiﬁed with chitosan

–

6–7

–

>90%

Salam et al. ()

Page 204

399

Renu et al.

Journal of Water Reuse and Desalination

Heavy metal removal from wastewater using various adsorbents

07.4

2017

Rad et al. ). Rad et al. () synthesised NaX nanozeolite

demonstrates more than 80% cadmium removal at an opti-

using a microwave heating method, and then polyvinylacetate

mum pH of 7–8. Table 14 summarises the removal

polymer/NaX nanocomposite nanoﬁbres were prepared using

parameters for the sequestering of cadmium using zeolite.

electrospinning method; the potential of these composite nanoﬁbres was then investigated for cadmium. The reported

Red mud

maximum adsorption capacity was 838.7 mg/g with 80% removal at an optimum pH of 5. Choi et al. () modiﬁed zeo-

Red mud is a waste material from the aluminium industry that

lite by replacing Si(IV)and Al(III) sites in the lattice with

may be converted into an efﬁcient adsorbent for cadmium

exchangeable cations such as sodium, magnesium, potassium,

removal from waste water (Gupta & Sharma ). Red mud

or calcium, leading to a net negative charge. Mg-modiﬁed zeo-

has the advantage of being cheap and available and possesses

lite has certain advantages such as non-toxicity, low cost,

a high capacity for cadmium removal; however, it also has

abundance (and hence availability) and large pore size of

some disadvantages including the difﬁculty of dealing with

40–50 nm compared to the non-modiﬁed adsorbent. This

the wastewater produced during red mud activation before

Mg-modiﬁed adsorbent has a cadmium removal of more

application, and regeneration/recovery of red mud is difﬁcult

than 98% at an optimum pH of 7. In addition, the adsorption

after application (Zhu et al. ). However, Zhu et al. ()

capacity of Mg-zeolite was found to be 1.5 times higher than

developed red mud as a novel adsorbent for cadmium removal

that of zeolite modiﬁed with Na or K and 1.5 to 2.0 times

from wastewater. In this regard, the adsorption process onto

higher than that of natural zeolite.

granular red mud was found to be spontaneous and endother-

Coal, which is used in many industries as a fuel, pro-

mic in nature. A maximum adsorption capacity of 52.1 mg/g

duces ﬂy ash as a by-product which causes air pollution

was reported at a pH of 3 to 6. Similarly, Gupta & Sharma

and presents disposal problems. Due to its low cost ﬂy ash

() also used red mud for cadmium removal from waste-

can be used for zeolite formation using the hydrothermal

water and complete removal was obtained at the lower

process (Hui et al. ). Javadian et al. () converted

concentration (1:78 × 10�5 to 1:78 × 10�4 Molar) while 60–

ﬂy ash into amorphous aluminosilicate adsorbent and

65% removal was obtained at the higher concentration

reported a maximum adsorption capacity for cadmium of

(1:78 × 10�4 to 1:78 × 10�3 Molar)

26.246 mg/g with 84% removal at an optimum pH of 5. Simi-

between 4 and 5. Ma et al. () used CaCO3-dominated

larly, Visa () converted ﬂy ash into zeolite for cadmium

red mud (red mud containing substantial amounts of CaCO3)

removal through a hydrothermal process using sodium

for cadmium removal from wastewater. With increase in satur-

hydroxide. These researchers reported that this product

ation degree of binding sites on red mud particles by the heavy

has high surface area and is rich in micropores and

metal, the proportion of HCH3COO-extractable Cu fraction

Table 14

optimum

Cadmium removal using using zeolite as an adsorbent Metal concentration

Optimum

Contact

Adsorbent

Removal

Adsorbent

(mg/L)

Best model ﬁt

time (min)

dose (g/L)

capacity (mg/g)

(%)

References

Synthetic zeolite A

100–2,000

–

Freundlich and D–R

180

1.0

315.65

–

El-Kamash et al. ()

Zeolite

25–100

Freundlich

–

76%

Rao et al. ()

Zeolite from ﬂy ash

1124.1–3372.3

6.6

Langmuir

1,440

57–195

95.6%

Izidoro et al. ()

Oil shale into zeolite

100

Sips

60–1,440

–

95.6

–

Shawabkeh et al. ()

Natural zeolite

9–90

Freundlich

1,440

–

71%

Hamidpour et al. ()

Page 205

400

Renu et al.

Journal of Water Reuse and Desalination

Heavy metal removal from wastewater using various adsorbents

07.4

2017

(acetic acid-extractable Cu fraction) increased accordingly.

and methanol to produce an adsorbent. The prepared

Cadmium increasingly bound with HCH3COO-extractable

DCB material behaved as a cation exchanger with 90%

forms until adsorption capacity of red mud was depleted. Ju

removal at an optimum pH 8. Azouaou et al. () used

et al. () mixed 2–8% w/w granular red mud with cement

waste material from cafeterias as an adsorbent for cad-

and reported an adsorption capacity of 9 mg/g. It was also

mium removal and reported an adsorption capacity of

found that an increase in temperature increases the equili-

15.65 mg/g with more than 80% removal at an optimum

brium adsorption which suggests that this adsorption process

pH of 7. Table 15 presents cadmium removal data for

is endothermic in nature.

coffee residue as an adsorbent.

Bio-adsorbents for the removal of cadmium

Rice husk

Coffee residue

Rice husk is an agricultural waste obtained from rice mills and it consists of cellulose, hemicelluloses, mineral ash,

Coffee residue has been reported as an efﬁcient adsorbent

lignin and a high percentage of silica (Rahman et al.

for the removal of cadmium from wastewater. For example,

). It contains groups such as –OH, Si-O-Si and -Si-H

Boonamnuayvitaya et al. () used coffee residues for

which have an afﬁnity for cadmium coordination and

cadmium removal and also blended them with clay to pre-

hence removal. It may be useful as an adsorbent for cad-

pare an adsorbent with a negative charge which promotes

mium removal because it is cheap and easily available.

cadmium complexation and removal. The prepared adsor-

Chemicals that are used for the modiﬁcation of rice husk

bent contains hydroxyl, carbonyl and amine groups and

in order to increase adsorption capacity include the bases

has a pyrolysis temperature of 500 C (this temperature

sodium hydroxide, epichlorohydrin and sodium carbonate

gives maximum adsorption capacity) and a particle size

(Kumar & Bandyopadhyay ). Ye et al. () modiﬁed

diameter of 4 mm. A weight ratio of coffee residue to

rice husk by constant stirring with sodium hydroxide for 24

clay of 80:20 was found to be the most suitable blend. Oli-

hours and reported an adsorption capacity for cadmium

veira et al. () employed coffee husks that comprise the

removal of 125.94 mg/g, which is higher than the non-

dry outer skin, pulp and parchment as these are likely to

modiﬁed rice husk at 73.96 mg/g, at an optimum pH of

represent the major residue obtained from the handling

6.5. Kumar & Bandyopadhyay () modiﬁed rice husk

and processing of coffee. For this material, the maximum

using epichlorohydrin, sodium hydroxide and sodium

adsorption capacity was reported to be 6.9 mg/g at an opti-

bicarbonate, and the adsorption capacity increased from

mum pH of 4 with a removal of 65–85%. Kaikake et al.

8.58 mg/g for raw rice husk to 11.12 mg/g, 20.24 mg/g

() soaked and degreased coffee beans (DCB) in water

and 16.18 mg/g, respectively, with the removal increasing

Table 15

Cadmium removal using coffee residue as an adsorbent Adsorbent

Metal concentration

Optimum

Best model

Contact

Adsorbent

capacity

Removal

Adsorbent

(mg/L)

ﬁt

time (min)

dose (g/L)

(mg/g)

per cent (%)

References

Coffee residues blended with clay

25–250

1.6–2.5

Langmuir

17.5–17.9

88–92%

Boonamnuayvitaya et al. ()

Coffee husks

50–100

Langmuir

4,320

6.7

6.9

65–85%

Oliveira et al. ()

Coffee beans

6–202

Langmuir

1,440

3.80

90%

Kaikake et al. ()

Coffee grounds from cafeterias

10–700

Langmuir

120

15.65

>80%

Azouaou et al. ()

Page 206

401

Renu et al.

Journal of Water Reuse and Desalination

Heavy metal removal from wastewater using various adsorbents

07.4

2017

from 75% to 86.2%, 97% and 97.2%, respectively, at an

anhydride in the presence of toluene in basic medium. The

optimum pH of 9. It was also reported that the equilibrium

adsorption capacity reported for this material was 200 mg/g

time was reduced from 600 min to 120 min, 240 min and

at an optimum pH of 4. Aziz et al. (b) modiﬁed olive

60 min, respectively. Ajmal et al. () treated rice husk

stones using concentrated sulphuric acid at room tempera-

using phosphate, and a maximum removal of 99% was

ture

reported at an optimum pH of 12. Srivastava et al. ()

hydroxide solution, and the maximum adsorption capacity

used mesoporous rice husk with an 80% pore area (ratio

was reported to be 128.2 mg/g at an optimum pH range of

followed

neutralisation

with

0.1 N

sodium

of rice husk’s unoccupied area to its total area) and

5–10. Blázquez et al. () used olive stones for cadmium

reported a 23.3% cadmium removal along with some

removal and observed the effect of different parameters on

other heavy metals at an optimum pH of 6. Sharma

the percentage removal. Thus it was found that for a smaller

et al. () used polyacrylamide grafted rice husk for cad-

size of adsorbent particles (250–355 nm) the removal

mium removal from wastewater, and 85% removal was

capacity increases up to 90% at an optimum pH of 11,

reported at an optimum pH of 9. Table 16 summarises

and the maximum adsorption capacity was reached within

the removal parameters for the sequestering of cadmium

20 minutes, which is fast compared to the equilibrium

using rice husk.

time achieved in cadmium removal using olive stones prepared by ZnCl2 activation (Kula et al. ) and by using

Powdered olive stones

olive cake (Al-Anber & Matouq ). Olive stone can also be used as an adsorbent by converting it into activated

Olive stones form part of the waste produced from the oleic

carbon using chemicals such as ZnCl2, H3PO4 and H2O2

industry and are available in olive oil producing countries

with a subsequent improvement in pore distribution that

(Bohli et al. ). Thus, the olive stone is a plentiful by-pro-

increases the surface area of the adsorbent. Kula et al.

duct of the olive oil industry and is a candidate for use as an

() used 20% zinc chloride as an olive stone activating

adsorbent for the removal of cadmium. Olive stones can be

agent for cadmium removal and 95% removal was reported

modiﬁed using succinic anhydride, sulphuric acid, nitric

and compared with raw olive stone (43%) at an optimum

acid or sodium hydroxide to increase adsorption (Blázquez

pH of 9. Obregón-Valencia & del Rosario Sun-Kou ()

et al. ; Aziz et al. a). Aziz et al. (a) modiﬁed

prepared activated carbon from carbon aguaje and olive

olive stones using succinic anhydride that chemically func-

fruit stone using phosphoric acid solution, and a maximum

tionalises it with succinate moieties that have an afﬁnity

adsorption capacity of 8.14 mg/g and 9.01 mg/g and a

for cadmium. This adsorbent was synthesised by esterifying

removal capacity of 61% and 68% was obtained for aguaje

the lignocellulosic matrix of the olive stone with succinic

and olive fruit stones, respectively. Hamdaoui ()

Table 16

Cadmium removal using rice husk as an adsorbent Metal

Adsorption

Removal

Contact time (min)

Adsorbent dose (g/L)

capacity (mg/g)

per cent (%)

Freundlich, Redlich– Peterson

1–10

3.04

29.8%

Srivastava et al. ()

Langmuir

1.0

41.15 and 38.76

–

El-Shafey ()

Freundlich, Langmuir and Dubinin– Radushkevish

4.0

–

97%

Akhtar et al. ()

Adsorbent

concentration (mg/L)

Optimum pH

Rice husk ash

10–100

Sulphuric acidtreated rice husk

50, 100

Activated rice husk

8.9–89 M

Best model ﬁt

Reference

Page 207

402

Renu et al.

Journal of Water Reuse and Desalination

Heavy metal removal from wastewater using various adsorbents

07.4

2017

compared the adsorption capacity of olive stone in the

because of these polyphenols, amine and carboxyl groups

absence of ultrasound (42.19 mg/g) and in the presence of

(Foo & Lu ; Lu & Foo ). Chand et al. () chemi-

ultrasound (55.87 mg/g) and with combined ultrasound/stir-

cally modiﬁed apple pomace with succinic anhydride via a

ring (64.94 mg/g). Ultrasound increases adsorption capacity

simple ring opening mechanism that provides a larger surface

of olive stone due to acoustic power which enhances mass

area on the material. The surface area is reported to increase

and heat transfer at ﬁlms and within the pore. Further, com-

by 18%, and consequently, 50 times less apple pomace was

bination of stirring with ultrasound leads to intensiﬁcation

required as an adsorbent. The adsorption capacity of modi-

of the removal of cadmium. Table 17 summarises the

ﬁed apple pomace (91.74 mg/g) was increased 20 times

removal parameters for the sequestering of cadmium using

compared to non-modiﬁed apple pomace (4.45 mg/g) and

powdered olive stones.

for the modiﬁed apple pomace a removal of 90% was obtained compared to 70% for non-modiﬁed apple pomace at an optimum pH of 4. Similarly, in other work, these

Apple pomace

researchers prepared an adsorbent by introducing the Apple pomace is a waste product from the apple juice indus-

xanthate moiety into apple pomace. The maximum adsorp-

try and is usually dumped at industrial sites in very large

tion capacity obtained for the xanthate modiﬁed material

quantities (Chand et al. ). An apple (solid residue part)

was reported to be 112.35 mg/g with a maximum removal

consists of the ﬂesh 95% (wt%), seed 2–4% (wt%) and stem

of 99.7% at an optimum pH of 4. This research suggests that

1% (wt%) (Chand et al. ). Apple pomace is the solid resi-

chemically modiﬁed apple pomace is best for cadmium

due part of the apple which is obtained during its processing

removal and the introduction of xanthate gives higher

�1

(Chand et al. ). Apple pomace contains 7.24 g kg

removal than succinic anhydride. Table 18 presents cadmium

removal data for apple pomace as an adsorbent.

total polyphenol which includes epicatechin (0.64 g/kg), caffeic acid (0.28 g/kg), 3-hydroxyphloridzin (0.27 g/kg),

Modiﬁed coconut waste

phloretin-20-xyloglucoside (0.17 g/kg), phloridzin (1.42 g/kg), quercetin-3-galactoside (1.61 g/kg), quercetin-3-galucoside (0.87 g/kg),

quercetin-3-xyloside

quercetin-

Seven and a half million tons of coconut per year is pro-

quercetin-3-rhamnoside

duced in India alone and the waste by-products have been

(0.47 g/kg). Thus, apple pomace behaves as a metal chelator

used as adsorbents for cadmium removal (Chadha ).

3-arbinoside

Table 17

(0.98 g/kg)

and

(0.53 g/kg),

Cadmium removal using powdered olive stone as an adsorbent

Adsorbent functionalised/

Metal

composite with/

concentration

Optimum

Contact

Adsorbent

Adsorption

Removal

modiﬁed

(mg/L)

Best model ﬁt

time (min)

dose (g/L)

capacity (mg/g)

per cent (%)

Reference

Olive cake

100

Langmuir and Freundlich

1,440

0.3

65.4

66%

Al-Anber & Matouq ()

Zinc chloride activated olive stone

15–45

Langmuir and Freundlich

–

95%

Kula et al. ()

Microwaved olive stone activated carbon

Langmuir

0.25–2

11.72

95.32%

Alslaibi et al. ()

Activated carbon from olive stones

56–562

Redlich– Peterson

200

17.665

23%

Bohli et al. ()

Olive stone waste

33–16,861

5.5–6

Langmuir and Freundlich

13.33

–

49.2%

Fiol et al. ()

Page 208

403

Table 18

Renu et al.

Journal of Water Reuse and Desalination

Heavy metal removal from wastewater using various adsorbents

07.4

2017

Cadmium removal using apple pomace as an adsorbent Metal concentration

Optimum

Best model

Contact

Adsorbent

Adsor bent

Removal per

Adsorbent

(mg/L)

ﬁt

time (min)

dose (g/L)

capacity (mg/g)

cent (%)

References

Succinic anhydride modiﬁed apple pomace

10–80

Langmuir

10–180

0.8 and 40

4.45, 91.74

70%, 90%

Chand et al. ()

Xanthate moiety into apple pomace

10–120

Langmuir

5–60

0.2–8

112.35

99.7%

Chand et al. ()

The sorption properties are due to the presence of functional

concentration range of 20 to 1,000 ppm with a maximum

groups such as hydroxyl and carboxyl and this material

adsorption capacity of 285.7 mg/g and 98% removal at pH 7.

demonstrates a high afﬁnity for metal ions (Tan et al.

Similarly, Sousa et al. () used green coconut shell for

). Coconut coir pith and coconut shell are waste by-pro-

cadmium removal, along with other heavy metals, and the

ducts that can be used for cadmium removal. Coir pith is a

maximum adsorption capacity found for the single com-

light ﬂuffy biomaterial generated during the separation of

ponent system (presence of cadmium only) was reported

the coconut ﬁbre from the husk (Namasivayam & Sureshku-

to be 37.78 mg/g and for the multicomponent system (pres-

mar ). Raw coir pith consists of 35% cellulose, 1.8%

ence of lead, nickel, zinc and copper along with cadmium),

fats, 25.2% lignin and resin, 7.5% pentosans, 8.7% ash,

11.96 mg/g at pH 5. Table 19 presents cadmium removal

11.9% moisture and 10.6% other substances (Dan ).

data for modiﬁed coconut waste as an adsorbent.

Kadirvelu & Namasivayam () prepared activated carbon using coconut coir pith and reported a maximum

Commercially available adsorbents for copper removal

adsorption capacity of 93.4 mg/g at a pH of 5. For cadmium

from wastewater

removal, along with some other heavy metals, Jin et al. () converted coconut into activated carbon and then grafted it

Magnetic adsorbents

with tetraoxalyl ethylenediamine melamine chelate using a pressure relief dipping ultrasonic method. The maximum

Various magnetic adsorbents have been used or show

adsorption capacity reported was 26.41 mg/g at an optimum

potential for the effective removal of copper from waste-

pH of 5.5. Pino et al. () used green coconut shell

water, including ‘magnetic’ adsorbent micro- and nano-

powder and reported removal of cadmium over a large

sized particles (Yin et al. ). These latter adsorbents

Table 19

Cadmium removal using modiﬁed coconut waste as an adsorbent Metal

Adsorbent

concentration (mg/L)

Optimum pH

Activated carbon from coconut shell

5–40

Activated carbon from coconut shell

1,124

Green coconut shell Green coconut shell

Adsorbent

Contact time (min)

Adsorbent dose (g/L)

capacity (mg/g)

Removal per cent (%)

Langmuir, Freundlich

120

0.3921

93.4

98%

Kadirvelu & Namasivayam ()

5.5

Langmuir

240

26.41

93.4%

Jin et al. ()

20–1,000

Langmuir

120

285.7

98%

Pino et al. ()

100

–

1.33– 9.98

1.620

37.38, 11.96

–

Sousa et al. ()

Best model ﬁt

References

Page 209

404

Renu et al.

Journal of Water Reuse and Desalination

Heavy metal removal from wastewater using various adsorbents

07.4

2017

show high adsorption capacity and can be harvested from

using an iron salt co-precipitation method followed by

aqueous solution in the presence of a suitable magnetic

direct encapsulation with a coating of pectin and in the

ﬁeld. In addition, such material is potentially reusable

absence of calcium cross linking. The experimental data

(Mehta et al. ). A problem with the use of unmodiﬁed

are reported to ﬁt both Langmuir and Freundlich models

magnetic particles is the formation of aggregates due to

and a maximum adsorption capacity of 48.99 mg/g was

magnetic dipolar attraction between the particles. To pre-

reported. The adsorbent can be further regenerated using

vent this, a layer of various polymer compounds or the

EDTA and a removal of 93.70% was obtained after the

inorganic oxide may be coated on the surface of the par-

ﬁrst regeneration cycle and a removal of 58.66% remained

ticles (Yin et al. ). Ren et al. () prepared a novel

even after a ﬁfth cycle. Hu et al. () used sulfonated gra-

adsorbent by using waste fungal mycelium obtained from

phene oxide for removal of copper from wastewater. The

industry (industries dealing with fungal products such as

introduction of the sulfo functional group to graphene

antibiotics, citric acid and enzymes), chitosan and iron

oxide is reported to increase the copper adsorption with

oxide nanoparticles utilising metal imprinting technology.

an adsorption capacity of 62.73 mg/g at pH 4.68 and the

Fungal mycelium has been used because of its low cost,

experimental data ﬁt the Langmuir isotherm.

abundance and high efﬁciency. However, its direct use is difﬁcult because of its limited reusability, relative low adsorption

and

low

mechanic

intensity

(mechanical

strength). Chitosan is considered useful since it is a poly-

and several authors have utilised alumina for this purpose

groups, which have an afﬁnity for copper removal, and

either in nanoparticulate form or via loading with cation

iron oxide is used because it is magnetic. In metal ion

exchangers (Mahmoud et al. ; Fouladgar et al. ).

imprinting technology, selective binding sites are made

For example, Fouladgar et al. () used Ɣ-alumina nano-

on synthetic polymer using metal ion templates, and after

particles for removal of copper along with nickel.

removal of these templates, polymer become more selec-

Nanoparticles are useful because of their high adsorption

tive for heavy metal removal from wastewater. Thus,

capacity due to the high number of metal coordination

binding of chitosan and industrial waste fungal mycelium

sites. These researchers have a best ﬁt for the Freundlich

on iron oxide nanoparticles produces a novel adsorbent

isotherm and a maximum adsorption capacity of 31.3 mg/g

known as magnetic Cu(II) ion imprinted composite adsor-

for copper removal from wastewater. Ghaemi () used

bent (Cu(II)-MICA). Ren et al. () reported that the

a phase inversion method to prepare a mixed matrix mem-

Langmuir isotherm ﬁts the experimental data well and a

brane using PES (polyethersulfone) and varying amounts

maximum adsorption capacity of 71.36 mg/g was reported.

(1 wt%) of alumina nanoparticles. Such mixed matrix

It was also shown that this adsorbent can be reused up to

membranes have shown higher water permeation com-

and

contains

-NH2

-OH

Alumina may be used for copper removal from wastewater

functional

saccharide

and

Alumina

ﬁve times with a regeneration loss of 14–15%. Lan et al.

pared to a pristine PES membrane that is facilitated by

() used hyaluronic acid supported magnetic micro-

the addition of small amounts of nanoparticles. This results

spheres for copper removal, and their adsorption capacity

in an increase in porosity and hydrophilicity. The mixed

is reported to increase from 6 mg/g to 12.2 mg/g as the

matrix membrane has shown the highest removal of

pH is increased from 2 to 6.8, and slowly decreases to

copper from wastewater of 60% compared to the PES

11.6 mg/g up to pH 8. The corresponding adsorption equi-

membrane (around 25%). Mahmoud et al. () removed

librium study showed that the copper adsorption of the

copper from wastewater using three new alumina adsor-

hyaluronic acid-supported magnetic microspheres had the

bents of acidic, neutral and basic nature and their surface

best ﬁt to the Freundlich isotherm model. Gong et al.

was modiﬁed by loading with 1-nitroso-2-napthol as a

() used a pectin-coated iron oxide magnetic nanocom-

cation exchanger. After modiﬁcation, alumina adsorbent

posite as an adsorbent for removal of copper from

become stronger towards acid leaching and thermal

wastewater. This nanocomposite adsorbent was synthesised

decomposition. The adsorption capacities obtained were

Page 210

405

Renu et al.

Journal of Water Reuse and Desalination

Heavy metal removal from wastewater using various adsorbents

07.4

2017

27.96 mg/g, 28.58 mg/g and 28.59 mg/g for the acidic, neu-

copper removal using chitosan immobilised on bentonite

tral and basic adsorbents, respectively. Conventional

clay (Futalan et al. ). Furthermore, the bed was regener-

porous solids such as ﬂy ash, clay and silica materials

ated using NaCl/HCl solution at pH 5 that gave 50% elution

have the disadvantage of having non-uniform pores and

efﬁciency. It increases removal capacity because the bed

low adsorption capacity. Thus, Rengaraj et al. () pre-

becomes free from heavy metals after contact with the

pared aminated and protonated mesoporous alumina for

eluent. Vengris et al. () modiﬁed clay using hydrochloric

copper removal from wastewater. Mesoporous alumina

acid followed by neutralisation of resultant solution with

have several advantages over conventional porous solids

sodium hydroxide for copper removal from wastewater.

such as a large surface area, uniform pore size distribution

Initially, the chemical composition (wt%) of clay was: iron

with a sponge-like interlinked pore system, high stability

oxide 6.9, silicon oxide 44.2, aluminium oxide 15.3, calcium

and high metal uptake capacity (Lee et al. ). Ion

oxide 13.8 and magnesium oxide 4. After treatment with

exchange takes place between copper and the hydrogen

hydrochloric acid, aluminium, iron and magnesium com-

ions that are present on the surface of mesoporous

pounds of clay had increased because acid treatment

alumina, and the maximum adsorption capacity obtained

causes dissolution of iron, calcium, magnesium and alu-

for aminated mesoporous alumina is 7.9239 mg/g com-

minium oxides and during the neutralisation process many

pared to 14.5349 mg/g for protonated mesoporous silica.

dissolved metals (except calcium) reprecipitate in the form of hydroxides and their amount in the modiﬁed adsorbent

Clay

increases. This leads to an increase in metal uptake capacity of modiﬁed clay compared to unmodiﬁed clay. This acidic

Clay may be used for removal of copper from wastewater

treatment led to the decomposition of the montmorillonite

and has a number of advantages over other adsorbents,

structure. The maximum adsorption capacity obtained for

such as high surface area, excellent physical (plasticity,

single component solutions was 0.75 mg/g, for ternary com-

bonding strength, shrinkage)/chemical properties (large

ponent solutions 0.80 mg/g and the experimental data ﬁtted

zeta potential, cation exchange property, shows monobasi-

the Langmuir isotherm. Similarly, Oubagaranadin et al.

city)

bearing

() modiﬁed montmorillonite-illite clay using sulphuric

strength, resistance to wear, resistance to chemical attack)

and

structural/surface

properties

(load

acid for copper removal from wastewater. The Brunauer–

(Singh et al. ; Krikorian & Martin ; Aşçı et al.

Emmett–Teller (BET) model ﬁtted well with the experimen-

). Thus, researchers have studied different types of

tal data and the shape of the isotherm indicated that copper

clay, either in raw form or after its modiﬁcation, for

adsorption was multilayer.

copper sequestration. For example, Bertagnolli et al. () employed bentonite clay after calcination at 400–500 C. W

Bio-adsorbents for copper removal from wastewater

Bentonite has several advantageous properties as an adsorbent including low cost, good ion exchange capacity,

Fungal biomass

selectivity and regenerability. After calcination, the chemical morphology and composition of clay does not change

Fungal biomass has been explored by several researchers for

although the resulting structural changes alter its behaviour

its potential to remove copper from wastewater. The use of

towards water and enables it to use inﬁxed bed columns

fungal biomass for such purposes has been hindered due to

with no expansion. This material showed a maximum

problems such as small particle size, poor mechanical

adsorption capacity of 11.89 mg/g. de Almeida Neto et al.

strength, low density and rigidity (Akar et al. ; McHale

() reported copper removal in a ﬁxed bed using Bofe

& McHale ; Volesky & Holan ). However, the use

bentonite calcinated clay, and a maximum adsorption

of a suitable matrix can potentially overcome these problems.

capacity of 19.0638 mg/g was reported. The equilibrium

Thus, Iqbal & Edyvean () used a low cost, physically

time was increased from 120 to 400 minutes which is

strong and highly porous matrix, namely ‘loofah sponge’ for

much less compared to equilibrium time obtained by

the immobilised biomass of Phanerochaete chrysosporium, Page 211

406

Renu et al.

Journal of Water Reuse and Desalination

Heavy metal removal from wastewater using various adsorbents

07.4

2017

Yeast

and a maximum adsorption capacity of 50.9 mg/g at pH 6 with 98% removal reported. Formaldehyde inactivated Cladosporium cladosporioides,

murorum and

Yeast has been successfully used as an adsorbent for the

Bjerkandera fungi, at optimum conditions, can also be used

Gliomastix

sequestration of copper. Yeast is a fungus and has a larger

for copper removal. These fungi are highly porous, their

size than bacteria and, like other eukaryotic organisms,

mesh structure provides ready access and a large surface

has a nucleus and associated cytoplasmic organelles. The

area for the biosorption of copper. Thus, Li et al. ()

cytoplasm present in living cells is important for the living

obtained maximum adsorption capacities of 7.74 mg/g,

cells because it interacts with metal ions and after entering

9.01 mg/g and 12.08 mg/g, and removals of 93.79%, 85.09%

into the cells, the heavy metal ions are separated into com-

and 81.96%, for C. cladosporioides, G. murorum and Bjer-

partments for removal (Wang & Chen ). Waste beer

kandera fungi, respectively. The biosorption data of all

yeast is a by-product of the brewing industry that is a

fungal species ﬁtted well with the Langmuir model. Ertugay

cheap and promising adsorbent for copper removal from

& Bayhan () used Agaricus bisporus fungi and 73.3%

wastewater (Han et al. ). These researchers reported a

removal was obtained at pH 5 with a preferred ﬁt to the

maximum uptake of copper of 1.45 mg/g with a preferred

Freundlich model compared to other adsorption models.

ﬁt to the Langmuir and Freundlich isotherms; bisorption

Table 20 summarises the parameters for the sequestration

was reached in equilibrium in 30 minutes. The sorption

of copper using fungal biomass.

capacity of beer yeast was found to be a function of the

Table 20

Copper removal using fungal biomass as an adsorbent Adsorption

Intial metal

Contact

concentration

time

Adsorbent

capacity

Removal per

Adsorbent

(mg/L)

Best model ﬁt

(min)

dose (g/L)

(mg/g)

cent (%)

References

Aspergillus niger

10–100

Langmuir and Freundlich

–

23.6

–

Mukhopadhyay ()

Mucor rouxii

10–1,000

5–6

Langmuir, adsorption

4,320

0.25

52.6

96.3%, 94.8%, 95.7%, 96.2%

Majumdar et al. ()

Fungal cells (dead) and (living)

20–100

5–9

–

4,320

0.2

–

95.27%

Hemambika et al. ()

Aspergillus niger

25–100

–

10, 200

15.6

–

Dursun et al. ()

Rhizopus oryazae ﬁlamentous fungus

20–200

4–6

Langmuir

200

19.4

–

Bhainsa & D’Souza ()

Pleurotus pulmonarius CCB019 and Schizophyllum commune

5–200

Langmuir

–

Veit et al. ()

Chlorella sp. and Chlamydomonas sp.

–

33.4

–

Maznah et al. ()

Trametes versicolor

37–80

5.51

Plackett– Burman

60.98

–

Şahan et al. ()

Aspergillus niger

10–100

Langmuir, Freundlich

2–5

23.62

30%

Mukhopadhyay et al. ()

Penicillium citrinum

10–90

Langmuir, Freundlich

1.5

76.2%

Verma et al. ()

Page 212

6.20, 1.52

–

407

Renu et al.

Journal of Water Reuse and Desalination

Heavy metal removal from wastewater using various adsorbents

07.4

2017

initial metal ion concentration, the adsorbent dose, pH, con-

Industrial algal waste has also been used for copper removal

tact time and the amount of salts added and the process best

with a maximum adsorption of 16.7 mg/g at pH 5.3 (Vilar

ﬁts the Langmuir and Freundlich adsorption models (Han

et al. ). Under hydrated and dehydrated conditions,

et al. ). Table 21 summarises the parameters for the

micro algae Spirulina platensis has also been reported to

sequestration of copper using yeast.

remove up to 90% of copper from aqueous solution (Solisio et al. ). The dried biomass of Spirogyra neglecta has a reported maximum adsorption capacity for copper of

Algal biomass

115.5 mg/g at pH 4.5–5 (Singh et al. ). Table 22 sumAlgae may be used for the removal of copper because of

marises the removal parameters for the sequestering of

their high capacity, low cost, renewability and ready abun-

copper using algal biomass as an adsorbent.

dance (Chen ). There are different types of marine algae, such as red algae, green algae and brown algae, that

Microbial (bacteria)

are used for copper removal from wastewater, and the main difference in these algae is in their respective cell

Bacteria and cyanobacteria remove heavy metal because

walls where biosorption occurs (Romera et al. ). The

the cell wall has the ability to capture the heavy metals

cell walls of brown algae contain cellulose (as a structural

due to negatively charged groups within its fabric (Uslu

support), alginic acid and polymers of mannuronic and

& Tanyol ). There are several processes to remove

guluronic acids complexed with metals such as sodium,

heavy metals, such as transport across the cell membrane,

magnesium, potassium, calcium and other polysaccharides

biosorption to cell walls, entrapment in extra cellular cap-

(Romera et al. ). Green algae mainly have cellulose in

sules, precipitation, complexation and oxidation/reduction

the cell wall with a high content of bonded proteins. There-

(Rai et al. ; Brady et al. ; Veglio et al. ). Bac-

fore, this material contains various functional groups such

teria

as carboxyl, amino, sulfate and hydroxyl. Red algae contain

microorganisms (Mann ) and bacteria species such as

are

the

most

abundant

and

versatile

cellulose in the cell wall, but their biosorption capacity is

Bacillus sp., Micrococcus luteus, Pseudomonas cepacia,

attributed mainly to the presence of sulfated polysaccharides

Bacillus subtilis and Streptomyces coelicolor have been

called galactans (Romera et al. ). Brown algae, Turbi-

used for copper removal from wastewater (Nakajima

naria ornate, and green algae, Ulothrix zonata, have shown

; Öztürk et al. ; Hassan et al. ). Veneu et al.

and

() used Streptomyces lumalinharesii for copper removal

147.06 mg/g from wastewater at pH 6 and pH 4.5, respect-

from wastewater and a removal of 81% was reported at an

ively (Nuhoglu et al. ; Vijayaraghavan & Prabu ).

optimum pH of 5 with best ﬁt to the Freundlich model.

maximum

Table 21

copper

removal

176.20 mg/g

Copper removal using yeast biomass as an adsorbent Initial metal concentration

Contact

Adsorbent

Adsorption capacity

Removal

Adsorbent

(mg/L)

Best model ﬁt

time (min)

dose (g/L)

(mg/g)

per cent

References

Caustic-treated Succharomyces cerevisiae yeast biomass

16–18

Freundlich, Langmuir

2,160

2.0

9.01

–

Lu & Wilkins ()

Saccharomyces cerevisiae biomass

25–200

3–4

Freundlich, Langmuir, Redlich– Peterson

–

2.59

43.08%

Cojocaru et al. ()

Baker’s yeast

100

2.7–6

Langmuir

250

–

Yu et al. ()

Page 213

408

Table 22

Renu et al.

Journal of Water Reuse and Desalination

Heavy metal removal from wastewater using various adsorbents

07.4

2017

Copper removal using algal biomass as an adsorbent Intial metal concentration

Contact time

Adsorbent

Adsorption capacity

Removal per cent

Adsorbent

(mg/L)

Best model ﬁt

(min)

dose (g/L)

(mg/g)

(%)

References

Sargassum sp., Padina sp., Ulva sp. and Gracillaria sp.

Langmuir

62.91, 72.44

90%

Sheng et al. ()

Padina sp.

127

Langmuir

50.87

90%

Kaewsarn ()

Sargassum

4–5

Equilibrium

–

1.2

2.3 meq/g

–

Kratochvil & Volesky ()

Macroalga, Sargassum muticum

15–190

4.5

Modiﬁed competitive Langmuir sorption

240

75%

Herrero et al. ()

Gelidium

317

5.3

Freundlich

1–20

31.137

97%

Vilar et al. ()

Cystoseira crinitophylla biomass

25, 40, 50

4.5

Langmuir, Freundlich

720

2.5

160

100%

Christoforidis et al. ()

Sargassum, Chlorococcum and GAC

1–100

4.5

Langmuir, Freundlich

60, 90, 300

0.1

71.4, 19.3, 11.4

87.3%

Jacinto et al. ()

Codium vermilara

10–150

Langmuir

120

0.5

16.521

–

Romera et al. ()

Spirogyra insignis

10–150

Langmuir

120

19.063

–

Romera et al. ()

Spirulina platensis

100–400

–

Langmuir, Freundlich

–

1–4

92.6–96.8

91%

Solisio et al. ()

Dried micro-algal/ bacterial biomass

10–1,000

Langmuir

120

0.4

18–31

80–100%

Loutseti et al. ()

Öztürk et al. () used S. coelicolor for copper removal

structure with nitrogen and oxygen as ligand atoms and

and reported 21.8% removal at an optimum pH of 5 with

most copper in bacterial cells is combined with amino

a good ﬁt to the Langmuir model. Uslu & Tanyol ()

acid residues present in cell surface protein. Table 23 sum-

used P. putida for copper removal as a single component

marises the removal parameters for the sequestering of

(in the presence of copper only) or as binary component

copper using bacteria as an adsorbent.

(in the presence of copper along with other heavy metal, i.e., lead here) and reported an endothermic and spontaneous

process

with

50%

copper

removal

from

wastewater. Lu et al. () used Enterobacter sp. J1 for

FACTORS AFFECTING ADOPTION OF HEAVY METALS

copper removal and an adsorption capacity of 32.5 mg/g and a removal of 90% of copper removal was reported at

There are many factors which affect heavy metal removal

pH >2. Even after four repeated adsorption and desorption

efﬁciency of adsorbents from wastewater. These factors are

cycles, the Enterobacter sp. J1 biomass achieved 79%

initial concentration, temperature, adsorbent dose, pH, con-

removal. Nakajima () studied removal of copper

tact time and stirring speed. Heavy metal removal per cent

using Arthrobacter nicotianae bacteria from wastewater

increases with increase in initial concentration, tempera-

by electron spin resonance method, and found that

ture, adsorbent dose, contact time and stirring speed (Sahu

copper ions present in bacterial cells are of octahedral

et al. ).

Page 214

409

Table 23

Renu et al.

Journal of Water Reuse and Desalination

Heavy metal removal from wastewater using various adsorbents

07.4

2017

Copper removal using bacteria as an adsorbent Initial metal concentration

Adsorbent

(mg/L)

Best model ﬁt

Contact time

Adsorbent

Adsorption capacity

Removal per cent

(min)

dose (g/L)

(mg/g)

(%)

References

Paenibacillus polymyxa

25–500

Langmuir

120

–

112, 1,602

–

Acosta et al. ()

Escherichia coli

32–64

–

8.846, 10.364

–

Ravikumar et al. ()

Pseudomonas stutzeri

30–100

Langmuir, Freundlich

36.2

–

Hassan et al. ()

Pseudomonas putida

0.1

Langmuir

6.6

80%

Pardo et al. ()

Sphaerotilus natans

100

Langmuir

150

–

Beolchini et al. ()

Sphaerotilus natans (Gram-negative bacteria)

–

Langmuir

44.48

–

Pagnanelli et al. ()

Bacillus sp. (bacterial strain isolated from soil)

100

Langmuir

16.25. 1.64

–

Tunali et al. ()

Lactobacillus sp. (DSM 20057)

0.398–39.8

3–6

Langmuir

5–1,440

0.3–10

0.046

–

Schut et al. ()

FUTURE PERSPECTIVE AND CHALLENGES IN REMOVAL OF HEAVY METALS

copper from wastewater. A wide range of adsorbents has been studied for removal of heavy metals from wastewater. A few adsorbents that stand out for their maximum adsorp-

In this review paper, the bioadsorbents used for removal of

tion capacities are: graphene sand composite (2,859.38 mg/g),

chromium, cadmium and copper are low cost adsorbents

composite of carbon nanotubes and activated alumina

and are beneﬁcial replacements for commercially available

(264.5 mg/g), PEI functionalised eggshell (160 mg/g) for

adsorbents. In some studies, removal efﬁciency of adsor-

chromium, chitosan/TiO2 composite (256.41 mg/g), chito-

bents for heavy metal removal from wastewater has been

san-coated ceramic alumina (108.7 mg/g), α-ketoglutaric

reported to increase after modiﬁcation. However, less

acid-modiﬁed magnetic chitosan (201.2 mg/g), electrospun

work has been carried out in this direction. Hence, our

nanoﬁbre membrane of PEO/chitosan (248.1 mg/g), NaX

future perspectives are to increase removal efﬁciency of

nanozeolite (838.7 mg/g), green coconut shell powder

bioadsorbents after modiﬁcation (at minimum requirements

(285.7 mg/g), succinic anhydride modiﬁed olive stones

of acid, bases and heat), regeneration of adsorbents, recov-

(200 mg/g) for cadmium, green coconut shell powder

ery of metal ions and application of bioadsorbents at

(285.7 mg/g), Paenibacillus polymyxa bacteria (1,602 mg/g)

commercial level. The challenge in heavy metal removal

for copper. Further, optimum values of parameters such as

from wastewater is that it may require large amounts of

pH, contact time and adsorbent dose were also compared

bioadsorbents and extra chemicals to maintain a pH that

for chromium, cadmium and copper removal from waste-

provides suitable conditions for adsorption.

water. It was found that the optimum value of pH is in the range of 1–2 for chromium, 4–7 for cadmium and 4.5–6 for copper. Similarly, the optimum value of contact

CONCLUSIONS

time for maximum removal is in the range of 120–9,900 minutes for chromium, 5–120 minutes for cadmium and 120

This review shows the potential of commercial and agricul-

minutes–12 hours for copper. However, the optimum

tural adsorbents for the removal of chromium, cadmium and

value of adsorbent dose is in the range of 0.75–10 g/L for Page 215

410

Renu et al.

Heavy metal removal from wastewater using various adsorbents

chromium, 0.01–4.5 g/L for cadmium and 0.25–1 g/L for copper. Overall, the adsorption data have been found to fit the Langmuir and Freundlich models, which indicates single and multilayer adsorption behaviour. Further, the cost of both commercial adsorbents and bioadsorbents was compared. The cost of commercial activated carbon is Rs. 500/kg; however, the cost of bioadsorbents is in the range of Rs. 4.4–36.89/kg, which is much less compared to the commercial adsorbents (Gupta & Babu ). Bioadsorbents have the benefits of being cheap, easily available, no sludge generation, can be regenerated, possess technical feasibility, engineering applicability and affinity for heavy metal removal.

ACKNOWLEDGEMENTS The authors wish to thank the Department of Chemical Engineering and Materials Research Centre, MNIT Jaipur for the ﬁnancial support to carry out my PhD research work. The authors declare that there are no conﬂicts of interest regarding the publication of this paper.

REFERENCES Acharya, J., Sahu, J. N., Sahoo, B. K., Mohanty, C. R. & Meikap, B. C.  Removal of chromium (VI) from wastewater by activated carbon developed from Tamarind wood activated with zinc chloride. Chemical Engineering Journal 150 (1), 25–39. Acosta, M. P., Valdman, E., Leite, S. G. F., Battaglini, F. & Ruzal, S. M.  Biosorption of copper by Paenibacillus polymyxa cells and their exopolysaccharide. World Journal of Microbiology and Biotechnology 21 (6–7), 1157–1163. Adam, F., Kandasamy, K. & Balakrishnan, S.  Iron incorporated heterogeneous catalyst from rice husk ash. Journal of Colloid and Interface Science 304 (1), 137–143. Agraﬁoti, E., Kalderis, D. & Diamadopoulos, E.  Ca and Fe modiﬁed biochars as adsorbents of arsenic and chromium in aqueous solutions. Journal of Environmental Management 146, 444–450. Aguado, J., Arsuaga, J. M., Arencibia, A., Lindo, M. & Gascón, V.  Aqueous heavy metals removal by adsorption on aminefunctionalized mesoporous silica. Journal of Hazardous Materials 163 (1), 213–221. Ahmad, W. A., Ahmad, W. H. W., Karim, N. A., Raj, A. S. & Zakaria, Z. A.  Cr (VI) reduction in naturally rich growth medium and sugarcane bagasse by Acinetobacter

Page 216

Journal of Water Reuse and Desalination

07.4

2017

haemolyticus. International Biodeterioration & Biodegradation 85, 571–576. Ai, Z., Cheng, Y., Zhang, L. & Qiu, J.  Efficient removal of Cr (VI) from aqueous solution with Fe@ Fe2O3 core � shell nanowires. Environmental Science & Technology 42 (18), 6955–6960. Ajmal, M., Rao, R. A. K., Anwar, S., Ahmad, J. & Ahmad, R.  Adsorption studies on rice husk: removal and recovery of Cd(II) from wastewater. Bioresource Technology 86 (2), 147–149. Akar, S. T., Akar, T., Kaynak, Z., Anilan, B., Cabuk, A., Tabak, Ö., Demir, T. A. & Gedikbey, T.  Removal of copper(II) ions from synthetic solution and real wastewater by the combined action of dried Trametes versicolor cells and montmorillonite. Hydrometallurgy 97 (1), 98–104. Akhtar, M., Iqbal, S., Kausar, A., Bhanger, M. I. & Shaheen, M. A.  An economically viable method for the removal of selected divalent metal ions from aqueous solutions using activated rice husk. Colloids and Surfaces B: Biointerfaces 75 (1), 149–155. Al-Anber, Z. A. & Matouq, M. A. D.  Batch adsorption of cadmium ions from aqueous solution by means of olive cake. Journal of Hazardous Materials 151 (1), 194–201. Aliabadi, M., Irani, M., Ismaeili, J., Piri, H. & Parnian, M. J.  Electrospun nanofiber membrane of PEO/Chitosan for the adsorption of nickel, cadmium, lead and copper ions from aqueous solution. Chemical Engineering Journal 220, 237–243. Al-Khaldi, F. A., Abusharkh, B., Khaled, M., Atieh, M. A., Nasser, M. S., Saleh, T. A., Agarwal, S., Tyagi, I. & Gupta, V. K.  Adsorptive removal of cadmium(II) ions from liquid phase using acid modified carbon-based adsorbents. Journal of Molecular Liquids 204, 255–263. Alothman, Z. A. & Apblett, A. W.  Metal ion adsorption using polyamine-functionalized mesoporous materials prepared from bromopropyl-functionalized mesoporous silica. Journal of Hazardous Materials 182 (1), 581–590. Al-Othman, Z. A., Ali, R. & Naushad, M.  Hexavalent chromium removal from aqueous medium by activated carbon prepared from peanut shell: adsorption kinetics, equilibrium and thermodynamic studies. Chemical Engineering Journal 184, 238–247. Alslaibi, T. M., Abustan, I., Ahmad, M. A. & Foul, A. A.  Cadmium removal from aqueous solution using microwaved olive stone activated carbon. Journal of Environmental Chemical Engineering 1 (3), 589–599. Anirudhan, T. S. & Sreekumari, S. S.  Adsorptive removal of heavy metal ions from industrial effluents using activated carbon derived from waste coconut buttons. Journal of Environmental Sciences 23 (12), 1989–1998. Ansari, M., Aroujalian, A., Raisi, A., Dabir, B. & Fathizadeh, M.  Preparation and characterization of nano-NaX zeolite by microwave assisted hydrothermal method. Advanced Powder Technology 25 (2), 722–727. Argun, M. E., Dursun, S., Ozdemir, C. & Karatas, M.  Heavy metal adsorption by modified oak sawdust: thermodynamics and kinetics. Journal of Hazardous Materials 141 (1), 77–85.

411

Renu et al.

Heavy metal removal from wastewater using various adsorbents

Arulkumar, M., Thirumalai, K., Sathishkumar, P. & Palvannan, T.  Rapid removal of chromium from aqueous solution using novel prawn shell activated carbon. Chemical Engineering Journal 185, 178–186. Aşçı, Y., Nurbaş, M. & Açıkel, Y. S.  Sorption of Cd(II) onto kaolin as a soil component and desorption of Cd(II) from kaolin using rhamnolipid biosurfactant. Journal of Hazardous Materials 139 (1), 50–56. Atieh, M. A.  Removal of chromium (VI) from polluted water using carbon nanotubes supported with activated carbon. Procedia Environmental Sciences 4, 281–293. Atieh, M. A., Bakather, O. Y., Tawabini, B. S., Bukhari, A. A., Khaled, M., Alharthi, M., Fettouhi, M. & Abuilaiwi, F. A.  Removal of chromium (III) from water by using modified and nonmodified carbon nanotubes. Journal of Nanomaterials 2010, 17. Aziz, A., Elandaloussi, E. H., Belhalfaoui, B., Ouali, M. S. & De Ménorval, L. C. a Efficiency of succinylated-olive stone biosorbent on the removal of cadmium ions from aqueous solutions. Colloids and Surfaces B: Biointerfaces 73 (2), 192–198. Aziz, A., Ouali, M. S., Elandaloussi, E. H., De Menorval, L. C. & Lindheimer, M. b Chemically modified olive stone: a low-cost sorbent for heavy metals and basic dyes removal from aqueous solutions. Journal of Hazardous Materials 163 (1), 441–447. Azouaou, N., Sadaoui, Z., Djaafri, A. & Mokaddem, H.  Adsorption of cadmium from aqueous solution onto untreated coffee grounds: equilibrium, kinetics and thermodynamics. Journal of Hazardous Materials 184 (1), 126–134. Babel, S. & Kurniawan, T. A.  Cr (VI) removal from synthetic wastewater using coconut shell charcoal and commercial activated carbon modified with oxidizing agents and/or chitosan. Chemosphere 54 (7), 951–967. Bansal, R. P., Donnet, J. P. & Stoeckli, F.  Active Carbon. Marcel Dekker Inc., New York, USA. Barnhart, J.  Occurrences, uses, and properties of chromium. Regulatory Toxicology and Pharmacology 26 (1), S3–S7. Beolchini, F., Pagnanelli, F., Toro, L. & Veglio, F.  Ionic strength effect on copper biosorption by Sphaerotilus natans: equilibrium study and dynamic modelling in membrane reactor. Water Research 40 (1), 144–152. Bertagnolli, C., Kleinübing, S. J. & Da Silva, M. G. C.  Preparation and characterization of a Brazilian bentonite clay for removal of copper in porous beds. Applied Clay Science 53 (1), 73–79. Bhainsa, K. C. & D’Souza, S. F.  Removal of copper ions by the filamentous fungus, Rhizopus oryzae from aqueous solution. Bioresource Technology 99 (9), 3829–3835. Bhattacharyya, S., Lelong, G. & Saboungi, M. L.  Recent progress in the synthesis and selected applications of MCM41: a short review. Journal of Experimental Nanoscience 1 (3), 375–395. Bilal, M., Shah, J. A., Ashfaq, T., Gardazi, S. M. H., Tahir, A. A., Pervez, A., Haroon, H. & Mahmood, Q.  Waste biomass

Journal of Water Reuse and Desalination

07.4

2017

adsorbents for copper removal from industrial wastewater – A review. Journal of Hazardous Materials 263, 322–333. Bingol, A., Ucun, H., Bayhan, Y. K., Karagunduz, A., Cakici, A. & Keskinler, B.  Removal of chromate anions from aqueous stream by a cationic surfactant-modified yeast. Bioresource Technology 94 (3), 245–249. Bishnoi, N. R., Bajaj, M., Sharma, N. & Gupta, A.  Adsorption of Cr (VI) on activated rice husk carbon and activated alumina. Bioresource Technology 91 (3), 305–307. Blázquez, G., Hernáinz, F., Calero, M. & Ruiz-Nunez, L. F.  Removal of cadmium ions with olive stones: the effect of some parameters. Process Biochemistry 40 (8), 2649–2654. Bohli, T., Ouederni, A., Fiol, N. & Villaescusa, I.  Evaluation of an activated carbon from olive stones used as an adsorbent for heavy metal removal from aqueous phases. Comptes Rendus Chimie 18 (1), 88–99. Boonamnuayvitaya, V., Chaiya, C., Tanthapanichakoon, W. & Jarudilokkul, S.  Removal of heavy metals by adsorbent prepared from pyrolyzed coffee residues and clay. Separation and Purification Technology 35 (1), 11–22. Borba, C. E., Guirardello, R., Silva, E. A., Veit, M. T. & Tavares, C. R. G.  Removal of nickel (II) ions from aqueous solution by biosorption in a fixed bed column: experimental and theoretical breakthrough curves. Biochemical Engineering Journal 30 (2), 184–191. Borsagli, F. G. M., Mansur, A. A., Chagas, P., Oliveira, L. C. & Mansur, H. S.  O-carboxymethyl functionalization of chitosan: complexation and adsorption of Cd (II) and Cr (VI) as heavy metal pollutant ions. Reactive and Functional Polymers 97, 37–47. Brady, D., Stoll, A. D., Starke, L. & Duncan, J. R.  Chemical and enzymatic extraction of heavy metal binding polymers from isolated cell walls of Saccharomyces cerevisiae. Biotechnology and Bioengineering 44 (3), 297–302. Burke, A. M., Hanrahan, J. P., Healy, D. A., Sodeau, J. R., Holmes, J. D. & Morris, M. A.  Large pore bi-functionalised mesoporous silica for metal ion pollution treatment. Journal of Hazardous Materials 164 (1), 229–234. Chadha, K. L.  Coconut research in India – a review. Indian Coconut Journal 36 (4), 13–19. Chand, P., Shil, A. K., Sharma, M. & Pakade, Y. B.  Improved adsorption of cadmium ions from aqueous solution using chemically modified apple pomace: mechanism, kinetics, and thermodynamics. International Biodeterioration & Biodegradation 90, 8–16. Chand, P., Bafana, A. & Pakade, Y. B.  Xanthate modified apple pomace as an adsorbent for removal of Cd (II), Ni (II) and Pb (II), and its application to real industrial wastewater. International Biodeterioration & Biodegradation 97, 60–66. Chen, J. P.  Decontamination of Heavy Metals: Processes, Mechanisms, and Applications. CRC Press, Boca Raton, FL, USA. Chen, G. C., Shan, X. Q., Zhou, Y. Q., Shen, X. E., Huang, H. L. & Khan, S. U.  Adsorption kinetics, isotherms and thermodynamics of atrazine on surface oxidized multiwalled

Page 217

412

Renu et al.

Heavy metal removal from wastewater using various adsorbents

carbon nanotubes. Journal of Hazardous Materials 169 (1), 912–918. Chen, A., Zeng, G., Chen, G., Hu, X., Yan, M., Guan, S., Chang, C., Lu, L., Zou, Z. & Xie, G.  Novel thiourea-modified magnetic ion-imprinted chitosan/TiO2 composite for simultaneous removal of cadmium and 2, 4-dichlorophenol. Chemical Engineering Journal 191, 85–94. Choi, H. J., Yu, S. W. & Kim, K. H.  Efficient use of Mgmodified zeolite in the treatment of aqueous solution contaminated with heavy metal toxic ions. Journal of the Taiwan Institute of Chemical Engineers 63, 482–489. Christoforidis, A. K., Orfanidis, S., Papageorgiou, S. K., Lazaridou, A. N., Favvas, E. P. & Mitropoulos, A. C.  Study of Cu(II) removal by Cystoseira crinitophylla biomass in batch and continuous flow biosorption. Chemical Engineering Journal 277, 334–340. Chuah, T. G., Jumasiah, A., Azni, I., Katayon, S. & Choong, S. T.  Rice husk as a potentially low-cost biosorbent for heavy metal and dye removal: an overview. Desalination 175 (3), 305–316. Clement, R. E., Eiceman, G. A. & Koester, C. J.  Environmental analysis. Analytical Chemistry 67 (12), 221R–255R. Cojocaru, C., Diaconu, M., Cretescu, I., Savić, J. & Vasić, V.  Biosorption of copper (II) ions from aqua solutions using dried yeast biomass. Colloids and Surfaces A: Physicochemical and Engineering Aspects 335 (1), 181–188. Cronje, K. J., Chetty, K., Carsky, M., Sahu, J. N. & Meikap, B. C.  Optimization of chromium (VI) sorption potential using developed activated carbon from sugarcane bagasse with chemical activation by zinc chloride. Desalination 275 (1), 276–284. Dan, T. K.  Development of light weight building bricks using coconut pith. Research and Industry 37 (1), 11–17. Daraei, H., Mittal, A., Noorisepehr, M. & Mittal, J.  Separation of chromium from water samples using eggshell powder as a low-cost sorbent: kinetic and thermodynamic studies. Desalination and Water Treatment 53 (1), 214–220. de Almeida Neto, A. F., Vieira, M. G. A. & da Silva, M. G. C.  Adsorption and desorption processes for copper removal from water using different eluents and calcinated clay as adsorbent. Journal of Water Process Engineering 3, 90–97. Dubey, R., Bajpai, J. & Bajpai, A. K.  Green synthesis of graphene sand composite (GSC) as novel adsorbent for efficient removal of Cr (VI) ions from aqueous solution. Journal of Water Process Engineering 5, 83–94. Duranoğlu, D., Trochimczuk, A. W. & Beker, Ü.  A comparison study of peach stone and acrylonitrile-divinylbenzene copolymer based activated carbons as chromium (VI) sorbents. Chemical Engineering Journal 165 (1), 56–63. Dursun, A. Y., Uslu, G., Cuci, Y. & Aksu, Z.  Bioaccumulation of copper(II), lead(II) and chromium(VI) by growing Aspergillus niger. Process Biochemistry 38 (12), 1647–1651. El-Kamash, A. M., Zaki, A. A. & El Geleel, M. A.  Modeling batch kinetics and thermodynamics of zinc and cadmium ions removal from waste solutions using synthetic zeolite A. Journal of Hazardous Materials 127 (1), 211–220.

Page 218

Journal of Water Reuse and Desalination

07.4

2017

El Nemr, A., El-Sikaily, A., Khaled, A. & Abdelwahab, O.  Removal of toxic chromium from aqueous solution, wastewater and saline water by marine red alga Pterocladia capillacea and its activated carbon. Arabian Journal of Chemistry 8 (1), 105–117. El-Shafey, E. I.  Sorption of Cd (II) and Se (IV) from aqueous solution using modified rice husk. Journal of Hazardous Materials 147 (1), 546–555. El-Sikaily, A., El Nemr, A., Khaled, A. & Abdelwehab, O.  Removal of toxic chromium from wastewater using green alga Ulva lactuca and its activated carbon. Journal of Hazardous Materials 148 (1), 216–228. Erdem, E., Karapinar, N. & Donat, R.  The removal of heavy metal cations by natural zeolites. Journal of Colloid and Interface Science 280 (2), 309–314. Ertugay, N. & Bayhan, Y. K.  The removal of copper (II) ion by using mushroom biomass (Agaricus bisporus) and kinetic modelling. Desalination 255 (1), 137–142. Farajzadeh, M. A. & Monji, A. B.  Adsorption characteristics of wheat bran towards heavy metal cations. Separation and Purification Technology 38 (3), 197–207. Feng, N., Guo, X., Liang, S., Zhu, Y. & Liu, J.  Biosorption of heavy metals from aqueous solutions by chemically modified orange peel. Journal of Hazardous Materials 185 (1), 49–54. Fiol, N., Villaescusa, I., Martínez, M., Miralles, N., Poch, J. & Serarols, J.  Sorption of Pb (II), Ni (II), Cu (II) and Cd (II) from aqueous solution by olive stone waste. Separation and Purification Technology 50 (1), 132–140. Foo, L. Y. & Lu, Y.  Isolation and identification of procyanidins in apple pomace. Food Chemistry 64 (4), 511–518. Fouladgar, M., Beheshti, M. & Sabzyan, H.  Single and binary adsorption of nickel and copper from aqueous solutions by γ-alumina nanoparticles: equilibrium and kinetic modeling. Journal of Molecular Liquids 211, 1060–1073. Friberg, L., Elinder, C. G. & Kjellstrom, T.  Environmental Health Criteria 134: Cadmium. World Health Organization, Geneva, Switzerland. Fu, F. & Wang, Q.  Removal of heavy metal ions from wastewaters: a review. Journal of Environmental Management 92 (3), 407–418. Futalan, C. M., Kan, C. C., Dalida, M. L., Pascua, C. & Wan, M. W.  Fixed-bed column studies on the removal of copper using chitosan immobilized on bentonite. Carbohydrate Polymers 83 (2), 697–704. Gao, Z., Wang, L., Qi, T., Chu, J. & Zhang, Y.  Synthesis, characterization, and cadmium (II) uptake of iminodiacetic acid-modified mesoporous SBA-15. Colloids and Surfaces A: Physicochemical and Engineering Aspects 304 (1), 77–81. Garg, V. K., Gupta, R., Kumar, R. & Gupta, R. K.  Adsorption of chromium from aqueous solution on treated sawdust. Bioresource Technology 92 (1), 79–81. Garg, U. K., Kaur, M. P., Sud, D. & Garg, V. K.  Removal of hexavalent chromium from aqueous solution by adsorption on treated sugarcane bagasse using response surface methodological approach. Desalination 249 (2), 475–479.

413

Renu et al.

Heavy metal removal from wastewater using various adsorbents

Ghaemi, N.  A new approach to copper ion removal from water by polymeric nanocomposite membrane embedded with γ-alumina nanoparticles. Applied Surface Science 364, 221–228. Gong, J. L., Wang, X. Y., Zeng, G. M., Chen, L., Deng, J. H., Zhang, X. R. & Niu, Q. Y.  Copper (II) removal by pectin–iron oxide magnetic nanocomposite adsorbent. Chemical Engineering Journal 185, 100–107. Gopalakrishnan, A., Krishnan, R., Thangavel, S., Venugopal, G. & Kim, S. J.  Removal of heavy metal ions from pharmaeffluents using graphene-oxide nanosorbents and study of their adsorption kinetics. Journal of Industrial and Engineering Chemistry 30, 14–19. Gueye, M., Richardson, Y., Kafack, F. T. & Blin, J.  High efficiency activated carbons from African biomass residues for the removal of chromium (VI) from wastewater. Journal of Environmental Chemical Engineering 2 (1), 273–281. Guo, X., Du, B., Wei, Q., Yang, J., Hu, L., Yan, L. & Xu, W.  Synthesis of amino functionalized magnetic graphenes composite material and its application to remove Cr (VI), Pb (II), Hg (II), Cd (II) and Ni (II) from contaminated water. Journal of Hazardous Materials 278, 211–220. Gupta, S. & Babu, B. V.  Economic feasibility analysis of low cost adsorbents for the removal of Cr (VI) from waste water. In: Proceedings of International Convention on Water Resources Development and Management (ICWRDM), BITS Pilani. Gupta, V. K. & Sharma, S.  Removal of cadmium and zinc from aqueous solutions using red mud. Environmental Science & Technology 36 (16), 3612–3617. Gupta, V. K., Agarwal, S. & Saleh, T. A.  Chromium removal by combining the magnetic properties of iron oxide with adsorption properties of carbon nanotubes. Water Research 45 (6), 2207–2212. Guru, P. S. & Dash, S.  Amino acid modified eggshell powder (AA-ESP) — A novel bio-solid scaffold for adsorption of some Styrylpyridinium dyes. Journal of Dispersion Science and Technology 34 (8), 1099–1112. Hajiaghababaei, L., Badiei, A., Ganjali, M. R., Heydari, S., Khaniani, Y. & Ziarani, G. M.  Highly efficient removal and preconcentration of lead and cadmium cations from water and wastewater samples using ethylenediamine functionalized SBA-15. Desalination 266 (1), 182–187. Hamadi, N. K., Chen, X. D., Farid, M. M. & Lu, M. G.  Adsorption kinetics for the removal of chromium (VI) from aqueous solution by adsorbents derived from used tyres and sawdust. Chemical Engineering Journal 84 (2), 95–105. Hamdaoui, O.  Removal of cadmium from aqueous medium under ultrasound assistance using olive leaves as sorbent. Chemical Engineering and Processing: Process Intensification 48 (6), 1157–1166. Hamidpour, M., Afyuni, M., Kalbasi, M., Khoshgoftarmanes, A. H. & Inglezakis, V. J.  Mobility and plant-availability of Cd (II) and Pb (II) adsorbed on zeolite and bentonite. Applied Clay Science 48 (3), 342–348.

Journal of Water Reuse and Desalination

07.4

2017

Han, R., Li, H., Li, Y., Zhang, J., Xiao, H. & Shi, J.  Biosorption of copper and lead ions by waste beer yeast. Journal of Hazardous Materials 137 (3), 1569–1576. Hassan, S. H., Kim, S. J., Jung, A. Y., Joo, J. H., Eun Oh, S. & Yang, J. E.  Biosorptive capacity of Cd(II) and Cu(II) by lyophilized cells of Pseudomonas stutzeri. Journal of General and Applied Microbiology 55 (1), 27–34. Heidari, A., Habibollah, Y. & Zahra, M.  Removal of Ni (II), Cd (II), and Pb (II) from a ternary aqueous solution by amino functionalized mesoporous and nano mesoporous silica. Chemical Engineering Journal 153 (1), 70–79. Hemambika, B., Rani, M. J. & Kannan, V. R.  Biosorption of heavy metals by immobilized and dead fungal cells: a comparative assessment. Journal of Ecology and the Natural Environment 3 (5), 168–175. Herrero, R., Lodeiro, P., García-Casal, L. J., Vilariño, T., ReyCastro, C., David, C. & Rodríguez, P.  Full description of copper uptake by algal biomass combining an equilibrium NICA model with a kinetic intraparticle diffusion driving force approach. Bioresource Technology 102 (3), 2990–2997. Hu, J., Chen, C., Zhu, X. & Wang, X.  Removal of chromium from aqueous solution by using oxidized multiwalled carbon nanotubes. Journal of Hazardous Materials 162 (2), 1542– 1550. Hu, X. J., Wang, J. S., Liu, Y. G., Li, X., Zeng, G. M., Bao, Z. L., Zeng, X. X., Chen, A. W. & Long, F.  Adsorption of chromium (VI) by ethylenediamine-modified cross-linked magnetic chitosan resin: isotherms, kinetics and thermodynamics. Journal of Hazardous Materials 185 (1), 306–314. Hu, X. J., Liu, Y. G., Wang, H., Chen, A. W., Zeng, G. M., Liu, S. M., Guo, Y. M., Hu, X., Li, T. T., Wang, Y. Q., Zhou, L. & Liu, S. H.  Removal of Cu(II) ions from aqueous solution using sulfonated magnetic graphene oxide composite. Separation and Purification Technology 108, 189–195. Hui, K. S., Chao, C. Y. H. & Kot, S. C.  Removal of mixed heavy metal ions in wastewater by zeolite 4A and residual products from recycled coal Fly ash. Journal of Hazardous Materials 127 (1), 89–101. Hummers, W. S. & Offeman, R. E.  Preparation of graphitic oxide. Journal of the American Chemical Society 80, 1339– 1339. Huo, H., Su, H. & Tan, T.  Adsorption of Agþ by a surface molecular-imprinted biosorbent. Chemical Engineering Journal 150 (1), 139–144. Hydari, S., Sharififard, H., Nabavinia, M. & Reza Parvizi, M.  A comparative investigation on removal performances of commercial activated carbon, chitosan biosorbent and chitosan/activated carbon composite for cadmium. Chemical Engineering Journal 193, 276–282. Ibrahim, H. S., Jamil, T. S. & Hegazy, E. Z.  Application of zeolite prepared from Egyptian kaolin for the removal of heavy metals: II. Isotherm models. Journal of Hazardous Materials 182 (1), 842–847. Igberase, E. & Osifo, P.  Equilibrium, kinetic, thermodynamic and desorption studies of cadmium and lead by polyaniline

Page 219

414

Renu et al.

Heavy metal removal from wastewater using various adsorbents

grafted cross-linked chitosan beads from aqueous solution. Journal of Industrial and Engineering Chemistry 26, 340–347. Iqbal, M. & Edyvean, R. G. J.  Biosorption of lead, copper and zinc ions on loofa sponge immobilized biomass of Phanerochaete chrysosporium. Minerals Engineering 17 (2), 217–223. Izidoro, J. D. C., Fungaro, D. A., Abbott, J. E. & Wang, S.  Synthesis of zeolites X and A from fly ashes for cadmium and zinc removal from aqueous solutions in single and binary ion systems. Fuel 103, 827–834. Jacinto, M. L. J., David, C. P. C., Perez, T. R. & De Jesus, B. R.  Comparative efficiency of algal biofilters in the removal of chromium and copper from wastewater. Ecological Engineering 35 (5), 856–860. Javadian, H., Sorkhrodi, F. Z. & Koutenaei, B. B.  Experimental investigation on enhancing aqueous cadmium removal via nanostructure composite of modified hexagonal type mesoporous silica with polyaniline/polypyrrole nanoparticles. Journal of Industrial and Engineering Chemistry 20 (5), 3678–3688. Javadian, H., Ghorbani, F., Tayebi, H. A. & Asl, S. H.  Study of the adsorption of Cd (II) from aqueous solution using zeolitebased geopolymer, synthesized from coal fly ash; kinetic, isotherm and thermodynamic studies. Arabian Journal of Chemistry 8 (6), 837–849. Jha, V. K., Matsuda, M. & Miyake, M.  Sorption properties of the activated carbon-zeolite composite prepared from coal fly ash for Ni(2þ), Cu(2þ), Cd(2þ) and Pb(2þ). Journal of Hazardous Materials 160 (1), 148–153. Jin, G. P., Zhu, X. H., Li, C. Y., Fu, Y., Guan, J. X. & Wu, X. P.  Tetraoxalyl ethylenediamine melamine resin functionalized coconut active charcoal for adsorptive removal of Ni (II), Pb (II) and Cd (II) from their aqueous solution. Journal of Environmental Chemical Engineering 1 (4), 736–745. Jing, X., Cao, Y., Zhang, X., Wang, D., Wu, X. & Xu, H.  Biosorption of Cr (VI) from simulated wastewater using a cationic surfactant modified spent mushroom. Desalination 269 (1), 120–127. Ju, S. H., Lu, S. D., Peng, J. H., Zhang, L. B., Srinivasakannan, C., Guo, S. H. & Wei, L. I.  Removal of cadmium from aqueous solutions using red mud granulated with cement. Transactions of Nonferrous Metals Society of China 22 (12), 3140–3146. Kadirvelu, K. & Namasivayam, C.  Activated carbon from coconut coirpith as metal adsorbent: adsorption of Cd (II) from aqueous solution. Advances in Environmental Research 7 (2), 471–478. Kaewsarn, P.  Biosorption of copper (II) from aqueous solutions by pre-treated biomass of marine algae Padina sp. Chemosphere 47 (10), 1081–1085. Kaikake, K., Hoaki, K., Sunada, H., Dhakal, R. P. & Baba, Y.  Removal characteristics of metal ions using degreased coffee beans: adsorption equilibrium of cadmium (II). Bioresource Technology 98 (15), 2787–2791. Karthikeyan, T., Rajgopal, S. & Miranda, L. R.  Chromium (VI) adsorption from aqueous solution by Hevea Brasilinesis

Page 220

Journal of Water Reuse and Desalination

07.4

2017

sawdust activated carbon. Journal of Hazardous Materials 124 (1), 192–199. Kaya, K., Pehlivan, E., Schmidt, C. & Bahadir, M.  Use of modiﬁed wheat bran for the removal of chromium (VI) from aqueous solutions. Food Chemistry 158, 112–117. Kobya, M.  Removal of Cr (VI) from aqueous solutions by adsorption onto hazelnut shell activated carbon: kinetic and equilibrium studies. Bioresource Technology 91 (3), 317–321. Kratochvil, D. & Volesky, B.  Biosorption of Cu from ferruginous wastewater by algal biomass. Water Research 32 (9), 2760–2768. Krikorian, N. & Martin, D. F.  Extraction of selected heavy metals using modiﬁed clays. Journal of Environmental Science and Health 40 (3), 601–608. Kula, I., Uğurlu, M., Karaoğlu, H. & Celik, A.  Adsorption of Cd (II) ions from aqueous solutions using activated carbon prepared from olive stone by ZnCl2 activation. Bioresource Technology 99 (3), 492–501. Kumar, U. & Bandyopadhyay, M.  Sorption of cadmium from aqueous solution using pretreated rice husk. Bioresource Technology 97 (1), 104–109. Kundu, A., Gupta, B. S., Hashim, M. A., Sahu, J. N., Mujawar, M. & Redzwan, G.  Optimisation of the process variables in production of activated carbon by microwave heating. RSC Advances 5 (45), 35899–35908. Lan, S., Wu, X., Li, L., Li, M., Guo, F. & Gan, S.  Synthesis and characterization of hyaluronic acid-supported magnetic microspheres for copper ions removal. Colloids and Surfaces A: Physicochemical and Engineering Aspects 425, 42–50. Larraza, I., López-Gónzalez, M., Corrales, T. & Marcelo, G.  Hybrid materials: magnetite–polyethylenimine– montmorillonite, as magnetic adsorbents for Cr (VI) water treatment. Journal of Colloid and Interface Science 385 (1), 24–33. Lee, B., Kim, Y., Lee, H. & Yi, J.  Synthesis of functionalized porous silicas via templating method as heavy metal ion adsorbents: the introduction of surface hydrophilicity onto the surface of adsorbents. Microporous and Mesoporous Materials 50 (1), 77–90. Lehmann, J.  A handful of carbon. Nature 447, 143–144. Li, X., Xu, Q., Han, G., Zhu, W., Chen, Z., He, X. & Tian, X.  Equilibrium and kinetic studies of copper (II) removal by three species of dead fungal biomasses. Journal of Hazardous Materials 165 (1), 469–474. Li, L., Fan, L., Sun, M., Qiu, H., Li, X., Duan, H. & Luo, C.  Adsorbent for chromium removal based on graphene oxide functionalized with magnetic cyclodextrin–chitosan. Colloids and Surfaces B: Biointerfaces 107, 76–83. Li, X., Ai, L. & Jiang, J.  Nanoscale zerovalent iron decorated on graphene nanosheets for Cr (VI) removal from aqueous solution: surface corrosion retard induced the enhanced performance. Chemical Engineering Journal 288, 789–797. Lingamdinne, L. P., Koduru, J. R., Choi, Y. L., Chang, Y. Y. & Yang, J. K.  Studies on removal of Pb (II) and Cr (III) using graphene oxide based inverse spinel nickel ferrite nanocomposite as sorbent. Hydrometallurgy 165, 64–72.

415

Renu et al.

Heavy metal removal from wastewater using various adsorbents

Liu, B. & Huang, Y.  Polyethyleneimine modified eggshell membrane as a novel biosorbent for adsorption and detoxification of Cr (VI) from water. Journal of Materials Chemistry 21 (43), 17413–17418. Liu, H., Liang, S., Gao, J., Ngo, H. H., Guo, W., Guo, Z., Wang, J. & Li, Y.  Enhancement of Cr(VI) removal by modifying activated carbon developed from Zizania caduciflora with tartaric acid during phosphoric acid activation. Chemical Engineering Journal 246, 168–174. Lo, S. F., Wang, S. Y., Tsai, M. J. & Lin, L. D.  Adsorption capacity and removal efficiency of heavy metal ions by Moso and Ma bamboo activated carbons. Chemical Engineering Research and Design 90 (9), 1397–1406. López-Téllez, G., Barrera-Díaz, C. E., Balderas-Hernández, P., Roa-Morales, G. & Bilyeu, B.  Removal of hexavalent chromium in aquatic solutions by iron nanoparticles embedded in orange peel pith. Chemical Engineering Journal 173 (2), 480–485. Loutseti, S., Danielidis, D. B., Economou-Amilli, A., Katsaros, C., Santas, R. & Santas, P.  The application of a micro-algal/ bacterial biofilter for the detoxification of copper and cadmium metal wastes. Bioresource Technology 100 (7), 2099–2105. Lu, Y. & Foo, L. Y.  Constitution of some chemical components of apple seed. Food Chemistry 61 (1), 29–33. Lu, Y. & Wilkins, E.  Heavy metal removal by caustic-treated yeast immobilized in alginate. Journal of Hazardous Materials 49 (2), 165–179. Lu, W. B., Shi, J. J., Wang, C. H. & Chang, J. S.  Biosorption of lead, copper and cadmium by an indigenous isolate Enterobacter sp. J1 possessing high heavy-metal resistance. Journal of Hazardous Materials 134 (1), 80–86. Lugo-Lugo, V., Barrera-Díaz, C., Ureña-Núñez, F., Bilyeu, B. & Linares-Hernández, I.  Biosorption of Cr (III) and Fe (III) in single and binary systems onto pretreated orange peel. Journal of Environmental Management 112, 120–127. Luo, C., Tian, Z., Yang, B., Zhang, L. & Yan, S.  Manganese dioxide/iron oxide/acid oxidized multi-walled carbon nanotube magnetic nanocomposite for enhanced hexavalent chromium removal. Chemical Engineering Journal 234, 256– 265. Lv, X., Xue, X., Jiang, G., Wu, D., Sheng, T., Zhou, H. & Xu, X.  Nanoscale zero-valent iron (nZVI) assembled on magnetic Fe3O4/graphene for Chromium (VI) removal from aqueous solution. Journal of Colloid and Interface Science 417, 51–59. Ma, Y., Lin, C., Jiang, Y., Lu, W., Si, C. & Liu, Y.  Competitive removal of water-borne copper, zinc and cadmium by a CaCO3-dominated red mud. Journal of Hazardous Materials 172 (2), 1288–1296. Ma, H. L., Zhang, Y., Zhang, L., Wang, L., Sun, C., Liu, P., He, L., Zeng, X. & Zhai, M.  Radiation-induced graft copolymerization of dimethylaminoethyl methacrylate onto graphene oxide for Cr(VI) removal. Radiation Physics and Chemistry 124, 159–163.

Journal of Water Reuse and Desalination

07.4

2017

MacCarthy, P., Klusman, R. W., Cowling, S. W. & Rice, J. A.  Water analysis. Analytical Chemistry 65 (12), 244R–292R. Machida, M., Mochimaru, T. & Tatsumoto, H.  Lead (II) adsorption onto the graphene layer of carbonaceous materials in aqueous solution. Carbon 44 (13), 2681–2688. Machida, M., Fotoohi, B., Amamo, Y., Ohba, T., Kanoh, H. & Mercier, L.  Cadmium (II) adsorption using functional mesoporous silica and activated carbon. Journal of Hazardous Materials 221, 220–227. Mahmoud, M. E., Osman, M. M., Hafez, O. F. & Elmelegy, E.  Removal and preconcentration of lead (II), copper (II), chromium (III) and iron (III) from wastewaters by surface developed alumina adsorbents with immobilized 1-nitroso-2naphthol. Journal of Hazardous Materials 173 (1), 349–357. Majumdar, S. S., Das, S. K., Saha, T., Panda, G. C., Bandyopadhyoy, T. & Guha, A. K.  Adsorption behavior of copper ions on Mucor rouxii biomass through microscopic and FTIR analysis. Colloids and Surfaces B: Biointerfaces 63 (1), 138–145. Mann, H.  Biosorption of heavy metals by bacterial biomass. In: Biosorption of Heavy Metals (B. Volesky, ed.). CRC Press, Boca Raton, FL, USA, pp. 93–137. Mantell, C. L.  Carbon and Graphite Handbook. Interscience, New York, USA. Marín, A. P., Ortuno, J. F., Aguilar, M. I., Meseguer, V. F., Sáez, J. & Lloréns, M.  Use of chemical modification to determine the binding of Cd (II), Zn (II) and Cr (III) ions by orange waste. Biochemical Engineering Journal 53 (1), 2–6. Maznah, W. W., Al-Fawwaz, A. T. & Surif, M.  Biosorption of copper and zinc by immobilised and free algal biomass, and the effects of metal biosorption on the growth and cellular structure of Chlorella sp. and Chlamydomonas sp. isolated from rivers in Penang, Malaysia. Journal of Environmental Sciences 24 (8), 1386–1393. McHale, A. P. & McHale, S.  Microbial biosorption of metals: potential in the treatment of metal pollution. Biotechnology Advances 12 (4), 647–652. Mehdipour, S., Vatanpour, V. & Kariminia, H.-R.  Influence of ion interaction on lead removal by a polyamide nanofiltration membrane. Desalination 362, 84–92. Mehta, D., Mazumdar, S. & Singh, S. K.  Magnetic adsorbents for the treatment of water/wastewater – a review. Journal of Water Process Engineering 7, 244–265. Min, Y. U. E., Zhang, M., Bin, L. I. U., Xing, X. U., Xiaoming, L. I., Qinyan, Y. U. E. & Chunyuan, M. A.  Characteristics of amine surfactant modified peanut shell and its sorption property for Cr (VI). Chinese Journal of Chemical Engineering 21 (11), 1260–1268. Mohan, D. & Pittman, C. U.  Activated carbons and low cost adsorbents for remediation of tri- and hexavalent chromium from water. Journal of Hazardous Materials 137 (2), 762–811. Mohan, D. & Pittman, C. U.  Arsenic removal from water/ wastewater using adsorbents – a critical review. Journal of Hazardous Materials 142 (1), 1–53. Mohanty, K., Jha, M., Meikap, B. C. & Biswas, M. N.  Removal of chromium (VI) from dilute aqueous solutions by

Page 221

416

Renu et al.

Heavy metal removal from wastewater using various adsorbents

activated carbon developed from Terminalia arjuna nuts activated with zinc chloride. Chemical Engineering Science 60 (11), 3049–3059. Mubarak, N. M., Alicia, R. F., Abdullah, E. C., Sahu, J. N., Haslija, A. A. & Tan, J.  Statistical optimization and kinetic studies on removal of Zn2þ using functionalized carbon nanotubes and magnetic biochar. Journal of Environmental Chemical Engineering 1 (3), 486–495. Mubarak, N. M., Sahu, J. N., Abdullah, E. C. & Jayakumar, N. S. a Removal of heavy metals from wastewater using carbon nanotubes. Separation & Puriﬁcation Reviews 43 (4), 311–338. Mubarak, N. M., Thines, R. K., Sajuni, N. R., Abdullah, E. C., Sahu, J. N., Ganesan, P. & Jayakumar, N. S. b Adsorption of chromium (VI) on functionalized and non-functionalized carbon nanotubes. Korean Journal of Chemical Engineering 31 (9), 1582–1591. Mubarak, N. M., Kundu, A., Sahu, J. N., Abdullah, E. C. & Jayakumar, N. S. c Synthesis of palm oil empty fruit bunch magnetic pyrolytic char impregnating with FeCl3 by microwave heating technique. Biomass and Bioenergy 61, 265–275. Mubarak, N. M., Sahu, J. N., Abdullah, E. C., Jayakumar, N. S. & Ganesan, P. a Novel microwave-assisted multiwall carbon nanotubes enhancing Cu (II) adsorption capacity in water. Journal of the Taiwan Institute of Chemical Engineers 53, 140–152. Mubarak, N. M., Sahu, J. N., Abdullah, E. C., Jayakumar, N. S. & Ganesan, P. b Microwave assisted multiwall carbon nanotubes enhancing Cd (II) adsorption capacity in aqueous media. Journal of Industrial and Engineering Chemistry 24, 24–33. Mubarak, N. M., Sahu, J. N., Wong, J. R., Jayakumar, N. S., Ganesan, P. & Abdullah, E. C. c Overview on the functionalization of carbon nanotubes. In: Chemical Functionalization of Carbon Nanomaterials: Chemistry and Applications (V. K. Thakur & M. K. Thakur, eds). CRC Press, Boca Raton, FL, USA, pp. 82–101. Mubarak, N. M., Sahu, J. N., Abdullah, E. C., Jayakumar, N. S. & Ganesan, P. a Microwave-assisted synthesis of multiwalled carbon nanotubes for enhanced removal of Zn (II) from wastewater. Research on Chemical Intermediates 42 (4), 3257–3281. Mubarak, N. M., Thobashinni, M., Abdullah, E. C. & Sahu, J. N. b Comparative kinetic study of removal of Pb2þ ions and Cr3þ ions from waste water using carbon nanotubes produced using microwave heating. C 2 (1), 7. Mubarak, N. M., Sahu, J. N., Abdullah, E. C. & Jayakumar, N. S. c Rapid adsorption of toxic Pb (II) ions from aqueous solution using multiwall carbon nanotubes synthesized by microwave chemical vapor deposition technique. Journal of Environmental Sciences 45, 143–155. Mukhopadhyay, M.  Role of surface properties during biosorption of copper by pretreated Aspergillus Niger biomass. Colloids and Surfaces A: Physicochemical and Engineering Aspects 329 (1), 95–99.

Page 222

Journal of Water Reuse and Desalination

07.4

2017

Mukhopadhyay, M., Noronha, S. B. & Suraishkumar, G. K.  Kinetic modeling for the biosorption of copper by pretreated Aspergillus niger biomass. Bioresource Technology 98 (9), 1781–1787. Muthukumaran, K. & Beulah, S.  Removal of chromium (VI) from wastewater using chemically activated Syzygium jambolanum nut carbon by batch studies. Procedia Environmental Sciences 4, 266–280. Nadeem, M., Mahmood, A., Shahid, S. A., Shah, S. S., Khalid, A. M. & McKay, G.  Sorption of lead from aqueous solution by chemically modified carbon adsorbents. Journal of Hazardous Materials 138 (3), 604–613. Nadeem, M., Shabbir, M., Abdullah, M. A., Shah, S. S. & McKay, G.  Sorption of cadmium from aqueous solution by surfactant-modified carbon adsorbents. Chemical Engineering Journal 148 (2), 365–370. Nakajima, A.  Electron spin resonance study of copper biosorption by bacteria. Water Research 36 (8), 2091–2097. Namasivayam, C. & Kadirvelu, K.  Uptake of mercury (II) from wastewater by activated carbon from an unwanted agricultural solid by-product: coirpith. Carbon 37 (1), 79–84. Namasivayam, C. & Sureshkumar, M. V.  Removal of chromium (VI) from water and wastewater using surfactant modified coconut coir pith as a biosorbent. Bioresource Technology 99 (7), 2218–2225. Naseem, R. & Tahir, S. S.  Removal of Pb (II) from aqueous/ acidic solutions by using bentonite as an adsorbent. Water Research 35 (16), 3982–3986. Ngah, W. W. & Hanafiah, M. A. K. M.  Removal of heavy metal ions from wastewater by chemically modified plant wastes as adsorbents: a review. Bioresource Technology 99 (10), 3935–3948. Nizamuddin, S., Kumar, J., Subramanian, N., Sahu, J. N., Ganesan, P., Mubarak, N. M. & Mazari, S. A.  Synthesis and characterization of hydrochars produced by hydrothermal carbonization of oil palm shell. The Canadian Journal of Chemical Engineering 93 (11), 1916–1921. Nizamuddin, S., Mubarak, N. M., Tiripathi, M., Jayakumar, N. S., Sahu, J. N. & Ganesan, P.  Chemical, dielectric and structural characterization of optimized hydrochar produced from hydrothermal carbonization of palm shell. Fuel 163, 88–97. Nuhoglu, Y., Malkoc, E., Gürses, A. & Canpolat, N.  The removal of Cu (II) from aqueous solutions by Ulothrix zonata. Bioresource Technology 85 (3), 331–333. Obregón-Valencia, D. & del Rosario Sun-Kou, M.  Comparative cadmium adsorption study on activated carbon prepared from aguaje (Mauritia flexuosa) and olive fruit stones (Olea europaea L.). Journal of Environmental Chemical Engineering 2 (4), 2280–2288. Oladoja, N. A., Ololade, I. A., Alimi, O. A., Akinnifesi, T. A. & Olaremu, G. A.  Iron incorporated rice husk silica as a sorbent for hexavalent chromium attenuation in aqueous system. Chemical Engineering Research and Design 91 (12), 2691–2702.

417

Renu et al.

Heavy metal removal from wastewater using various adsorbents

Oliveira, W. E., Franca, A. S., Oliveira, L. S. & Rocha, S. D.  Untreated coffee husks as biosorbents for the removal of heavy metals from aqueous solutions. Journal of Hazardous Materials 152 (3), 1073–1081. Oubagaranadin, J. U. K., Murthy, Z. V. & Mallapur, V. P.  Removal of Cu (II) and Zn (II) from industrial wastewater by acid-activated montmorillonite-illite type of clay. Comptes Rendus Chimie 13 (11), 1359–1363. Owlad, M., Aroua, M. K. & Daud, W. M. A. W.  Hexavalent chromium adsorption on impregnated palm shell activated carbon with polyethyleneimine. Bioresource Technology 101 (14), 5098–5103. Oyaro, N., Ogendi, J., Murago, E. N. & Gitonga, E.  The contents of Pb, Cu, Zn and Cd in meat in Nairobi, Kenya. Journal of Food, Agriculture and Environment 5, 119–121. Özer, A. & Özer, D.  The adsorption of Cr (VI) on sulphuric acid-treated wheat bran. Environmental Technology 25 (6), 689–697. Öztürk, A., Artan, T. & Ayar, A.  Biosorption of nickel (II) and copper (II) ions from aqueous solution by Streptomyces coelicolor A3 (2). Colloids and Surfaces B: Biointerfaces 34 (2), 105–111. Pagnanelli, F., Esposito, A., Toro, L. & Veglio, F.  Metal speciation and pH effect on Pb, Cu, Zn and Cd biosorption onto Sphaerotilus natans: langmuir-type empirical model. Water Research 37 (3), 627–633. Pardo, R., Herguedas, M., Barrado, E. & Vega, M.  Biosorption of cadmium, copper, lead and zinc by inactive biomass of Pseudomonas putida. Analytical and Bioanalytical Chemistry 376 (1), 26–32. Pérez-Quintanilla, D., del Hierro, I., Fajardo, M. & Sierra, I.  Cr (VI) adsorption on functionalized amorphous and mesoporous silica from aqueous and non-aqueous media. Materials Research Bulletin 42 (8), 1518–1530. Pino, G. H., de Mesquita, L. M. S., Torem, M. L. & Pinto, G. A. S.  Biosorption of cadmium by green coconut shell powder. Minerals Engineering 19 (5), 380–387. Politi, D. & Sidiras, D.  Wastewater treatment for dyes and heavy metals using modiﬁed pine sawdust as adsorbent. Procedia Engineering 42, 1969–1982. Rad, L. R., Momeni, A., Ghazani, B. F., Irani, M., Mahmoudi, M. & Noghreh, B.  Removal of Ni2þ and Cd2þ ions from aqueous solutions using electrospun PVA/zeolite nanoﬁbrous adsorbent. Chemical Engineering Journal 256, 119–127. Rahman, I. A., Ismail, J. & Osman, H.  Effect of nitric acid digestion on organic materials and silica in rice husk. Journal of Materials Chemistry 7 (8), 1505–1509. Rai, L. C., Gaur, J. P. & Kumar, H. D.  Phycology and heavymetal pollution. Biological Reviews 56 (2), 99–151. Rao, G. P. C., Satyaveni, S., Ramesh, A., Seshaiah, K., Murthy, K. S. N. & Choudary, N. V.  Sorption of cadmium and zinc from aqueous solutions by zeolite 4A, zeolite 13X and bentonite. Journal of Environmental Management 81 (3), 265–272. Rao, K. S., Mohapatra, M., Anand, S. & Venkateswarlu, P.  Review on cadmium removal from aqueous solutions.

Journal of Water Reuse and Desalination

07.4

2017

International Journal of Engineering, Science and Technology 2 (7), 81–103. Ravat, C., Dumonceau, J. & Monteil-Rivera, F.  Acid/base and Cu (II) binding properties of natural organic matter extracted from wheat bran: modeling by the surface complexation model. Water Research 34 (4), 1327–1339. Ravikumar, S., Yoo, I. K., Lee, S. Y. & Hong, S. H.  Construction of copper removing bacteria through the integration of twocomponent system and cell surface display. Applied Biochemistry and Biotechnology 165 (7–8), 1674–1681. Ren, Y., Zhang, M. & Zhao, D.  Synthesis and properties of magnetic Cu (II) ion imprinted composite adsorbent for selective removal of copper. Desalination 228 (1), 135–149. Rengaraj, S., Kim, Y., Joo, C. K. & Yi, J.  Removal of copper from aqueous solution by aminated and protonated mesoporous aluminas: kinetics and equilibrium. Journal of Colloid and Interface Science 273 (1), 14–21. Renge, V. C., Khedkar, S. V. & Pandey Shraddha, V.  Removal of heavy metals from wastewater using low cost adsorbents: a review. Scientiﬁc Reviews and Chemical Communications 2 (4), 580–584. Rohim, R., Ahmad, R., Ibrahim, N., Hamidin, N. & Abidin, C. Z. A.  Characterization of calcium oxide catalyst from eggshell waste. Advances in Environmental Biology 8, 35–39. Romera, E., González, F., Ballester, A., Blázquez, M. L. & Munoz, J. A.  Comparative study of biosorption of heavy metals using different types of algae. Bioresource Technology 98 (17), 3344–3353. Rosen, M. J. & Kunjappu, J. T.  Surfactants and Interfacial Phenomena. John Wiley & Sons, Hoboken, NJ, USA. Ruthiraan, M., Abdullah, E. C., Thines, K. R. & Mubarak, N. M. a Synthesis of magnetic biochar from Garcinia Mangostana peel using mufﬂe furnace for adsorption of Zn2þ ions from aqueous solution. International Journal of Chemical Engineering 2, 18–21. Ruthiraan, M., Mubarak, N. M., Thines, R. K., Abdullah, E. C., Sahu, J. N., Jayakumar, N. S. & Ganesan, P. b Comparative kinetic study of functionalized carbon nanotubes and magnetic biochar for removal of Cd2þ ions from wastewater. Korean Journal of Chemical Engineering 32 (3), 446–457. Sabzoi, N., Yong, E. K., Jayakumar, N. S., Sahu, J. N., Ganesan, P., Mubarak, N. M. & Mazari, S. A.  An optimisation study for catalytic hydrolysis of oil palm shell using response surface methodology. Journal of Oil Palm Research 27 (4), 339–351. Ş ahan, T., Ceylan, H., Ş ahiner, N. & Aktaş, N.  Optimization of removal conditions of copper ions from aqueous solutions by Trametes versicolor. Bioresource Technology 101 (12), 4520–4526. Sahu, J. N., Acharya, J. & Meikap, B. C.  Response surface modeling and optimization of chromium (VI) removal from aqueous solution using Tamarind wood activated carbon in batch process. Journal of Hazardous Materials 172 (2), 818–825. Saifuddin, M. & Kumaran, P.  Removal of heavy metal from industrial wastewater using chitosan coated oil palm shell charcoal. Electronic Journal of Biotechnology 8 (1), 43–53.

Page 223

418

Renu et al.

Heavy metal removal from wastewater using various adsorbents

Salah, T. A., Mohammad, A. M., Hassan, M. A. & El-Anadouli, B. E.  Development of nano-hydroxyapatite/chitosan composite for cadmium ions removal in wastewater treatment. Journal of the Taiwan Institute of Chemical Engineers 45 (4), 1571–1577. Salam, M. A., Makki, M. S. & Abdelaal, M. Y.  Preparation and characterization of multi-walled carbon nanotubes/chitosan nanocomposite and its application for the removal of heavy metals from aqueous solution. Journal of Alloys and Compounds 509 (5), 2582–2587. Sankararamakrishnan, N., Jaiswal, M. & Verma, N.  Composite nanoﬂoral clusters of carbon nanotubes and activated alumina: an efﬁcient sorbent for heavy metal removal. Chemical Engineering Journal 235, 1–9. Sardella, F., Gimenez, M., Navas, C., Morandi, C., Deiana, C. & Sapag, K.  Conversion of viticultural industry wastes into activated carbons for removal of lead and cadmium. Journal of Environmental Chemical Engineering 3 (1), 253–260. Schut, S., Zauner, S., Hampel, G., König, H. & Claus, H.  Biosorption of copper by wine-relevant lactobacilli. International Journal of Food Microbiology 145 (1), 126–131. Selomulya, C., Meeyoo, V. & Amal, R.  Mechanisms of Cr (VI) removal from water by various types of activated carbons. Journal of Chemical Technology and Biotechnology 74 (2), 111–122. Selvi, K., Pattabhi, S. & Kadirvelu, K.  Removal of Cr (VI) from aqueous solution by adsorption onto activated carbon. Bioresource Technology 80 (1), 87–89. Sharma, D. C. & Forster, C. F.  A preliminary examination into the adsorption of hexavalent chromium using low-cost adsorbents. Bioresource Technology 47 (3), 257–264. Sharma, N., Kaur, K. & Kaur, S.  Kinetic and equilibrium studies on the removal of Cd2þ ions from water using polyacrylamide grafted rice (Oryza sativa) husk and (Tectona grandis) saw dust. Journal of Hazardous Materials 163 (2), 1338–1344. Shawabkeh, R., Al-Harahsheh, A., Hami, M. & Khlaifat, A.  Conversion of oil shale ash into zeolite for cadmium and lead removal from wastewater. Fuel 83 (7), 981–985. Shen, Y. S., Wang, S. L., Tzou, Y. M., Yan, Y. Y. & Kuan, W. H.  Removal of hexavalent Cr by coconut coir and derived chars – The effect of surface functionality. Bioresource Technology 104, 165–172. Sheng, P. X., Wee, K. H., Ting, Y. P. & Chen, J. P.  Biosorption of copper by immobilized marine algal biomass. Chemical Engineering Journal 136 (2), 156–163. Shin, K. Y., Hong, J. Y. & Jang, J.  Heavy metal ion adsorption behavior in nitrogen-doped magnetic carbon nanoparticles: isotherms and kinetic study. Journal of Hazardous Materials 190 (1), 36–44. Shukla, A., Zhang, Y. H., Dubey, P., Margrave, J. L. & Shukla, S. S.  The role of sawdust in the removal of unwanted materials from water. Journal of Hazardous Materials 95 (1), 137–152.

Page 224

Journal of Water Reuse and Desalination

07.4

2017

Singh, S. P., Ma, L. Q. & Harris, W. G.  Heavy metal interactions with phosphatic clay. Journal of Environmental Quality 30 (6), 1961–1968. Singh, A., Kumar, D. & Gaur, J. P.  Copper(II) and lead(II) sorption from aqueous solution by non-living Spirogyra neglecta. Bioresource Technology 98 (18), 3622–3629. Singh, K. K., Hasan, S. H., Talat, M., Singh, V. K. & Gangwar, S. K.  Removal of Cr (VI) from aqueous solutions using wheat bran. Chemical Engineering Journal 151 (1), 113–121. Solisio, C., Lodi, A., Torre, P., Converti, A. & Del Borghi, M.  Copper removal by dry and re-hydrated biomass of Spirulina platensis. Bioresource Technology 97 (14), 1756–1760. Sousa, F. W., Oliveira, A. G., Ribeiro, J. P., Rosa, M. F., Keukeleire, D. & Nascimento, R. F.  Green coconut shells applied as adsorbent for removal of toxic metal ions using fixed-bed column technology. Journal of Environmental Management 91 (8), 1634–1640. Srivastava, V. C., Mall, I. D. & Mishra, I. M.  Characterization of mesoporous rice husk ash (RHA) and adsorption kinetics of metal ions from aqueous solution onto RHA. Journal of Hazardous Materials 134 (1), 257–267. Srivastava, V. C., Mall, I. D. & Mishra, I. M.  Removal of cadmium (II) and zinc (II) metal ions from binary aqueous solution by rice husk ash. Colloids and Surfaces A: Physicochemical and Engineering Aspects 312 (2), 172–184. Sugashini, S. & Begum, K. M. M. S.  Preparation of activated carbon from carbonized rice husk by ozone activation for Cr (VI) removal. New Carbon Materials 30 (3), 252–261. Sumathi, K. M. S., Mahimairaja, S. & Naidu, R.  Use of low-cost biological wastes and vermiculite for removal of chromium from tannery effluent. Bioresource Technology 96 (3), 309–316. Tan, W. T., Ooi, S. T. & Lee, C. K.  Removal of chromium (VI) from solution by coconut husk and palm pressed fibres. Environmental Technology 14 (3), 277–282. Thangalazhy-Gopakumar, S., Al-Nadheri, W. M. A., Jegarajan, D., Sahu, J. N., Mubarak, N. M. & Nizamuddin, S.  Utilization of palm oil sludge through pyrolysis for bio-oil and bio-char production. Bioresource Technology 178, 65–69. Thangavel, S. & Venugopal, G.  Understanding the adsorption property of graphene-oxide with different degrees of oxidation levels. Powder Technology 257, 141–148. Thines, R. K., Mubarak, N. M., Ruthiraan, M., Abdullah, E. C., Sahu, J. N., Jayakumara, N. S., Ganesan, P. & Sajuni, N. R.  Adsorption isotherm and thermodynamics studies of Zn(II) on functionalized and non-functionalized carbon nanotubes. Advanced Science, Engineering and Medicine 6 (9), 974–984. Thines, K. R., Abdullah, E. C., Ruthiraan, M. & Mubarak, N. M.  Production of magnetic biochar derived from durian’s rind at vacuum condition for removal of methylene blue pigments from aqueous solution. Environment 14, 15. Toro, P., Quijada, R. & Yazdani-Pedram, M.  Eggshell, a new bio-filler for polypropylene composites. Materials Letters 61, 4347–4350.

419

Renu et al.

Heavy metal removal from wastewater using various adsorbents

Tunali, S., Cabuk, A. & Akar, T.  Removal of lead and copper ions from aqueous solutions by bacterial strain isolated from soil. Chemical Engineering Journal 115 (3), 203–211. Ullah, I., Nadeem, R., Iqbal, M. & Manzoor, Q.  Biosorption of chromium onto native and immobilized sugarcane bagasse waste biomass. Ecological Engineering 60, 99–107. Uslu, G. & Tanyol, M.  Equilibrium and thermodynamic parameters of single and binary mixture biosorption of lead (II) and copper (II) ions onto Pseudomonas putida: effect of temperature. Journal of Hazardous Materials 135 (1), 87–93. Veglio, F., Beolchini, F. & Gasbarro, A.  Biosorption of toxic metals: an equilibrium study using free cells of Arthrobacter sp. Process Biochemistry 32 (2), 99–105. Veit, M. T., Tavares, C. R. G., Gomes-da-Costa, S. M. & Guedes, T. A.  Adsorption isotherms of copper (II) for two species of dead fungi biomasses. Process Biochemistry 40 (10), 3303–3308. Veneu, D. M., Torem, M. L. & Pino, G. A.  Fundamental aspects of copper and zinc removal from aqueous solutions using a Streptomyces lunalinharesii strain. Minerals Engineering 48, 44–50. Vengris, T., Binkien, R. & Sveikauskait, A.  Nickel, copper and zinc removal from waste water by a modiﬁed clay sorbent. Applied Clay Science 18 (3), 183–190. Verma, A., Singh, A., Bishnoi, N. R. & Gupta, A.  Biosorption of Cu (II) using free and immobilized biomass of Penicillium citrinum. Ecological Engineering 61, 486–490. Vijayaraghavan, K. & Prabu, D.  Potential of Sargassum wightii biomass for copper (II) removal from aqueous solutions: application of different mathematical models to batch and continuous biosorption data. Journal of Hazardous Materials 137 (1), 558–564. Vilar, V. J., Botelho, C. M. & Boaventura, R. A.  Copper desorption from Gelidium algal biomass. Water Research 41 (7), 1569–1579. Vilar, V. J., Botelho, C. M. & Boaventura, R. A.  Copper removal by algae Gelidium, agar extraction algal waste and granulated algal waste: kinetics and equilibrium. Bioresource Technology 99 (4), 750–762. Visa, M.  Synthesis and characterization of new zeolite materials obtained from Fly ash for heavy metals removal in advanced wastewater treatment. Powder Technology 294, 338–347. Volesky, B. & Holan, Z. R.  Biosorption of heavy metals. Biotechnology Progress 11 (3), 235–250. Wan, M. W., Petrisor, I. G., Lai, H. T., Kim, D. & Yen, T. F.  Copper adsorption through chitosan immobilized on sand to demonstrate the feasibility for in situ soil decontamination. Carbohydrate Polymers 55 (3), 249–254. Wang, J. & Chen, C.  Biosorbents for heavy metals removal and their future. Biotechnology Advances 27 (2), 195–226.

Journal of Water Reuse and Desalination

07.4

2017

Wang, J. S., Peng, R. T., Yang, J. H., Liu, Y. C. & Hu, X. J.  Preparation of ethylenediamine-modified magnetic chitosan complex for adsorption of uranyl ions. Carbohydrate Polymers 84 (3), 1169–1175. Woolf, D., Amonette, J. E., Street-Perrott, F. A., Lehmann, J. & Joseph, S.  Sustainable biochar to mitigate global climate change. Nature Communications 1, 56. Wu, Y., Luo, H., Wang, H., Wang, C., Zhang, J. & Zhang, Z.  Adsorption of hexavalent chromium from aqueous solutions by graphene modified with cetyltrimethylammonium bromide. Journal of Colloid and Interface Science 394, 183–191. Xiao, Y., Liang, H., Chen, W. & Wang, Z.  Synthesis and adsorption behavior of chitosan-coated MnFe2O4 nanoparticles for trace heavy metal ions removal. Applied Surface Science 285, 498–504. Xu, X., Cao, X. & Zhao, L.  Comparison of rice husk-and dairy manure-derived biochars for simultaneously removing heavy metals from aqueous solutions: role of mineral components in biochars. Chemosphere 92 (8), 955–961. Yang, S., Li, L., Pei, Z., Li, C., Lv, J., Xie, J., Wen, B. & Zhang, S. a Adsorption kinetics, isotherms and thermodynamics of Cr(III) on graphene oxide. Colloids and Surfaces A: Physicochemical and Engineering Aspects 457, 100–106. Yang, G., Tang, L., Lei, X., Zeng, G., Cai, Y., Wei, X., Zhou, Y., Li, S., Fang, Y. & Zhang, Y. b Cd(II) removal from aqueous solution by adsorption on α-ketoglutaric acidmodified magnetic chitosan. Applied Surface Science 292, 710–716. Ye, H., Zhu, Q. & Du, D.  Adsorptive removal of Cd (II) from aqueous solution using natural and modified rice husk. Bioresource Technology 101 (14), 5175–5179. Yin, D., Du, X., Liu, H., Zhang, Q. & Ma, L.  Facile one-step fabrication of polymer microspheres with high magnetism and armored inorganic particles by Pickering emulsion polymerization. Colloids and Surfaces A: Physicochemical and Engineering Aspects 414, 289–295. Yu, J., Tong, M., Sun, X. & Li, B.  Enhanced and selective adsorption of Pb2þ and Cu2þ by EDTAD-modified biomass of baker’s yeast. Bioresource Technology 99 (7), 2588–2593. Yu, Y. L., Zhuang, Y. T., Song, X. Y. & Wang, J. H.  Lyophilized carbon nanotubes/graphene oxide modified cigarette filter for the effective removal of cadmium and chromium from mainstream smoke. Chemical Engineering Journal 280, 58–65. Zhu, C., Luan, Z., Wang, Y. & Shan, X.  Removal of cadmium from aqueous solutions by adsorption on granular red mud (GRM). Separation and Purification Technology 57 (1), 161–169.

First received 3 June 2016; accepted in revised form 28 August 2016. Available online 3 November 2016

Page 225

264

07.3

2017

Inﬂuence of nitrite on the removal of Mn(II) using pilot-scale bioﬁlters Qingfeng Cheng, Lichao Nengzi, Dongying Xu, Junyuan Guo and Jing Yu

ABSTRACT Two pilot-scale biofilters were used to systematically investigate the influence of nitrite on biological Mn(II) removal. Gibbs free energy change (ΔG) of the redox reaction between MnO2 and NO–2 was 122.28 kJ mol–1 in 298 K, suggesting that MnO2 could not react with NO–2. When nitrite in the influent was increased from 0.05 to 0.5 mg L–1, manganese oxides did not react with nitrite in anaerobic conditions; nitrite was quickly oxidized and biological Mn(II) removal was slightly affected in 2 h in aerobic conditions. When nitrite was accumulated in the biofilter by increasing ammonia concentration,

Qingfeng Cheng (corresponding author) Lichao Nengzi Dongying Xu Junyuan Guo Jing Yu College of Resources and Environment, Chengdu University of Information Technology, Chengdu 610225, China E-mail: chqf185@163.com

nitrite existed for more than 3 d and biological Mn(II) removal was affected in 3 d. When Mn(II) and ammonia in the influent were about 2 and 1.5 mg L–1, respectively, both of them were completely removed and the oxidation-reduction potential was increased with the depth of the filter from 16 to 122 mV. Biological Mn(II) removal followed the first-order reaction, and the k-value was 0.687 min–1. Key words

| bioﬁlter, groundwater, kinetics, manganese removal, nitrite

INTRODUCTION Groundwater is often mildly acidic and devoid of dissolved

), Fe(II) and Mn(II) are objectionable for the following

oxygen (DO) (Azher et al. ), so when groundwater ﬂows

reasons: (a) Fe(II) and Mn(II) give a metallic taste in

through soils, minerals and rocks, soluble Fe(II) and Mn(II)

water systems (Azher et al. ); (b) iron and manganese

are present ( Jusoh et al. ), either in dissolved mineral

deposits build up in pipelines reducing the pipe diameter

form, or associated with various organics, minerals or che-

in the distribution systems and eventually clog the pipe; (c)

lating agents. In addition, the predominant form of Mn(II)

in water distribution systems, Fe(II) and Mn(II) are

at low or neutral pH values is Mn2þ, which occurs primarily

substrates for bacteria growth (Azher et al. ), hence

as a free cation in natural waters (Nealson et al. ).

when the bacteria die and slough off, bad odors and unplea-

Continuously increasing ammonia concentration in ground-

sant tastes may be produced (Kontari ; Gouzinis et al.

water has been observed in the past years, owing to the

). In addition, when Mn(II) exceeds the permitted

discharge of waste from both industry and bank-side resi-

limit, Mn(II) has been found to affect the central nervous

dents without adequate pre-treatment and sub-optimal

system (Sharma et al. ). The presence of ammonia in

conditions of the catchments (Okoniewska et al. ;

drinking water treatment could affect the chlorination pro-

Akkera et al. ).

cess and Mn(II) bioﬁltration system (Hasan et al. ).

When Fe(II) and Mn(II) are present in drinking water at

Ammonia will react with chlorine to form disinfection

concentrations exceeding the permitted limits of 0.2 and

by-products (Richardson & Postigo ), which could

–1

0.05 mg L , respectively (Tekerlekopoulou & Vayenas

damage the human nervous system (Nieuwenhuijsen et al. ), cause a deterioration in the taste and odor of water

This is an Open Access article distributed under the terms of the Creative Commons Attribution Licence (CC BY 4.0), which permits copying,

(Richardson et al. ), and reduce the disinfection efﬁ-

adaptation and redistribution, provided the original work is properly cited

ciency (WHO ). Furthermore, ammonia can interfere

(http://creativecommons.org/licenses/by/4.0/).

with the Mn(II) bioﬁltration process by consuming excessive

doi: 10.2166/wrd.2016.210

Page 227

265

Q. Cheng et al.

Inﬂuence of nitrite on biological Mn(II) removal

Journal of Water Reuse and Desalination

07.3

2017

oxygen during nitriﬁcation, resulting in moldy and earthy

In this study, the ΔG of the redox reaction between

tasting water (WHO ). So groundwater which contains

MnO2 and NO–2 was calculated. Two pilot-scale bioﬁlters

high concentrations of Fe(II), Mn(II) and ammonia needs to

were established to investigate the redox reaction between

be treated before it used for industry and humans. Chemical

MnO2 and HNO2 in anaerobic conditions, and the inﬂuence

methods could be used to oxidize Fe(II), Mn(II) and/or

of added and generated nitrite on biological Mn(II) removal

ammonia, but chemical oxidation may produce potential

in a bioﬁlter. The main objectives of this study were to verify

hazardous by-products and/or introduce other pollutants

whether the reaction between MnO2 and NO–2 could occur,

into the produced water. Moreover, it is difﬁcult to simul-

and ﬁnd the reason why the start-up period of bioﬁlters for

taneously remove ammonia and manganese with only one

Mn(II) and ammonia removal was much longer than the

chemical (Han et al. ). The advantages of the biological

bioﬁlters for Mn(II) removal.

oxidation process over the conventional physicochemical methods include high ﬁltration rates, and low operation and maintenance costs. As single stage ﬁltration is used, it

MATERIALS AND METHODS

is not necessary to provide additional chemicals, and the volume of the generated sludge is appreciably smaller and

Two pilot-scale bioﬁlters were developed in a groundwater

easier to handle (Pacini et al. ; Han et al. ).

treatment plant (GWTP), which is located in Harbin city,

It is reported that achieving simultaneous removal of

P.R. China. The height and diameter of ﬁlters 1 and 2

ammonia and Mn(II) would be very difﬁcult (Hasan et al.

were 300 × 25 and 300 × 15 cm, respectively, and the effec-

; Han et al. ), since biological Mn(II) removal can

tive working volume was 74 and 26 L, respectively

only take place after complete nitriﬁcation because of the

(Figure 1). At the top of each ﬁlter, the incoming waters

necessary evolution of the oxidation-reduction potential

were ﬁrstly mixed in the mixing chamber and then they

(ORP) (Frischherz et al. ; Vandenabeele et al. ;

ﬂowed into the ﬁlter. Meanwhile, at the bottom of each

Hasan et al. ). Thus the start-up period of bioﬁlters for

ﬁlter, an underdrain system was used to collect the treated

Mn(II) and ammonia removal needs 3–4 months (Frisch-

water and any biological solids which detached from the

herz et al. ), while the bioﬁlters for Mn(II) removal is

media. Along each ﬁlter depth there were 20 sampling

only 1–2 months (Frischherz et al. ; Vayenas et al.

ports at 10 cm intervals for Fe(II), Mn(II), ammonia, nitrite

). It is not clear how biological Mn(II) removal and bio-

and ORP concentration measurements in the bulk liquid.

logical ammonia removal are linked. A better understanding

Tank 1 (volume was 2,000 L) was used to collect raw

of the interactions between the two phenomena is important

groundwater (DO was about 0.2 mg L–1) or aerated raw

from an economic point of view (Vandenabeele et al. ).

groundwater (DO was about 8.5 mg L–1), which were

Nitriﬁcation (biological ammonia oxidation) is carried out

obtained from the GWTP. Tank 2 (volume was 200 L) was

by two different consecutive microbial processes, nitritiﬁca-

used to collect the efﬂuent water of ﬁlter 1. In order to

tion and nitratiﬁcation (Vayenas et al. ). In nitritiﬁcation

increase the concentration of Mn(II), nitrite and ammonia

process, ammonia is oxidized to nitrite by the bacterial

in the inﬂuent stock solutions of 20 g L–1 Mn(II), 2 g L–1

genera Nitrosomonas; and in nitratiﬁcation process, Nitro-

NO–2-N and 20 g L–1 NHþ 4 -N were prepared in tanks 3, 4

bacter converts nitrite to nitrate. In addition, Nitrosomonas

and 5 (volume was all 50 L) by diluting MnSO4·H2O,

and Nitrobacter are aerobic and autotrophic bacteria. The

NaNO2 and NH4Cl, respectively.

nitritiﬁcation rate is faster than the nitratiﬁcation rate,

Filters 1 and 2 were packed with manganese sand at a

resulting in nitrite accumulating in the bioﬁlters during the

height of 150 cm and a diameter of 0.8–1 mm. It should be

start-up period. Vandenabeele et al. () investigated the

noted that the bioﬁlters were operating about 18 months

inﬂuence of nitrite on Mn(II) removal using PYM medium

before this experiment. Real groundwater, which was

(Ehrlieh & Zapkin ); however, few researchers have

extracted from the wells with a depth of 40–50 m, in

investigated the inﬂuence of nitrite on Mn(II) removal

Harbin city, P.R. China, was used throughout this exper-

using a bioﬁlter.

iment. The compounds in the real groundwater are

Page 228

266

Q. Cheng et al.

Figure 1

Journal of Water Reuse and Desalination

Inﬂuence of nitrite on biological Mn(II) removal

07.3

2017

Schematic drawing of the pilot-scale bioﬁlters.

shown in Table 1. The concentration of total iron, Mn(II)

was 8:30 am every day, and the backwashing time was

and ammonia in the raw groundwater was about 8–13,

10:00 am.

1.1 and 1.2 mg L–1, respectively. The temperature was approximately 8 C. Downward gravity ﬂow was adopted W

The redox reaction between nitrite and Mn(IV)

in the bioﬁlters, and the amount of the ﬂow was controlled at the entry point. Due to pore clogging from bacteria

In order to verify whether the reaction between MnO2

growth and iron and manganese precipitation on the sup-

and NO–2 could occur under the conditions of this exper-

port

was

iment, the ΔG of the redox reaction between MnO2 and

performed approximately every 2 days using high-water

NO–2 was calculated with a thermodynamic temperature

up ﬂow velocities to wash out dead bacteria and maintain

of 298 and 281 K and an atmospheric pressure of

the activity of the system at a high level. The sampling time

100 kPa.

materials

surfaces,

regular

backwashing

The raw groundwater in tank 1 was pumped into the Table 1

Physicochemical characteristics of the raw groundwater

mixing chamber of filter 1, and then flowed into the filter; simultaneously DO in the raw groundwater in filter 1 was

Properties

Value

Properties

Value

Temperature

∼8 C

NO–2

∼0.002 mg/L

∼6.9

NHþ 4

∼1.2 mg/L

Total Fe

∼10.5 mg/L

CODMn

∼1.9 mgO2/L

Mn(II)

∼1.1 mg/L

0.2 mg L–1. The efﬂuent water of ﬁlter 1 in tank 2 and the

Color

stock solutions of Mn(II) and nitrite in tanks 3 and 4,

Total As

<0.002 mg/L

∼0.2 mg/L

SO2– 4

∼5 mg/L

respectively, were pumped to the mixing chamber of ﬁlter

Alkalinity

210.0 mg CaCO3/L

F–

∼0.5 mg/L

Turbidity

∼1.1 NTU

Cl–

∼2 mg/L

ﬂow rates of the stock solutions of Mn(II) and nitrite were

Total Hardness

140–170 mg CaCO3/L

regulated to ensure that Mn(II) in the inﬂuent was approxi-

NO–3

∼0.03 mg/L

increased to about 3 mg L–1. The concentration of total iron, Mn(II) and ammonia in efﬂuent of ﬁlter 1 was about 0.1, 1.5 and 0.7 mg L–1, respectively, and DO was about

2 which was sealed, mixed and ﬂowed into ﬁlter 2. The

mately 2 mg L-1 and nitrite was approximately 0.05, 0.1, 0.2 and 0.5 mg L–1, respectively. Page 229

267

Q. Cheng et al.

Journal of Water Reuse and Desalination

Inﬂuence of nitrite on biological Mn(II) removal

07.3

2017

The inﬂuence of added and generated nitrite on Mn(II)

ammonia, total iron, Mn(II) and nitrite concentration

removal in a bioﬁlter

measurements were according to the Standard Methods for the Examination of Water and Wastewater (Rittmann

In order to investigate the inﬂuence of added nitrite on

& Snoeyinck ).

Mn(II) removal using a biofilter, the experiment was operated as in the previous section except that the DO in the influent of filter 2 was above 10 mg L–1. In order to investigate the

RESULTS AND DISCUSSION

influence of generated nitrite on Mn(II) removal using a biofilter, the concentration of ammonia in influent of filter 1 –1

was increased from approximately 1.1 to 1.5 mg L

adding the stock solution of ammonia in tank 5 (DO in the

The redox reaction between nitrite and Mn(IV) 2þ þ 2NO� MnO2 þ 2NO� 2 ¼ Mn 3

influent of filter 1 was approximately 11 mg L–1). An equation of Gibbs free energy change: The concentration profiles of ORP in simultaneous Mn(II) and ammonia removal

ΔG ¼ ΔH � T ΔS

The aerated raw groundwater in tank 1 was pumped into

where ΔG is Gibbs free energy change (kJ mol–1), ΔH is free

ﬁlter 1, and the concentration of total iron, Mn(II) and

enthalpy change (kJ mol–1), T is thermodynamic tempera-

ammonia in efﬂuent water of ﬁlter 1 was lower than 0.1,

(1)

ture (K), and ΔS is entropy change (kJ mol–1 K–1).

0.05 and 0.1 mg L–1, respectively. The efﬂuent water of

When the thermodynamic temperature was 298 (25 C)

ﬁlter 1 in tank 2 was aerated and DO was increased to

and 281 K (8 C), the ΔG was 122.28 and 120.92 kJ mol–1 in

approximately 11 mg L–1, then the efﬂuent water in tank 2

100 kPa, respectively, which suggested that MnO2 cannot

and the stock solutions of Mn(II) and ammonia in tanks 3

react with NO–2 under these conditions.

and 5, respectively, were pumped to filter 2. The concentration of Mn(II) and ammonia in the influent of filter 2 was approximately 2 and 1.5 mg L–1, respectively.

When nitrite (approximately 0.05, 0.1, 0.2 and 0.5 mg L–1) and Mn(II) (approximately 2 mg L–1) were added to ﬁlter 2 and DO in the inﬂuent was approximately 0.2 mg L–1, only a small part of nitrite was oxidized by DO in depths of 0–0.1 m of the

Kinetics of biological Mn(II) oxidation

ﬁlter, while Mn(II) was obviously decreased in depths of 0–0.2 m (Figure 2). The activity of manganese oxidizing bac-

The efﬂuent water in tank 2 (as in the previous section) and

teria (MnOB) was higher than nitrite oxidizing bacteria

the stock solution of Mn(II) in tank 3 were pumped into

(NOB) in very low DO conditions. In depths of 0.2–1.5 m,

ﬁlter 2. In addition, the concentration of Mn(II) in the inﬂu-

DO was lower than 0.1 mg L–1, therefore nitrite and Mn(II)

ent was approximately 4 mg L–1. The determination of the

could not be oxidized by DO. Nitrite and Mn(II) were not

empty ﬁlter contacted time (EFCT) of the groundwater in

varied in depths of 0.2–1.5 m, suggesting that nitrite did not

the bioﬁlter was based on the following equation:

react with manganese oxides in the bioﬁlter in anaerobic conditions, which corresponds with the result of the ΔG.

EFCT ¼ filter height ðmÞ=linear velocity ðm h�1 Þ

The inﬂuence of added and generated nitrite on Mn(II) removal in a bioﬁlter

Analysis methods When the concentration of nitrite and Mn(II) in the inﬂuent The pH, ORP and DO measurements were conducted using

of ﬁlter 2 was approximately 0.05 and 2 mg L–1, respectively,

a pH meter (Ultra BASIC UB-10), an ORP meter (pH 315i-

nitrite and Mn(II) were completely oxidized in depths of

WTW) and a DO meter (Oxi 315i-WTW), respectively. The

0–0.1 m of the ﬁlter after 1 d (Figure 3). The ﬁlter was

Page 230

268

Q. Cheng et al.

Inﬂuence of nitrite on biological Mn(II) removal

Journal of Water Reuse and Desalination

07.3

2017

used to remove Mn(II) and ammonia before this experiment, and an abundance of MnOB and NOB existed in the filter, which quickly oxidized Mn(II) and ammonia. Then the concentration of nitrite was increased to about 0.1, 0.2 and 0.5 mg L–1, respectively, and nitrite and Mn(II) were also completely removed in depths of 0–0.1 m after 1 d. Before nitrite was added to filter 2, Mn(II) was completely removed in depths of 0–0.1 m of the filter. When nitrite was added to filter 2 with a concentration of approximately 0.05 mg L–1 (Figure 4(a)), Mn(II) was increased to 0.23 mg L–1 after 1 h in depths of 0.1 m (Figure 4(b)), and then decreased to lower than 0.05 mg L–1 after 2 h. When the concentration of added nitrite was increased to approximately 0.5 mg L–1 (Figure 4(c)), Mn(II) was increased to 0.18 mg L–1 after 1 h in depths of 0.1 m (Figure 4(d)), and then decreased to lower than 0.05 mg L–1 after 2 h. When nitrite was added to filter 2, biological Mn(II) removal was Figure 2

Nitrite (a) and Mn(II) (b) concentration proﬁles along depth of ﬁlter 2 for nitrite feed concentration of approximately 0.05, 0.1, 0.2 and 0.5 mg L–1, respectively, and Mn(II) feed concentration of approximately 2 mg L–1 in anaerobic conditions.

affected slightly, this is attributed to the presence of NOB, which quickly oxidized the added nitrite, and the MnOB adapted to the nitrite presented conditions. When ammonia in the influent of filter 1 was increased from approximately 1.1 to 1.5 mg L–1, nitrite accumulated in the filter. Nitrite rapidly increased in depths of 0–0.4 m of the filter, and increased in depths of 0.4–0.8 m, then decreased to approximately 0.1 mg L–1 in depths of 1.5 m after 1 d (Figure 5(d)). The concentration of nitrite in the effluent was decreased to 0.045 and 0.02 mg L–1 after 2 and 3 d, respectively. Ammonia in depths of 0–0.8 m was obviously increased after 1 d (Figure 5(b)) and then quickly decreased. Mn(II) in depths of 0.8 m was 0.046, 0.159, 0.091 and 0.046 mg L–1 after 0, 1, 2 and 3 d, respectively (Figure 5(a)), while total iron was almost unchanged along the filter depth (Figure 5(c)). When ammonia in the influent was suddenly increased and nitrite was accumulated, biological Mn(II) removal was obviously affected, while Fe(II) removal was almost not affected. The reasons are as follows: Fe(II) was chemically and biologically removed in depths of 0–0.2 m where the nitrite was relatively low; while most of the Mn(II) was removed in depths of 0.2–0.8 m where the nitrite was high. When nitrite was generated in the filter, Mn(II) removal was affected in 3 days; however, when nitrite

Figure 3

Mn(II) (a) and nitrite (b) concentration proﬁles along depth of ﬁlter 2 for Mn(II)

was added to the ﬁlter, even the concentration of nitrite

feed concentration of approximately 2 mg L–1, and nitrite feed concentration of approximately 0.05, 0.1, 0.2 and 0.5 mg L–1, respectively, in aerobic

was much higher, Mn(II) removal was affected in only

conditions.

2 h. The reason was because the added nitrite was quickly Page 231

269

Figure 4

Q. Cheng et al.

Inﬂuence of nitrite on biological Mn(II) removal

Journal of Water Reuse and Desalination

07.3

2017

The variation of nitrite concentration profiles along depth of filter 2 for nitrite feed concentrations of approximately 0.05 (a) and 0.5 mg L–1 (c), respectively, in 1, 2 and 3 h after nitrite is added to the filter, respectively, and Mn(II) concentration profiles in 0, 1, 2 and 3 h ((b) nitrite was approximately 0.05 mg L–1) and ((d) nitrite was approximately 0.5 mg L–1), respectively.

Figure 5

The variation of Mn(II) (a), ammonia (b) and total iron (c) concentration profiles along depths of filter 1 in 0, 1, 2 and 3 d after ammonia increased from approximately 1.1 to 1.5 mg L–1, respectively, and nitrite (d) concentration profiles in 1, 2 and 3 d, respectively.

oxidized to nitrate by NOB, but the generated nitrite

and ammonia removal, the suitable inoculated bacteria

needed much longer to be completely oxidized. So in

were the biomass obtained from biological Mn(II) and

order to shorten the start-up period of bioﬁlter for Mn(II)

ammonia removal ﬁlter, because the presence of NOB

Page 232

270

Q. Cheng et al.

Journal of Water Reuse and Desalination

Inﬂuence of nitrite on biological Mn(II) removal

07.3

2017

could quickly oxidize nitrite to nitrate, and MnOB adapted to the nitrite presented conditions. The concentration profiles of ORP in simultaneous Mn(II) and ammonia removal When the concentration of Mn(II) and ammonia in the influent was approximately 2 and 1.5 mg L–1, respectively, ORP was 16 mV in influent, quickly increased to 75 mV in depths of 0.2 m of the filter, increased to 94 mV in depths of 0.4 m, and slowly increased to 122 mV in the effluent; Mn(II) and ammonia were decreased to 0.069 and 0.086 mg L–1 in depths of 0.1 m, respectively. Tekerlekopoulou & Vayenas () investigated the ORP profiles along the depth of the biofilters for Fe(II), Mn(II) and ammonia removal, and found that ORP increased along the filter depth from 150 to 600 mV, depending on the feeding concentrations. In their investigation, ORP was much higher than in filter 2, because DO in their filter was 7–8 mg L–1; however, DO in the effluent of filter 2 was lower than 1 mg L–1. Kinetics of biological manganese oxidation The removal kinetics of contaminants during water treatment

Figure 6

Mn(II) concentration proﬁles along depth of ﬁlter 2 for Mn(II) feed concentration of approximately 4 mg L–1 (a), linear regression analysis of Mn(II)

is considered an important issue, because it can provide infor-

depletion in relation with the empty bed contact time (b).

mation about the required time that the speciﬁc contaminant needs to be removed efﬁciently, which is necessary in sizing

results indicated that the ln{[M(II)]t/[Mn(II)o]} versus time

treatment units (Katsoyiannis & Zouboulis ). The con-

(EFCT) was linear. The value of k was 0.687 min–1 and the

centration of Mn(II) in groundwater in China was normally

half-life time for the depletion of Mn(II) was 1.010 min in the

lower than 3.5 mg L–1. In this experiment, Mn(II) in inﬂuent

pilot-scale bioﬁlter. The experiment was carried out at the

was 4.17 mg L

–1

and decreased to 0.069 and 0.000 mg L

–1

actual pH value of the groundwater, i.e. 7.0.

depths of 0.4 and 0.5 m of the ﬁlter (Figure 6(a)), respectively. From the obtained results, the kinetics of Mn(II) oxidation could be calculated by assuming that all the soluble Mn(II) was oxidized, and then removed by ﬁltration. By keeping DO constant and at a constant pH value, the Mn(II) depletion

CONCLUSIONS ΔG of the redox reaction between MnO2 and NO–2 in 298 (25 C) and 281 K (8 C) was calculated and the results W

rate would be ﬁrst order, i.e.

�

d½Mn(II)� ¼ K½Mn(II)� dt

suggested that MnO2 cannot react with NO–2. In the bioﬁlter, nitrite could not react with manganese oxides in anaerobic (2)

conditions. Biological Mn(II) removal was affected by nitrite, and the longer the nitrite was present in the bioﬁlter, the

A plot of ln{[M(II)]t/[Mn(II)o]}versus time (EFCT) would be

longer the Mn(II) removal was affected. In the start-up

linear if the kinetics of Mn(II) oxidation were indeed ﬁrst order,

period, the presence of nitrite in the bioﬁlter was the main

and the slope of such a plot would be –k-value (Figure 6(b)). The

reason for the start-up period of bioﬁlters for Mn(II) and Page 233

271

Q. Cheng et al.

Inﬂuence of nitrite on biological Mn(II) removal

ammonia removal being much longer than biofilters for Mn(II) removal. When Mn(II) and ammonia in influent were 2 and 1.5 mg L–1, respectively, ORP increased along the filter depth from 16 to 122 mV. Biological Mn(II) removal followed the first-order reaction, the k-value was 0.687 min–1 and the halflife time for the depletion of Mn(II) was 1.010 min.

ACKNOWLEDGEMENTS This work was kindly supported by the Scientiﬁc Research Foundation of CUIT (KYTZ201511) and the Program of Education Department of Sichuan Province (16ZB0221).

REFERENCES Akker, B., Holmes, M., Cromar, N. & Fallowfield, H.  Application of high rate nitrifying trickling filters for potable water treatment. Water Res. 42, 4514–4524. Azher, N. E., Gourich, B., Vial, C., Soulami, M. B., Bouzidi, A. & Ziyad, M.  Effect of alcohol addition in gas hold-up, liquid velocity and mass transfer in split-rectangular airlift bioreactor. Biochem. Eng. J. 23, 161–167. Azher, N. E., Gourich, B., Vial, C., Soulami, M. B. & Ziyad, M.  Study of ferrous iron oxidation in Morocco drinking water in an airlift reactor. Chem. Eng. Process. 47, 1877–1886. Ehrlieh, H. L. & Zapkin, M. A.  Manganese-rich layers in calcareous deposits along the Western shore of the Dead Sea may have a bacterial origin. Geomicrobiol J. 4 (2), 207–221. Frischherz, H., Zibuschka, F., Jung, H. & Zerobin, W.  Biological elimination of iron and manganese. Water Supply 3, 125–136. Gouzinis, A., Kosmidis, N., Vayenas, D. V. & Lyberatos, G.  Removal of Mn and simultaneous removal of NH3, Fe and Mn from potable water using a trickling filter. Water Res. 32, 2442–2450. Han, M., Zhao, Z. W., Gao, W. & Cui, F. Y.  Study on the factors affecting simultaneous removal of ammonia and manganese by pilot-scale biological aerated filter (BAF) for drinking water pre-treatment. Bioresour. Technol. 145, 17–24. Hasan, H. A., Abdullah, S. R. S., Kofli, N. T. & Kamarudin, S. K.  Effective microbes for simultaneous bio-oxidation of ammonia and manganese in biological aerated filter system. Bioresour. Technol. 124, 355–363. Hasan, H. A., Abdullah, S. R. S., Kamarudin, S. K. & Kofli, N. T.  On–off control of aeration time in the simultaneous removal of ammonia and manganese using a biological aerated filter system. Process Saf. Environ. Prot. 91, 415–422.

Journal of Water Reuse and Desalination

07.3

2017

Jusoh, A. B., Cheng, W. H., Low, W. M., Nora’aini, A., Megat, M. J. & Noor, M.  Study on the removal of iron and manganese in groundwater by granular activated carbon. Desalination 182, 347–353. Katsoyiannis, I. A. & Zouboulis, A. I.  Biological treatment of Mn(II) and Fe(II) containing groundwater: kinetic considerations and product characterization. Water Res. 38, 1922–1932. Kontari, N.  Groundwater, iron and manganese: an unwelcome trio. Water Eng. Manage. 135, 25–26. Nealson, K. H., Teho, B. M. & Rosson, R. A.  Occurrence and mechanisms of microbial oxidation of manganese. Adv. Appl. Microbiol. 33, 279–318. Nieuwenhuijsen, M. J., Toledano, M. B., Eaton, N. E., Fawell, J. & Elliott, P.  Chlorination disinfection byproducts in water and their association with adverse reproductive outcomes: a review. Occup. Environ. Med. 57, 73–85. Okoniewska, E., Lach, J., Kacprzak, M. & Neczaj, E.  The removal of manganese, iron and ammonium nitrogen on impregnated activated carbon. Desalination 206, 251–258. Pacini, A. V., Ingallinella, M. A. & Sanguinetti, G.  Removal of iron and manganese using biological roughing up flow filtration technology. Water Res. 39, 4463–4475. Richardson, S. D. & Postigo, C.  Chapter 4, Drinking water disinfection by-products. In: The Handbook of Environmental Chemistry: Emerging Organic Contaminants and Human Health, vol. 20. Springer, New York, pp. 93–137. Richardson, S. D., Plewa, M. J., Wagner, E. D., Schoeny, R. & DeMarini, D. M.  Occurrence, genotoxicity, and carcinogenicity of regulated and emerging disinfection byproducts in drinking water: a review and roadmap for research. Mutat. Res. Rev. Mutat. 636, 178–242. Rittmann, B. E. & Snoeyinck, V. L.  Achieving biologically stable drinking water. J. AWWA 76, 156–174. Sharma, S. K., Kappelhof, J., Groenendijk, M. & Schippers, J. C.  Comparison of physicochemical iron removal mechanisms in filters. J. Water Supply Res. Technol. Aqua 50, 187–198. Tekerlekopoulou, A. G. & Vayenas, D. V.  Ammonia, iron and manganese removal from potable water using trickling filters. Desalination 210, 225–235. Tekerlekopoulou, A. G. & Vayenas, D. V.  Simultaneous biological removal of ammonia, iron and manganese from potable water using a trickling filter. Biochem. Eng. J. 39, 215–220. Vandenabeele, J., Woestyne, M. V., Houwen, F., Germonpr, R., Vandesande, D. & Verstraete, W.  Role of autotrophic nitrifiers in biological manganese removal from groundwater containing manganese and ammonium. Microb. Ecol. 29, 83–98. Vayenas, D. V., Pavlou, S. & Lyberatos, G.  Development of a dynamic model describing nitritification and nitratification trickling filters. Water. Res. 31, 1135–1147. WHO  Ammonia in Drinking Water. Guidelines for DrinkingWater Quality. Health Criteria and Other Supporting Information, vol. 2, 2nd edn. World Health Organization, Geneva.

First received 12 December 2015; accepted in revised form 8 April 2016. Available online 13 May 2016

Page 234

Journal of

Water, Sanitation & Hygiene for Development

ISSN 2043-9083 iwaponline.com/washdev

The Journal of Water, Sanitation and Hygiene for Development covers the science, policy and practice of drinking-water supply, sanitation and hygiene at local, national and international levels. The journal’s scope includes: • Water supply: intermittent supply, community and utility water supplies, water treatment, distribution, storage • Sanitation: collection, transport, treatment, use, discharge, on-site and off-site sanitation, resources recovery • Hygiene: behaviours, education, change • Technical and managerial issues: characteristics of and constraints to conventional and innovative approaches, technical options and boundaries of technical application, emerging issues, emergencies and disasters, impacts on health, poverty and development, sustainability, demand, marketing, organizing supply chains • Institutional development: roles of public and private sector, capacity building, governance, education and training • Financing and economic analysis: cost-effectiveness and cost-benefits, role and impact of subsidies, user fees, financial instruments, innovations in financing • Policy: aspects/developments in the role of national policy on service provision, human rights and rights-based approaches policy, developing appropriate and scaleable legal and regulatory approaches, norms and standards • International policy: aid and aid effectiveness; international targets, conventions and agreements, policy For more details, visit iwaponline.com/washdev

Page 238

Review Paper

07.1

2017

Review Paper The elimination of open defecation and its adverse health effects: a moral imperative for governments and development professionals Duncan Mara

ABSTRACT In 2015 there were 965 million people in the world forced to practise open defecation (OD). The adverse health effects of OD are many: acute effects include infectious intestinal diseases, including diarrheal diseases which are exacerbated by poor water supplies, sanitation and hygiene; adverse pregnancy outcomes; and life-threatening violence against women and girls. Chronic effects include soil-transmitted helminthiases, increased anaemia, giardiasis, environmental enteropathy and small-

Duncan Mara Emeritus Professor of Civil Engineering, Institute for Public Health and Environmental Engineering, School of Civil Engineering, University of Leeds, Leeds LS2 9JT, UK E-mail: d.d.mara@leeds.ac.uk

intestine bacterial overgrowth, and stunting and long-term impaired cognition. If OD elimination by 2030 is to be accelerated, then a clear understanding is needed of what prevents and what drives the transition from OD to using a latrine. Sanitation marketing, behaviour change communication, and ‘enhanced’ community-led total sanitation (‘CLTS þ ’), supplemented by ‘nudging’, are the three most

likely joint strategies to enable communities, both rural and periurban, to become completely

OD-free and remain so. It will be a major Sanitation Challenge to achieve the elimination of OD by 2030, but helping the poorest currently plagued by OD and its serious adverse health effects should be our principal task as we seek to achieve the sanitation target of the Sustainable Development Goals – indeed it is a moral imperative for all governments and development professionals. Key words

| child health, diarrhea, environmental enteropathy, impaired cognition, open defecation, stunting

INTRODUCTION In 2015 965 million people had no sanitation facility and

only 2% of the richest quintile (Figure 2). However, in

were therefore forced to defecate in the open (WHO/

low-income urban areas the number of open defecators

UNICEF ) (Figure 1). The average proportion of ‘open

can also be very high: for example, in India Gupta et al.

defecators’ in developing countries is 16%, and in the

() found that 35–47% of poor households in Delhi,

least-developed countries 20%. Table 1 lists those countries

Indore, Meerut and Nagpur did not have any toilet facility.

with more than 15% open defecators and highlights those

Part of the sanitation target of the Sustainable Development

with more than 50%. Most of these open defecators are

Goals is to eliminate open defecation (OD) by 2030 (United

poor and live in rural areas – for example, in India, which

Nations General Assembly ). If the same proportion of

had a total of 564 million open defecators in 2015, 61% of

‘open defecators’ to the total without improved sanitation in

the rural population were open defecators vs only 10% of

2015 (965 million to 2.4 billion, i.e. 42%) is assumed for 2030,

the urban population (WHO/UNICEF ), and 95% of

then 42% of the 2016–2030 population increase of 1.1 billion

the poorest quintile in rural areas were open defecators vs

(UNDESA ), plus the current number of open defecators,

doi: 10.2166/washdev.2017.027

Page 239

D. Mara

Journal of Water, Sanitation and Hygiene for Development

The elimination of open defecation

07.1

2017

i.e. 1.069 billion people. Subtracting from this the 965 million open defecators in 2015 gives the number of people removed from OD during the 5-year period 2011– 2015, i.e. 104 million, equivalent to 57,000 people per day. This is better than that achieved during 1991–2015, but it is still far short, by a factor of 4, of the requirement for 2030. However, some countries have done very well in reducing OD: for example, in rural Vietnam 43% of the population practised OD in 1990, but by 2015 this had been reduced to 1%; in Bangladesh the corresponding ﬁgures were 40 and 2%; and in Mexico they were 51 and 4% (WHO/UNICEF ). Given that there are ‘no solutions Figure 1

OD by a young boy in periurban India (photograph courtesy of Professor Barbara Evans, University of Leeds).

are required to move from OD to ﬁxed-point defecation, prefer-

without political solutions’, the exceptionally good progress in these and some other countries may have been due, at least in part, to their politicians and senior civil servants

ably (in the new terminology of JMP b) to ‘basic’ sanitation

‘thinking clean’, i.e. deciding that OD was not ‘clean’ and

or ideally ‘safely-managed’ sanitation, i.e. a total of nearly 1.4 bil-

that therefore something had to be done to reduce or elimin-

lion people, or some 260,000 per day during 2016–2030.

ate it, and then transposing this decision into action.

In 1990, 31% of the then developing-country population of

At the current rate of global progress, the target of no

4.1 billion were open defecators, and in 2015 16% of the then

OD by 2030 is unlikely to be realised. Thus to achieve the

developing-country population of 6 billion were open defeca-

SDG target of ‘No OD by 2030’ requires a huge global

tors, i.e. 1.29 billion and 965 million, respectively (WHO/

step-change in addressing and reducing to zero the preva-

UNICEF ). Thus, during the whole of the 25-year period

lence of OD in developing countries. To do this, Ministry

1991–2015 there was a reduction in OD of 325 million

of Health ofﬁcials and development professionals need to

people, equivalent to only 36,000 per day; this was due in

be fully aware of the major adverse health consequences

part to the large population increase during this period.

of OD, and how best to eliminate OD – in particular, what

In 2010, 19% of the then developing-country population of 5.6 billion were open defecators (WHO/UNICEF ),

Table 1

mix of sanitation ‘hardware’, social-science ‘software’, and ﬁnancial support is appropriate.

Countries with more than 15% and more than 50% of their populations practising OD in 2015 (WHO/UNICEF 2015)

Region

Countries with >15% ODa and percentages of populations practising ODb,c

Africa

Angola (30%), Benin (53%), Burkina Faso (55%), Cabo Verde (24%), Central African Republic (22%), Chad (64%), Côte d’Ivoire (26%), Djibouti (20%), Eritrea (77%), Ethiopia (29%), Ghana (15%), Guinea (22%), Guinea-Bissau (17%), Lesotho (33%), Liberia (48%), Madagascar (40%), Mauritania (35%), Mozambique (39%), Namibia (48%), Niger (73%), Nigeria (25%), São Tome e Principe (54%), Sierra Leone (24%), South Sudan (74%), Togo (52%), Zimbabwe (28%)

Asia Paciﬁc

Cambodia (47%), India (44%), Indonesia (20%), Kiribati (36%), Laos (33%), Nepal (32%), Solomon Islands (54%), Timor-Leste (26%)

Latin America & Caribbean

Bolivia (17%), Haiti (19%)

The average 2015 OD rate for developing countries was 16%, and for the least developed countries 20%.

b c

Some countries with high OD rates in 1990 reported in WHO/UNICEF (2015) have no reported OD rates for 2015 (and are thus excluded from this table). Countries with >50% open defecators are shown in bold.

Page 240

D. Mara

Journal of Water, Sanitation and Hygiene for Development

The elimination of open defecation

07.1

2017

losses) (IHME ). The World is not good at handwashing: Freeman et al. () estimated that globally 81% of people do not practise safe handwashing. A further acute health effect of OD is adverse pregnancy outcomes, such as increases in low birth weights, preterm births, stillbirths, and spontaneous abortions (Padhi et al. ). Finally, there is violence against women and girls, which is often life-threatening. Violence against women and girls of all ages in LICs and LMICs caused a DALY loss of 7.8 Figure 2

Percentage of rural population in India practising OD, by wealth quintile ( JMP 2015a).

million years in 2015 (IHME ). Physical violence, which may include murder, rape, stabbing and other bodily harm, is a not uncommon experience for women

ADVERSE HEALTH EFFECTS OF OD

and girls as they journey to a place of OD, especially at night (Gómez et al. ). Bhalla () reported the occur-

The adverse health effects of OD can be divided into acute

rence of two ‘open-defecation murders’ in rural India:

effects and chronic effects. Both cause a high burden of disease and a large number of premature deaths, especially in

‘The two [girl] cousins, who were from a low-caste Dalit

children under ﬁve years of age. These adverse health effects

community and aged 14 and 15, went missing from

of OD occur because OD results in massive faecal contami-

their village home in Uttar Pradesh’s Budaun district

nation of the local environment; consequently, open

when they went out to go to the toilet [in a neighbouring

defecators are repeatedly exposed to faecal bacteria and

ﬁeld]. The following morning, villagers found the bodies

faecal pathogens, and this is particularly serious for young

of the two teenagers hanging from a mango tree in a

children whose immune systems and brains are not yet

nearby orchard.’

fully developed.

It transpired that the two girls had been attacked and gang-

Acute health effects of OD

raped by ﬁve local men before they were hanged. Unfortu-

The principal acute adverse health effect of OD is infectious

et al. () reported that many women in Bhopal and

excreta-related intestinal disease, of which diarrheal diseases (DD) are the most common. DD were the third cause of death in children under ﬁve years of age (U5) in 2015 in low-income and lower-middle-income countries (LICs and LMICs), resulting in 499,000 deaths (8.6% of all U5-deaths), and a disability-adjusted life year (DALY) loss

nately, such incidents are not at all uncommon: Gosling Delhi, India, and Kampala, Uganda experienced violence and harassment on a daily basis. Such violence may often induce longer-term psychological damage. To help counter such violence House et al. () have prepared a practitioner’s toolkit on ‘Violence, Gender and WASH’.

of 45.1 million years (8.5% of total U5-DALY losses) (IHME ). One of the commonly ascribed reasons for

Chronic health effects of OD

high incidences of DD is a poor water supply, poor sanitation, and poor hygiene, especially poor hand-hygiene

There are ﬁve principal widespread chronic health effects

(WHO ). The burden of U5-disease in LICs and

most probably due to OD: soil-transmitted helminthiases

LMICs in 2015 due to no handwashing-with-soap was a

(STHs),

DALY loss of 26.4 million years (5.7% of total U5-DALY

enteropathy

losses); the corresponding ﬁgure for unsafe sanitation was

(SIBO), and stunting (low height-for-age) with accompany-

a DALY loss of 26.6 million years (5.7% of total U5-DALY

ing impaired cognition.

increased and

anaemia,

giardiasis,

small-intestine

environmental

bacterial

overgrowth

Page 241

D. Mara

Journal of Water, Sanitation and Hygiene for Development

The elimination of open defecation

Soil-transmitted helminthiases

07.1

2017

such as school attendance (Coffey & Geruso ). Irondeﬁciency anaemia caused an all-age both-sex DALY loss

The most common STHs are ascariasis (caused by the

in LICs and LMICs of 36.1 million years in 2015 (IHME

human roundworm, Ascaris lumbricoides), trichuriasis

). In a study on anaemia in Nepal, Coffey & Geruso

(caused by the human whipworm, Trichuris trichiura), and

() found that ‘poor local sanitation and, speciﬁcally,

human hookworm disease (caused by Ancylostoma duode-

OD cause lower hemoglobin and higher rates of anemia in

nale and Necator americanus). Globally, an estimated 439

children’.

million people were infected with hookworm in 2010, 819 million with A. lumbricoides and 465 million with T. tri-

Giardiasis

chiura (Pullan et al. ). The burdens of disease associated with these STHs are high: in 2015 ascariasis in

The long-term post-infection consequences of giardiasis

LICs and LMICs caused an all-age both-sex DALY loss of

include low height-for-age, low weight-for age, small mid-

878,000 years, trichuriasis 340,000 years, and human hook-

upper-arm-circumference-for-age, low serum-levels of zinc

worm disease 2.2 million years (IHME ).

and iron, chronic and persistent diarrhea with consequent

Ascariasis, trichuriasis and hookworm disease cause impaired

cognition,

notably

school-aged

children

malabsorption, irritable bowel syndrome deﬁciencies, and impaired cognition (Halliez & Buret ).

(Nokes et al. ; Partovi et al. ; Spears & Haddad ). The areas most affected are verbal ﬂuency, short-

Environmental enteropathy and SIBO

term memory, and speed of information processing, which are precisely the areas most needed for people to be able

There has been considerable research on the association

to contribute effectively to socio-economic development.

between stunting (see ‘Stunting’ below) and environmental

Infection with two or more of these helminths impairs cog-

enteropathy (also called tropical enteropathy and environ-

nition to a greater extent than infection with only one

mental enteric dysfunction). Environmental enteropathy is

( Jardim-Botelho et al. ).

a condition which results in the malabsorption of nutrients

Trichuriasis is associated with ‘anaemia (see “Increased

in the small intestine and this leads to stunting; some or

anaemia” below), growth retardation (i.e. stunting – see

many of the nutrients in a child’s foods are not absorbed

“Environmental enteropathy and SIBO” below) and intesti-

and so are unavailable for the child’s growth. The term

nal leakiness’ (Cooper et al. ). In a study of 9,860

‘environmental enteropathy’ was used by Fagundes-Neto

refugees in Texas, latent tuberculosis infection was found

et al. () to describe a common syndrome in which

to be positively associated in those refugees with hookworm

there are non-speciﬁc histopathological and functional

infection (Board & Suzuki ).

changes of the small intestine in children of poor families

The World Health Organization has a global target to

living in conditions lacking basic sanitary facilities and

eliminate morbidity due to STHs in preschool and school-

chronically exposed to faecal contamination. They studied

age children by 2020 (WHO ). This is to be achieved

112 children and found that carbohydrate load tests

by regularly treating (deworming at school) at least 75% of

revealed 49% lactose malabsorption, 30% sucrose malab-

the children in endemic areas – an estimated 873 million

sorption and 5% glucose malabsorption, and that small

children.

bowel biopsy showed partial villous atrophy in 94% of the samples studied.

Increased anaemia

More recent research has conﬁrmed these ﬁndings. Humphrey () reported that a key cause of child under-

In adults, anaemia reduces productivity and is associated

nutrition was environmental enteropathy, and that this

with higher maternal mortality; in children, it impairs phys-

enteropathy is caused by faecal bacteria ingested in large

ical and cognitive development directly, and it also affects

quantities by young children living in conditions of poor

human capital accumulation via impacts on behaviours

sanitation and hygiene. She postulated that provision of

Page 242

D. Mara

Journal of Water, Sanitation and Hygiene for Development

The elimination of open defecation

toilets and promotion of handwashing after faecal contact could reduce or prevent environmental enteropathy and its

07.1

2017

example, a z score of �2 means that a child’s height is two

standard deviations below the median height for that

adverse effects on growth; and she noted that prevention

child’s age and sex, and the child is therefore considered

of this enteropathy, which afﬂicts almost all children in

stunted; for severe stunting the z score is �3 or lower.) In

the developing world, will be crucial to normalise child

developing countries as a whole stunting is decreasing –

growth, and that this will not be possible without the pro-

from 251 million children under ﬁve in 1990 to 156 million

vision of toilets. Mbuya & Humphrey () endorsed this

children in 2014, except in Africa where it is increasing –

by stating that the unhygienic environments in which infants

from 47 million children in 1990 to 58 million in 2014

and young children live and grow must contribute to, if not

(UNICEF ). Stunting affects poor children much more

be the overriding cause of, this environmental enteric dys-

than children from rich families: for example, in least devel-

function. They suggested that a household-level package of

oped countries, 49% of the poorest children are stunted vs

‘baby-WASH’ interventions (sanitation and water improve-

26% of the richest children; boys are more stunted than

ment, handwashing with soap, ensuring a clean play and

girls (43 vs 38%), and children living in rural areas are

infant-feeding environment, and food hygiene) that inter-

more stunted than those in urban areas (43 vs 32%)

rupted

speciﬁc

pathways

through

which

feco-oral

transmission occurs in the ﬁrst two years of a child’s life may be central to global stunting-reduction efforts. Donowitz & Petri () found that:

(UNICEF ). In 2015 stunting caused a U5-DALY loss in LICs and LMICs of 21.4 million years (IHME ). Stunting is exacerbated by (a) the density of OD – the number of people practising OD per km2 (Spears ); (b) environmental enteropathy and SIBO (see ‘Environmental

‘Small-intestine bacterial overgrowth (SIBO) occurs

enteropathy and SIBO’ above); and (c) DD and STHs (see

when colonic quantities of commensal bacteria are pre-

‘Soil-transmitted helminthiases’ above) (Spears & Haddad

sent in the small bowel. SIBO is associated with

). In a 10-year study of 119 slum children in northeast

conditions of disrupted gastrointestinal (GI) motility

Brazil, Moore et al. () found that children who had

leading to stasis of luminal contents. Recent data show

had a high burden (∼9 episodes) of DD in their ﬁrst two

that SIBO is also found in children living in unsanitary

years of life were on average 3.6 cm shorter at age seven

conditions who do not have access to clean water.

than other children, and those children who had also had

SIBO leads to impaired micronutrient absorption and

an early childhood helminthiasis were on average a further

increased GI permeability, both of which may contribute

4.6 cm shorter at the same age. In a study of children living

to growth stunting in children.’

in a periurban shanty town in Lima, Peru, Berkman et al. () found that:

Stunting ‘During the ﬁrst two years of life, 46 (32%) of 143 children Target #2.2 of the Sustainable Development Goals includes

were stunted. Children with severe stunting in the second

‘achieving, by 2025, the internationally agreed targets on

year of life scored 10 points lower on the WISC-R [‘Wechs-

stunting and wasting in children under ﬁve years of age’

ler Intelligence Scales for Children – Revised’ (Wechsler

(United Nations General Assembly ). The ‘internation-

)] test at age nine than children without severe stunting

ally agreed target’ for stunting is to reduce by 2025 the

[in their second year of life]. Children with more than one

number of stunted children under the age of 5 in 2010 by

episode of Giardia lamblia per year scored 4.1 points

40% (de Onis et al. ). Stunting is deﬁned as a height

lower than children with one episode or fewer per year.

that is two or more standard deviations below the median

Neither

height for the child’s age and sex. (The World Health Organ-

parvum infection was associated with WISC-R scores’.

diarrhea

prevalence

nor

Cryptosporidium

ization publishes charts and tables for boys’ and girls’ median heights-for-age and values of the appropriate stan-

Eppig et al. (), in their study on the prevalence of infec-

dard deviations (WHO ). A ‘z score’ is used: for

tious-disease agents and cognitive ability, postulated that the Page 243

D. Mara

The elimination of open defecation

Journal of Water, Sanitation and Hygiene for Development

07.1

2017

bodies of young children face a competition for energy

sanitation was the second highest risk factor for stunting, with

(derived from their nutrient intake) between the develop-

7.2 million attributable cases (out of a total of 44.1 million

ment and use of their brain and the development and use

cases – i.e. 16%); the highest risk factor was foetal growth

of their immune system. Children repeatedly exposed to

restriction (10.8 million attributable cases), and the third

infectious-disease agents are seriously disadvantaged:

highest was DD (5.8 million attributable cases). In summary: (a) OD → violence against women and girls

‘[They] must activate [their] immune system to ﬁght off

as they walk to OD sites, including murder, rape, stabbing,

the infection, at energetic expense. Of these, diarrheal

other serious bodily harm, and any resulting longer-term

diseases may impose the most serious cost on their

psychological/psychosocial damage; and (b) high OD den-

hosts’ energy budget. First, diarrheal diseases are the

sity → extreme

most common category of disease on every continent,

environment → frequent ingestion of large numbers of faecal

[…] Second, diarrhea can prevent the body from acces-

bacteria and faecal pathogens, and frequent percutaneous

faecal

contamination

the

local

sing any nutrients at all. If exposed to diarrheal

entry of hookworm larvae, by young children → high inci-

diseases during their ﬁrst ﬁve years, individuals may

dence of infectious intestinal disease and helminthiases, and

experience lifelong detrimental effects to their brain

mass development of SIBO and environmental enteropathy →

development, and thus intelligence’.

high levels of nutrient malabsorption and childhood stunting, and all the cognitive and physical consequences thereof.

To this ‘brain’ scenario can be added stunting: the more nutrients children do not get through exposure to infectious-disease agents or, in the reasoning of environmental

SOCIAL PREFERENCE FOR OD

enteropathy given above, through continuous exposure to faecal bacteria, the more they will be stunted.

Despite these associated adverse health outcomes, OD is

The long-term consequences of childhood stunting include

often a preferred practice, notably in rural India, where

adverse effects on cognitive development, school achievement,

61% of the population are open defecators (WHO/

economic productivity in adulthood, and maternal reproduc-

UNICEF ), Coffey et al. () found robust evidence

tive outcomes (Dewey & Begum ). Adverse ‘maternal

that supported a preference for OD, with many respondents

reproductive outcomes’ include not only adverse neonatal

in their survey in rural India claiming that OD was more

and infant outcomes, but also chronic diseases in adulthood

pleasurable and desirable than latrine use. Devine & Kull-

for the surviving children in their later life – for example,

mann () found that in rural East Java, Indonesia, many

increased cardiovascular disease, high blood pressure, respirat-

men considered OD ‘normal’, and that it had distinct

ory diseases, and Paget’s disease (Barker ).

beneﬁts such as social interaction and physical comfort

Hoddinott et al. () make the economic case for redu-

(especially in the case of defecation in a river). Tiwaril

cing stunting. Using ‘credible estimates of beneﬁt-cost ratios

() reported that in rural Uttar Pradesh, India, because

(BCRs) for a plausible set of nutritional interventions to

they were used to the ‘comfortable ﬁelds’, 90 families quietly

reduce stunting’, they found that in 17 high-burden countries

demolished the toilets inside their house that were built

these BCRs ranged from 3.6 (Democratic Republic of the

under the Swachh Bharat Abhiyaan (see below), as they pre-

Congo) to 48 (Indonesia), with a median value of 18 (Bangla-

ferred to resume OD.

desh). Thus reducing stunting is a very good economic

Figure 2 shows that even some of the two richest wealth

proposition, and so investment in sanitation to reduce stunt-

quintiles in India practise OD, presumably because they

ing is also a very good economic proposition (Augsburg

prefer this to using a toilet (which they could easily afford).

et al. ). The importance of this has been conﬁrmed by

Of course, in other countries where OD is common

Danaei et al. (), who studied the risk factors for childhood

(Table 1), a social preference for OD may not exist. People

stunting at age two in 137 developing countries. They found

in these countries may be practising OD because they

that 36% of two-year olds were stunted, and that unimproved

cannot afford a latrine (Augsburg et al. ), or because, if

Page 244

D. Mara

The elimination of open defecation

Journal of Water, Sanitation and Hygiene for Development

07.1

2017

they live in urban slums, there is no space available to con-

‘An innovative methodology for mobilising communities

struct latrines.

to completely eliminate open defecation (OD). Communities are facilitated to conduct their own appraisal and analysis of open defecation and take their own action

SWACHH BHARAT ABHIYAAN – ‘CLEAN INDIA MISSION’ In his 2014 Independence Day speech, the Prime Minister of India, Shri Narendra Modi, spoke about OD and the need for toilets (Modi a):

to become open-defecation free (ODF).’ In Bangladesh, the success in reducing rural OD from 40% in 1990 to 2% in 2015 (WHO/UNICEF ), and to <1% in 2016 (Ministry of Local Government Rural Development and Co-operatives ), has long been ascribed to properlydesigned and well-executed CLTS (Sanan & Moulik ).

‘Has it ever pained us that our mothers and sisters have

Further information on CLTS and the elimination of OD

to defecate in open? Whether dignity of women is not

is given by Kar & Chambers () and Bongartz et al.

our collective responsibility? The poor womenfolk of

(). Importantly, CLTS does not prescribe the adoption

the village wait for the night; until darkness descends,

of any one particular sanitation technology; thus all appro-

they can’t go out to defecate. What bodily torture they

priate sanitation options should be considered with the

must be feeling, how many diseases that act might engen-

beneﬁciary communities, recognising that the available tech-

der. Can’t we just make arrangements for toilets for the

nical options are likely to be different in urban and rural

dignity of our mothers and sisters?’

areas. WSP/MDWS () details some of the best practices in rural sanitation in India.

On 2 October 2014 Prime Minister Modi launched ‘Swachh Bharat Abhiyaan’ (SBA, ‘Clean India Mission’), one objective of which is to end OD by 2 October 2019, the 150th anniver-

ACCELERATING THE ELIMINATION OF OD

sary of Mahatma Gandhi’s birth (Modi b). This is clearly a very ambitious ﬁve-year target, given that India has 565

If progress towards OD elimination is to be accelerated,

million open defecators; this is the largest country-number

then a clear understanding of what prevents and what

in the world (by over an order of magnitude) and represents

drives the transition from OD to using a latrine is necessary.

54% of all open defecators (WHO/UNICEF ).

Augsburg et al. () found that cost was the principal con-

SBA followed on from the Total Sanitation Campaign

sideration that militated against latrine adoption in both

(TSC) instituted in 1999. A review of TSC by WaterAid India

India and Nigeria; this indicates that subsidies and access

() found much variability in results from state to state,

to credit (e.g. subsidized microﬁnance loans) are clearly

especially in states where the approach was centralized,

important (see, for example, Evans et al. ; Newman

rather than being decentralized to the community level.

et al. ).

Menon () criticized SBA for this reason, stating that sub-

Augsburg & Rodríguez-Lesmes (), working in low-

sidy-driven Swachh Bharat was a failed, old idea, and that a

income urban areas and slums and rural areas in India,

community-driven approach was needed to stop OD. This is

found that there was a strong correlation of toilet ownership

in agreement with WaterAid India’s () ﬁnding that commu-

with perceived health, with households that owned a toilet

nity-led total sanitation (CLTS) could be one of the approaches

believing themselves and their family to be healthier than

explored for faster and more sustainable results on the ground.

their peers who did not – thus suggesting that, contrary to often held views, health considerations play at least some role in the decision to acquire sanitation.

THE CLTS APPROACH TO ENDING OD

Village-wide and slum-wide elimination of OD depends for its success on: (1) the selection and community-wide

IDS () describes CLTS as:

installation, both with the participation of the beneﬁciary Page 245

D. Mara

The elimination of open defecation

Journal of Water, Sanitation and Hygiene for Development

07.1

2017

community, of a locally-suitable sanitation technology,

demand for improved sanitation facilities. While forma-

which the local community understands and agrees to use

tive research is the foundation of any sanitation

sustainably; and (2) the selection, installation (again with

marketing program, essential to understanding what pro-

community participation) and correct use of a locally appro-

ducts the target population desires and what price they’re

priate handwashing-with-soap facility.

willing to pay for them, components such as the market-

It is very important that the whole community becomes

ing mix, communications campaign, and implementation

‘open defecation free’ (ODF). Andrés et al. (), in a study

are also critical to the design and implementation of

involving 209,762 children under the age of four in rural

effective program.’

India, which investigated the potential beneﬁts, in terms of a reduction in diarrhea, to children living in households

Devine & Kullmann () recommend CLTS and behaviour

with ‘improved’ sanitation facilities, found that there was

change communication (BCC) as useful adjuncts to SM

no improvement at all until 30% coverage was achieved

because, while CLTS focuses on changing community prac-

(i.e. 30% of all households in the village community

tices, BCC focuses on changing individual or household

having their own improved sanitation facility), and that

behaviours. Thus BCC can be used to sustain and sup-

half of the potential beneﬁts were only reached when cover-

plement CLTS in motivating individuals to become open-

age was approximately 75%. Vyas et al. () found a

defecation-free and sustain this behaviour over time. Perez

similar relationship between stunting and ODF status in

() reported on research carried out in Bangladesh

rural Cambodia: children living in completely ODF villages

which examined the long-term sustainability of sanitation

had z-scores above �1.5 during the whole of their ﬁrst ﬁve

behaviours and facilities in areas that were declared ODF;

years of life, whereas those living in villages where everyone

one of the main ﬁndings was that the BCC campaign

practised OD had z-scores below �2 from age 20 months

directed at households to stop practising OD was very perva-

onwards; those children living in villages where some

sive: campaign messages were communicated through

people practised OD had z-scores close to �2 from age

various channels and settings, including messaging by mem-

two onwards. Such externalities (external, that is, to each

bers and ofﬁcers of the local Union Parishad (the smallest

individual household) reﬂect the relative importance of

rural administrative unit) at meetings, rallies, over loudspea-

faeco-oral disease transmission in the ‘public’ and ‘private’

ker announcements, and through household visits by Union

domains, as discussed by Cairncross et al. (). In order

Parishad members or NGO workers.

to interrupt transmission, interventions are needed in both the private domain (individual household-level improved sanitation) and in the public domain (all of one’s co-villagers

ODFþ and CLTS þ

having their own improved sanitation facility). CLTS seeks

There is currently a move, at least in thinking, from ODF to

to establish a social norm for eliminating OD in the whole

‘ODF þ ’ – that is, to develop sound models to ensure that,

community such that it, as a unit, realises all the disadvantages of OD (especially those for women and girls), so that every household in the community has and uses a safelymanaged latrine.

once ODF status has been achieved, it is sustained for all time, and how CLTS might be modiﬁed (and perhaps described as ‘CLTS þ ’) to encourage this to happen, includ-

ing such topics as locally correct latrine selection, latrine ﬁnancing and possible subsidies, sufﬁcient water supplies

Sanitation marketing and behaviour change

for personal and domestic hygiene (handwashing with

communication

soap, and cleansing used cooking and eating utensils), and household- and community-level operation and mainten-

WSP () deﬁnes sanitation marketing (SM) as:

ance (Bongartz et al. ). ‘Nudging theory’ has been recommended as a means to change OD practice to

‘An emerging ﬁeld that applies social and commercial

ODFþ (Neal et al. ) – ‘nudges’ are small changes to

marketing approaches to scale up the supply and

the mental environment that can channel decision-making

Page 246

D. Mara

The elimination of open defecation

Journal of Water, Sanitation and Hygiene for Development

07.1

2017

and behaviour in new ways. Nudging is based on scientiﬁc

urban slums, which are home to some 881 million people

ﬁndings from psychology, cognitive science and behavioural

(30% of the urban population in developing countries, up

economics, on which Neal et al. () proposed a frame-

to 56% in Sub-Saharan Africa) (UN-Habitat ), house-

work of eight principles to support the initiation and

hold-level sanitation is infeasible due to space constraints.

maintenance of OD behaviour change: (1) ensure critical

Safely-managed shared sanitation is, however, a feasible

sanitation products and infrastructure are immediately and

and tested sanitation option to replace OD in low-income

consistently physically available for the users; (2) create or

high-density urban areas (Burra et al. ; Mara ).

capitalize on context change to drive new behaviour of

In addition, there is a need in CLTSþ for local

toilet use; (3) piggyback on other existing behaviours and

businesses and tradesmen to be trained in latrine selection,

cues (e.g. washing clothes, water gathering); (4) strategically

construction, and ﬁnancing, and also, where appropriate,

increase friction for the undesired behaviour (OD) and

the provision of locally-produced and locally-suitable pour-

lessen it for the desired one (sustained toilet use); (5) support

ﬂush squat-pans or pedestal-seat units (Sy et al. ), hard-

context-stable repetition for latrine use; (6) embed ritualized

ware for urine-diverting eThekwini latrines, pipework and

elements in the change process (e.g. integrate OD messaging

accessories for condominial sewerage, and also facilities

into already ritualized cultural practices); (7) leverage point-

for handwashing with soap ( Jenkins et al. ).

of-action reminders and cues (e.g. use of coloured agents to clean latrine slabs); and (8) highlight descriptive and localized norms that reduce cognitive demands (e.g. develop

CONCLUDING REMARKS

systems to address the whole community or a women’s group, rather than individual households). CLTS þ , supplemented with ‘nudging’, would enable

1. This paper has sought to review and collate key evidence on OD, especially the numbers of people practising OD,

rural households to move directly from OD to ‘safely-mana-

the health effects of OD, and how best OD might be

ged’ on-site sanitation and hygiene – which is the SDG

eliminated.

target ( JMP b). The technologies for safely-managed

2. The adverse health consequences of OD are so extreme

on-site sanitation are well established – for example, arbor-

that, if ODFþ status in not reached in rural villages, small

loos (which are especially suitable in low-density rural

towns and low-income periurban areas, including slums,

areas; fruit or medicinal trees are planted in the shallow

there will be more ‘lost generations’ of physically-impaired

pits when full to provide food and income) (Morgan ),

and cognitively-challenged children and adults. All Minis-

single-pit VIP latrines, urine-diverting eThekwini latrines

try of Health ofﬁcials and development professionals

(which, because they are wholly above-ground, are suitable

need to be aware of the physical and mental outcomes of

in areas subject to ﬂooding or with high groundwater

OD in young children, some of which are irreversible.

tables and where pit emptying is difﬁcult or not well prac-

3. The elimination of OD is primarily a complex sociocul-

tised) (WIN-SA ), and single-pit or alternating twin-pit

tural and sociopolitical task. It is not a major technical

pour-ﬂush latrines.

or ﬁnancial challenge as CLTS, with its option to con-

In low-income urban areas it is more difﬁcult to move to

sider all types of sanitation and handwashing facilities,

safely-managed sanitation as faecal-sludge management is

does not require the development of new technologies

more complex and more expensive than in rural areas. How-

speciﬁcally for OD elimination as several existing tech-

ever, safely-managed sanitation can be readily achieved with

nologies are already ﬁt-for-purpose; nor does it always

off-site systems such as condominial sewerage (Melo ,

necessitate the provision of subsidies. The further devel-

); household ﬁnancial costs for this sanitation system

opment and rigorous ﬁeld-testing of ‘CLTS þ ’ is needed

are low – for example, in the state of Rio Grande do Norte

to ensure that there is no reversion to OD in communities

in Brazil (where the system was developed in the early

which have become OD-free.

1980s) the monthly charge is only BRL 2.18 (GBP 0.50,

4. SM and BCC are very valuable techniques and should be

USD 0.63) per household per month (CAERN ). In

applied as the ﬁrst steps in CLTS/CLTSþ – i.e. these Page 247

D. Mara

The elimination of open defecation

three techniques should be used in sequence for best results. 5. It will be a major sanitation challenge to achieve the elimination of OD by 2030, but it is a challenge that governments and development professionals should stand up to and embrace. Helping the poorest plagued by OD should be our principal task as we all seek to achieve the sanitation target of the Sustainable Development Goals – indeed it is our moral imperative.

REFERENCES Andrés, L. A., Briceño, B., Chase, C. & Echenique, J. A.  Sanitation and Externalities: Evidence from Early Childhood Health in Rural India (Policy Research Working Paper No. 6737). World Bank, Washington, DC. Augsburg, B. & Rodríguez-Lesmes, P.  Sanitation Dynamics: Toilet Acquisition and Its Economic and Social Implications (IFS Working Paper W15/15). Institute for Fiscal Studies, London. Augsburg, B., Caeyers, B. & Oteiza, F.  The Costs and Beneﬁts of Investing in A Toilet: Views From Indian and Nigerian Households and Their Policy Implications. Institute for Fiscal Studies, London. Barker, D. J. P.  Mothers, Babies and Disease in Later Life. BMJ Publishing Group, London. Berkman, D. S., Lescano, A. G., Gilman, R. H., Lopez, S. L. & Black, M. M.  Effects of stunting, diarrhoeal disease, and parasitic infection during infancy on cognition in late childhood: a follow-up study. Lancet 359, 564–571. Bhalla, N.  Teen Girls Gang-Raped and Hanged From A Tree – Police. Thomson Reuters Foundation, New Delhi. Available from: http://in.reuters.com/article/uk-india-killingsidINKBN0E91AI20140529 (accessed 6 February 2016). Board, A. R. & Suzuki, S.  The interrelation between intestinal parasites and latent TB infections among newly resettled refugees in Texas. Int. Health 8 (1), 67–72. Bongartz, P., Vernon, N. & Fox, J.  Sustainable Sanitation for All: Experiences, Challenges, and Innovations. Practical Action Publishing, Rugby. Burra, S., Patel, S. & Kerr, T.  Community-designed, built and managed toilet blocks in Indian cities. Environ. Urban. 15 (2), 11–32. CAERN  Tabela Tarifária Única 2015. Compahnia de Águas e Esgotos do Rio Grande do Norte, Natal. Available from: www.caern.rn.gov.br (accessed 18 November 2016). Cairncross, S., Blumenthal, U., Kolsky, P., Moraes, L. & Tayeh, A.  The public and domestic domains in the transmission of disease. Trop. Med. Int. Health 1 (1), 27–34.

Page 248

Journal of Water, Sanitation and Hygiene for Development

07.1

2017

Coffey, D. & Geruso, M.  Sanitation, Disease, and Anemia: Evidence From Nepal. College of Liberal Arts, University of Texas at Austin, Austin, TX. Available from: https://laits. utexas.edu/∼mlg2296/images/SanitationAnemia.pdf (accessed 18 January 2016). Coffey, D., Gupta, A., Hathi, P., Khurana, N., Spears, D., Srivastav, N. & Vyas, S.  Revealed Preference for Open Defecation: Evidence from a New Survey in Rural North India (SQUAT Working Paper No. 1). RICE Institute, Amston, CT. Cooper, E. S., Whyte-Alleng, C. A. M., Finzi-Smith, J. S. & MacDonald, T. T.  Intestinal nematode infections in children: the pathophysiological price paid. Parasitology 104, S91–S103. Danaei, G., Andrews, K. G., Sudfeld, C. R., Fink, G., McCoy, D. C., Peet, E., Sania, A., Smith Fawzi, M. C., Ezzati, M. & Fawzi, W. W.  Risk factors for childhood stunting in 137 developing countries: a comparative risk assessment analysis at global, regional, and country levels. PLoS Med. 13 (11), e1002164. de Onis, M., Dewey, K. G., Borghi, E., Onyango, A. W., Blössner, M., Daelmans, B., Piwoz, E. & Bianca, F.  The World Health Organization’s global target for reducing childhood stunting by 2025: rationale and proposed actions. Matern. Child Nutr. 9 (Suppl. 2), 6–26. Devine, J. & Kullmann, C.  Introductory Guide to Sanitation Marketing. Water and Sanitation Program, World Bank, Washington, DC. Dewey, K. G. & Begum, K.  Long-term consequences of stunting in early life. Matern. Child Nutr. 7(Suppl. 3), 5–18. Donowitz, J. R. & Petri Jr, W. A.  Pediatric small intestine bacterial overgrowth in low-income countries. Trends Mol. Med. 21 (1), 6–15. Eppig, C., Fincher, C. L. & Thornhill, R.  Parasite prevalence and the worldwide distribution of cognitive ability. Proc. R. Soc. B Biol. Sci. 277, 3801–3808. Evans, B., van der Voorden, C. & Peal, A.  Public Funding for Sanitation: The Many Faces of Sanitation Subsidies. Water Supply & Sanitation Collaborative Council, Geneva. Fagundes-Neto, U., Viaro, T., Wehba, J., da Silva Patrício, F. R. & Machado, N. L.  Tropical enteropathy (Environmental enteropathy) in early childhood: a syndrome caused by contaminated environment. J. Trop. Pediatr. 30 (4), 204–209. Freeman, M. C., Stocks, M. E., Cumming, O., Jeandron, A., Higgins, J. P. T., Wolf, J., Prüss-Üstün, A., Bonjour, S., Hunter, P. R., Fewtrell, L. & Curtis, V.  Hygiene and health: systematic review of handwashing practices worldwide and update of health effects. Trop. Med. Int. Health 19, 906–916. Gómez, M., Gomez, S., Kabajuni, A., Kaur, P., King, W. & Ricciardi, V.  Women, Slums and Urbanisation: Examining the Causes and Consequences. Centre on Housing Rights and Evictions, Geneva. Gosling, L., Irvine, C., Schechtman, L. & Velleman, Y.  Nowhere to Go: How a Lack of Safe Toilets Threatens to Increase Violence Against Women in Slums. WaterAid Sverige, Stockholm.

D. Mara

The elimination of open defecation

Gupta, K., Arnold, F. & Lhungdim, H.  Health and Living Conditions in Eight Indian Cities (Report for National Family Health Survey – 3, India, 2005–06). International Institute for Population Sciences, Mumbai. Halliez, M. C. M. & Buret, A. G.  Extra-intestinal and longterm consequences of Giardia duodenalis infections. World J. Gastroenterol. 19 (47), 8974–8985. Hoddinott, J., Alderman, H., Behrman, J. R., Haddad, L. & Horton, S.  The economic rationale for investing in stunting reduction. Matern. Child Nutr. 9(Suppl. 2), 69–82. House, S., Ferron, S., Sommer, M. & Cavill, S.  Violence, Gender and WASH: A Practitioner’s Toolkit. WaterAid, London. Available from: http://violence-wash.lboro.ac.uk (accessed 21 January 2016). Humphrey, J. H.  Child undernutrition, tropical enteropathy, toilets, and handwashing. Lancet 374, 1032–1035. IDS  Maximising the Potential of Community-led Total Sanitation. Institute of Development Studies, University of Sussex, Falmer, Brighton. http://www.ids.ac.uk/project/ sharing-lessons-improving-practice-maximising-the-potentialof-community-led-total-sanitation (accessed 24 February 2016). IHME  GBD [Global Burden of Disease] Compare 1990– 2015. Institute for Health Metrics and Evaluation, University of Washington, Seattle, WA. Available from: http://vizhub. healthdata.org/gbd-compare (accessed 16 November 2016). Jardim-Botelho, A., Raff, S., Rodrigues, Rde Á., Hoffman, H. J., Diemert, D. J., Corrêa-Oliveira, R., Bethony, J. M. & Gazzinelli, M. F.  Hookworm, Ascaris lumbricoides infection and polyparasitism associated with poor cognitive performance in Brazilian schoolchildren. Trop. Med. Int. Health 13 (8), 994–1004. Jenkins, M. W., Anand, A. R., Revell, G. & Sobsey, M. D.  Opportunities to improve domestic hygiene practices through new enabling products: a study of handwashing practices and equipment in rural Cambodia. Int. Health 5 (4), 295–301. JMP a India: JMP Wealth Quintiles. WHO/UNICEF Joint Monitoring Programme for Water Supply and Sanitation, World Health Organization, Geneva. http://www.wssinfo.org/ documents/?tx_displaycontroller[type]¼wealth_quintiles (accessed 18 January 2016). JMP b WASH Post-2015: Proposed Indicators for Drinking Water, Sanitation and Hygiene (revised edition). WHO/ UNICEF Joint Monitoring Programme for Water Supply and Sanitation, World Health Organization, Geneva. Kar, K. & Chambers, R.  Handbook on Community-led Total Sanitation. CLTS Knowledge Hub, Institute for Development Studies, University of Sussex, Falmer, Brighton. Mara, D.  Shared sanitation: to include or to exclude? Trans. R. Soc. Trop. Med. Hyg. 110 (5), 265–267. Mbuya, M. N. & Humphrey, J. H.  Preventing environmental enteric dysfunction through improved water, sanitation and hygiene: an opportunity for stunting reduction in developing countries. Matern. Child Nutr. 12(Suppl. 1), 106–120.

Journal of Water, Sanitation and Hygiene for Development

07.1

2017

Melo, J. C.  The Experience of Condominial Water and Sewerage Systems in Brazil: Case Studies from Brasília, Salvador, and Parauapebas. Water and Sanitation Program, World Bank, Washington, DC. Melo, J. C.  Sistema Condominial: Uma Resposta ao Desaﬁo da Universalização do Saneamento. Ministério das Cidades, Brasília. Menon, S.  Swachh Bharat Campaign: More Money Down the Drain. Governance Now (Noida, UP), 10 December. Available from: www.governancenow.com/news/regularstory/more-money-drain (accessed 18 January 2016). Ministry of Local Government, Rural Development and Co-operatives  Bangladesh Country Paper. Report presented at the Sixth South Asian Conference on Sanitation (SACOSANVI), 11–13 January, Dhaka. Available from: www.sacosanvi.gov. bd/data/frontImages/Bangladesh_Country_Paper.pdf (accessed 15 November 2016). Modi, N. a Prime Minister Narendra Modi’s Speech on 68th Independence Day at Red Fort, 16 August. Available from: www.narendramodi.in/text-of-pms-speech-at-red-fort-2 (accessed 18 January 2016). Modi, N. b Text of Prime Minister, Shri Narendra Modi, at the launch of Swachhata Mission, 2 October. Available from: www.narendramodi.in/text-of-prime-minister-shri-narendramodi-at-the-launch-of-swachhata-mission-2872 (accessed 18 January 2016). Moore, S. R., Lima, A. A. M., Conaway, M. R., Schorling, J. B., Soares, A. M. & Guerrant, R. L.  Early childhood diarrhoea and helminthiases associate with long-term linear growth faltering. Int. J. Epidemiol. 30 (6), 1457–1464. Morgan, P.  The Arborloo Book: How to Make A Simple Pit Toilet and Grow Trees or Make Humus for the Garden. Stockholm Environment Institute, Stockholm. Neal, D., Vujcic, J., Burns, R., Wood, W. & Devine, J.  Nudging and Habit Change for Open Defecation: New Tactics From Behavioral Science. Water and Sanitation Program, World Bank, Washington, DC. Newman, A., Smets, S. & Kov, P.  Making Toilets More Affordable for The Poor Through Microﬁnance. Water and Sanitation Program, World Bank, Washington, DC. Nokes, C., Grantham-McGregor, S. M., Sawyer, A. W., Cooper, E. S., Robinson, B. A. & Bundy, D. A. P.  Moderate to heavy infections of Trichuris trichiura affect cognitive function in Jamaican school children. Parasitology 104 (3), 539–547. Padhi, B. K., Baker, K. K., Dutta, A., Cumming, O., Freeman, M. C., Satpathy, R., Das, B. S. & Panigrahi, P.  Risk of adverse pregnancy outcomes among women practicing poor sanitation in rural India: a population-based prospective cohort study. PLoS Med. 12 (7), e1001851. Partovi, F., Khalili, G., Kariminia, A. & Mahmoudzadeh-Niknam, H.  Effect of Giardia lamblia infection on the cognitive function of school children. Iran. J. Public Health 36 (1), 73–78.

Page 249

D. Mara

The elimination of open defecation

Perez, E.  What Does It Take to Scale Up Rural Sanitation? Water and Sanitation Program, World Bank, Washington, DC. Pullan, R. L., Smith, J. L., Jasrasariaand, R. & Brooker, S. J.  Global numbers of infection and disease burden of soil transmitted helminth infections in 2010. Parasit. Vectors 7, 37. Sanan, D. & Moulik, S. G.  Community-Led Total Sanitation in Rural Areas: An Approach That Works. Water and Sanitation Program – South Asia, World Bank, New Delhi. Spears, D.  How Much International Variation in Child Height Can Sanitation Explain? (Policy Research Working Paper No. 6351). World Bank, Washington, DC. Spears, D. & Haddad, L. J.  The Power of WASH: Why Sanitation Matters for Nutrition. In: 2014–2015 Global Food Policy Report, pp. 19–23. International Food Policy Research Institute, Washington, DC. Sy, J., Warner, R. & Jamieson, J.  Tapping the Markets: Opportunities for Domestic Investments in Water and Sanitation for the Poor. World Bank, Washington, DC. Tiwaril, M.  UP villagers prefer open fields, raze Swachh loos. The Times of India, New Dehli, 14 January. http:// timesofindia.indiatimes.com/city/bareilly/UP-villagersprefer-open-fields-raze-Swachh-loos/articleshow/50582495. cms (accessed 6 February 2017). UNDESA  World Population Prospects: The 2015 Revision – Data Booklet. United Nations Department of Economic and Social Affairs, Population Division, New York, NY. UN-Habitat  Slum Almanac 2015–2016: Tracking Improvement in the Lives of Slum Dwellers. United Nations Human Settlements Programme, Nairobi. UNICEF  Child Nutrition Interactive Dashboard – 2015 Edition. UNICEF, New York, NY. Available from: http:// data.unicef.org/resources/child-nutrition-interactivedashboard-2015-edition.html (accessed 18 January 2016). United Nations General Assembly  Transforming Our World: The 2030 Agenda for Sustainable Development (Document A/RES/70/1, 21 October). United Nations, New York, NY. Vyas, S., Spears, D., Kov, P. & Smets, S.  Investing in the Next Generation: Growing Tall and Smart with Toilets – Stopping

Journal of Water, Sanitation and Hygiene for Development

07.1

2017

Open Defecation Improves Children’s Height in Cambodia. Water and Sanitation Program, World Bank, Washington, DC. WaterAid India  Feeling the Pulse: A Study of the Total Sanitation Campaign in Five States. WaterAid India, New Delhi. Wechsler, D.  Wechsler Intelligence Scales for Children – Revised. Psychological Corporation, Harcourt Brace Jovanovich Inc., San Antonio, TX. WHO  Preventing Diarrhoea through Better Water, Sanitation and Hygiene: Exposures and Impacts in Lowand Middle-Income Countries. World Health Organization, Geneva. WHO  Child Growth Standards: Length/height for Age. World Health Organization, Geneva. Available from: www.who.int/ childgrowth/standards/height_for_age/en/ (accessed 17 January 2016). WHO  Soil-Transmitted Helminth Infections (Fact sheet). World Health Organization, Geneva. Available from: www. who.int/mediacentre/factsheets/fs366/en/# (accessed 9 March 2016). WHO/UNICEF  Progress on Drinking Water and Sanitation: 2012 Update. Joint Monitoring Programme for Water Supply and Sanitation, World Health Organization, Geneva. WHO/UNICEF  Progress on Sanitation and Drinking-water – 2015 Update and MDG Assessment. Joint Monitoring Programme for Water Supply and Sanitation, World Health Organization, Geneva. WIN-SA  eThekwini’s Water and Sanitation Programme (Lessons Series No. 2). Water Information Network – South Africa, Pretoria. WSP  Sanitation Marketing Toolkit. Water and Sanitation Program, World Bank, Washington, DC. Available from: https://wsp.org/toolkit/toolkit-home (accessed 17 November 2016). WSP/MDWS  Pathway to Success: Compendium of Best Practices in Rural Sanitation. Water and Sanitation Program, World Bank and Ministry of Drinking Water and Sanitation, New Delhi.

First received 2 February 2016; accepted in revised form 17 December 2016. Available online 4 February 2017

Page 250

196

Review Paper

07.2

2017

Review Paper Qualitative comparative analysis for WASH research and practice Jessica Kaminsky and Elizabeth Jordan

ABSTRACT Qualitative comparative analysis is an established research method that has been underutilized in water, sanitation, and hygiene (WASH) research. It has immense potential for addressing the complexity inherent to WASH projects, and can produce robust and transparent results from intermediate or large numbers of cases. The method enables researchers and practitioners to blend quantitative and qualitative metrics to build more nuanced contextual knowledge, and is able to detect combinations of causal conditions that lead to outcomes of interest. This means that the method is uniquely positioned for building empirically founded theories of change that reﬂect contextual complexity. In this review paper we use hypothetical data and a review of the existing literature to showcase where and how the method can be productively applied in WASH research

Jessica Kaminsky (corresponding author) Department of Civil and Environmental Engineering, University of Washington, 121H More Hall, Seattle, WA 98195, USA E-mail: jkaminsk@uw.edu Elizabeth Jordan USAID, 1300 Pennsylvania Ave, NW, Washington, DC 20523, USA

and practice. Key words

| hygiene, methods, QCA, sanitation, WASH, water

INTRODUCTION Recent years have seen the water, sanitation, and hygiene

The gold standard for research methodology is often

(WASH) community increase its focus on evidence-based

considered to be the randomized controlled trial (RCT),

approaches, monitoring, and evaluation. This move is

which emulates clinical trial research methods. For

intended to improve accountability and results for both

example, a handful of recent RCTs have tested the impact

donors and the communities where projects take place.

of community-led total sanitation (CLTS) methods (Clasen

In part, this trend towards measurement is a reaction to

et al. ; Patil et al. ; Guiteras et al. ; Pickering

the most fundamental and important questions for global

et al. ). Like any tool, however, RCTs are not perfect.

development: why is it that some projects succeed while

One practical problem is the considerable expense of

others fail? And what, exactly, do we mean by success

implementation. Other methodological issues are common

and failure in WASH projects? While occasionally there

to any quantitative approach; closed-ended questionnaires

may be simple answers to these questions, more often the

allow statistical analysis but force individual responses

answers themselves are complex systems of dynamic,

into pre-determined schema that may or may not be appro-

contextual factors and decoupled impacts (Meyer &

priate. In a related methodological issue, quantitative

Rowan ).

methods prove relationships but struggle to discover how or why variables contribute to the outcome of interest. For

This is an Open Access article distributed under the terms of the Creative Commons Attribution Licence (CC BY 4.0), which permits copying,

example, why did the previously referenced and well-

adaptation and redistribution, provided the original work is properly cited

designed CLTS studies show differing impacts on outcomes

(http://creativecommons.org/licenses/by/4.0/).

like stunting, incidence of diarrhea, and rates of change in

doi: 10.2166/washdev.2017.240

Page 251

197

J. Kaminsky & E. Jordan

Journal of Water, Sanitation and Hygiene for Development

Qualitative comparative analysis

07.2

2017

latrine ownership? Different and complementary research

QCA is founded in set theory. The simplest set relation is

methods – like the qualitative comparative analysis (QCA)

a subset – for example, water projects are a subset of WASH

method described in this paper – are needed to answer

projects. This type of set relation deﬁnes both water projects

these important questions.

(as one kind of a WASH project) and also deﬁnes WASH

In contrast to statistical methods, traditional qualitative

projects (as including, among other things, water projects).

research approaches allow local knowledge to emerge

While deﬁnitional sets can be trivial, more interesting set

from the data and are well suited to discovering how and

relations emerge when researchers seek causal relationships

why WASH interventions work in a particular context.

between various phenomena; the causality claim, of course,

However,

statistically

is founded on theory and sector knowledge. For example,

generalized, and the very nuance of the answers that

and as discussed below, there are both sustainable and

qualitative methods generate can mean they are relatively

unsustainable school sanitation projects, and we suspect

difﬁcult to communicate and apply in different contexts.

there are reasons why projects turn out to belong to one

Because of these differing strengths and weaknesses,

or the other of these subsets of school sanitation projects.

high quality quantitative and qualitative research deeply

Causal conditions are the reasons that the researcher

qualitative

ﬁndings

cannot

complement each other. Each type of approach – and

believes may inﬂuence the outcome of interest. While

there are many methods within these broad categories –

these are similar to independent variables in a statistical

can answer different kinds of research questions. For

analysis, they do not take on many of the assumptions of

example, the very complexity of factors discovered through

variables in a statistical analysis. As we discuss below, set

qualitative cases may provide an explanation for why it is

relationships are importantly different than correlational

so difﬁcult to statistically link WASH interventions and

relationships, in part because of the underlying assumptions

health outcomes (Schmidt ). Or, quantitative studies

about the symmetry of theorized relationships between

may statistically validate qualitative ﬁndings, discovering

causes and outcomes (see Table 4 and the related discussion

the importance of each factor relative to the outcome of

for an example of this difference). The thought structure of

interest.

this paragraph parallels that of the ﬁrst chapters of Ragin’s

QCA (Ragin ) is a research method that blends the

() book. The reader is referred to that book for more

strengths of qualitative and quantitative methods. QCA is a

details on the fundamentals of set theoretic thought, which

set-theoretic method that seeks combinations of causal con-

are not limited to the introductory examples provided

ditions, or pathways, that lead to an outcome of interest. To

here. Given the difference and utility of the QCA research

do so, deep qualitative and quantitative case knowledge is

method, this paper is a methodological contribution

explicitly represented by calibrated, quantitative measure-

intended to describe how QCA may be useful to the

ments in a truth table. This truth table is then simpliﬁed

WASH community.

using either Boolean algebra or fuzzy logic in a fully reproducible, generalized set theoretic analysis. The method lives between qualitative and quantitative analysis, and can

QCA FOR WASH

handle either intermediate or large numbers of cases. While a relatively new research method (Ragin )

In recent years a handful of researchers have begun to apply

which was originally used in the areas of comparative

QCA to WASH research, which we deﬁne as research

politics and historical sociology, over recent decades it has

interested in drinking water supply, sanitation and hygiene

made signiﬁcant inroads to a wide range of research

for developing nations, communities, and households

communities,

and

worldwide. While limited in number, the existing studies

engineering. QCA allows us to rigorously analyze different

showcase the wide variety of applications in which QCA

types, quantities, and combinations of qualitative and quan-

can be valuable. To identify the examples referenced here

titative data; thus we suggest it is an important addition to

(which necessarily represent a subset of those examples pre-

the WASH toolkit.

sent in the literature), we searched the literature for ‘QCA’ in

Page 252

including

economics,

management

198

J. Kaminsky & E. Jordan

Journal of Water, Sanitation and Hygiene for Development

Qualitative comparative analysis

07.2

2017

combination with either ‘WASH,’ ‘water,’ ‘sanitation,’ or

(Welle et al. ). As will be described in more detail

‘hygiene.’ We included papers that considered WASH pro-

below, many of these authors identiﬁed an intermediate

jects as some cases in a larger dataset. These searches

number of cases and a need for nuanced yet rigorous quanti-

were carried out on article databases including Academic

ﬁcation of complexity as rationales for choosing the QCA

Search Complete, Engineering Village, Web of Science,

method.

the Environmental Science Collection, SCOPUS, and WorldCat. We also reviewed the ﬁrst 100 results for each search on the GoogleScholar search engine. A limitation

QCA VARIANTS

of this approach is that most non-academic publications are not archived in these databases. As such, for each com-

There are three variants of QCA analysis. These variants

bination of search terms we also reviewed the ﬁrst 100

describe the way that the causal conditions and outcome

results returned from standard Google searches, as well as

of interest are deﬁned. In any of these three variants, and

searching the SuSanA knowledgebase (http://www.susana.

if the analytic decisions discussed here are fully documen-

org/en/resources/library) for ‘QCA’. This approach ident-

ted,

iﬁed resources such as QCA-based evaluation protocols

transparent measures of reliability. The simplest is crisp set

and toolkits (Annamalai et al. ; OpenIDEO ). This

QCA (csQCA), in which each causal condition is measured

search identiﬁed 17 key documents, including four prac-

as being either fully present or fully absent. For example,

titioner-published

Kaminsky & Javernick-Will () use csQCA to describe

reports,

two

dissertations,

and

academic journal articles.

QCA

enables

fully

replicable

analysis

with

household toilets that were or were not functional on the

The majority of the key articles we identiﬁed dealt with

day of a research visit. Similarly, Chatterley et al. ()

water, while a few treated sanitation and hygiene. The cases

use csQCA to analyze schools with and without well-

analyzed in the articles varied widely in scale. For example,

maintained toilets. In a slightly more complex variant,

several papers analyzed individual households (Spencer

multiple value QCA (mvQCA) permits the inclusion of

; Kaminsky & Javernick-Will ) or schools (Chatter-

non-dichotomous

ley et al. , ), others analyzed development projects

multiple values. mvQCA is best suited for studies in which

(Boudet et al. ; Santosh Kumar Delhi et al. ), and

the variables can be summarized into a small number of

another analyzed public private partnerships for urban

discrete options (Gross & Garvin ). For example, in a

water supply (House ). Similarly, there is a wide range

study attempting to understand the effect of water supply

of research topics in this set of papers. The most common

operators, we might wish to consider three variants: (1)

measurements

which

can

take

is an emphasis on sustainability (Chatterley et al. ,

community-managed supply, (2) private operator, and (3)

2014; Kaminsky & Javernick-Will ; Welle et al. ).

public operator. An mvQCA study can also include variables

Interestingly, and as opposed to past trends in the broader

that are dichotomized.

sanitation literature (Rosenqvist et al. ), these papers

Finally, in the most conceptually complex variant,

use sustainability to refer to the sustained use of WASH ser-

fuzzy set QCA (fsQCA) allows for each variable to be

vices with reference to social systems rather than targeting

assigned a value between zero and one corresponding to

environmental sustainability. Other work studies methods

its degree of membership in a set. In an fsQCA study, a

of project delivery such as private participation in WASH

score of 1 represents full membership in a set and a score

infrastructure construction (Santosh Kumar Delhi et al.

of 0 represents full non-membership, with 0.5 as the point

; House ) or drivers of conﬂict regarding these pro-

of maximum ambiguity of set membership. Values between

jects (Boudet et al. ). Several of the papers used QCA

0 and 1 represent varying degrees of membership and non-

in combination with either qualitative or statistical methods

membership. These scales are not linear, and the way in

for mixed-methods analysis (House ; Welle et al. ).

which the set is deﬁned will affect the values. fsQCA is

This included one of the few identiﬁed publications

useful in cases where restricting all conditions to dichoto-

coming from practice rather than academic researchers

mous values would result in a loss of detail from data as Page 253

199

J. Kaminsky & E. Jordan

Journal of Water, Sanitation and Hygiene for Development

Qualitative comparative analysis

it can represent small, but meaningful, differences between cases. For example, in Chatterley et al. () the set of schools with well-managed sanitation is deﬁned as schools with toilets that are both functional and clean (Table 2 provides details of Chatterley’s deﬁnitions). In this paper, schools with excellent performance on both of these metrics are fully in the set of schools with well-managed sanitation; schools with moderate performance on these metrics are partially in the set of schools with wellmanaged sanitation; schools with poor performance on these metrics are fully out of the set of schools with wellmanaged sanitation.

Table 2

07.2

2017

Excerpted fsQCA coding scheme from Chatterley et al. (2014)

Well-managed sanitation services

Minimum of the following two measures: Reliably functional toiletsa: 1: students have reliable access to functional services; repairs timely addressed 0.67: all toilets usually function, but repair needs are not always timely addressed 0.33: some toilets are frequently unusable; repairs are not timely addressed 0: students do not have reliable access; repairs are rarely addressed and Reliably clean toiletsb: 1: all toilets are almost always clean and quickly cleaned when dirty

THE QCA PROCESS

0.67: usually more or less clean, with some instances where they remain dirty

Deﬁning outcomes and conditions

0.33: frequently unclean and are usually considered unclean by students 0: rarely clean and students label them as dirty

For all three variants of QCA, the ﬁrst step is to deﬁne the outcome(s) of interest to the research question. The outcome of interest is whatever the study intends to measure as an outcome of the intervention; examples in the WASH sector could include open defecation-free status in a community, functionality of a handpump, or household practice of a

‘Functional’ ¼ waste is easily ﬂushed, the building structure, doors and locks function providing privacy, water is available, and soap is available in or near the toilet. ‘Repairs timely addressed’ ¼ minor critical repairs (needed for use), such as a door lock or clogged toilet, are repaired within 24 hours, major critical repairs, such as a broken pan or door,

are repaired within 1 week, minor non-critical repairs, such as a broken tap, are repaired within 1 week, and major non-critical repairs, such as a broken water pump, are repaired within 1 month. b

‘Clean’ ¼ no visible feces on the ﬂoor/walls/seat, no ﬂies, and no foul smell.

hygiene behavior. This step informs the process of case selection, as it is necessary to purposefully identify a set of cases that demonstrate a range of the outcomes for the analysis. For example, in Table 1 the hypothetical outcome of interest is Sustained Water Services, meaning cases with and without sustained water service would be needed for the dataset.

The next step is to identify conditions that are expected to inﬂuence the outcome(s) of interest. Conditions are the variables that distinguish one case from another. For example, in Table 1 we provide a hypothetical example where the causal conditions are Community Participation and a Municipal Utility. The selection of conditions for any QCA study is iterative. Conditions are logically con-

Table 1

Hypothetical example of cases and variables

structed and should generally be grounded in theory. However, one of QCA’s strengths is the ability to build

Outcome of interest: Sustained Water Service

theory from the analysis. Thus, some of the conditions

Community

Municipal

Set notationa

may be selected for inductive reasons, meaning additional

participation

utility

representing

Observed cases

Condition A

Condition B

this combination

conditions may emerge during the data collection. Indeed,

Cases 1–4

Yes

A* ∼ B

it is likely that a large number of conditions will be identiﬁed

Cases 5–8

Yes

∼A*B

(Amenta & Poulsen ). However, each new condition

No observed cases

∼A* ∼ B

adds complexity to the logic space (the space deﬁned by

No observed cases

Yes

A*B

all of the possible value-combinations of the conditions

(Ragin )), so it is practically important to limit the

Note: hypothetical data.

number of conditions.

In set notation, the symbol * signiﬁes and, while ∼ signiﬁes not.

Page 254

200

J. Kaminsky & E. Jordan

Journal of Water, Sanitation and Hygiene for Development

Qualitative comparative analysis

07.2

2017

A large number of conditions will likely result in a

scale of traditional qualitative analyses. Similarly, while

unique explanation for each case, making it difﬁcult to inter-

there is no upper bound to the number of cases QCA can

pret and generalize the results. As such, there are various

consider, the method does not require minimum sample

documented techniques to reduce the causal conditions in

sizes. This combination means QCA ﬁlls the methodological

a QCA analysis (Ragin ; Rihoux & Ragin ). For

need for a way to rigorously handle an intermediate num-

example, it is possible that some of the identiﬁed causal con-

bers of cases.

ditions will not vary signiﬁcantly between the cases selected

Unobserved conﬁgurations are theoretical combinations

because of the research context. As in statistical analysis, the

of conditions that are not found in any of the cases analyzed,

non-varying conditions cannot be included in the analysis.

and will appear in any QCA study. For example, in the

Such variables are called domain conditions. While these

hypothetical dataset represented in Table 1, we do observe

domain conditions cannot be analyzed, they may have an

cases with the outcome of Sustained Water Service when

important inﬂuence on the outcome through their presence

we observe Community Participation without a Municipal

and interactions with other causal conditions. It is impor-

Utility (Cases 1–4), and vice versa (Cases 5–8). However,

tant to clearly describe any domain conditions, as these

while they are logically possible we do not observe any

limit the generalizability of the results. It may also be poss-

cases with neither Community Participation nor a Munici-

ible to combine initially hypothesized conditions if inter-

pal

relationships between the conditions can be identiﬁed. For

Participation and a Municipal Utility. These are called unob-

example, discriminant analysis can be used to identify

served conﬁgurations, regardless of the likelihood of their

strong bivariate relationships, or composite conditions can

existence. During QCA analysis, unobserved conﬁgurations

be created through techniques such as factor analysis

may be handled in three standard ways; the researcher must

( Jordan et al. ).

determine which of these is most appropriate to the research

Utility,

any

cases

with

both

Community

question and data. The different assumptions in each of Case selection

these three methods should be expected to result in different answers, which are called the complex, intermediate, and

For QCA analysis, cases are selected to exhibit the greatest

parsimonious solutions. These are discussed in more detail

possible variety of configurations (a configuration is defined

in the Pathway analysis section below.

by each case’s set of condition and outcome values). Although many criticize the conscious selection of cases

Data collection

as an improper manipulation of the dataset, this practice is appropriate for QCA because the method’s logic is not prob-

Data must be collected for each condition in each case. Gen-

abilistic: that is, it does not matter if only a few cases exhibit

erally, data will be collected on more conditions than are

certain conditions (Berg-Schlosser et al. ). Rather, the

actually used in the analysis, given the iterative nature of

selection of cases exhibiting maximum variation in con-

the QCA process. Both qualitative and quantitative data

dition and outcome values will result in the richest

can be used in the analysis, but the researcher must have suf-

possible explanations of relationships among the variables

ﬁcient knowledge of each case to make determinations

(Gross ). The number of cases included in the analysis

about the variable calibration described below.

should be driven by the size of the logic space (the number of all possible conﬁgurations) and the feasibility of

Variable calibration

data collection. A key strength of QCA analysis is that it allows researchers to handle an intermediate number of

Once the cases, outcomes, and preliminary conditions have

cases, too many for qualitative analysis and too few for

been determined, raw data for each case (qualitative and/or

statistical analysis. As we will discuss below, the use of

quantitative) must be collected and calibrated according to

mathematics to search for patterns in case data reduces

set deﬁnitions relevant to the research questions, underpin-

the data processing demands that logistically limit the

ning theory, and the dataset itself. These set deﬁnitions Page 255

201

J. Kaminsky & E. Jordan

Qualitative comparative analysis

Journal of Water, Sanitation and Hygiene for Development

07.2

2017

rigorously describe each case in terms of the causal con-

of poor countries. As such, the scores for both these

ditions that are hypothesized to be relevant to the

nations would be set to 0 (indicating non-membership in

outcome of interest. For example, and as described in

the set of poor countries).

more detail below, set deﬁnitions might include what kind

The second method for calibrating fsQCA data is indir-

of management scheme is used for water system manage-

ect calibration. This can be done for either quantitative or

ment, the volume of water used by each household, or

qualitative data by creating groupings of cases. To calibrate

how wealthy communities are.

qualitative data the researcher develops a list of operationa-

The method of calibration depends on the variant of

lized measures for each of the conditions and outcomes.

QCA being undertaken. For a csQCA study, all conditions

Then, qualitative anchors are deﬁned for full membership

must be calibrated as either a zero or one. Qualitative data

and non-membership in the set and the case data are eval-

are calibrated by deﬁning the features of what is within

uated based on these operationalized deﬁnitions. As for

and outside of the set. For example, Kaminsky & Javer-

direct calibration, these points should be anchored in exter-

nick-Will () coded toilets as either socially sustainable

nal criteria and theoretical and case knowledge. For

(1, deﬁned as owner maintenance post-construction and unbro-

example, Table 2 shows an example of indirect calibration

ken slab, pit rings, and superstructure on the day of the visit) or

from the literature (Chatterley et al. ), where schools

unsustainable (0, if either of the two criteria were not met)

that virtually always have clean and functional toilets are

on the day of a research visit. Quantitative data are dichoto-

deﬁned as being fully in the set of schools with well-mana-

mized through the determination of a numeric cutoff point.

ged sanitation.

For example, water samples might be coded as having either

For any of these methods of calibration, a clear cali-

positive or negative fecal coliform test results, based on

bration protocol and inter-calibrator reliability checks are

international standards for water quality. In contrast, for

needed to support the validity of the ﬁndings. Through this

an mvQCA study a small number of discrete options are

process, it is likely that the calibration methods will be itera-

deﬁned for each condition. For example, community water

tively improved to ensure that real differences between cases

supply could be coded as community managed (0), having

are captured accurately.

a private operator (1), or having a public operator (2). Calibration for an fsQCA study can be more complex, as each variable is represented on a continuous scale

Constructing and analyzing the truth table

between zero and one. The ﬁrst method for calibration, direct calibration, can be used for quantitative data. How-

The calibrated data are used to populate a truth table that

ever, quantitative data cannot simply be normalized to

represents the calibrated conditions and outcomes. The

values between 0 and 1, as the calibrated values represent

truth table (Table 3) consists of columns for each con-

the degree of membership in a set and must be based on

dition and outcome, with rows representing each case.

the set deﬁnitions. To perform direct calibration, the

Once the truth table is generated, the researcher may

research must specify three breakpoint values: full mem-

ﬁnd contradictory conﬁgurations, or cases with identical

bership, full non-membership and the crossover point

conditions and differing outcomes. These can be resolved

(equal to a 0.5 score). These points should be anchored

by considering the conditions included to see whether

in external criteria and theoretical and case knowledge.

(for example) there is a missing condition that explains

For example, in examining country level data for GNP to

the difference between the two cases. Alternatively, cali-

assess membership in the set of poor countries, the vari-

bration cutoffs may be re-examined to determine if an

ation between countries that are clearly outside of the set

important difference between the two cases was obscured

of poor countries is unimportant to the analysis, and the

in the initial calibration. Researchers intending to use

anchor points must be set accordingly. For example, both

QCA should note that the creation of a contradiction-free

Sweden and Norway are non-poor, and any variation

truth table is extremely time consuming and requires deep

between the two is unimportant to the set classiﬁcation

case knowledge.

Page 256

202

Table 3

J. Kaminsky & E. Jordan

Journal of Water, Sanitation and Hygiene for Development

Qualitative comparative analysis

07.2

2017

Truth table example Condition A

Condition B

Condition C

Condition D

Outcome

Case 1

0.67

Case 2

0.33

0.67

Case 3

0.33

…

Case N

0.33

0.67

Note: hypothetical data.

The ﬁrst step in QCA analysis determines which individ-

Figure 1

Necessity and sufﬁciency.

ual conditions are necessary or sufﬁcient to achieve the outcome. The second step determines which combinations

intermediate solution for the hypothetical example given

of conditions combine to lead to the achievement (or non-

in Table 1, the presence of Community Participation was

achievement) of an outcome. These two steps are discussed

enough to achieve Sustained Water Services, regardless of

in more detail below. Although these analyses can be per-

the presence or absence of a Municipal Utility. Another

formed by hand, software can facilitate analysis. One

way to say this is to say that Community Participation is suf-

option is the open source fs/QCA software developed by

ﬁcient to achieve Sustained Water Services. In contrast,

Charles Ragin, which can be used for csQCA or fsQCA.

necessity measures the degree to which the outcome is a

Other options are Tosmana, a software package designed

subset of individual causal conditions, meaning that, if all

for mvQCA and csQCA studies, or various and constantly

(or nearly all) cases where the outcome is present have a

evolving options for STATA and R ( Jordan et al. ;

particular condition present, we would consider that con-

Schneider & Wageman ; Ragin et al. )

dition necessary. For example, Bogler & Meierhofer () ﬁnd that both trouble-free production and high demand are necessary for sustainable colloidal silver ﬁlter businesses

Necessity and sufﬁciency of individual conditions

in Nepal.

In QCA, sufﬁciency is a measure of the degree to which an

resented by Equations (1) and (2) respectively, where Xi

individual causal condition is a subset of the outcome (see

and Yi represent single conditions. Typically, researchers

Figure 1). If a speciﬁc condition always (or nearly always)

require a necessity score of at least 0.9 to call a condition

results in a positive outcome, that condition would be

necessary for the outcome of interest, and a sufﬁciency

deemed sufﬁcient. For example, when we described the

score of at least 0.8 to call a condition sufﬁcient for the

Mathematically, necessity and sufﬁciency may be rep-

Table 4

Symmetric and non-symmetric relationships Outcome of interest: Unsustained Water Service

Outcome of interest: Sustained Water Service

Causal condition: Community Participation absent

Causal condition: Community Participation present

Causal condition: Community Participation absent

Causal condition: Community Participation present

Symmetric relationship (chi-squared p < 0.000)

Non-symmetric relationship (chi-squared p ¼ 0.12)

Note: hypothetical data.

Page 257

203

J. Kaminsky & E. Jordan

Journal of Water, Sanitation and Hygiene for Development

Qualitative comparative analysis

outcome of interest. P ðminðXi , Yi ÞÞ P Xi P (minðXi , Yi Þ) P Sufficiency ¼ Yi

Necessity ¼

07.2

2017

generally considered acceptable. According to Ragin, ‘Consistency, like signiﬁcance, signals whether an empirical (1)

connection merits the close attention of the investigator. If a hypothesized subset relation is not consistent, then the

(2)

The analysis of combinations of conditions, a key

strength of QCA, is discussed below.

researcher’s theory or conjecture is not supported.’ In contrast, coverage is a measure of how much a particular pathway accounts for the instance of the outcome, giving a measure of the importance of that pathway. Sufﬁciency and coverage use the same equation (Equation (2)) but coverage describes a particular combination of conditions rather than

Pathway analysis of combinations of conditions

considering individual conditions. As such, to measure coverage using Equation (2), Xi represents the membership in a

The truth table is analyzed using Boolean algebra (for

conﬁguration, and Yi represents the membership in the out-

csQCA and mvQCA studies) and fuzzy logic (for fsQCA

come condition. High coverage scores indicate that a given

studies). For more details on the mathematics behind these

pathway represents many of the represented cases. However,

analyses see Ragin (, ). Regardless of which math-

this does not mean that pathways with low coverage are

ematical approach is used, the analysis of the truth table

unimportant, as QCA is not probabilistic. Despite this, know-

results in the discovery of combinations of conditions

ing which pathways to a given result are seen more frequently

(often called pathways) that lead to a particular outcome

can help guide practitioners to interventions that may be

of interest, with quantitative scores that describe how well

more likely to apply to many cases.

each of these pathways describes the dataset. For example, in Table 1 there were multiple cases that showed the con-

Complex, parsimonious, and intermediate solutions

ditions of Community Participation, no Municipal Utility, and the outcome of Sustained Water Services; these cases

The truth table pathways analysis results in three different

share a pathway.

solutions: complex, parsimonious, and intermediate (as

Two metrics are employed to assess QCA pathway out-

described below). These different solutions are based on

puts: consistency and coverage. Consistency is a measure

different assumptions made about the unobserved conﬁgur-

of the degree to which cases sharing the same combination

ations discussed above in the section entitled ‘Case

of conditions have the same outcome. In other words, con-

selection’. QCA uses counterfactual analysis to transparently

sistency is a measure of the extent to which the observed

compare the impacts of assumptions regarding unobserved

cases align with each other. High consistency means a

conﬁgurations, and to obtain more parsimonious solutions

given pathway almost always leads to a certain outcome,

based on these unobserved conﬁgurations. However, it is

while low consistency means a given pathway only some-

up to the researcher to use her theoretical and substantive

times leads to the outcome of interest. Necessity and

knowledge to decide to what degree these unobserved con-

consistency use the same equation (Equation (1)) but con-

ﬁgurations should be included in the analysis. For example,

sistency describes a particular combination of conditions

in the hypothetical example given in Table 1 we might not

rather than considering individual conditions. As such, to

expect to see cases of Sustained Water Services with the

measure consistency using Equation (1), Xi represents the

absence of both Community Participation and a Municipal

membership in a conﬁguration, and Yi represents the mem-

Utility, but we probably would suspect that there are unob-

bership in the outcome condition. A consistency score of 1

served cases with the presence of both of these conditions.

would indicate perfect consistency, where all cases with a

To validate this intuition, published literature from academic

given set of causal conditions have membership in the out-

journals or practice can be used. Alternatively, more research

come set to a greater degree than membership in the

would be needed to seek out additional cases and better

conﬁguration; however, consistency scores above 0.8 are

populate the logic space.

Page 258

204

J. Kaminsky & E. Jordan

Journal of Water, Sanitation and Hygiene for Development

Qualitative comparative analysis

07.2

2017

As such, there are three possible solutions from each

This asymmetry exists because the question of what con-

QCA run, depending on how unobserved conﬁgurations

ditions lead to a positive outcome is not necessarily the

are used during simpliﬁcation (ranging from none to all

same question as which conditions lead to the negation of

those logically possible). Firstly, the complex solution does

that outcome. Therefore, this should be treated as a separate

not incorporate any counterfactuals and is based entirely

truth table analysis, and interpreted using the same pro-

on the observed cases. This will often be a highly compli-

cedures as the analysis for the positive outcome. For

cated solution, sometimes with a unique pathway for each

example, Chatterley et al. () report pathways to both

observed case. It is not typically used by researchers as it

well-maintained school toilets (the positive outcome) and

does not take into account any theoretical knowledge

pathways to poorly maintained school toilets (the negative

about the link between conditions and outcomes.

outcome). In that study, poor construction is a causal con-

In contrast, the intermediate solution uses ‘easy’ counterfactuals,

which

are

researcher-speciﬁed

dition that appears in all pathways to poorly maintained

theoretical

school toilets, while its inverse (quality construction) is a

assumptions. For example, a researcher may believe that

condition in only some of the pathways to well-maintained

three conditions (A, B and C) are likely related to the posi-

school toilets. In other words, poor quality construction is

tive instance of an outcome, but only observes cases where

present in all cases with the poorly maintained school

A and B are present, but C is absent (i.e. A*B* ∼ C). If the

toilet outcome. However, good quality construction is not

researcher has strong knowledge that the presence of C

present in all cases with well-maintained school toilets.

should contribute to the outcome under the scenario, then

This suggests poor construction can be overcome, given

the assumption that A*B*C would lead to the outcome

the presence of a number of other conditions such as (for

would be an ‘easy’ counterfactual. It should be noted that

example) a local sanitation champion.

these types of assumptions are common, but usually implicit, in traditional comparative case study analysis. A strength of QCA is that these assumptions are clearly docu-

CRITIQUES OF QCA

mented throughout the analysis process. Thirdly and ﬁnally, the parsimonious solution is

As for any method, there have been important critiques

obtained by using all of the unobserved conﬁgurations as

made of QCA. Most recently, there has been a ﬂurry of

potential simplifying assumptions in the truth table analysis.

research attention that uses simulations to examine the

In this solution, the researcher does not specify which

robustness of QCA ﬁndings. For example, Hug ()

assumptions are reasonable, but rather allows the software

claims that QCA does not allow researchers to directly

to ﬁnd the mathematically simplest solution. Clearly, the

account for measurement error and uses a quasi-Monte

researcher must evaluate each of the assumptions that

Carlo analysis to demonstrate the implications of this for

result from the parsimonious solution algorithm to ensure

research conclusions. In another example, Braumoeller

that they are theoretically plausible. It is quite likely that

() makes the strong claim that QCA does not permit

(as in the example just given) at least some of the assump-

researchers to discover whether or not their ﬁndings are

tions leading to the parsimonious solution would be

the result of chance, noting that QCA does not generate stat-

difﬁcult to justify. As such, the intermediate solution

istical signiﬁcance tests. Similarly, Krogslund et al. ()

should be reported unless there are strong theoretical

note that the results of QCA are sensitive to researcher

reasons to accept the parsimonious solution.

decisions such as cutoffs for the minimum frequency of

Note that it is also recommended to perform an analysis

cases that are included in analysis and the minimum and

of which conditions lead to the lack of attainment of an

maximum sufﬁciency scores required for the analysis.

outcome. Because QCA accounts for conﬁgurational

More troublingly, and related to Braumoeller’s critique,

complexity and asymmetrical relationships (which are

they also claim that QCA suffers from conﬁrmation bias.

discussed in more detail below), this will not necessarily

As might be expected, these various critiques have been

be the negation of the conditions that led to the outcome.

answered by other QCA simulation analyses that claim Page 259

205

J. Kaminsky & E. Jordan

Journal of Water, Sanitation and Hygiene for Development

Qualitative comparative analysis

07.2

2017

(for example) that many of the issues observed in the critical

ground between traditional case study methods and large

simulations stem from a fundamental misunderstanding of

n statistical methods, enabling cross-case comparisons to

the QCA method and analysis procedures (for one example,

identify patterns across cases while retaining sensitivity to

see Rohlfing ).

contextual detail. It relies on set-theoretical descriptions of

It is not our purpose here to undertake quantitative mod-

cases, and can be used for exploratory analysis intended to

eling to contribute to this scholarly conversation. Instead, we

build theory or to test existing theories. Many qualitative

would note that similar critiques could be leveled at most

researchers already implicitly use set theory to describe

research methods. Any analysis can be undermined by unde-

ﬁndings of case studies by examining cases that share an

tected measurement errors; QCA’s dependence on deep,

outcome and determining what causal conditions they

qualitative case knowledge is an important answer to this

share; QCA provides a numeric approach to discovering

problem and is one of the key strengths of the method. It is

and documenting these relationships, with transparent

also true that QCA does not generate measures of statistical

documentation of each step in the analysis. For a related dis-

signiﬁcance. However, there are links between statistical sig-

cussion of common pitfalls in the use of QCA, we refer the

niﬁcance and the quantitative consistency values generated

reader to Jordan et al. ().

by QCA, as discussed by Ragin (). The required values

For WASH research, an important attribute of QCA is

for consistency, sufﬁciency, etc. are indeed researcher

its ability to handle combinations of qualitative and quanti-

selected in QCA, much as the required minimum p value

tative data. For example, in their fsQCA analysis of the

for statistical signiﬁcance is researcher selected in regression

challenges facing the production and marketing of colloidal

analysis. However, and once again paralleling good research

silver water ﬁlters in Nepal, Bogler & Meierhofer ()

practice in statistics, past research establishes guidance for

were able to consider quantitative data, like population den-

what acceptable values for these cutoffs are and what the

sity and percentage of people treating water, alongside

risks of deviating from these standards are.

qualitative data, such as the reasons customers gave for

An equally important critique coming from qualitative

not buying ﬁlters and strategies used for new customers.

research traditions would emphasize that the calibration

A key conceptual difference between QCA and more tra-

scales required for QCA analysis are deeply – and poten-

ditional statistical methods is the different assumption

tially

regarding symmetry of relationships. Table 4 uses hypotheti-

reality. As such, there is a risk that the numeric scores pro-

cal data and a highly simpliﬁed example to describe why this

vide a sense of false precision. In this sense, QCA may be

assumption is important. The uppermost portion of the table

seen as overly positivistic and reductive. In response to

shows an example of a perfectly symmetric relationship.

these important critiques, we acknowledge that the use of

Here, we see that when Community Participation is present,

any particular research method is extremely unlikely to

we achieve the outcome of Sustained Water Services. In

enable perfect project outcomes (as deﬁned by either the

contrast, when Community Participation is absent, we see

problematically

–

simpliﬁed

representations

development community or as deﬁned by the end users).

abandoned water systems. Given these data, both statistical

However, we do argue that methodological diversity can

analysis and QCA ﬁnd a strong link between Community

help us move towards more sustainable WASH infrastruc-

Participation and Sustained Water Services. However, it

ture, by which we mean infrastructure that is used and

would also be possible to ﬁnd an asymmetric relationship

maintained by communities over time.

between these variables, as shown in the lower portion of Table 4. Here, we see that all cases with Community Participation achieve the Sustained Water Services outcome.

WHY AND WHEN TO USE QCA

However, when Community Participation is absent some cases experienced Sustained Water Services and others

The preceding sections described how to perform a QCA

showed Unsustained Water Services. Using statistical analy-

analysis; the following sections outline why and when

sis techniques such as a chi squared test, we would believe

QCA is appropriate. Generally, QCA provides a middle

that Community Participation is not a statistically signiﬁcant

Page 260

206

J. Kaminsky & E. Jordan

Qualitative comparative analysis

Journal of Water, Sanitation and Hygiene for Development

07.2

2017

factor in Sustained Water Services; the p value resulting from

interested in learning if there is more than one way to

a chi squared test is 0.12, which is above typical cutoffs for

attain an outcome of interest. Similarly, QCA requires

statistical signiﬁcance. In contrast, in set-based analysis like

explicit calibration deﬁnitions for both conditions and out-

QCA, we would describe these as set relationships and

comes; this ability to deal with and document nuanced

gain the insight that while Community Participation does

differences is a key strength of the method. In addition,

seem to be sufﬁcient for achieving Sustained Water Services,

the process of condition calibration is a systematic way

it is not necessary for achieving it. In other words, Commu-

for researchers to incorporate unexpected complexities

nity Participation is a valuable strategy, but not the only one.

and nuances that emerge during the analysis. As such,

The example in Table 4 also shows the importance of

QCA is well-suited for examining not only if a particular

considering conﬁgurational complexity in socially inﬂuenced

intervention results in the outcome we expect, but how

research topics like WASH. While regression methods exam-

and why such an intervention does (or does not) work.

ine the relative contribution of variables, holding other

Results of QCA studies include analysis of what variables

modeled variables equal, QCA seeks combinations of vari-

are necessary and/or sufﬁcient to achieve an outcome

ables that lead to an outcome and recognizes that there are

and pathways demonstrating what possible combinations

likely several different combinations of factors that may

of variables may lead to an outcome. To enable reproduci-

result in an outcome. This allows us to handle situations

bility and transparency of analysis, documentation of the

where uniformity of causal effects cannot be assumed. For

various analytic decisions detailed in this paper should be

example, in the hypothetical example shown in Table 4 we

reported for every QCA analysis. To date, the majority of

might add another causal condition such as the presence

QCA studies have been from academics, but QCA has

of a municipal utility (as we did earlier in Table 1). This

strong potential to be useful to practitioners as well.

resolves the excerpted hypothetical data in Table 4 showing

Often, evaluations of WASH projects rely on qualitative

that to achieve Sustained Water Services a community

methods; QCA offers a complementary approach to prac-

requires some combination of Community Participation, a

titioners who wish to rigorously evaluate the success of

Municipal Utility, or both working together. This hypotheti-

their projects across a limited number of cases in order

cal ﬁnding means that Community Participation and the

to gain an understanding of why interventions may have

presence of a Municipal Utility are substitutable conditions,

succeeded or failed in different contexts.

or conditions that are interchangeable in terms of achieving the outcome of Sustained Water Service.

ACKNOWLEDGEMENTS CONCLUSION

The authors are grateful to Dr Rachel Peletz and the journal’s anonymous reviewers for their comments on

QCA has not been used frequently in studies of water, sani-

drafts of this manuscript.

tation and hygiene interventions to date. However, given its ability to account for conﬁgurational complexity through the use of Boolean algebra or fuzzy logic, it is a

REFERENCES

promising method for WASH researchers that strongly complements more commonly used methods. QCA is a useful method to apply when there is a desire to examine the complex interactions between conditions that may inﬂuence the success of interventions, and particularly when there are strong reasons for considering both quantitative and qualitative data. In addition, QCA is useful in complex situations where researchers or practitioners are

Amenta, E. & Poulsen, J. D.  Where to begin a survey of ﬁve approaches to selecting independent variables for qualitative comparative analysis. Sociological Methods & Research 23 (1), 22–53. Annamalai, T. R., Narayanan, S., Devkar, G., Kumar, V. S., Devaraj, R. & Mahalingam, A.  Protocol for a Systematic Portfolio Review: To What Extent Have the Water, Sanitation, and Hygiene (WASH) Sub-Sectors Incorporated the Life-Cycle Approach into Policies, Programmes and

Page 261

207

J. Kaminsky & E. Jordan

Qualitative comparative analysis

Projects during Teh MDG Period? New Delhi, India: International Initiative for Impact Evaluation. Retrieved from http://www.3ieimpact.org/en/evidence/systematicreviews/details/337/. Berg-Schlosser, D., De Meur, G., Rihoux, B. & Ragin, C. C.  Qualitative comparative analysis (QCA) as an approach. In: Configurational Comparative Methods, vol. 51. Sage Publications, Los Angeles, pp. 1–18. Bogler, A. & Meierhofer, R.  The challenge of producing and marketing colloidal silver water filters in Nepal. Water 7 (7), 3599–3612. Boudet, H., Jayasundera, D. & Davis, J.  Drivers of conflict in developing country infrastructure projects: experience from the water and pipeline sectors. Journal of Construction Engineering and Management 137 (7), 498–511. http://doi. org/10.1061/(ASCE)CO.1943-7862.0000333. Braumoeller, B. F.  Guarding against false positives in qualitative comparative analysis. Political Analysis 23 (4), 471–487. http://doi.org/10.1093/pan/mpv017. Chatterley, C., Linden, K. G. & Javernick-Will, A.  Identifying pathways to continued maintenance of school sanitation in Belize. Journal of Water Sanitation and Hygiene for Development 3 (3), 411–422. Chatterley, C., Javernick-Will, A., Linden, K. G., Alam, K., Bottinelli, L. & Venkatesh, M.  A qualitative comparative analysis of well-managed school sanitation in Bangladesh. BMC Public Health 14 (1), 1. Clasen, T., Boisson, S., Routray, P., Torondel, B., Bell, M., Cumming, O., Ensink, J., Freeman, M., Jenkinds, M., Odagiri, M., Ray, S., Sinha, A., Suar, M. & Schmidt, W.  Effectiveness of a rural sanitation programme on diarrhoea, soil-transmitted helminth infection, and child malnutrition in Odisha, India: a cluster-randomised trial. The Lancet Global Health 2 (11), e645–e653. Gross, M.  Aligning Public-Private Partnership Contracts with Public Objectives for Transportation Infrastructure. Virginia Polytechnic Institute and State University. Retrieved from http://vtechworks.lib.vt.edu/handle/10919/28785. Gross, M. & Garvin, M.  Structuring PPP toll-road contracts to achieve public pricing objectives. Engineering Project Organization Journal 1 (2), 143–156. http://doi.org/10.1080/ 21573727.2011.572256. Guiteras, R., Levinsohn, J. & Mobarak, A. M.  Encouraging sanitation investment in the developing world: a clusterrandomized trial. Science 348 (6237), 903–906. House, R. S.  Public Utilities in the Age of Partnership: Lessons from Private Participation in Urban Water Supply. National University of Singapore, Singapore. Retrieved from http://scholarbank.nus.edu.sg/bitstream/handle/10635/ 113292/RSH_WaterPPP_Dissertation_Nov2014.pdf? sequence=1. Hug, S.  Qualitative comparative analysis: how inductive use and measurement error lead to problematic inference. Political Analysis 21 (2), 252–265. http://doi.org/10.1093/ pan/mps061.

Page 262

Journal of Water, Sanitation and Hygiene for Development

07.2

2017

Jordan, E., Gross, M., Javernick-Will, A. & Garvin, M.  Use and misuse of qualitative comparative analysis. Construction Management & Economics 29, 1159–1173. http://dx.doi.org/ 10.1080/01446193.2011.640339. Kaminsky, J. & Javernick-Will, A.  The internal social sustainability of sanitation infrastructure: a set theoretic approach. Environmental Science & Technology 48 (17), 10028–10035. http://doi.org/10.1021/es501608p. Krogslund, C., Choi, D. D. & Poertner, M.  Fuzzy sets on shaky ground: parameter sensitivity and confirmation bias in fsQCA. Political Analysis 23 (1), 21–41. http://doi.org/10. 1093/pan/mpu016. Meyer, J. W. & Rowan, B.  Institutionalized organizations: formal structure as myth and ceremony. American Journal of Sociology 83 (2), 340–363. OpenIDEO  Appraisal Tools for Sanitation Micro Finance. Retrieved October 3, 2016, from https://challenges.openideo. com/challenge/water-and-sanitation/second-evaluation/ appraisal-tools-for-sanitation-micro-finance. Patil, S. R., Arnold, B. F., Salvatore, A. L., Briceno, B., Ganguly, S., Colford Jr, J. M. & Gertler, P. J.  The effect of India’s total sanitation campaign on defecation behaviors and child health in rural Madhya Pradesh: a cluster randomized controlled trial. PLoS Medicine 11 (8), e1001709. Pickering, A. J., Djebbari, H., Lopez, C., Coulibaly, M. & Alzua, M. L.  Effect of a community-led sanitation intervention on child diarrhoea and child growth in rural Mali: a clusterrandomised controlled trial. The Lancet Global Health 3 (11), e701–e711. Ragin, C.  Fuzzy-Set Social Science. University of Chicago Press, Chicago. Ragin, C.  Redesigning Social Inquiry: Fuzzy Sets and Beyond. University Of Chicago Press, Chicago. Ragin, C.  The Comparative Method: Moving beyond Qualitative and Quantitative Methods, 2nd edn. University of California Press, Oakland, CA. Ragin, C., Drass, K. & Davey, S.  Fuzzy-Set/Qualitative Comparative Analysis (Version 2.0). Department of Sociology, University of Arizona, Tuscon, Arizona. Retrieved from www.fsqca.com. Rihoux, D. B. & Ragin, C.  Configurational Comparative Methods: Qualitative Comparative Analysis (QCA) and Related Techniques. Sage Publications, Los Angeles. Rohlfing, I.  Mind the gap: A review of simulation designs for Qualitative Comparative Analysis. Research & Politics 2 (4), 1–4. http://doi.org/10.1177/2053168015623562. Rosenqvist, T., Mitchell, C. & Willetts, J.  A short history of how we think and talk about sanitation services and why it matters. Journal of Water, Sanitation and Hygiene for Development 6 (2), 298–312. http://doi.org/10.2166/ washdev.2016.118. Santosh Kumar Delhi, V., Mahalingam, A. & Palukuri, S.  Governance issues in BOT based PPP infrastructure projects in India. Built Environment Project and Asset Management 2 (2), 234–249.

208

J. Kaminsky & E. Jordan

Qualitative comparative analysis

Schmidt, W.-P.  The elusive effect of water and sanitation on the global burden of disease. Tropical Medicine & International Health 19 (5), 522–527. http://doi.org/10. 1111/tmi.12286. Schneider, C. Q. & Wageman, C.  How To … Use Software for Set-Theoretic Analysis: Online Appendix to ‘Set-Theoretic Methods for the Social Sciences’. Cambridge University Press, Cambridge, MA. Retrieved from www.cambridge.org/US/ download_ﬁle/134448/.

Journal of Water, Sanitation and Hygiene for Development

07.2

2017

Spencer, J. H.  Household strategies for securing clean water the demand for piped water in Vietnam’s peri-urban settlements. Journal of Planning Education and Research 28 (2), 213–224. Welle, K., Williams, J., Pearce, J. & Befani, B.  Testing the Waters: A Qualitative Comparative Analysis of the Factors Affecting Success in Rendering Water Services Sustainable Based on ICT Reporting. WaterAID, Uganda. Retrieved from http://opendocs.ids.ac.uk/opendocs/handle/123456789/ 7099.

First received 3 October 2016; accepted in revised form 18 January 2017. Available online 8 March 2017

Page 263

Oï¬&#x192;cial Journal of the World Water Council

Water Policy

ISSN 1366-7017 iwaponline.com/wp

Water management and water infrastructure are preconditions for civilization, and demands on our water resources are increasing. Throughout the world there is therefore a growing need to build a capacity for integrated water management. But this calls for a new dialogue between different private and public communities – policy making, diplomatic, administrative, financial, legal and technical/scientific – along with the traditional water communities. Water Policy provides a forum for this dialogue. The journal’s scope includes: • • • • • • • • • • • • • • •

Ecosystems, engineering, management and restoration Engineering and design River-basin and watershed management Multiple uses of water Pollution monitoring and control Management, use and sharing of trans-boundary waters, treaties and allocation agreements Capacity building Flood control and disaster management Groundwater remediation and the conjunctive use of groundwater and surface water Public participation, consensus building and confidence building Conflict management and negotiations of water resources Demand management Commercialization of water Integrated water resources management Allocation of risks among stakeholders For more details, visit iwaponline.com/wp

Page 266

Water Policy 19 (2017) 358–375

Linking environmental flows to sediment dynamics Diego García de Jalóna, Martina Bussettinib, Massimo Rinaldic, Gordon Grantd, Nikolai Friberge, Ian G. Cowxf, Fernando Magdalenog and Tom Buijseh a

Corresponding author. Dept of Natural Systems & Resources, ETSI Montes, Forestales y Medio Natural, Universidad Politécnica de Madrid, Ciudad Universitaria, 28040 Madrid, Spain. E-mail: diego.gjaoln@upm.es b Institute for Environmental Protection and Research (ISPRA), Rome, Italy c Dipartimento di Scienze della Terra (UNIFI), 50139 Firenze, Italy d Oregon State University/USDA, Corvallis, USA e Norwegian Institute for Water Research (NIVA), Oslo, Norway f Hull International Fisheries Institute, The University of Hull, HU6 7RX Hull, UK g Centro de Estudios de Técnicas Aplicadas (CEDEX), Madrid, Spain h Department of Freshwater Ecology & Water Quality (Deltares), 3584 CB Utrecht, The Netherlands

Abstract This is a policy discussion paper aimed at addressing possible alternative approaches for environmental flows (eFlows) assessment and identification within the context of best strategies for fluvial restoration. We focus on dammed rivers in Mediterranean regions. Fluvial species and their ecological integrity are the result of their evolutionary adaptation to river habitats. Flowing water is the main driver for development and maintenance of these habitats, which is why e-Flows are needed where societal demands are depleting water resources. Fluvial habitats are also shaped by the combined interaction of water, sediments, woody/organic material, and riparian vegetation. Water abstraction, flow regulation by dams, gravel pits or siltation by fine sediments eroded from hillslopes are pressures that can disturb interactions among water, sediments, and other constituents that create the habitats needed by fluvial communities. Present e-Flow design criteria are based only on water flow requirements. Here we argue that sediment dynamics need to be considered when specifying instream flows, thereby expanding the environmental objectives and definition of e-Flows to include sediments (extended e-Flows). To this aim, a hydromorphological framework for e-Flows assessment and identification of best strategies for fluvial restoration, including the context of rivers regulated by large dams, is presented. Keywords: Ecological status; Environmental flows (e-Flows); Flow regulation; Hydromorphology (HYMO); Large dam; River management; Sediments This is an Open Access article distributed under the terms of the Creative Commons Attribution Licence (CC BY-NC-SA 4.0), which permits copying, adaptation and redistribution for non-commercial purposes, provided the contribution is distributed under the same licence as the original, and the original work is properly cited (http://creativecommons.org/licenses/by-nc-sa/4.0/). doi: 10.2166/wp.2016.106 © 2017 The Authors Page 267

D. G. de Jalón et al. / Water Policy 19 (2017) 358–375

359

Introduction We are living in the Anthropocene Era and we are becoming increasingly aware of the large body of evidence showing that human interactions with the hydrological cycle have serious consequences for rivers and ecosystems (Vörösmarty et al., 2003). Gerten et al. (2013) pointed out the existence of a ‘planetary boundary’ for fresh water used by humans, and proposed ways forward to refine and reassess it. They suggested that a key element involves quantifying local water availabilities taking account of environmental flow (e-Flow) requirements. Human populations have water demands that are prioritized according to various needs: (1) vital water (drinking water, hygiene, sanitation), (2) social water (gardens, swimming pools) and especially (3) commercial water (required for hydropower, intensive agriculture, industrial processes, tourism infrastructure). In warmer climates all these demands are greater than in colder climates. Typically, commercial water use represents more than 80% of all demand, which is often even greater than water availability. Socio-economic drivers, such as agriculture, energy production and land development, and the pressures they create on water resources (e.g. through construction and operation of dams, irrigation systems) have important effects on hydromorphology (HYMO) and ecosystems. This is omnipresent across the whole of Europe, but particularly evident in Mediterranean rivers, due to their combination of a strong external water demand and HYMO characteristics related to low specific runoff. As temperature and rainfall are out of phase with each other in semi-arid climate regimes, i.e. higher summer temperatures and low river flows, and vice versa, Mediterranean rivers cannot naturally satisfy water demands. This situation has justified the construction of a huge number of large reservoirs. According to the International Commission of Large Dams (ICOLD), the European member states with the largest number of reservoirs are: Spain (1,082), Turkey (976), France (713), the UK (607) and Italy (542) (ICOLD, 2007). Southern Mediterranean countries (excluding Turkey) are clearly the ones with the largest numbers of large dams, followed by Western countries and Eastern countries (including Russia and Ukraine) (Figure 1). Of all European regions, Mediterranean countries also use the most water stored in reservoirs for irrigation. In addition, historical land overexploitation and today’s intensive agriculture on slopes cause high catchment erosion, sediment yield, and transport. This latter problem is widespread and shared by continental lowland basins too. Dams and other pressures, such as weirs and water abstraction, have important effects on the HYMO and ecosystems. The environmental effects of dams and the reservoirs they impound vary greatly with their regional or environmental setting, which controls the natural flow regime, and their size (morphometry and capacity) and purpose, which affect dam outlet and reservoir characteristics and operational procedures of the dam and its reservoir (see Figure 2). The impacts of large dams have a global dimension and there are many comprehensive reviews of the effects and ecological impacts downstream of dams (e.g. Ward & Stanford, 1979; Petts, 1984; Williams & Wolman, 1984; Vörösmarty et al., 1997; Grant et al., 2003; Grant, 2012). Lloyd et al. (2004) estimated that the maximum water storage behind 746 of the world’s largest dams was equivalent to 20% of global mean annual runoff and the median water residence time behind those impoundments was 0.40 years. However, dams do not only regulate water flow. More recently, Vörösmarty et al. (2003) estimated that more than 50% of the basin-scale sediment flux in regulated basins is trapped Page 268

360

D. G. de Jalón et al. / Water Policy 19 (2017) 358–375

Fig. 1. (a) Number of large dams per European countries (data from ICOLD). (b) Water abstraction used for irrigation by European countries (European Environmental Agency, 2010).

in artificial impoundments (Figure 2(a)) based on discharge-weighting large reservoirs trap 30% and small reservoirs an additional 23%. Water Framework Directive and environmental flows This paper draws on the Water Framework Directive (WFD) common implementation strategy (CIS) Guidance Document ‘Ecological flows in the implementation of the Water Framework Directive’ (European Commission, 2015) as a starting point. Our main objective is to emphasize the importance of HYMO, especially for rivers that are heavily regulated by large dams. Thus, we adopt the perspective of the Guidance Document that environmental flows (e-Flows) are more than just minimum flows, and have to include all the components of the hydrological regime. E-Flows play different roles in different fluvial settings. Ideally we can view e-Flows as restoration measures since their aim is to support the achievement of good ecological status in rivers subject to hydrological pressures. When these pressures are exerted by major infrastructures such as large dams, however, the changes caused in the river ecosystem can be so profound that e-Flows can only be considered as mitigation measures. In addition, River Basin Management Plans (RBMPs) often consider e-Flows as preventive measures for many river sections that are not regulated or affected by water abstraction. Furthermore, for fluvial segments under some form of protection, e-Flows represent Page 269

D. G. de Jalón et al. / Water Policy 19 (2017) 358–375

361

Fig. 2. (a) Reservoirs are sediment traps. Barasona Reservoir in R. Esera (Ebro Basin). Over 80% of reservoir capacity has been lost. (b) Armouring river bed (R. Pas). Incision caused by smaller substrate size selective erosion. (c) Lateral bank erosion in meander. R. Gallego. (d) Sediment deposits in lateral banks. R. Guadalete (photograph by D. G. De Jalon, 2015).

conservation measures. With such different objectives of e-Flows, the question arises: should all types of e-Flows be quantified in a single manner? In the context of the WFD, ecological flows are defined as a flow regime consistent with the achievement of the environmental objectives of a water body (i.e. good ecological status – for natural water bodies; good ecological potential – for heavily modified – HMWB – and artificial water bodies; and good quantitative and chemical status for groundwater bodies). Ecological flows represent therefore a ‘potential’ measure to reach the objectives, as the real measure will derive from the evaluation of all the physical, legal and socio-economic constraints related to the water body. As a potential measure, ecological flows come into play when the results of the WFD Art.5, risk analysis on a catchment, show that some water bodies are at risk of failing their objectives due to an inadequate (in terms of magnitude and timing) flow regime (e.g. a reach downstream of a reservoir). Identifying whether it is possible to manage such a regime to make it consistent with the environmental objectives set, requires determining the current natural or anthropic constraints on the catchment (HYMO economic, social, etc.) through analysis of scenarios. Such scenarios need to evaluate remedial measures not only in terms of their impacts on the status of water bodies, but also on the uses of water in Page 270

362

D. G. de Jalón et al. / Water Policy 19 (2017) 358–375

the actual system. This is crucial when addressing HMWBs, as they are designated on the basis of their legitimate use. Problem description The CIS Guidance Document ‘Ecological flows in the implementation of the Water Framework Directive’ (European Commission, 2015) presents an overview of methodologies for e-Flows implementation. However, it does not address in depth certain crucial aspects, among which is the definition of a HYMO regime consistent with a desired ecological state and relevant scales to be used in the assessment. The e-Flows concept: only water? Rivers and their ecosystems reflect hierarchies of control. We can consider rivers as complex organisms, whose functioning needs both water flowing and its particular metabolites (sediments; woody debris; particulate organic matter; dissolved solids and gases). However, a large dam on a river disturbs not only the natural water flow regime, but often to a greater extent, the natural fluxes of these metabolites. Therefore, when we use environmental flows as an instrument to improve the ecological status of water bodies, we should also consider the fluxes of all ‘metabolites’ that allow the existence of biological communities. Such holistic methodologies considering the many interacting components of aquatic systems, including sediments, are increasingly recommended although in many cases assessment of e-Flows is mainly based on hydrological and hydraulic assessment (Anderson et al., 2006; Meitzen et al., 2013). But, other approaches defining e-Flows integrating diverse disciplines should be mentioned, like the Holistic Approach (Arthington et al., 1992). The Building Block method developed in South Africa (King & Louw, 1998) considers geomorphology as the physical template for biological processes. Sediment assessments for e-Flows have been carried out by others, such as King et al. (2003) who developed the Downstream Response to Imposed Flow Transformation methodology. However, the sediment-flow component is not included in e-Flows: they remain e-Flows confined to only water and incorporate geomorphology as an effect of flow (‘large floods mobilize coarse sediments’). The Building Block method has been widely applied in Southern and Eastern Africa and in many other parts of the world, and is suggested as the basic methodology to be applied in the UK (Acreman & Dunbar, 2004). In the context of the WFD, e-Flows represent a possible measure to reach the objectives of good ecological status or potential. There is still too little experience with the implementation of e-Flow based measures; a review of the hydrological measures applied at the EU level, based on the information derived from the RBMPs showed that they have been established according to the ‘minimum flow’ concept (Sánchez-Navarro & Schmidt, 2012). As such, no consideration has been given to the morphological evolution of the affected reaches or channels, which could have caused a consistent channel conveyance change. Beyond the flow regime, sediment transport plays a fundamental role in determining and maintaining channel morphology and related habitats. A river habitat is the result of a balance between interacting geomorphological forces: water, sediments and riparian vegetation in a spatial template (fluvial reach). Water flow has the hydraulic Page 271

D. G. de Jalón et al. / Water Policy 19 (2017) 358–375

363

energy able to erode, transport and deposit sediments and riparian vegetation growth is able to consolidate deposited sediments, but old vegetation stands may reduce water erosion capacity. Thus, habitat morphology depends also on the structure and composition of the riparian stand and on its present interaction with the hydro-geomorphic pattern of the river. The importance of sediment transport and related geomorphic processes as key components to evaluate has only recently begun to be acknowledged. Meitzen et al. (2013) emphasized how fluvial geomorphology and riverine ecology represents an ideal confluence to examine the contribution of the geomorphic field tradition to environmental flows. They developed a question-based framework that will facilitate holistic and interdisciplinary environmental flow assessments. Definition of environmental objectives and monitoring the efficiency of measures According to the WFD, the environmental objective of a river water body coincides with its ecological status, mainly given by the combination of the status of the relevant biological quality elements, each assessed through indicators. However, the majority of these commonly used indicators do not respond, with a necessary degree of sensitivity, to hydrological and morphological pressures or to multi-stressor systems, as acknowledged by the scientific community (Friberg et al., 2011, 2013). Therefore, objectives can hardly be defined and/or measured in terms of current biological indicators. Among biological quality elements, fish is the most reactive one to HYMO pressures, but no efficient/official method to assess their status is currently available for use in Mediterranean countries like Spain or Italy. The notions of ‘ecological status’ and ‘ecological potential’ are highly dependent upon generally negotiated choices of metrics and thresholds, the definitions of which are constrained by the limits of assessment methods, their interpretability and ability to accurately assign a given system to a particular class (Friberg et al., 2011). Therefore, there is a need to develop (HYMO) pressures – specific indicators based on HYMO responsive biological elements, including alternative sampling strategies. Moreover, because of the strong nexus between HYMO and biology, where hydrology is the main pressure affecting the status of water bodies, it is suggested to define objectives also in the context of a HYMO restoration action and measure it through hydrological and morphological parameters. The e-Flows in Mediterranean streams: the Bonsai river syndrome Mediterranean and semi-arid countries are heavily affected by large dams and thus the implementation of adequate e-Flows is strongly needed. We have seen that dam impacts have wider ecosystem effects, and to design e-Flows we need to determine the drivers of flow and sediment changes below dams. A framework summarizing the effects of large dams on fluvial processes and HYMO variables is shown in Figure 3 (García de Jalón et al., 2013). Besides the changes to instream flows, we find other significant fluvial processes altered by dams, like sediment fluxes, bank stabilization, substrate armouring, riparian vegetation encroachment, and even water physico-chemical degradation. These processes are responsible for changing habitats that often are unable to maintain reference communities and often cause a decline in biodiversity and an invasion of exotic species. Page 272

364

D. G. de Jalón et al. / Water Policy 19 (2017) 358–375

Fig. 3. Conceptual framework of large dams and reservoirs effects on HYMO and physico-chemical (PHYCHE) processes and variables (POM ¼ particulate organic matter; LWD ¼ large woody debris) (Source: García de Jalón et al., 2013).

Water releases downstream of a dam entrain sediments through a size selective process that causes river substrate evolution with different stages (Collier et al., 2000):

• Over many years there is a ‘wave’ of sediment deficit that moves downstream along the river, changing its substrate traits: sediment calibre increases, as does armouring. • Later, substrate comes to equilibrium between the regulated flow regime and sediment input by tributaries. • The effects on the biota vary in space and time according to these stages of substrate change. Therefore, setting e-Flows (including water and sediments) must take into account this substrate evolution for each reach of the river. In order to clarify concepts, we introduce an alternative term, Environmental Water & Sediment flows (EWS-Flows), leaving traditional term e-Flows for only water environmental flows. E-Flows are assessed by a variety of methods and approaches, but are rarely applied in Mediterranean countries. Most of the e-Flows proposed in the RBMPs represent a very low percentage of the mean annual flows (Figure 4). Those particular e-Flow regimes may be supported by modelling but empirical data proving their positive effect on downstream waterbody status enhancement is still missing (European Commission, 2015). E-Flows should also be able to maintain essential geomorphological processes responsible for the habitat required by native species, and this is not adequately emphasized in the CIS Guidance Document or in most methods currently used in Europe to set e-Flow levels. We should remark on the great influence of riparian vegetation dynamics (Hupp, 1999; Corenblit et al., 2007) that must be considered in specifying e-Flows below large dams, especially in warm climates that promote intensive growth and recruitment. This vegetation encroachment stabilizes the new channel, even within extraordinary floods. Page 273

D. G. de Jalón et al. / Water Policy 19 (2017) 358–375

365

Fig. 4. Percentage of natural flow represented by designated e-Flows for the river water bodies in the Spanish Duero Basin District from its RBMP.

Prevention of vegetation encroachment could be a basic objective of effective e-Flows, particularly in Mediterranean streams, where common irrigation reservoirs release high summer flows, thus supporting maximum plant growth potential as the normal summer drought is eliminated (Magdaleno & Fernández, 2011; Stella et al., 2013; González del Tánago et al., 2015; Lobera et al., 2015). Ultimately, regulated river dimensions are so reduced that they develop into a small remnant: a ‘Bonsai river’. Policy options Water and sediment transport in rivers are intrinsically linked and actions on one component will interact with the other. Therefore, managing environmental flows without considering sediment dynamics will not yield the desired positive effects. By contrast, the combined management of the two components may have more cost-benefit impacts, from reduced water releases to temperature mitigation or pollutant abatement. We propose a policy where water and sediments are considered together when dealing with the impacts of reduced flows and the use of e-Flows as a possible mitigation action, expanding the present definition of e-Flows and coupling flows with sediment dynamics. Benefits from this policy proposal come from both the ecological perspective as well as from reservoir management as sediments cause problems through siltation of reservoirs and loss of their functionality and ability to regulate, and by silting up river beds. With this objective, we define a HYMO framework to assess the status and impacts of sediment and flow management and an e-Flows toolbox adapted to couple flows to sediment. HYMO framework for e-Flows Although the importance of sediment and geomorphological processes has been acknowledged in the CIS Guidance Document ‘Ecological flows in the implementation of the Water Framework Directive’ (European Commission, 2015), the links between hydrology and channel morphology are still only marginally considered in the evaluation of e-Flows. Within the context of the REFORM project (http://www.reformrivers.eu/), a multi-scale, spatialtemporal geomorphological framework has been developed. This framework can be used to assess Page 274

366

D. G. de Jalón et al. / Water Policy 19 (2017) 358–375

HYMO conditions and to identify suitable restoration measures. In this section, we will set e-Flows within the context of the HYMO assessment framework. The aim of this section is to provide a ‘road map’ on possible future developments with wider inclusion of geomorphological processes. The basic hypothesis (paradigm) is that enhancing morphological conditions will promote a positive ecological response. Figure 5 illustrates three different groups of possible actions (hydrological regime, sediment and woody debris transport, together with direct morphological enhancement) producing morphological change (enhancement) and ecological response. Because hysteresis affects HYMO and ecological processes, complementary actions may be needed to speed up the habitat recovery processes. Measures like direct morphological reconstruction, removing mature riparian forests, and eliminating or reducing transversal and longitudinal barriers, are examples of these complementary measures. In fact, it is now widely recognized that the geomorphological dynamics of a river and the functioning of natural physical processes are essential to create and maintain habitats and ensure ecosystem integrity (e.g. Kondolf et al., 2003; Wohl et al., 2005; Fryirs et al., 2008; Habersack & Piégay, 2008). The current approach to setting e-Flows is to focus on the hydrological regime in anticipation of promoting some ecological response. However, two other types of actions are also possible: focusing on the sediment transport regime (e.g. releasing sediments downstream of dams or other obstructions) or directly manipulating channel morphology. Any of these actions may induce morphological channel changes, therefore promoting habitat recovery and diversity. The choice of the best option to be considered in combination with changes in the hydrological regime (i.e. sediment transport versus morphological enhancement) depends on the specific context, for example the reach sensitivity and morphological potential (see below). Therefore, selecting the appropriate measures requires setting the river reach within a wider spatial-temporal framework. We make use of a HYMO assessment framework to provide a stronger foundation for determining eFlows. The spatial and temporal contexts are based on the multi-scale, process-based, hierarchical framework developed in the REFORM project (Gurnell et al., 2014). The framework is structured into a sequence of procedural stages and steps to assess river conditions and to support the selection of appropriate management actions (REFORM Deliverable 6.2; Rinaldi et al., 2015). The overall framework incorporates four stages (Figure 6): I. delineation and characterization of the river system; II. assessment of past temporal changes and current river conditions; III. assessment of future trends; and IV. identification of management actions.

Fig. 5. Potential e-Flows actions involving possible modifications of the hydrological regime, sediment transport, or morphological reconstruction (Rinaldi et al., 2015). Note that the current e-Flows approach linking flows directly to ecological response ignores such complex interactions. Page 275

D. G. de JalĂłn et al. / Water Policy 19 (2017) 358â&#x20AC;&#x201C;375

367

Fig. 6. Structure of the overall REFORM HYMO framework (Rinaldi et al., 2015). On the right side, the graph emphasizes that the present state of the river system represents a spot within a long trajectory of evolution that needs to be known to understand current conditions and possible future trends. On the left side, the multi-scale hierarchical framework used for delineation and characterization of the fluvial system is presented (Gurnell et al., 2014).

Stage I: delineation and characterization of the river system. Stage I aims to provide a catchment-wide delineation, characterization and analysis of the river system. This is fundamental to properly set the existing HYMO pressures (dams, weirs, water abstraction, etc.) within a catchment-wide context, and to better understand the factors controlling channel morphology and processes in the current condition. Relevant aspects for e-Flows include: identification of main sediment sources, delivery processes, and sediment transport along the river network to set the existing alteration (e.g. dam) in the catchment context; evaluation of effective discharge and of the specific flow needed to initiate sediment transport; and evaluation of impacts of existing alterations on sediment budget. Stage II: assessment of past temporal changes and current river conditions. After setting the stream and causes of alteration in an appropriate spatial context, it is fundamental to investigate past conditions and factors influencing changes. A first step is to identify the major changes in controlling variables (e.g. factors influencing flow and sediment transport) that may have determined changes in the channel and river corridor conditions over the last decades or centuries. These steps aim to reconstruct trajectories of morphological changes of the potentially impacted reaches. Relevant aspects for e-Flows include understanding how HYMO alterations (e.g. dams, weirs, water abstraction) have impacted channel morphology, and the spatial and temporal extent of any alteration. Page 276

368

D. G. de JalĂłn et al. / Water Policy 19 (2017) 358â&#x20AC;&#x201C;375

Stage III: assessment of future trends. Stage III applies methods and procedures to assess river conditions and the degree of HYMO alteration related to existing pressures. This type of assessment requires knowledge of past and current conditions. Three types of assessment are carried out: (1) Hydrological assessment: pre-impact and post-impact periods are analysed and the deviation of the hydrological regime from unaltered conditions quantified. (2) Sediment budget assessment: pre-impact and post-impact periods are analysed and the deviation of the sediment regime from unaltered conditions quantified. (3) Morphological assessment: consists of a geomorphological evaluation of river conditions including assessment of channel forms and processes, geomorphological adjustments, and human alterations. The assessments enable classification of the state (e.g. good, poor) of each investigated river reach to identify portions of the river system that potentially require different types of management actions (e.g. preservation or enhancement). Relevant aspects for e-Flows include: hydro peaking, modification of effective discharge, impacts on sediment budgets downstream of barriers to sediment transport. Stage IV: identification of management actions. Stage IV includes an assessment of potential morphological changes, identification of potential restoration measures, and evaluation of their impacts on future morphological trends. The first step diagnoses the condition and sensitivity of specific reaches to changes in hydrological and sediment conditions that can be associated with e-Flows. Adoption of some restoration action requires an evaluation of the likelihood that river change will take place, and of the morphological potential that could be achieved in response to a given modification of flows. This assessment is based on the knowledge gained during the previous stages, i.e. on current conditions and past changes. Based on the assessment of sensitivity and of morphological potential, target reaches and possible morphological conditions that can be achieved are identified. The next steps are aimed at identifying possible restoration actions, and assessing scenario-based possible future trends related to selected actions. Relevant aspects for e-Flows include: identification of flows needed to initiate transport, coupling peak flows with sediment availability, determining and maintaining channel morphology and related habitats, quantification of sediment deficit or surplus, release of sediments downstream of barriers, removal of barriers and evaluation of effectiveness of different measures. Sediment flow management: the particular case of sediment replenishment Managing hydrological and sediment regimes together to meet geo-ecological objectives in dynamic riverscapes deals with measures such as: (a) to modify flow regime; (b) to modify sediment transport regime; (c) to modify sediment supply; and (d) to engineer channels and habitat. Any decision making on the measure(s) to be used (single or a combination of them) needs to diagnose the state of the channel and to predict its response to such measures. These diagnoses and predictions are based on the comparative assessment of upstream sediment supply to channel transport capacity. Suitable and detailed methods for this assessment are shown in Grant et al. (2003) and Schmidt & Wilcox (2008). Too much sediment in the channel must be managed primarily by reducing the production at source or intercepting it before it reaches the channel. The lack of sediment in a river reach is a more common Page 277

D. G. de Jalón et al. / Water Policy 19 (2017) 358–375

369

problem than excess sediment. The reintroduction of sediments in a reach with sediment deficit can be carried out by means of upstream dam removal, or by mitigating a dam’s trapping effects, or by adding sediments directly to the river. Below dams, fluvial systems need sediments for recovering their natural forms and functioning. In addition, water managers need to recover reservoir storage capacity lost to sedimentation. Thus, a win-win option requires recovering connectivity of sediment flow, from the reservoir basin to the river downstream of the dam. To address these two issues, the accumulated sediments must be relocated below the dam either through flushing from the reservoir (White, 2001) or by replenishment below the tail water (Figure 7). This latter process has been implemented in Japan, the USA and Switzerland (Cajot et al., 2012). Sediment replenishment basically consists of dredging or excavating the accumulation of sediments in a dam’s reservoir and transporting them to the reach just below the dam, where natural or artificial floods will distribute them along the riverbed. In order to improve downstream ecological status, optimal sediments for replenishment must be selected, as coarser substrates are more beneficial for benthic communities than silt, which may impact interstitial habitats by clogging bed sediments, and causing high turbidity (Ock et al., 2013). The construction of check-dams, located upstream of reservoirs, where

Fig. 7. Scheme of a sediment bypass tunnel system associated with a reservoir designed with the sediment intake located at the reservoir head under free surface conditions (based on Auel & Boes, 2011). Page 278

370

D. G. de JalĂłn et al. / Water Policy 19 (2017) 358â&#x20AC;&#x201C;375

coarse particles settle, may trap larger sediments before they enter the functional reservoir and facilitate their removal by land-based excavation, and do not require any water level modification in the larger reservoir (Okano et al., 2004). In order to relocate sediments in the riverbed efficiently, we need to know the effects of grain size, the amount of sediments replenished, the frequency of operations and when the sediment should be deposited (Cajot et al., 2012). Another effective measure to limit sediment trapping by reservoirs and to decrease the reservoir sedimentation involves constructing sediment bypass tunnels. These tunnels route sediments (both bed load and suspended load) around the reservoir into the tail water during flood events, thereby reducing sediment accumulation. The number of actual sediment bypass tunnels globally is, however, limited (six in Switzerland and five in Japan) due to high capital and maintenance costs. The design of a bypass tunnel consists of a guiding structure in the reservoir, an intake structure with a gate, a short and steep acceleration section, a long and smooth bypass tunnel section, and an outlet structure (Auel & Boes, 2011; Figure 7). Fukuda et al. (2012) demonstrated the recovery of riffles and pools and the grain size distribution in the downstream reaches below Asahi Dam reservoir, after the construction of a sediment flushing tunnel. Other methods to eliminate the sediments accumulation in the reservoirs are based on a floating platform with hydraulic equipment that dredges the compacted sediment and pumps it through a piping system to be released near the bottom outlet of the dam, where it can be eroded and passed through the outlet (Figure 8). This hydraulic system can be set to move sediment to the dam downstream section at a rate similar to the sediment yield reaching the reservoir, in order to maintain its storage capacity (Bartelt et al., 2012). It should be noted that in gravel bed rivers, this method has a major drawback, as it is only able to remove fine sediments, whose release in the downstream reaches can degrade benthic communities. Conclusions and recommendations Estimation of e-Flows that are necessary to maintain the desired river ecological state is not straightforward, as the quantitative links between HYMO and biology are not yet well known, due to the

Fig. 8. Hydraulic pumping system to remove accumulated sediments from reservoir tails into the bottom outlet of the dam, in order to be flushed downstream of the dam (Bartelt et al., 2012). Page 279

D. G. de JalĂłn et al. / Water Policy 19 (2017) 358â&#x20AC;&#x201C;375

371

insufficient number of consistent data and to the weak response of current biological metrics to HYMO pressures. Geomorphic dynamics of a river and the functioning of natural physical processes are essential to create and maintain habitats and ensure ecosystem integrity and the links between hydrology and geomorphology are generally well known. Therefore, one approach to estimating e-Flows is to identify those flows required to maintain certain geomorphic processes and forms that directly contribute to aquatic habitat and ecosystem functioning. Such an approach would broaden the current strategy for setting e-Flows, which is to focus on the hydrologic regime in anticipation of promoting some ecological response. Elements of this broadened approach include other types of actions beyond specifying flows alone, such as focusing on the sediment transport regime (e.g. releasing sediments downstream of dams or other obstructions), or directly manipulating channel morphology (i.e. morphological reconstruction). Any of these actions (HYMO-based measures) may induce morphological channel changes, therefore promoting habitat recovery and diversity. The choice of the best option to be considered in combination with changes in the hydrologic regime (i.e. sediment transport vs. morphological reconstruction) depends on the specific context, for example the reach sensitivity and morphological potential. Therefore, selecting the appropriate measures requires setting the river reach within a wider spatial-temporal framework. It is also important to state that sediment releases should be timed with natural sediment fluxes â&#x20AC;&#x201C; just as is the case for water releases. Within the context of the project REFORM, such a multi-scale framework has been developed and can be used as a strong methodological foundation for determining e-Flows, dealing with hydrological, morphological and ecological processes in concert. Constraints on flow and sediment management The implementation of e-Flows is constrained by our understanding of the ecological processes, of the services they provide, and by the socio-economic requirements on water resources. Whilst the latter issue is clearly recognized, ecosystem services related to natural processes and to eFlow releases are yet not sufficiently acknowledged by managers and stakeholders. Therefore, ecological benefits that can be achieved by e-Flows, and in particular the added value of considering sediment related e-Flows, should be clearly justified, both from an ecological perspective (maintenance and development of habitats) and from an economic one (e.g. less water discharge if combined with sediment delivering strategy, mitigate incision problems, etc.). Precepts to re-policy Water and sediments are intrinsically interconnected in natural river systems. Fluvial communities have evolved to adapt to this interaction, and thus many of their habitat requirements depend on HYMO dynamics. Setting e-Flows (EWS-Flows including water and sediments) must take into account past morphological evolution and trajectories as well as the current status of the river system, and the water and sediment fluxes in the network, to inform the possible scenarios, prior to implementation of measures in each targeted reach of the river. Sediment management therefore needs to be built into the analytical and decision-making framework. Page 280

372

D. G. de Jalón et al. / Water Policy 19 (2017) 358–375

Environmental flows, including sediment EWS-Flows, should be implemented and monitored within an adaptive management framework. Monitoring the outcomes of e-Flows is needed because our understanding of water and sediment requirements by key aquatic biota and ecosystem functions is not precise and often critical decisions are made with relatively weak ecological evidence to support them. E-Flow monitoring programmes should have a practical approach to ensure that e-Flow implementation achieves its objectives and, in any case, identifies gaps and provides recommendations for relevant improvements. Recommendations for future actions The management of intensive flow regulated water bodies must be framed in a policy that includes sediment management strategies into e-Flows, which are what we have called EWS-Flows. We recommend starting with demonstration or pilot projects in large reservoirs where sediment release mechanisms (e.g. in association with high flow events) can be planned and implemented. These experimental EWS-Flow releases should be assessed evaluating the ecological recovery of downstream reaches together with the benefits of reservoir operation (avoiding dams silting up and maintaining reservoir capacity). The basic issue to which a dam manager must respond is how much sediment needs to be transferred downstream and how frequently during a certain period. Once cases have been implemented and practical experience has been developed it will possible to generate a plan for a widespread application of EWS-Flows. However, we must mention that e-Flow and EWS-Flow implementation (via RBMP linked to relicensing dams) implies changes in the conditions of water concessions and previously acquired rights, making it necessary to consider the legal constraints and possible socio-economic compensation. There is a need for long-term research (that should incorporate existent experiences, including the outcomes from the REFORM project), based on the following specific critical points:

• Ecological benefits of e-Flows, although acknowledged, are not well supported by quantitative evi• • • •

•

dence and too few well-documented cases exist. As the majority of current biological methods do not detect the impact of HYMO pressures or the effects of HYMO-based measures, including e-Flows, with a necessary degree of precision, revision of such methods should be promoted. Alternative biological methods should be developed, accounting for HYMO functionality measurement, riparian zones and stressor-specific deviation estimation. Riparian vegetation should be considered as a quality element per se as well as HYMO traits. Long-term experiments are needed to implement and validate the revised/new approaches. In the meantime, as process-based HYMO assessment methods can easily and directly assess HYMO alteration, they should be used along the whole gradient to support ecological assessment. Moreover, assessment of spatio-temporal alteration of local HYMO (physical habitat) could be used as a proxy for ecological status. Experimental use of reservoirs for research could provide empirical data that link instream flows with biological elements and their ecological status. Also, these experiments could provide valuable data on how coupling flow and sediments create adequate habitats to be colonized by aquatic biota. Page 281

D. G. de Jalón et al. / Water Policy 19 (2017) 358–375

373

The research agenda should include, as a priority, all the necessary steps to develop new alternative, multi-scale approaches to ecological monitoring and assessment, so that the WFD can be better implemented and possible enhancement proposed for its coherent implementation in time for the next revision of the RBMPs and of the Directive.

Acknowledgements This paper has been produced under the REFORM project (REstoring rivers FOR effective catchment Management), which has received funding from the European Union’s Seventh Programme for research, technological development and demonstration under Grant Agreement No. 282656.

References Acreman, M. C. & Dunbar, M. J. (2004). Defining environmental river flow requirements – a review. Hydrology and Earth System Sciences 8(5), 861–876. Anderson, K. E., Paul, A. J., McCauley, E., Jackson, L. J., Post, J. R. & Nisbet, R. M. (2006). Instream flow needs in streams and rivers: the importance of understanding ecological dynamics. Frontiers in Ecology and the Environment 4, 309–318. Arthington, A. H., Bunn, S. E., Pusey, B. J., Bluhdorn, D. R., King, J. M., Day, J. A., Tharme, R. E. & O’Keeffe, J. H. (1992). Development of an holistic approach for assessing environmental flow requirements of riverine ecosystems. In: Proceedings of an International Seminar and Workshop on Water Allocation for the Environment, November 1991. Pigram, J. J. & Hooper, B. P. (eds). The Centre for Water Policy Research, University of New England, Armidale, Australia. Auel, C. & Boes, R. M. (2011). Sediment bypass tunnel design – review and outlook. In: Dams and Reservoirs under Changing Challenges. Schleiss, A. J. & Boes, R. M. (eds). Taylor & Francis Group, London, pp. 403–412. Bartelt, D., Bundesmann, R. & Sevis, I. (2012). Continuous sediment transfer (ConSed-Method). A way to reduce the risk of sedimentation and siltation, flooding, soil and coast erosion, and siltation of groundwater. In: ACWUA’s 5th Best Practices Conference on Water Resources Management in the Arab Region, Muscat, Oman. Cajot, S., Schleiss, A., Sumi, T. & Kantoush, S. (2012). Reservoir sediment management using replenishment: a numerical study of Nunome Dam. In: Procs. Int. Symp. On Dams for Changing World, Kyoto. Collier, M., Webb, R. H. & Schmidt, J. C. (2000). Dams and Rivers: A Primer on the Downstream Effects of Dams (Vol. 1126). DIANE Publishing, Collingdale. Corenblit, D., Tabacchi, E., Steiger, J. & Gurnell, A. M. (2007). Reciprocal interactions and adjustments between fluvial landforms and vegetation dynamics in river corridors: a review of complementary approaches. Earth Science Reviews 84, 56–86. European Commission (2015). Ecological Flows in the Implementation of the Water Framework Directive. Technical Report. Guidance Document No. 31. p. 106. European Environmental Agency (2010). Use of freshwater resources. CSI 018/WAT 001. Friberg, N., Bonada, N., Bradley, D. C., Dunbar, M. J., Edwards, F. K., Grey, J., Hayes, R. B., Hildrew, A. G., Lamouroux, N., Trimmer, M. & Woodward, G. (2011). Biomonitoring of human impacts in freshwater ecosystems: the good, the bad and the ugly. Advances in Ecological Research 44, 1–68. Friberg, N., O’Hare, M. & Poulsen, A. M. (eds) (2013). Impacts of hydromorphological degradation and disturbed sediment dynamics on ecological status. D3.1 Impacts of HyMo degradation on ecology, EU FP7 project (Restoring rivers for effective catchment Management). Grant Agreement 282656FP7 REFORM. Fryirs, K. A., Arthington, A. & Grove, J. (2008). Principles of river condition assessment. In: River Futures: An Integrative Scientific Approach to River Repair. Brierley, G. & Fryirs, K. A. (eds). Society for Ecological Restoration International, Island Press, Washington, DC, USA, pp. 100–124. Fukuda, T., Yamashita, K., Osada, K. & Fukuoka, S. (2012). Study on flushing mechanism of dam reservoir sedimentation and recovery of riffle-pool in downstream reach by a flushing bypass tunnel. In: International Symposium on DAMS FOR A CHANGING WORLD – Need for Knowledge Transfer across the Generations & the World, Kyoto, Japan. Page 282

374

D. G. de Jalón et al. / Water Policy 19 (2017) 358–375

García de Jalón, D., Alonso, C., González del Tango, M., Martinez, V., Gurnell, A., Lorenz, S., Wolter, C., Rinaldi, M., Belletti, B., Mosselman, E., Hendrik, D. & Geerling, G. (2013). Review on pressure effects on hydromorphological variables and ecologically relevant processes. REFORM Deliverable (D 1.2). http://www.reformrivers.eu/ (accessed 12 October 2015). Gerten, D., Hoff, H., Rockström, J., Jägermeyr, J., Kummu, M. & Pastor, A. V. (2013). Towards a revised planetary boundary for consumptive freshwater use: role of environmental flow requirements. Current Opinion in Environmental Sustainability 5, 551–558. González del Tánago, M., Bejarano, M. D., García de Jalón, D. & Schmidt, J. C. (2015). Biogeomorphic responses to flow regulation and fine sediment supply in Mediterranean streams (the Guadalete River, southern Spain). Journal of Hydrology 528, 751–762. doi:10.1016/j.jhydrol.2015.06.065. Grant, G. E. (2012). The geomorphic response of gravel-bed rivers to dams: perspectives and prospects. In: Gravel-bed Rivers: Processes, Tools, Environments. Church, M., Biron, P. & Roy, A. (eds). John Wiley & Sons, Ltd, pp. 165–179. Grant, G. E., Schmidt, J. C. & Lewis, S. L. (2003). A geological framework for interpreting downstream effects of dams on rivers. In: A Peculiar River: Geology, Geomorphology, and Hydrology of the Deschutes River, Oregon, Water Sci. Appl., vol. 7. O’Connor, J. E. & Grant, G. E. (eds). AGU, Washington, DC, pp. 203–219. Gurnell, A. M., Belletti, B., Bizzi, S., Blamauer, B., Braca, G., Buijse, A. D., Bussettini, M., Camenen, B., Comiti, F., Demarchi, L., García De Jalón, D., González Del Tánago, M., Grabowski, R. C., Gunn, I. D. M., Habersack, H., Hendriks, D., Henshaw, A., Klösch, M., Lastoria, B., Latapie, A., Marcinkowski, P., Martínez-Fernández, V., Mosselman, E., Mountford, J. O., Nardi, L., Okruszko, T., O’Hare, M. T., Palma, M., Percopo, C., Rinaldi, M., Surian, N., Weissteiner, C. & Ziliani, L. (2014). A hierarchical multi-scale framework and indicators of hydromorphological processes and forms. Deliverable 2.1, Part 1, of REFORM (REstoring rivers FOR effective catchment Management), a Collaborative project (large-scale integrating project) funded by the European Commission within the 7th Framework Programme under Grant Agreement 282656. Habersack, H. & Piégay, H. (2008). River restoration in the Alps and their surroundings: past experience and future challenges. In: Gravel-bed Rivers VI – From Process Understanding to River Restoration. Developments in Earth Surface Processes. Habersack, H., Piégay, H. & Rinaldi, M. (eds). Elsevier, Amsterdam, The Netherlands, pp. 703–738. Hupp, C. R. (1999). Relations among riparian vegetation, channel incision processes and forms, and large woody debris. In: Incised River Channels. Darby, S. E. & Simon, A. (eds). Wiley, London, pp. 219–245. International Commission on Large Dams (2007). World Register of Dams. Paris. (www.icold-cigb.net) (accessed 2 September 2015). King, J. & Louw, D. (1998). Instream flow assessments for regulated rivers in South Africa using the Building Block Methodology. Aquatic Ecosystem Health and Management 1, 109–124. King, J., Brown, C. & Sabet, H. (2003). A scenario-based holistic approach to environmental flow assessments for rivers. River Research and Applications 19(5-6), 619–639. Kondolf, G. M., Piégay, H. & Sear, D. (2003). Integrating geomorphological tools in ecological and management studies. In: Tools in Fluvial Geomorphology. Kondolf G. M. & Piégay H. (eds). John Wiley & Sons, Chichester, pp. 633–660. Lloyd, N., Quinn, G., Thoms, M., Arthington, A., Gawne, B., Humphries, P. & Walker, K. (2004). Does Flow Modification Cause Geomorphological and Ecological Response in Rivers? A Literature Review from an Australian Perspective. Technical report 1/2004, CRC for Freshwater Ecology, Canberra. 51 pp. Lobera, G., Besné, P., Vericat, D., López-Tarazón, J. A., Tena, A., Aristi, I., Díez, J. R., Ibisate, A., Larrañaga, A., Elosegi, A. & Batalla, R. J. (2015). Geomorphic status of regulated rivers in the Iberian Peninsula. Science of the Total Environment 508, 101–114. doi:10.1016/j.scitotenv.2014.10.058. Magdaleno, F. & Fernández, J. A. (2011). Hydromorphological alteration of a large Mediterranean river: relative role of high and low flows on the evolution of riparian forests and channel morphology. River Research and Applications 27(3), 374–387. Meitzen, K. M., Doyle, M. W., Thoms, M. C. & Burns, C. E. (2013). Geomorphology in the interdisciplinary context of environmental flows. Geomorphology 200, 143–154. Ock, G., Sumi, T. & Takemon, Y. (2013). Sediment replenishment to downstream reaches below dams: implementation perspectives. Hydrological Research Letters 7, 54–59. doi:10.3178/hrl.7.54. Okano, M., Kikui, M., Ishida, H. & Sumi, T. (2004). Reservoir sedimentation management by coarse sediment replenishment below dams. In: Proceedings of the Ninth International Symposium on River Sedimentation, Yichang, China. Petts, G. E. (1984). Impounded Rivers. John Wiley & Sons, Ltd, Chichester. Page 283

D. G. de Jalón et al. / Water Policy 19 (2017) 358–375

375

Rinaldi, M., Belletti, B., Bizzi, S., Blamauer, B., Brabec, K., Braca, G., Bussettini, M., Comiti, F., Demarchi, L., García de Jalón, D., Giełczewski, M., Golfieri, B., González del Tánago, M., Grabowski, R., Gurnell, A. M., Habersack, H., Hellsten, S., Kaufman, S., Klösch, M., Lastoria, B., Mao, L., Marchese, E., Marcinkowski, P., Martínez-Fernández, V., Mosselman, E., Muhar, S., Nardi, L., Okruszko, T., Paillex, A., Percopo, C., Poppe, M., Rääpysjärvi, J., Schirmer, M., Stelmaszczyk, M., Surian, N., Van de Bund, W., Vezza, P. & Weissteiner, C. (2015). Final report on methods, models, tools to assess the hydromorphology of rivers. Deliverable 6.2, a report in five parts of REFORM (REstoring rivers FOR effective catchment Management), a Collaborative project (large-scale integrating project) funded by the European Commission within the 7th Framework Programme under Grant Agreement 282656. Sánchez-Navarro, R. & Schmidt, G. (2012). Environmental flows as a tool to achieve the WFD objectives. Study for the European Commission. 43 pp. Schmidt, J. C. & Wilcock, P. R. (2008). Metrics for assessing the downstream effects of dams. Water Resources Research 44, 4. Stella, J. C., Rodríguez-González, P. M., Dufour, S. & Bendix, J. (2013). Riparian vegetation research in Mediterranean-climate regions: common patterns, ecological processes, and considerations for management. Hydrobiologia 719(1), 291–315. Vörösmarty, C. J., Fekete, B. & Sharma, K. (1997). The potential impact of neo-Castorization on sediment transport by the global network of rivers. In: Proceedings Human Impact on Erosion and Sedimentation. AHS Publ. 245, 261–273. Vörösmarty, C. J., Meybeck, M., Fekete, B., Sharma, K., Green, P. & Syvitski, J. P. (2003). Anthropogenic sediment retention: major global impact from registered river impoundments. Global and Planetary Change 39, 169–190. Ward, J. V. & Stanford, J. A. (1979). The Ecology of Regulated Streams. Plenum Press, New York. White, R. (2001). Evacuation of Sediments from Reservoirs. Thomas Telford Publishing, London. Williams, G. P. & Wolman, M. G. (1984). Downstream Effects of Dams on Alluvial Rivers. United States Geological Survey, Professional Paper 1286. Wohl, E., Angermeier, P. L., Bledsoe, B., Kondolf, G. M., McDonnell, L., Merritt, D. M., Palmer, M. A., Poff, M. L. & Tarboton, D. (2005). River restoration. Water Resources Research 41, W10301, doi: 10.1029/2005WR003985. Received 6 January 2016; accepted in revised form 15 August 2016. Available online 5 December 2016

Page 284

Water Policy 19 (2017) 1063–1080

Canadian and Australian researchers’ perspectives on promising practices for implementing Indigenous and Western knowledge systems in water research and management R. D. Stefanellia,*, H. Castledenb, A. Cunsoloc, D. Martind, S. L. Harpere and C. Harta a

Department of Geography and Planning, Queen’s University, Mackintosh Corry Hall, D201, 68 University Ave, Kingston, ON, Canada K7L3N6 *Corresponding author. E-mail: robert.stefanelli@queensu.ca b Departments of Geography and Planning, and Public Health Sciences, Queen’s University, Mackintosh Corry Hall, D201, 68 University Ave, Kingston, ON, Canada K7L3N6 c Labrador Institute of Memorial University, Room 110, College of the North Atlantic Building. P.O. Box 490, Station B., Happy Valley-Goose Bay, NL, Canada A0P 1E0 d School of Health and Human Performance Dalhousie University Stairs House, P.O. Box 15000, 6230 South Street, Halifax, NS, Canada B3H 4R2 e University of Guelph, Stewart Building, 2524, 50 Stone Road E., Guelph, ON, Canada N1G 2W1

Abstract National and international policies have called for the inclusion of Indigenous peoples and the uptake of Indigenous knowledge alongside Western knowledge in natural resource management. Such policy decisions have led to a recent proliferation of research projects seeking to apply both Indigenous and Western knowledge in water research and management. While these policies require people with knowledge from both Western and Indigenous perspectives to collaborate and share knowledge, how best to create and foster these partnerships is less understood. To elicit this understanding, 17 semi-structured interviews were completed with academic researchers from Canada and Australia who conduct integrative water research. Participants, most of whom were non-Indigenous, were asked to expand on their experiences in conducting integrative water research projects, and findings were thematically analyzed. Our findings suggest that Indigenous and Western knowledge systems influence how one relates to water, and that partnerships require a recognition and acceptance of these differences. We learned that community-based participatory research approaches, and the associated tenets of fostering mutual trust and community ownership for such an approach, are integral to the meaningful engagement that is essential for developing collaborative partnerships to implement both Indigenous and Western knowledge systems and better care for water.

Page 285

1064

R. D. Stefanelli et al. / Water Policy 19 (2017) 1063–1080

Keywords: Academic researchers; Australia; Canada; Indigenous knowledge; Indigenous people; Integrative water research; Western knowledge

Introduction Water is one of the most abundant natural resources, yet access to safe and sufficient water sources for all, and ensuring this same level of security for future generations, is one of the most pressing challenges humanity is facing today. Per the United Nations (2016), on a global scale, 1.8 billion people lack access to contaminant-free drinking water, 1.7 billion people live in areas where water demand exceeds source supply, and nearly 2.5 billion people are without access to wastewater and sanitation services. This water crisis is exacerbated by source water pollution (McDonald et al., 2016), inadequate wastewater and sanitation infrastructure (Daley et al., 2015), ice melt (Hansen et al., 2016), flooding (Gersonius et al., 2013), drought (Cook et al., 2015), aquatic resource depletion (Eero et al., 2012), over allocation (Duncan, 2014), climate change (Vörösmarty et al., 2000), unequal economic distribution (Butler et al., 2016), and mismanagement of our water resources (Barlow, 2015). These issues represent a few of the current challenges contributing to the global water crisis. Nowhere is this global water crisis more apparent than in Indigenous contexts (White et al., 2012). While Western science has led to many innovations in the treatment and management of our water resources, the exclusive use of Western science and methods has not adequately addressed the higher frequency and persistence of waterrelated challenges in Indigenous communities as compared to non-Indigenous communities (White et al., 2012; Jackson et al., 2014). In examining data trends from 1990 to 2014, both Canada and Australia scored in the top ten globally when analyzing levels of human development, per the United Nations Human Development Index (HDI) (United Nations Development Programme, 2015). Both countries also have large Indigenous populations, and these populations have consistently reported substantially lower health and human development scores (Cooke et al., 2007). In their report on HDI scores for 1990–2000, both Indigenous and Settler (non-Indigenous peoples) Canadians showed improved scores (although a gap still existed between the two), while Indigenous Australians showed a decrease in HDI scores – further widening the disparity between Settler Australians, whose scores steadily rose during that time (Cooke et al., 2007). In 2006, the HDI disparity between Settler and Indigenous populations in both Canada and Australia was only marginally better than scores reported in 1981 (Mitrou et al., 2014). In Canada and Australia, the higher frequency and persistence of water-related challenges in Indigenous communities is strongly apparent ( Jackson et al., 2014; Health Canada, 2016). Drinking water advisories1 in Indigenous communities across Canada are rampant (Health Canada, 2016; First Nations Health Authority, 2017); they burden Indigenous communities in Canada at a significantly higher rate than Settler communities (White et al., 2012). Whereas in Australia, resource depletion and diversion away from Indigenous territories for agriculture represent two of the most pressing water-related challenges that Indigenous communities experience (Jackson et al., 2014). 1

Drinking water advisories are classified into one of three tiers, based on real or perceived risk: (1) Boil Water Advisory – Most common advisory issued when water is known or suspected to contain disease-causing bacteria. (2) Do Not Consume – Issued when water supply is known to be contaminated and should not be ingested. (3) Do Not Use – Issued when contact with the water may pose mild to severe health risks (First Nations Health Authority, 2017). Page 286

R. D. Stefanelli et al. / Water Policy 19 (2017) 1063–1080

1065

Recognizing the value of place-based Indigenous knowledge, and the inadequacies of Western science to assist in reducing this disparity between Indigenous and non-Indigenous communities in both countries, researchers and managers (who remain largely members of Settler society) have begun to seek out ways to implement Indigenous and Western knowledge systems to research and manage our shared waters. In addition to the benefits of using multiple perspectives, Indigenous peoples in Canada and Australia have rights to autonomy that support the implementation of both Indigenous and Western knowledge systems. As both Canada and Australia have agreed to the (non-binding) terms of the United Nations Declaration on the Rights of Indigenous Peoples (although they, along with New Zealand and the United States, were the only four member states to originally oppose the Declaration), there is an expectation, albeit non-enforceable, that researchers respect Indigenous selfdetermination and Indigenous knowledge when conducting research with Indigenous peoples2. As such, implementation of both knowledge systems represents a promising shift towards viable solutions for addressing the water crisis, and this is reflected in the growing number of publications resulting from integrative water research projects in Canada and Australia (Stefanelli et al., 2017). Attempts at, and challenges to, the implementation of these complementary systems of knowledge in water research and management is an emerging area of study for researchers in Canada (e.g., Castleden et al., 2017) and Australia (e.g., Barber & Jackson, 2011; Finn & Jackson, 2011). Although both Australian and Canadian researchers have demonstrated successes in integrative3 Indigenous and Western knowledge mobilization in various water-related fields (e.g., Ayre & Mackenzie, 2013; von der Porten & de Loë, 2013a), uncertainty remains about how best to implement the expertise of Indigenous peoples with the expertise of Western-based water researchers and managers, in practice. Given this uncertainty, we wanted to explore this approach in more detail by conducting interviews with authors of studies in this field. What follows is a description of the methods we used to recruit water researchers from Australia and Canada to qualitatively explore their experiences in attempting to implement Indigenous and Western knowledge systems into their water research and management practices. This study is a corollary of a larger program of research, funded by the Canadian Water Network, which sought to determine the most promising methods and models for engaging in integrative water research and water management with First Nations, Inuit, and Métis4 peoples in Canada (Castleden et al., 2015). A National Advisory Committee5 of Indigenous and non-Indigenous knowledge-holders guided that program design and its implementation. Drawing from the original program of research, which uses Canadian data involving First Nations, Inuit, and Métis, this study is a natural progression to internationalize that work.

To read the UNDRIP in its entirety, please see http://www.un.org/esa/socdev/unpfii/documents/DRIPS_en.pdf. By ‘integrative’, we draw upon the work of Bartlett et al. (2015), which describes ‘integrative’ as ‘bringing together of the scientific knowledges and ways of knowing from Indigenous and Western worldviews’ (p. 3), while avoiding the past-tense ‘integrated’, to signify that sharing knowledge is an ongoing process, not an endpoint. 4 First Nations, Inuit, and Métis are the three distinct Indigenous populations recognized in the Canadian Constitution as ‘Aboriginal’. We have chosen to use ‘Indigenous’ to represent these populations in accordance with the United Nations Declaration on the Rights of Indigenous People (United Nations General Assembly, 2007). 5 The establishment of a National Advisory Committee was integral to our research project as it allowed for Indigenous and Western knowledge-holders to collaboratively design the research project and provide recommendations at all stages of the process. This committee afforded us the opportunity to conduct our research project in a manner that was consistent with the topic of our study – to implement Indigenous and Western knowledge systems in a meaningful and respectful manner. 3

Page 287

1066

R. D. Stefanelli et al. / Water Policy 19 (2017) 1063–1080

Methods Interview protocol To move beyond the data provided through a systematic-realist review in Part One of this study (Stefanelli et al., 2017), the research team conducted interviews with the lead authors of Canadian and Australian research projects that were selected as ‘exemplars’ in the field of integrative water research to unveil any additional insights that were not apparent within the published literature. Exemplars were determined based on author frequency within the included records of our systematic-realist literature review as well as novel theoretical, methodological, and/or substantive findings within the article. We had intended to examine literature and to contact first authors from four English-speaking countries with a shared, though different, history of British colonialism: Canada, Australia, New Zealand, and the United States. However, upon review of the integrative Indigenous and Western knowledge literature related to water, we found that Canada (45 records) and Australia (26 records) produced a substantially larger body of included literature than New Zealand (14 records) and the United States (12 records) (Stefanelli et al., 2017). Interviews were semi-structured, and followed a series of questions from an interview guide that was originally developed with the National Advisory Committee, and then adapted Canada-specific language to reflect an Australian research context. Additionally, a section of questions regarding the Australian National Water Initiative6 were added due to recurring references made to this Initiative in the Australian academic literature. The questions flowed from four broad areas: (1) general experiences in conducting integrative research; (2) detailed accounts of specific integrative research projects; (3) researcher definitions and/or descriptions of the terminology used within this field (such as Indigenous/Western knowledge and methodologies); and (4) prospects for success in implementing both Indigenous and Western knowledge systems into the realm of water research and management. Participant recruitment There were 24 exemplars from Canada (15) and Australia (9) (see Table 1), selected based on their contributions to the literature in the field of integrative water research, and each of the first authors were contacted to participate in semi-structured interviews7. Of those, 17 individuals consented to participate in a 60 minute, semi-structured interview (12 Canadian-based researchers and five Australian-based researchers), completed over a period of 12 months (February 2015–2016). Only two of the 17 interviewees self-identified as Indigenous, while the remainder identified as Settlers8,9. 6

The National Water Initiative, which has since been abolished (2015), emerged in 2004 to coordinate water resource extraction between the states of Australia. Of importance to our project is a subsection of that initiative that required states to work with Indigenous people and governments to address those community-level concerns. For further reading, see http://www.mdba.gov.au/kid/files/2022-NWI2011-BiennialAssessment-full_report.pdf. 7 We acknowledge that the convention of authorship varies between disciplines. We opted for contacting the first author because, despite their rank on the research team, they are often the ones that contributed most in the research project. 8 We return to the importance of this limitation later in the paper with respect to the systemic whiteness of the academy, the epistemic dominance of Western science in universities, and how it relates to the broader issue of exclusion of Indigenous peoples and their knowledge systems in academic research and publication. 9 We did not ask participants to identify along the gender spectrum, nor did we undertake a sex- or gender-based analysis of our data. Page 288

1067

R. D. Stefanelli et al. / Water Policy 19 (2017) 1063–1080 Table 1. Exemplar identiﬁcation and recruitment. Canada Included records Exemplars identiﬁed Interviews conducted

Australia a

45 (42 Unique First Authors (UFA) ) 15 12 (29% of UFAs)

26 (16 UFAa) 9 5 (31% of UFAs)

Although there were 45 and 26 records included from Canada and Australia, respectively, several authors had numerous articles included in the systematic-realist review (Stefanelli et al., 2017).

Canadian recruitment. Recruitment emails were sent to 15 researchers, of whom 12 researchers consented and participated in our study, while two did not respond, and one declined to participate. Of the 12 interviewees, 11 stated a history of working with First Nations populations, three of 12 researchers described their work with Inuit communities, and one of 12 interviewees identified a previous research partnership with a Métis community. Participants reported that their integrative projects had been conducted in the Canadian North (i.e., Yukon, Northwest Territories, and/or Nunavut), British Columbia, Saskatchewan, Ontario, and Newfoundland/Labrador.

Australian recruitment. Like Canada, Australian researchers were identified through the completion of the systematic realist review based on their experiences in conducting integrative research. Nine researchers were identified for their integrative work and were recruited via email. Of these, six researchers agreed to participate, although only five researchers were interviewed (the sixth did not respond to multiple attempts to set an interview date). Of the five researchers interviewed, all had worked with Australian Aboriginal communities, while only one discussed working with Torres Strait Islander communities. Geographically, participants discussed their experiences of working with communities in Western Australia, Northern Territory, South Australia, Victoria, New South Wales, and Queensland.

Data analysis Interviews were audio recorded (with permission) and then transcribed verbatim, coded, and thematically analyzed (Dunn, 2016). We began our analysis using the process of open coding whereby transcripts were read in full and codes were derived from the data. Identified latent and manifest codes were noted and defined in a codebook, which was then used in the second stage of the analysis process. In this stage, each transcript was again read in full, and sections of text were highlighted per the appropriate code. It should be noted that codes were not mutually exclusive. The final stage of the analysis process was thematic separation in which all participant quotations from each selected code were placed in separate documents according to code. These documents allowed for overarching themes and sub-themes to be identified across, between, and within each country. Interview participants were given an opportunity to review the transcription of their interview to ensure clarity and appropriate use of data, and they were given an opportunity to review the use of their quotations in the context of the findings to validate the conclusions. Page 289

1068

R. D. Stefanelli et al. / Water Policy 19 (2017) 1063–1080

Results As noted in the Methods section, we remind readers that our findings represent the perspectives of mostly Settler-researcher and their experiences in conducting integrative water research. Indigenous perspectives may have revealed additional and/or different themes; however, from our data, three broad themes emerged, and within these, seven sub-themes were identified. Not surprisingly, there was substantial overlap between them. For the purposes of presenting coherent findings, they have been disarticulated from each other below. The first broad theme is relating to water, which comprised two sub-themes: (1) coming to ‘know’ our relationship to water and (2) viewing water as a right or a relationship. The second broad theme, power, included two sub-themes: (3) power dynamics in the socio-political context and (4) power dynamics in the researcher–community relationship. The final theme was integrative knowledge implementation, which included three sub-themes: (5) support (or lack thereof); (6) implementation without an equal benefit; and (7) participants’ ‘lessons learned’. Interviewees are identified by number and home country with ‘A’ for Australian researchers and ‘C’ for Canadian researchers. Relating to water The theme of ‘relating to water’ encompassed the many ways in which interview participants discussed how they understood, and/or how they perceived their community partners to have understood, their personal and/or professional relationship to water. Coming to ‘know’ our relationship to water. Participants were asked to describe Indigenous and Western knowledge systems as a way of beginning our interviews about their research. In doing so, we developed a baseline understanding of the ways in which Indigenous and Western epistemologies shaped their relationship(s) with water. Interviewees generally shared their understandings of the differences between Indigenous and Western knowledge systems as being experiential and objectivityfocused, respectively. For example, one participant stated their interpretation of Indigenous knowledge as: ‘Knowledge that encompasses ways of thinking but also ways of being and doing. It is knowledge that is enacted in practice and passed on in informal contexts as well as formal ones. I think it is knowledge that … is less commonly systematically described, at least to outsiders.’ (A3) Whereas another participant described Western knowledge as: ‘Very structured in … that there’s very set steps of how that knowledge is produced, who produces knowledge that’s considered valued, and … it’s very rooted in kind of a colonial history of how we consider what is ‘valued knowledge’. Western science is often produced through sort of measurement means, replication of specific results.’ (C9) Participants often chose to discuss the incongruences between Indigenous and Western knowledge to differentiate between the two systems: Page 290

R. D. Stefanelli et al. / Water Policy 19 (2017) 1063–1080

1069

‘Non-Indigenous researchers typically use the scientific methodology and remove the emotion, perspective and opinion from the data; and that is how you understand how a natural system operates. Indigenous people in my experience would think that is only a partial understanding of how a natural system operates, and that you have to understand the cultural contexts.’ (A1) While participants largely conceded that it may be useful to define or describe systems of knowledge through a comparison to other systems, they also thought it was important for researchers to spend more time understanding their own system of knowledge, to know exactly what their knowledge system was being compared to: ‘There is also a degree to which [Indigenous knowledge] is often contrasted with scientific or Western knowledge … Indigenous knowledge is ‘this’ because it’s not ‘that’, or it’s different from ‘that’. But sometimes what it is that it’s being compared to isn’t particularly well described or understood in its own sense.’ (A3) Although the knowledge systems differ, they are not incommensurable. Participants cautioned against viewing Indigenous and Western knowledge as two distinct entities that could be picked apart and selectively added to one another as doing so would support the false dichotomy that exists between the two knowledges. As one participant stated, ‘When we set up binaries, one is given the status and the authority and the power, and the other gets marginalized’ (C8). Participants reiterated numerous times that both systems of knowledge are different, but not entirely opposite or conflicting, and that these systems have developed, and should continue to develop together. At the same time, as noted above, participants themselves used binaries to distinguish between them (e.g., referring to Indigenous knowledge as experiential and Western knowledge as objective). Our relationship to water: rights and responsibilities. Here is one point of divergence between Australian and Canadian participants, as the former referred more often to rights, while the latter discussed responsibilities to water. Participants noted that Indigenous and Western knowledge systems differed in how human relationships to water were viewed – as a right to water, or a responsibility to care for water. One Canadian participant explicitly referred to this being an obstacle that researchers must consider when attempting integrative water research: ‘[It is important] to note that many [Indigenous nations] don’t just see their rights to their land and water as a right, but also as a responsibility. So it’s not just, ‘This is our land, and we have the right to it,’ but, ‘We have the responsibility that’s been passed down from Elders, or that we have been taught that it’s our job to take care of these lands and waters.’’ (C1) Other Canadian participants referred to cultural protocols that exist in many Indigenous communities: ‘It’s personal in that it was driven by my personal responsibility, and I had the Elders tell me that maybe it was my responsibility as an [Indigenous person that] was driving that, that I feel the need that something needs to be done about this … the crisis, the water crisis.’ (C3) Page 291

1070

R. D. Stefanelli et al. / Water Policy 19 (2017) 1063–1080

In Australia, circumstances are similar to Canada in that rights to water are an important, though partial, aspect of water for communities that researchers may consider in their work10. However, as one participant noted, the existence of Title rights did not necessarily clarify the ambiguity in how those rights were to be asserted, particularly in the case of water rights: ‘Australia has a Native Title Act essentially that protects Native Title rights on country or areas where people have a right to access. The Act also protects the right to harvest resources for traditional purposes in those areas. Because water extraction upstream for agricultural purposes can erode that right, Indigenous rights to water is actually less clear in the Australia context than the Native Title right to hunt and fish.’ (A1) Ultimately, participants overall indicated that both rights and responsibilities to water must be considered when conducting integrative knowledge implementation research. Power dynamics The second major theme that emerged from the analysis was the importance of understanding and reconciling the unequal power dynamics that had arisen through colonialism and widespread systemic racism, and that continues to disadvantage Indigenous peoples in Canada and Australia in water-related contexts. Power dynamics in the socio-political context. In talking about their general experiences and respective projects, participants often referred to the power held by institutions and/or state governing bodies in relation to the management and governance of water resources. In Australia, much of this discussion focused on the implementation of the 2004 National Water Initiative, which attempted to reform past practices that had led to over-allocations of water resources, ‘… typically for agricultural purposes’ (A3). Participants highlighted fragmentation in governance as a primary cause of the mismanagement of water: ‘Part of this is the enduring problem in Australia of federalism, where each of the states have jurisdiction to some extent over water. Partly what the National Water Initiative did was to try to wrestle some of that policy control back from the states, but those water plans continue to be instituted under state and territory legislation. So that disjunction in terms of jurisdictional control meant that [implementation] was going to be at the discretion of the states and the territories.’ (A4) To meaningfully engage with Indigenous people and Indigenous knowledge systems about water, participants stressed the need to recognize that current colonial government decision-making structures value Western scientific knowledge over all other knowledge systems. As one participant described: ‘Federally and provincially they’re pretty biased in terms of a preference towards [Positivist] data … in their decision-making. And those decisions they’re making based on that one kind of evidence and 10

The Indigenous rights contexts of the two countries are vastly different and beyond the scope of this article.

Page 292

R. D. Stefanelli et al. / Water Policy 19 (2017) 1063–1080

1071

that one kind of truth impacts people that don’t necessarily identify the same way with that truth. They’ve got policies in this country founded and based and supported and justified from one knowledge system that are influencing and very directly impacting the lives of the First Peoples of this country.’ (C10) In addition to recognizing the current dominance of Western knowledge systems in policy-making, participants indicated that governing bodies and researchers alike must also better understand the power dynamic that exists, arising from our colonial context as well as the ‘everyday’ reality of socio-political power relations. These dynamics create barriers to facilitating meaningful engagement and collaboration. A Canadian researcher noted: ‘When you have something like a collaborative process that is premised on genuine speech and dialogue and consensus decision-making and you’re assuming that everybody’s equal, you’re immediately out to lunch. Because if you’ve got some little collaborative watershed process and it is, ‘Hi, I’m Joe, I’m a retired school teacher,’ and ‘I’m Mary. I’m a housewife who’s interested in the environment,’ and ‘Yeah, I’m Steve and I represent a $400 billion mining corporation, but we’re all equal here!’ No, we’re not. And so power is a critical factor that I think we need to get much better at understanding when it comes to what facilitates effective governance.’ (C6) As this example illustrates, effective implementation of knowledge systems is much more than an inclusion of Indigenous peoples or having provided a seat at the table to discuss – from only a Western epistemology – the best practices in managing natural resources. Participants noted that status quo water research and management, with the implementation of Indigenous peoples’ knowledge within a Western framework was insufficient. Doing so would not address the socio-political power relations, much of which involves embedded colonialism in this context. Power dynamics in the researcher–community relationship. Within the broader milieu of socio-political power relations, researcher–community partner power dynamics also surfaced from the data, including questions regarding who decides when and where collaboration is to take place, or who decides the details of the research process? Participants emphatically stressed the importance of relationship building as an effective strategy to mitigate these power imbalances: ‘It is important to form relationships first and maintain them. Because without those strong relationships I don’t think what we did could have worked. We spent up to four months a year living in Indigenous communities. But without that, this sort of research sort of skims off the top of the proper understanding.’ (A1) While relationship building is a key process, another participant discussed the difficulties related to grant procurement when attempting to build a relationship and design a research project together with a community partner: ‘I came in all ready with all my natural science funding to do my natural science research, and then the Traditional Knowledge component came after, which is probably more typical, because how often, especially in northern communities, are you in a situation where you just pull some money Page 293

1072

R. D. Stefanelli et al. / Water Policy 19 (2017) 1063–1080

from somewhere and head up and try to engage with a community from the beginning and have it all in sync from the get-go? That’s what we wanted to occur, but there often isn’t the mechanism in place to allow that to happen.’ (C7) Even when funding constraints were not an issue, a Canadian participant acknowledged the difficulties that may occur in designing a research project together with a community, as Western-style research methods may not be consistent with the cultural practices of the community partners: ‘One way of looking at it is just even the assumption that we would look or sit down to talk about something. In some cultures, decisions are made out on the land, and people talk as they are doing whatever they’re doing. Just that process of who makes decisions and when and how it’s discussed differs. So sitting down in a boardroom immediately biases the conversations to a Western way.’ (C1) While it is important to establish meaningful relationships with community partners and to design research projects that have a shared benefit for both parties, there are structural obstacles that hamper these relationship-building efforts. As one participant stated, ‘things don’t matter ‘til it matters to the ‘money people’’’ (C6), and this economics-driven way of thinking privileges the academic researcher over the community participants – as it is the academic that often receives the research funding. Participants indicated that the effective implementation of Indigenous knowledge systems in water research and management was unlikely to occur without wholesale changes to the Western-based broader academic framework (and that subsuming one into the other is not an appropriate goal). However, within current frameworks, they indicated that alterations to timing and funding structures represented the most likely areas for change to allow researchers the freedom to co-design projects with community partners. Indigenous and Western knowledge implementation The final key overarching theme that emerged from the data related to researchers’ attempts to implement both knowledge systems in water research and management, and are discussed below. Support (or lack thereof) for Indigenous and Western knowledge implementation. Many participants spoke about various forms of resistance (including resistance from government officials, industry professionals, and other researchers) that they had experienced during their research careers. In some cases, they stressed that potential partnerships between conservation managers and Indigenous communities had collapsed when both parties had felt that the differences that existed between them regarding resource management initiatives were too great to find a middle ground to build from. As stated by one participant, ‘some conservationists won’t have anything to do with [Indigenous] resource management because the two groups have very different goals in their management of resources.’ (A2). Systematic discrimination and institutional racism were issues raised that could hinder implementation of Indigenous and Western knowledge; however, participants also stressed that knowledge implementation may be inappropriate in all areas of the research study – at least while the current Western research paradigm holds sway. Most researchers were hesitant to promote the implementation of Page 294

R. D. Stefanelli et al. / Water Policy 19 (2017) 1063–1080

1073

Indigenous and Western knowledge in instances where they recognized there was selective extraction and insertion of Indigenous knowledge to ‘tick a box’. As one participant noted: ‘I think that’s part of the risk, right, is that one gets subsumed into the other, and I think that is what happens … when we set up binaries, like this one is given the status and the authority and the power, and the other gets marginalized.’ (C8) When this type of selective extraction occurred, participants cautioned that it threatened to devalue the entire Indigenous knowledge system in favor of a Western knowledge system: ‘I find a lot of the so-called efforts to integrate Western science and traditional knowledge are very much of that flavour. When you look at what integration means, it’s like the really important questions get answered by the biologists, and then there’s this thin, sort of politically correct layer of Indigenous knowledge that gets put on top of it. And I think that’s a bullshit kind of approach.’ (C6) As this participant alluded to, tokenistic engagement is an obstacle in the design and implementation of water research and management. Implementation without equal benefits. Integrative Indigenous and Western knowledge research requires the development of honest and respectful relationships built on the premise of mutually beneficial partnerships. However, participants cited examples, either from their own research or from the work of others, where these relationships did not develop and attempts to implement Indigenous knowledge in water research and management were not conducted in an appropriate manner. They referred to such attempts as lacking meaningful engagement with Indigenous communities, and they noted that an asymmetrical distribution of benefits had led to a lack of trust in the partnership. As one participant discussed in a general sense: ‘The context in which Indigenous research takes place is often one in which there’s very little trust going in. And there’s been very little gain in the past as far as research not serving Indigenous peoples but rather objectifying, pathologizing, downright ripping off and using what’s been there, as opposed to serving the communities in a decolonizing capacity.’ (C2) In discussions surrounding their own work, one participant cited the restrictions that occur in academia that prevent researchers and community members from working towards the same goal: ‘I find that the ways in which we’re encouraged to write about these things are often not really serving the communities that we are working with and that we’re hopefully intending to serve. So we’re ending up serving our own disciplinary requirements and career paths rather than producing work that is truly beneficial for communities, and that’s a larger structural problem in terms of the academy and academics working within Indigenous communities, is that our work is often measured … in terms of scientific and academic processes that do little to assist community.’ (C12) Another cause for the unequal distribution of research benefits between researchers or managers and communities that was purported by several Canadian participants was a form of tokenistic engagement Page 295

1074

R. D. Stefanelli et al. / Water Policy 19 (2017) 1063–1080

that occurs when Indigenous communities were treated as one of many stakeholder groups with a vested interest in the topic at hand to be included in the decision-making process. Indigenous communities are not merely stakeholders, as one Canadian participant stated: ‘[Indigenous peoples] are absolutely not a ‘minority’ of Canada, and I think it’s common just to lump them in with other peoples. Some people really think that that’s a really progressive step forward – to include [Indigenous peoples] in this conversation. But that’s nowhere near the level of understanding or respect that’s needed to create a relationship where something could be done in terms of real action.’ (C1) While Australian participants did not discuss the implications of viewing Indigenous peoples as one of many stakeholders, they noted that a common misunderstanding existed in the water research and management communities: that Indigenous peoples’ values always aligned with those of conservation organizations. Thus, there had been an assumption that consultation with Indigenous peoples was unnecessary: ‘We assumed for a long time that if we get conservation objectives right, then Indigenous people would just live off the natural system that is protected anyway. More and more we are starting to understand that the objectives of conservation and other groups can be quite different than those of Indigenous people, so we need to consider those differences explicitly.’ (A1) Even still, once these differences had been considered and understood, there existed the possibility that engagement may not be meaningful, and Indigenous knowledge implementation may not occur: ‘We had identified those Indigenous priorities, but whether they were going to be given equal weight, finally, whether we brought the [government] department at all to the point of thinking that it was actually important that the plan reflected the communities’ wishes, is somewhat doubtful.’ (A2) Participants from both countries indicated the importance of designing research projects that were mutually beneficial for the research team and the community partners as integral for the creation of meaningful partnerships that supported the authentic implementation of both Indigenous and Western knowledge systems. Participants’ ‘lessons learned’. Participants were asked to discuss some ‘successes’ in integrative water research and management, be that at a national level, or at an individual level through their work with Indigenous community partners. Australian participants pointed to the enactment of the National Water Initiative as an important starting point in the effective implementation of Indigenous knowledge. Such policy initiatives were highlighted as a potential pathway to providing support for researchers that were attempting to implement Indigenous and Western knowledge systems as they relate to water: ‘I think the strength [of the National Water Initiative] is, like I’ve said, that it specifically mentions the need to include Indigenous perspectives when you manage water. So it is explicitly stated at the Page 296

R. D. Stefanelli et al. / Water Policy 19 (2017) 1063–1080

1075

national level that Indigenous people are important users, and there is a better understanding that the Native Title right can be eroded if water isn’t carefully managed.’ (A1) Canadian participants noted that a transition is occurring more at an individual level whereby water researchers and managers are recognizing that using only a Western perspective to view water ignores the potential contributions from people that have lived in relationship with, and developed an understanding of, the lands and waters around them since time immemorial: ‘If we’re talking in Western science that nature is something that can be managed and controlled and predicted, then we’re really not listening very well to what it is that people who have spent a great deal of time on the land and who relate to it in different ways and ways that are informed by Indigenous knowledge and action.’ (C12) Overall, participants stressed that engaging with Indigenous knowledge-holders requires a commitment of relational authenticity; it is a complex task not to be taken lightly. It requires a significant time and resource investment from both the community and the researcher(s). An Australian participant stated: ‘I found it difficult because I wasn’t living in community. I was … travelling probably on average once every two or three weeks to the Islands. So that is not an ideal situation … it is difficult to develop those trusting relationships and understand the nuances of the context if you’re visiting.’ (A4) Despite some of the challenges of integrative Indigenous and Western knowledge implementation and the resource constraints associated with community-based participatory research (CBPR)11 methods, many of the participants’ ‘lessons learned’ aligned with the tenets of this approach. While CBPR was derived from a Western frame, participants offered practical guidance for other water researchers who want to use this approach as one that can involve either or both Indigenous and Western research approaches, strategies, and methods. For instance, when discussing their CBPR project, an Australian participant stated that an Indigenous ontology was used to better understand river flows, while remaining consistent with community values: ‘Even the concept of calendar months are fine for a scientific understanding, but when you try to tie a calendar month into a river flow understanding, it is quite meaningless because of the variability each year. It doesn’t matter what month you are in, it is when a river hits a particular flow rate that the fish start biting and you start catching them.’ (A5) 11

CBPR, as we have used here, refers to the co-design, and co-completion of research projects where community partners and academic researchers respect each other’s strengths. This research style transcends mere participation in projects, and requires collaboration and active participation from both parties through the entirety of the research process. Both Canada (Canadian Institutes of Health Research, Natural Sciences and Engineering Research Council of Canada, and Social Sciences and Humanities Research Council of Canada, 2014) and Australia (Australian Institute of Aboriginal and Torres Strait Islander Studies, 2012) have federal research policies that support this type of research. For further reading on CBPR, see Castleden et al. (2017). Page 297

1076

R. D. Stefanelli et al. / Water Policy 19 (2017) 1063–1080

Such examples speak to the careful considerations that these participants urge other researchers to consider when attempting integrative water research and management, and many of the considerations emerged from meaningful discussions with community partners about aspects of the water system that were highly prioritized. Participants emphatically encouraged Settler researchers to involve Indigenous peoples throughout the entirety of the research process to ensure the research had meaning for both parties. Discussion The findings from this research suggest that integrative approaches to Indigenous and Western knowledge implementation in the field of water research and management are influenced by our personal and professional relationships with water, the structures of power in State–Indigenous and academic–community relationships, and the need to overcome the challenges preventing implementation from taking place. The concept of place (though not stated explicitly was implied in many of the participants’ interviews), including both its physical and socio-cultural components, is integral to understanding the development of Indigenous knowledge systems. This physical knowledge of place is hundreds of generations in the making, while the socio-cultural components are continually strengthened through existence and relationality in the space of a place. The importance of place as a concept is a point of connection between Indigenous and Western knowledge systems, rather than a point of divergence. From colonization, onward, both Indigenous and Western knowledge systems have continued to develop simultaneously with each other through (albeit violent) co-existence in place – a co-existence predicated on Indigenous exploitation and Settler domination (Regan, 2006; Wolfe, 2006). Despite attempts at assimilation, this co-existence in place has led to the expansion of different, yet not incommensurable systems of knowledge, including knowledge systems as they relate to the same water sources (Woodward, 2008). To implement systems of knowledge that include both Indigenous and Western strengths in water research and management, we found that academic researchers must respect Indigenous research protocols, spend time with Indigenous peoples on the land, in their traditional territories, and develop relationships and understandings of the importance of the territories to the communities whose cultural, spiritual, mental, and physical health and identity depend on access to the lands and resources (Richmond & Ross, 2009). We also found that rights to water, as opposed to responsibilities to water, is a point of divergence not just with respect to the experiences of Canadian and Australian researchers, but also across Indigenous and Western knowledge systems. To view water from a rights perspective is to view water through a Western, legal framework – which is not always consistent with the way Indigenous communities relate to their water resources (White et al., 2012). Several of the Canadian participants in this study noted that many Indigenous worldviews include the notion that they have been given a responsibility from The Creator to take care of water and the land, and that the responsibility aspect is lost when access to water is viewed only through a rights lens. Australian participants did not explicitly use ‘responsibility’ as a concept although it was implied through their stories of how community research partners relate to their water resources (see also, for example, Jackson et al., 2014). Rather than discussing a responsibility relationship to water in detail, respondents from Australia chose to focus more intently on the importance of government and policy-makers in fulfilling their legal requirements under the Native Title Act. Page 298

R. D. Stefanelli et al. / Water Policy 19 (2017) 1063–1080

1077

Central to the tenets of meaningful engagement of local Indigenous communities and the implementation of Indigenous Ways of Knowing is an understanding of power structures, and a subsequent redistribution of power within the research process (Nadasdy, 1999; Latulippe, 2015a, 2015b). These structures of power are evident within both the socio-political realm, and the researcher–community partnership realm. Indigenous communities in both Canada and Australia face similar challenges in relation to power dynamics, where only recently have governance structures and academic researchers begun the process of redistributing power in research and management (von der Porten & de Loë, 2013b; Jackson et al., 2014). In Canada and Australia, participants referred to socio-political dynamics related to the history of colonialism. This history, and its ongoing manifestations, has created an environment in which Western science is heavily favored, and Indigenous knowledge and knowledge-holders are often included in a tokenistic manner in resource governance discussions (Wilson, 2008). Within this argument are the complexities associated with stakeholders and stakeholder engagement (von der Porten & de Loe, 2013a). In Canada, for example, Indigenous peoples are not just stakeholders with an interest in the results of a project – they are rights-holders vis-à-vis our Constitution and through Treaties. Within this Canadian context, the recently elected federal government has acknowledged the need to enter a Nation-to-Nation relationship with Indigenous peoples to rectify the imbalance of power that currently defines Indigenous–State relations, although time will tell if these campaign promises materialize. To help remedy power imbalances in research relationships, participants from both Canada and Australia unsurprisingly referenced the tenets of a CBPR approach that uses Indigenous and Western methods as an appropriate avenue to explore when undertaking research on water resource management. Such tenets include establishing meaningful relationships with communities, and working together as equals in the development of research from the proposal stage through to knowledge mobilization through publication and action (Castleden et al., 2017). Participants in both Canada and Australia noted that taking a CBPR approach to their water research proved to be a positive way to support the implementation of Indigenous and Western Ways of Knowing. For the strengths of Indigenous knowledge to be operationalized alongside the strengths of Western knowledge in an integrative way (see Bartlett et al., 2015), government officials, researchers, and community members must account for and overcome barriers to implementation. The ‘lessons learned’ and philosophical guidance offered by Australian and Canadian participants stressed the importance of relationship building, equitable partnerships, and equal distribution of benefits, among others to meaningfully engage with Indigenous community partners, and to implement both Indigenous and Western Ways of Knowing to take care of water. Implications While the focus of these interviews was on the topic of water, it should be noted that many of the conclusions that emerged from this research have implications beyond the realm of water and often provide a commentary on the relationship that exists between Indigenous and Settler populations and our lived experiences in place. Academic and community-based researchers alike seeking to conduct research that implements both Indigenous and Western knowledge systems in integrative ways should consider what this means in the context of any research project (Latulippe, 2015a, 2015b). While researchers within this field generally have a solid understanding of just how much time, Page 299

1078

R. D. Stefanelli et al. / Water Policy 19 (2017) 1063–1080

energy, and resources are required to successfully implement Indigenous and Western knowledge systems, participants stated that there is a need for this same level of recognition from a broader audience – particularly those within government and academic environments such as policy-makers, granting agencies, and publishers. At present, these institutions have begun to recognize the value of using multiple knowledge systems to overcome water-related challenges (Jackson et al., 2014); however, recognition that knowledge implementation should occur and enacting policies that require this implementation have not led to a clear articulation of how multiple knowledge systems can be implemented. Improving institutional understanding as well as providing funding and resources to researchers working to address these questions of how to implement Indigenous and Western knowledge would allow for researchers to better overcome the many obstacles that impede the meaningful engagement between community and researcher as it relates to water research and management. Limitations Like most qualitative studies that use interviewing as a data collection method, this research was subject to the level of interest and willingness to share information on the part of the research participants. Another key limitation to this project was that most (15 of 17) of the participants that were recruited were of a Settler heritage, and all were academically trained (i.e., trained in Western science). While this study allows us to better understand the perspectives of mainly Settler academic researchers working at the leading edge in this field, the logical next step for this work would be to examine the experiences of Indigenous community researchers and knowledge-holders to further our understandings. Conclusion The implementation of Indigenous and Western knowledge systems occurs within large structures of power and disparate understandings of our relationship with water. The purpose of this paper has been to explore researchers’ experiences in conducting integrative water research and management in Canada and Australia by providing them with an opportunity to expand on the details of their approach beyond what they published in the water literature. Water is a transboundary resource and its protection requires cooperation between many levels of government. However, as Indigenous peoples the world over continue to remind us, water is more than just a resource that needs to be managed – water is life. Without water, neither Indigenous nor Settler peoples can survive. We all have a stake in ensuring water is protected, and to date, the exclusive use of Western sciences has failed to provide access to safe water for Indigenous peoples in Canada, Australia, and beyond. This is not meant as a critique of Western science; instead, we posit that the exclusive use of Western knowledge limits the questions we can ask of water, and the answers we can hear. To account for this, we should first acknowledge that Indigenous communities have, and continue to develop research protocols, and to follow those protocols where they exist. Next, we must approach research partnerships from a position of humility (a reflexive awareness that there are limits to our own knowledge) and respect for both Indigenous and Western ontologies around water. From there we can collaboratively determine appropriate strategies to address these water-related questions. In sum, we need to establish authentic partnerships designed in a manner that will support the strengths Page 300

R. D. Stefanelli et al. / Water Policy 19 (2017) 1063–1080

1079

of both Indigenous and Western knowledge systems to take care of water for present and future generations.

References Australian Institute of Aboriginal and Torres Strait Islander Studies (2012). Guidelines for Ethical Research in Australian Indigenous Studies, 2012. https://aiatsis.gov.au/research/ethical-research/guidelines-ethical-research-australian-indigenousstudies (accessed February 13 2017). Ayre, M. & Mackenzie, J. (2013). ‘Unwritten, unsaid, just known’: the role of Indigenous knowledge(s) in water planning in Australia. Local Environment 18(7), 753–768. Barber, M. & Jackson, S. (2011). Aboriginal water values and resource development pressures in the Pilbara region of northwest Australia. Australian Aboriginal Studies 2, 32–49. Barlow, M. (2015). The water crisis comes home: thanks to growing inequality, climate change, and the mis-management of natural resources, the world’s water crisis has finally reached the Global North. Nation 300(31), 12–14. Bartlett, C., Marshall, M., Marshall, A. & Iwama, M. (2015). Integrative science and two-eyed seeing: enriching the discussion framework for healthy communities. In: Ecologically Emancipated Communities. Hallstrom, L. (ed.), UBC Press, Vancouver, BC, Canada. Butler, L. J., Scammell, M. K. & Benson, E. B. (2016). The Flint, Michigan, water crisis: a case study in regulatory failure and environmental injustice. Environmental Justice 9(4), 93–97. Canadian Institutes of Health Research, Natural Sciences and Engineering Research Council of Canada, and Social Sciences and Humanities Research Council of Canada (2014). Tri-Council Policy Statement: Ethical Conduct for Research Involving Humans, December 2014. Castleden, H., Cunsolo, A., Harper, S. L., Martin, D., Hart, C., Stefanelli, R., Day, L. & Lauridsen, K. (2015). Living with water: Examining methods and models for integrative Indigenous and Western Knowledge to inform – and transform – water research and management in Canada. Report prepared for the Canadian Water Network. Castleden, H., Hart, C., Cunsolo, A., Harper, S. & Martin, D. (2017). Reconciliation and relationality in water research and management in Canada: Implementing indigenous ontologies, epistemologies, and methodologies. In: Water Policy and Governance in Canada. Springer International Publishing, Switzerland, pp. 69–95. Cook, B. I., Ault, T. R. & Smerdon, J. E. (2015). Unprecedented 21st century drought risk in the American Southwest and Central Plains. Science Advances 1(1), 1–7. Cooke, M., Mitrou, F., Lawrence, D., Guimond, E. & Beavon, D. (2007). Indigenous well-being in four countries: an application of the UNDP’s human development index to indigenous peoples in Australia, Canada, New Zealand, and the United States. BMC International Health and Human Rights 7(1), 9. Daley, K., Castleden, H., Jamieson, R., Furgal, C. & Ell, L. (2015). Water systems, sanitation, and public health risks in remote communities: Inuit resident perspectives from the Canadian Arctic. Social Science & Medicine 135, 124–132. Duncan, R. (2014). Regulating agricultural land use to manage water quality: the challenges for science and policy in enforcing limits on non-point source pollution in New Zealand. Land Use Policy 41, 378–387. Dunn, K. (2016). Interviewing. In Qualitative Research Methods in Human Geography, 4th edn. Hay, I. (ed.), Oxford University Press, Toronto, ON, Canada, pp. 149–188. Eero, M., Vinther, M., Haslob, H., Huwer, B., Casini, M., Storr-Paulsen, M. & Köster, F. W. (2012). Spatial management of marine resources can enhance the recovery of predators and avoid local depletion of forage fish. Conservation Letters 5(6), 486–492. Finn, M. & Jackson, S. (2011). Protecting indigenous values in water management: a challenge to conventional environmental flow assessments. Ecosystems 14(8), 1232–1248. First Nations Health Authority (FNHA) (2017). Drinking Water Safety Program. http://www.fnha.ca/what-we-do/environmental-health/drinking-water-safety-program (accessed February 12 2017). Gersonius, B., Ashley, R., Pathirana, A. & Zevenbergen, C. (2013). Climate change uncertainty: building flexibility into water and flood risk infrastructure. Climatic Change 16(2), 411–423. Hansen, J., Sato, M., Hearty, P., Ruedy, R., Kelley, M., Masson-Delmotte, V., Russell, G., Tselioudis, G., Cao, J., Rignot, E., Velicogna, I., Tormey, B., Donovan, B., Kandiano, E., von Schuckmann, K., Kharecha, P., Legrande, A. N., Bauer, M. & Page 301

1080

R. D. Stefanelli et al. / Water Policy 19 (2017) 1063–1080

Lo, K.-W. (2016). Ice melt, sea level rise and superstorms: evidence from paleoclimate data, climate modeling, and modern observations that 2 °C global warming could be dangerous. Atmospheric Chemistry and Physics 16(6), 3761–3812. Health Canada (2016). Drinking Water Advisories in First Nations Communities. Ottawa, ON. http://www.hc-sc.gc.ca/fniahspnia/promotion/public-publique/water-dwa-eau-aqep-eng.php (accessed February 13 2017). Jackson, S., Douglas, M. M., Kennard, M. J., Pusey, B. J., Huddleston, J., Harney, B., Liddy, L., Liddy, M., Liddy, R., Sullivan, L., Huddleston, B., Banderson, M., McMah, A. & Allsop, Q. (2014). ‘We like to listen to stories about fish’: integrating Indigenous ecological and scientific knowledge to inform environmental flow assessments. Ecology and Society 19(1), 43. Latulippe, N. (2015a). Bridging parallel rows: epistemic difference and relational accountability in cross-cultural research. International Indigenous Policy Journal 6(2), 1–17. Latulippe, N. (2015b). Bringing governance into the conversation: introducing a typology of traditional knowledge literature. AlterNative 11(2), 118–131. McDonald, R. I., Weber, K. F., Padowski, J., Boucher, T. & Shemie, D. (2016). Estimating watershed degradation over the last century and its impact on water-treatment costs for the world’s large cities. Proceedings of the National Academy of Sciences 113(32), 9117–9122. Mitrou, F., Cooke, M., Lawrence, D., Povah, D., Mobilia, E., Guimond, E. & Zubrick, S. R. (2014). Gaps in Indigenous disadvantage not closing: a census cohort study of social determinants of health in Australia, Canada, and New Zealand from 1981–2006. BMC Public Health 14(201), 1–9. Nadasdy, P. (1999). The politics of TEK: power and the ‘integration’ of knowledge. Arctic Anthropology 36(1/2), 1–18. Regan, P. Y. L. (2006). Unsettling the Settler Within: Canada’s Peacemaker Myth, Reconciliation, and Transformative Pathways to Decolonization. Doctoral Dissertation, University of Victoria, Victoria, Canada. Richmond, C. A. & Ross, N. A. (2009). The determinants of First Nation and Inuit health: a critical population health approach. Health & Place 15(2), 403–411. Stefanelli, R. D., Castleden, H., Harper, S. L., Martin, D., Cunsolo, A. & Hart, C. (2017). Experiences with integrative Indigenous and Western knowledge in water research and management: a systematic realist review of literature from Canada, Australia, New Zealand, and the United States. Environmental Reviews 999, 1–11. United Nations (2016). United Nations Water and Sanitation: Sustainable Development Goals – Goal 6: Ensure Access to Water and Sanitation for All. New York. http://www.un.org/sustainabledevelopment/water-and-sanitation/ (accessed February 12 2017). United Nations Development Programme (2015) United Nations Human Development Report 2015: Work for Human Development. New York, pp. i–182. http://hdr.undp.org/sites/default/files/2015_human_development_report.pdf (accessed February 12 2017). United Nations General Assembly (2007). United Nations Declaration on the Rights of Indigenous Peoples: Resolution / adopted by the General Assembly, 2 October 2007, A/RES/61/295. United Nations General Assembly, Geneva. von der Porten, S. & de Loë, R. C. (2013a). Collaborative approaches to governance for water and Indigenous peoples: a case study for British Columbia, Canada. Geoforum 50, 149–160. von der Porten, S. & de Loë, R. C. (2013b). Water governance and Indigenous governance: towards a synthesis. Indigenous Policy Journal 23(4), 1–12. Vörösmarty, C. J., Green, P., Salisbury, J. & Lammers, R. B. (2000). Global water resources: vulnerability from climate change and population growth. Science 289(5477), 284–288. White, J. P., Murphy, L. & Spence, N. (2012). Water and Indigenous peoples: Canada’s paradox. International Indigenous Policy Journal 3(3), 1–28. Wilson, S. (2008). Research is Ceremony: Indigenous Research Methods. Fernwood Publishing, Black Point, NS, Canada. Wolfe, P. (2006). Settler colonialism and the elimination of the native. Journal of Genocide Research 8(4), 387–409. Woodward, E. L. (2008). Creating the Ngan’gi Seasons calendar: reflections on engaging Indigenous knowledge authorities in research. Learning Communities 2, 125–137. Received 6 December 2016; accepted in revised form 6 April 2017. Available online 19 July 2017

Page 302

Water Practice & Technology

ISSN Online 1751-231X iwaponline.com/wpt

Water Practice and Technology covers all practical aspects of water and wastewater treatment and management throughout the water cycle. Papers are primarily aimed at practitioners, but will also be of interest to scientists, managers and those active in training and professional development. The journal’s scope includes: • Solutions to practical problems in the design, operation or management of water supply, wastewater treatment, drainage and flood protection • Environmental management, including social, economic and public participation aspects of water This includes informative case studies, practical “know-how” reports, full-scale applications of new technologies, “best practice” and applied management concepts, including lessons learnt from unsuccessful experiences. For more details, visit iwaponline.com/wpt

Page 306

Water Practice & Technology Vol 12 No 4 doi: 10.2166/wpt.2017.087

Removal of pharmaceuticals with ozone at 10 Swedish wastewater treatment plants F. Nilssona,b,*, M. Ekblada,c, J. la Cour Jansena and K. Jönssona a

Water and Environmental Engineering at the Department of Chemical Engineering, Lund University, P.O. Box 124, Lund SE-221 00, Sweden

Primozone Production AB, Terminalvägen 2, Löddeköpinge SE-246 42, Sweden

Sweden Water Research AB, Ideon Science Park, Scheelevägen 15, Lund 223 70, Sweden

*Corresponding author. E-mail: ﬁlip.nilsson@primozone.com

Abstract Pilot-scale tests were run with ozonation for reduction of 24 pharmaceuticals at 10 full-scale wastewater treatment plants in southern Sweden. Reduction was evaluated based on doses of 3, 5 and 7 g O3/m3 at all plants. The reduction of pharmaceuticals reached on average 65% at 3 g O3/m3, 78% at 5 g O3/m3 and 88% for 7 g O3/m3 in terms of total concentration of pharmaceuticals. Speciﬁc ozone dose (ratio O3:TOC) was found to be highly inﬂuential on pharmaceutical removal. At two WWTPs, the pharmaceutical removal was severely reduced. Key words: ozonation, pharmaceuticals, pilot-scale

INTRODUCTION Many countries are considering the need for reduction of pharmaceuticals and other organic micropollutants in wastewater. In Switzerland the legal framework is already in place (Eggen et al. 2014), in the EU, the list of priority pollutants already include organic micropollutants and the ‘watch list’ has recently been extended with a number of pharmaceuticals (2013/39/EU). In Sweden, the first fullscale installation based on ozonation, is under construction (IVL 2016) even though the needs and requirements of such a treatment step are still debated. Ozonation and activated carbon treatment, or a combination seems to be the winning technologies for reduction of organic micropollutants. Full-scale installations have only been reported in a few countries (Cimbritz et al. 2016) but pilot-scale installations have been running at several places in order to test the technology and to give guidelines for design (Hollender et al. 2009; Wert et al. 2009; Ibáñez et al. 2013; Margot et al. 2013). Such guidelines are problematic as long as the substances included in the control program and the limits and control methods are not selected at the same time. Typically, a number of substances in high concentration for which reasonable analytical methods exist are selected and the final effluent concentration or the percentage reduction is used as evaluation criteria (Huber et al. 2005; Hansen et al. 2010; Antoniou et al. 2013). For design of equipment and estimation of the economy in ozonation, guidelines for the needed ozone dose is typically based on the content of organic material in the treated wastewater. As the major part of the ozone is consumed by organic matter left after the normal treatment only a minor part is used to oxidize micropollutants. In addition, pH, alkalinity and a number of substances that might be present in treated wastewater are known to have a significant impact on ozone Page 307

872

Water Practice & Technology Vol 12 No 4 doi: 10.2166/wpt.2017.087

consumption and consequently on the ozone dose needed (Gottschalk et al. 2010; Hansen et al. 2010; Antoniou et al. 2013; Hey et al. 2014). The Swedish debate about the need for reduction of organic micropollutants, especially pharmaceuticals, might end up with a general requirement of wastewater treatment plants to include reduction of pharmaceuticals in the near future. Consequently, a deeper understanding of the expected need for ozone addition is required before the economic and environmental consequences can be evaluated. Testing at single treatment plants might not be representative for other WWTPs. Therefore, the main objective of the present study was to test whether ozone can be used in the same manner and reach comparable results in terms of pharmaceutical reduction regardless of the WWTP conﬁguration. The secondary objective was to study how the concentration of TOC in the treated wastewater impacts the efﬁciency of pharmaceutical reduction and evaluate whether it can be used in a general model to control the amount of ozone being produced in a full-scale ozone installation.

MATERIALS AND METHODS Pilot plant

A schematic representation of the ozone equipment used throughout the trials is depicted in Figure 1. The objective of the system was to produce and dissolve ozone into wastewater in a measurable and repeatable way at 10 WWTPs. All equipment was housed in a 20 feet container. A submerged centrifugal pump delivered 18–20 m3/h of treated wastewater into a drum filter. The purpose of the drum filter was to reduce turbidity and minimize the impact of fluctuations in WWTP performance (such as an underperforming clarifier) on the ozone pilot plant. The flow entering the drum filter was monitored by a flow meter at the inlet. The reduction of turbidity across the filter was monitored by two turbidity meters positioned before and after the filter. Filtered wastewater entered a holding tank to equalize the flow before the ozone injection. The excess flow (about 12 m3/h) was discharged back into the wastewater stream, downstream of the submerged centrifugal pump. Ozone was produced from onsite generated oxygen. A main PLC was connected to the flow meter, ozone generator and ozone concentration meter, which enabled the ozone dose (displayed as g O3/m3 on the main PLC screen) to be monitored manually throughout the trials. The wastewater

Figure 1 | Schematic overview of the equipment used. 1: Compressor (AirSep, Topaz Plus), 2: PSA oxygen supply (AirSep, Topaz Plus), 3: Chiller (Lauda, UC Mini), 4: Ozone generator (Primozone, GM2), 5: Ozone concentration meter (BMT, 964-C), 6: Submerged centrifugal pump (Mecana, TF2), 7: Flowmeter (Mecana, TF2), 8: Turbidity meter (Mecana, TF2), 9: Drum filter (Mecana, TF2), 10: Turbidity meter (Mecana, TF2), 11: 1 m3 equalization tank, 12: Booster pump (Grundfos, CM5-4), 13: Flow meter (Honsberg), 14: Venturi injector (Mazzei, 1583), 15: 500 l pressurized reaction vessel (HRT: 5 min), 16: Sludge from drum filter, 17: Excess flow, 18: Discharge from ozonation, 19: Main PLC (Schneider, Modicon M251).

Page 308

873

Water Practice & Technology Vol 12 No 4 doi: 10.2166/wpt.2017.087

from the equalization tank (6 m3/h) was pumped with a booster pump through a venturi injector and mixed with the ozone, the flow of water being monitored by a flow meter. The ozonated wastewater then entered a pressurized reaction tank (5 min HRT at 6 m3/h) before being discharged back into the wastewater stream downstream of the submerged centrifugal pump. There were a total of 3 sampling locations, at the inlet of the drum filter (IN), after the filter (AF) and outlet (OUT) of the pressurized reaction tank.

Operation of the pilot plant

The trials were run in the same manner throughout all 10 WWTPs. The submerged pump was lowered into the discharge stream at the WWTP, wastewater was pumped through the system for at least 24 hours prior to the commencement of the trials. This was done to safeguard that no residuals from the previous trial were present. The ozone production was started and adjusted manually until the production corresponded to the lowest ozone dose (3 g O3/m3 wastewater). After the ozone ﬂow had reached the required dosage, it was kept running for 20 minutes (4 HRT) before the ﬁrst samples were taken. Samples were collected in glass bottles from points IN (0.5 L), AF (0.5 L) and OUT (1.5 L) every 10 minutes for a total of 60 minutes which resulted in 3.5 L sample from point IN, 3.5 L from point AF and 10.5 L from point OUT (3 g O3/m3). The next ozone doses (5 and 7 g O3/m3 wastewater) were then introduced to the system in the same manner. The samples from points IN and AF were taken as composite samples for the entire trial run.

Analysis

All samples were analyzed in the lab at Lund University. The chemical analysis was conducted with a spectrophotometer (Hach-Lange DR 2800): COD (Hach-Lange, LCK 314), TOC (LCK 385), Tot-P (LCK 349), PO4-P (LCK 349), Tot-N (LCK 138), NH4-N (LCK 303), NO3-N (LCK 339), NO2-N (LCK 341). The other analysis were conducted using the standard procedures: SS (according to SSEN 872:2005) and pH (WTW pH 320). SUVA 254 measurement was conducted with a modiﬁcation of the standard method published by USEPA (2005), absorbance was measured at 254 nm and the results were normalized with regards to TOC. The SS content in the samples taken from point OUT were so low that TOC can be regarded as dissolved organic carbon. SUVA 254 was not conducted for the ﬁrst three WWTPs. Samples were also sent to IVL (Swedish Environmental Research Institute) for pharmaceutical analysis (liquid chromatography-tandem mass spectrometry) were carried out in accordance with Gros et al. (2006). The pharmaceuticals analyzed are listed in Table 1.

Wastewater treatment plants

The trials were run at 10 different WWTPs in southern Sweden (Table 2), all designed for more than 10,000 PE. There are differences in the geographical location of the plants, as well as the configuration and industries connected to them. The plants are briefly described in Table 2. Sjölunda WWTP has an unusual implementation of BOD and nitrogen removal. A detailed description of the plant can be found in (Hanner et al. 2003). In short, BOD is removed in a high loaded activated sludge plant. Nitrification takes place in trickling filters with plastic carriers followed by denitrification in a two-stage MBBR system. Final separation takes place in a flotation plant. Page 309

Water Practice & Technology Vol 12 No 4 doi: 10.2166/wpt.2017.087

874 Table 1 | The 24 pharmaceuticals included in the analysis Name

Type

Name

Type

Amlodipine

Antihypertensive

Metoprolol

Antihypertensive

Atenolol

Antihypertensive

Naproxen

Anti-inﬂammatory

Bisoprolol

Antihypertensive

Oxazepam

Sedative

Caffeine

Stimulant

Paracetamol

Anti-inﬂammatory

Carbamazepine

Sedative

Propranolol

Antihypertensive

Ciproﬂoxacin

Antibiotic

Ranitidine

Antiulcer

Citalopram

Antidepressant

Sertralin

Antidepressant

Diclofenac

Anti-inﬂammatory

Sulfamethoxazole

Antibiotic

Furosemide

Diuretic

Terbutaline

Asthma medication

Hydrochlorothiazide

Antihypertensive

Tetracycline

Antibiotic

Ibuprofen

Anti-inﬂammatory

Trimetoprim

Antibiotic

Ketoprofen

Anti-inﬂammatory

Warfarin

Anticoagulant

Table 2 | General description of the WWTPs in this trial. WWTP

PE (connected)

BOD removal

Nitriﬁcation

Denitriﬁcation

Sand ﬁltration

Sternö

21,200

Sjöhög

33,900

Nyvångsverket

11,800

Torekov

12,900

Sjölunda

317,000

MBBR

Källby

98,600

Ellinge

20,100

Kävlinge

29,000

Svedala

12,000

Västra Stranden

70,000

AS: activated sludge, TF: trickling filter, MBBR: moving bed biofilm reactor. X in the final column denotes that the WWTP utilizes a sand filter

RESULTS AND DISCUSSION Pharmaceuticals

There are at present no set concentration limits for speciﬁc pharmaceuticals within the EU. Therefore, the reduction of pharmaceuticals in this paper is presented as reduction of the total concentration of all analyzed pharmaceuticals. Furthermore, as there are no reduction criteria in use in the EU, the Swiss limit (Eggen et al. 2014) of 80% reduction of pharmaceuticals is designated as the target value. The sum of all 24 pharmaceuticals is depicted in Table 3. If a pharmaceutical concentration was below the detection limit, the concentration of such compound was set to half the detection limit. The total concentration of pharmaceuticals that entered the pilot plant varied between 4,600 and 18,700 ng/L. As the analysis for pharmaceuticals were conducted on dissolved compounds only and no precipitating agents were employed, the drum ﬁlter is not considered to have had any impact on the pharmaceutical removal in these trials. Figure 2 shows the total pharmaceutical removal at the three applied ozone doses. On average, the total concentration of measured pharmaceuticals was reduced by 65% at 3 g O3/m3, 78% at 5 g O3/m3 Page 310

Water Practice & Technology Vol 12 No 4 doi: 10.2166/wpt.2017.087

875

Table 3 | Sum of the pharmaceutical concentrations (ng/L) at points IN and OUT at the 10 WWTPs. OUT Ozone dose WWTP

3 g O3/m3

5 g O3/m3

7 g O3/m3

Sternö

10,651

5,064

2,505

1,093

Sjöhög

15,321

9,270

3,487

1,418

Nyvångsverket

8,359

2,284

1,597

674

Torekov

4,603

1,223

458

240

Sjölunda

12,420

2,969

2,027

860

Källby

12,095

2,308

1,048

711

Ellinge

17,860

6,680

5,040

1,953

Kävlinge

10,896

2,253

640

472

Svedala

18,702

10,189

10,258

6,283

Västra Stranden

7,838

2,453

2,156

1,683

Figure 2 | Total pharmaceutical removal at the three ozone doses (3, 5 and 7 g O3/m3).

and 88% at 7 g O3/m3. However, these figures are the average removal from all the WWTPs, if the last two WWTPs were to be removed from consideration, said figures reaches 67%, 83% and 92% for 3, 5 and 7 g O3/m3 doses. This increase in averages points to the impact of the lower removal at Svedala and Västra Stranden and elicits further analysis, as something clearly caused the pharmaceutical removal to be less effective at those WWTPs. In the case of Svedala WWTP, the total concentration of pharmaceuticals entering the pilot plant reached 18,702 ng/L which is the highest of all the WWTPs, however, it is considered to be comparable to Ellinge WWTP at 17,860 ng/L. Furthermore, the inlet concentration to the pilot plant at Västra Stranden (7,838 ng/L) is comparable to Nyvångsverket WWTP (8,359 ng/L). Thus, the inlet concentrations of pharmaceuticals at Svedala and Västra Stranden are not considered to be responsible for the lower reduction at those plants. When applying the criteria set for removal of pharmaceuticals (.80%), 3 g O3/m3 is sufficient only at two WWTPs (Källby and Kävlinge). At 5 g O3/m3 the number of plants reaching the criteria increases to 5 (Nyvång, Torekov, Sjölunda, Källby and Kävlinge). At 7 g O3/m3 all but the last two WWTPs reaches 80% removal. The pharmaceutical compounds are divided into groups in Table 4. The compounds which were removed above 80% at more than 6 WWTPs are grouped according to the ozone dose required to reach that criteria. The column ‘Removed ,80% or at ,6 plants’ contains the pharmaceuticals Page 311

Water Practice & Technology Vol 12 No 4 doi: 10.2166/wpt.2017.087

876

Table 4 | Individual pharmaceuticals divided into groups depending on to which extent they were removed. Removed at 3 g O3/m3

Removed at 5 g O3/m3

Removed at 7 g O3/m3

Removed ,80% or at ,6 plants

Not found in sufﬁcient concentration

Diclofenac

Citalopram

Hydrochlorothiazide

Ibuprofen

Ciproﬂoxacin

Furosemide

Sulfamethoxazole

Warfarin

Tetracycline

Naproxen

Atenolol

Caffeine

Amlodipine

Carbamazepine

Bisoprolol

Ketoprofen

Propranolol

Metoprolol

Oxazepam

Ranitidine

Sertralin

Paracetamol Terbutaline Trimetoprim

33%

2.8%

36.7%

26.5%

1.5%

which were not removed above 80% or removed above 80% but in fewer than 6 WWTPs and are considered difficult to remove. The last column contain the pharmaceuticals that were not found in sufficient concentration at sufficient number of WWTPs. The average relative inlet concentration of the compound groups (%) are listed at the bottom of each group. The first two groups in Table 4 (‘Removed at 3 g O3/m3’ and ‘Removed at 5 g O3/m3’) corresponds reasonably well with the findings of Antoniou et al. (2013) and Margot et al. (2013). The last two groups (‘Removed at 7 g O3/m3’ and ‘Removed ,80% or at ,6 plants’) corresponds well with the findings of Hey et al. (2014) and Hollender et al. (2009). The average removal of individual pharmaceutical compounds is depicted in Figure 3 along with the standard deviation of each compound. As ciproflaxin, tetracycline and amlodipine were only found in a handful of WWTPs they are excluded from the graph.

Figure 3 | Average pharmaceutical removal of individual pharmaceutical compounds at all 10 WWTPs.

As is apparent from Figure 3, the standard deviation is quite large for some of the pharmaceutical compounds, especially at the lowest ozone dose (3 g O3/m3). For instance, ibuprofen exhibits a standard deviation of 38% at the lowest ozone dose. The reason for this rather large standard deviation is unknown, however, it is not surprising since the trials were conducted in pilot-scale at several WWTPs with varying wastewaters. A radar chart of the removal of diclofenac, from the group ‘Removed at 3 g O3/m3’ in Table 4 is depicted in Figure 4. Diclofenac is removed completely at all WWTPs at the lowest ozone dose Page 312

877

Water Practice & Technology Vol 12 No 4 doi: 10.2166/wpt.2017.087

Figure 4 | Radar chart of the removal of diclofenac at 10 WWTPs.

except at Svedala and Västra Stranden, where it takes an ozone dose of 7 g O3/m3 to remove this compound with more than 80%. The removal of citalopram (from the group ‘Removed at 5 g O3/m3’ in Table 4) is depicted in Figure 5. This compound was not found in Västra Stranden so that WWTP is excluded from this chart. An ozone dose of 5 g O3/m3 is required to remove this compound above 80% at all plants except Svedala where it is not removed above 40% at any ozone dose. When the removal of hydrochlorothiazide (from the group ‘Removed at 7 g O3/m3’ in Table 4) is plotted in the same way (Figure 6) it becomes apparent that the more difﬁcult a compound is to remove the larger the variations in removal becomes. For instance, in both Nyvångsverket and Torekov this compound is removed by 80% with the lowest ozone dose (3 g O3/m3). Whereas in Sternö, Sjöhög, Sjölunda and Ellinge the lowest ozone dose does not remove this compound to more than 40%. None of the ozone doses were sufﬁcient to remove hydrochlorothiazide at neither Svedala nor Västra Stranden.

Figure 5 | Radar chart of the removal of citalopram at 9 WWTPs.

Figure 6 | Radar chart of the removal of hydrochlorothiazide of at 10 WWTPs.

Page 313

Water Practice & Technology Vol 12 No 4 doi: 10.2166/wpt.2017.087

878

In most of the samples, the increasing dose of ozone (3 to 7 g O3/m3) oxidized nitrite to below the detection limit. As ozone reacts well with nitrite this was not unexpected, however, the samples from Svedala and Västra Stranden did not show the same trend and nitrite was still measurable after ozonation. The lack of nitrite oxidation at those plants indicates that ozone scavenging took place. The ingoing nitrite concentrations were so low at those plants (0.14 for Svedala and 0.09 mg/L for Västra Stranden) that the nitrite is not considered to be the reason for the poor removal of pharmaceuticals. COD and TOC

COD was measured in all samples and ranged between 21 mg/L and 35 mg/L, but since COD analysis is not in general available in Sweden anymore TOC is used instead. The concentration of total organic carbon discharged from the WWTPs (Table 5), ranged between 8.4 and 13.9 mg TOC/L. Ozone did remove some of the TOC albeit not to a high degree. A closer look at the TOC ﬁgures reveals that the TOC values for Svedala and Västra Stranden WWTPs are not so high as to explain the discrepancy in pharmaceutical removal at those plants. Table 5 | TOC measurements (mg TOC/L) in points IN, AF and OUT at the 10 WWTPs. OUT Ozone dose WWTP

IN (mg/L)

AF (mg/L)

3 g O3/m3

5 g O3/m3

7 g O3/m3

Sternö

13.9

13.5

13.4

Sjöhög

10.6

10.1

9.8

9.9

Nyvångsverket

11.4

8.1

Torekov

9.5

8.7

8.5

8.6

8.4

Sjölunda

11.0

10.8

9.3

10.6

9.9

Källby

9.8

8.9

8.1

8.0

8.2

Ellinge

13.1

12.9

12.5

Kävlinge

8.4

8.2

7.3

7.7

7.4

Svedala

13.3

12.3

12.7

12.5

Västra Stranden

11.4

10.9

10.8

11.0

No results available due to lacking samples.

SUVA 254 nm

The SUVA 254 values are depicted in Figure 6. A decline in SUVA 254 can be considered to be an indicator of increasing pharmaceutical removal as well as a decline in the total concentration of aromatic substances (Wittmer et al. 2015). The majority of the SUVA 254 values in Figure 7 declines as the ozone dose increases. The SUVA 254 results obtained from Svedala follow the same trend as the other WWTPs, however, with a much lower reduction of aromatics at the higher ozone doses. In Västra Stranden, the SUVA 254 does not decline at all even at 7 g O3/m3. The behavior displayed at Västra Stranden WWTP indicates that ozone scavenging took place to a high degree. Pharmaceutical removal as a function of TOC

The relationship between TOC and pharmaceutical removal is depicted in Figure 8. TOC and ozone dose are combined into speciﬁc ozone dose (ratio O3:TOC). Page 314

879

Water Practice & Technology Vol 12 No 4 doi: 10.2166/wpt.2017.087

Figure 7 | The speciﬁc UV absorbance at 254 nm at the different measuring points.

Figure 8 | A combined graph of speciﬁc ozone dose (ratio O3:TOC) and pharmaceutical removal.

As the ratio of ozone to TOC increases so does the pharmaceutical removal. The removal efficiency increases rapidly when O3:TOC is increased from approximately 0.2 to 0.4 after which the total removal levels off. This overall behavior is not surprising as the compounds which are easily removed (removed at 3 and 5 g O3/m3, Table 4) are removed well above 80% at the lower doses (O3:TOC 0.2–0.4). Followed by the more difficult compounds (removed at 7 g O3/m3 and Removed ,80% and/or at ,6 plants, Table 4) which requires a higher specific ozone dose, eventually leaving only the compounds which are not susceptible to ozone oxidation in this range of ozone doses. The spread in the data points at the lower region (O3:TOC 0.2–0.4) is quite substantial while being more clustered together in the higher region (O3:TOC 0.4–0.8). The apparent trend seen in Figure 8 can be useful in the early stages when an ozone system is to be sized. However, the high degree of spread between the data points (especially between O3:TOC 0.2– 0.4) and the low number of samples (3 samples at 10 WWTPs) points to that further testing is needed before the ratio O3:TOC can be used as an online control parameter in full-scale. However, if tests were to be performed at one WWTP instead of 10 and run for a longer time, it is very likely that a model of how specific ozone dose impacts pharmaceutical removal could be found for that specific WWTP. Page 315

880

Water Practice & Technology Vol 12 No 4 doi: 10.2166/wpt.2017.087

In these trials, the only ozone dose able to meet the .80% removal criteria at a majority (8 out of 10) of the plants is 7 g O3/m3. A detailed cost calculation has not been made for this paper. However, a recent publication by Mulder et al. (2015) calculated the cost of running an ozone installation with a dose of 7.7 g O3/m3 to 0.16 €/m3 treated wastewater (+0.03 €) for a 300,000 p.e. WWTP. The cost of running a comparable PAC (powdered activated carbon) installation reached 0.18 €/m3 (+0.03 €) treated wastewater (Mulder et al. 2015). Svedala and Västra Stranden WWTPs

The removal of pharmaceuticals was in general quite high when excluding the last two WWTPs. However, when the last two WWTPs are included, the averages drops signiﬁcantly. The fact that nitrite was detected in the samples from these plants after ozonation was surprising since nitrite is highly reactive with ozone. Therefore, the most probable reasons for the lower pharmaceutical reduction at these plants are either an equipment failure leading to lower ozone doses or ozone scavenging of an unknown compound. No failures of the ozone equipment was detected at either of these trial runs which enhances the probability of an unknown ozone scavenging compound.

CONCLUSIONS The main purpose of the pilot-scale trials was to evaluate the practical application of ozone at different WWTPs without considering the differences at the plants. A criteria of 80% total removal of pharmaceuticals was established as a benchmark. This criteria was met at all WWTPs at 7 g O3/m3 except at Svedala and Västra Stranden, therefore the process can be said to remove pharmaceuticals efficiently and with reasonably comparable results but only at the higher ozone doses. The reason for the lower removal efficiency at Svedala and Västra Stranden WWTPs was not found. A link between specific ozone dose (ratio O3:TOC) and pharmaceutical removal efficiency was found to exist but it is not accurate enough to be integrated as a parameter to control the output of ozone as of yet. Further work is clearly needed to acquire a general model which can be implemented at any WWTP.

ACKNOWLEDGEMENTS This work was ﬁnancially supported by: The Swedish Water & Wastewater Association through VAteknik Södra, Tillväxtverket, Primozone Production AB and the municipalities of Karlshamn, Ystad, Malmö, Lund, Kävlinge, Halmstad, Svedala, Eslöv, Åstorp and Båstad. The authors thank all who participated in the study and wish to extend special thanks to the personnel at the WWTPs who made this study possible.

REFERENCES Antoniou, M. G., Hey, G., Vega, S. R., Spiliotopoulou, A., Fick, J., Tysklind, M., la Cour Jansen, J. & Andersen, H. R. 2013 Required ozone doses for removing pharmaceuticals from wastewater. Science of the Total Environment 456–457, 42–49. Cimbritz, M., Tumlin, S., Hagman, M., Dimitrova, I., Hey, G., Mases, M., Åstrand, N. & la Cour Jansen, J. 2016 Treatment of Pharmaceutical Residues and Other Micropollutants – a Literature Survey. Report 2016-04. Svenskt Vatten AB, Stockholm. Eggen, R. I. L., Hollender, J., Joss, A., Schärer, M. & Stamm, C. 2014 Reducing the discharge of micropollutants in the aquatic environment: the beneﬁts of upgrading wastewater treatment plants. Environmental Science & Technology 48, 7683–7689. EU/2013/39, OJ L 226, 24.8.2013, pp. 1–17.

Page 316

881

Water Practice & Technology Vol 12 No 4 doi: 10.2166/wpt.2017.087

Gottschalk, C., Libra, J. A. & Saupe, A. 2010 Ozonation of Water and Wastewater – A Practical Guide to Understanding Ozone and its Application, 2nd edn. Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim. Gros, M., Petrović, M. & och Barceló, D. 2006 Development of a multi-residue analytical methodology based on liquid chromatography-tandem mass spectrometry (LC-MS/MS) for screening and trace level determination of pharmaceuticals in surface and wastewaters. Talanta 70, 678–690. Hanner, N., Aspegren, H., Nyberg, U. & Andersson, B. 2003 Upgrading the Sjölunda WWTP according to a novel process concept. Water Science & Technology 47(12), 1–7. Hansen, K. M. S., Andersen, H. R. & Ledin, A. 2010 Ozonation of estrogenic chemicals in biologically treated sewage. Water Science & Technology 62(3), 649–657. Hey, G., Vega, S. R., Fick, J., Tysklind, M., Ledin, A., la Cour Jansen, J. & Andersen, H. R. 2014 Removal of pharmaceuticals in WWTP effluents by ozone and hydrogen peroxide. Water SA 40(1), 1–10. Hollender, J., Zimmermann, S. G., Koepke, S., Krauss, M., Mcardell, C. S., Ort, C., Singer, H., Von Gunten, U. & Siegrist, H. 2009 Elimination of organic micropollutants in a municipal wastewater treatment plant upgraded with a full-scale postozonation followed by sand filtration. Environmental Science & Technology 43, 7862–7869. Huber, M. M., Göbel, A., Joss, A., Hermann, N., Löffler, D., Mcardell, C. S., Ried, A., Siegrist, H., Ternes, T. A. & von Gunten, U. 2005 Oxidation of pharmaceuticals during ozonation of municipal wastewater effluents: a pilot study. Environmental Science & Technology 39, 4290–4299. Ibáñez, M., Gracia-Lor, E., Bijlsma, L., Morales, E., Pastor, L. & Hernández, F. 2013 Removal of emerging contaminants in sewage water subjected to advanced oxidation with ozone. Journal of Hazardous Materials 260, 389–398. IVL 2016 Sweden’s first purification plant for removal of wastewater pharmaceutical residues under construction. IVL news bulletin. http://www.ivl.se. 30th of November (accessed 25 January 2017). Margot, J., Kienle, C., Magnet, A., Weil, M., Rossi, L., de Alencastro, L. F., Abegglen, C., Thonney, D., Chèvre, N., Schärer, M. & Barry, D. A. 2013 Treatment of micropollutants in municipal wastewater: ozone or powdered activated carbon? Science of the Total Environment 461–462, 430–498. Mulder, M., Antakyali, D. & Ante, S. 2015 Costs of Removal of Micropollutants From Effluents of Municipal Wastewater Treatment Plants – General Cost Estimates for the Netherlands Based on Implemented Full Scale Post Treatments of Effluents of Wastewater Treatment Plants in Germany and Switzerland. STOWA and Waterboard the Dommel, Netherlands. SS-EN 872: 2005 Water quality - Determination of suspended solids - Method by filtration through glass fiber filters (In Swedish). USEPA. 2005 Determination of total organic carbon and specific UV absorbance at 254 nm in source water and drinking water. USEPA Document # EPA/600/R-05/055. Wert, E. C., Rosario-Ortiz, F. & Snyder, S. A. 2009 Effect of ozone exposure on the oxidation of trace organic contaminants in wastewater. Water Research 43, 1005–1014. Wittmer, A., Heisele, A., Mcardell, C. S., Böhler, M., Longree, P. & Siegrist, H. 2015 Decreased UV absorbance as an indicator of micropollutant removal efficiency in wastewater treated with ozone. Water Science & Technology 71(7), 980–985.

Page 317

Water Practice & Technology Vol 12 No 4 doi: 10.2166/wpt.2017.101

Aerobic granular biomass technology: advancements in design, applications and further developments Mario Pronka, Andreas Giesenb,*, Andrew Thompsonc, Struan Robertsonc and Mark van Loosdrechta a

Delft University of Technology, Biotechnology, Maasweg 9, Delft 2629 HZ, The Netherlands

Royal HaskoningDHV, P.O. Box 1132, Amersfoort 3800 BC, The Netherlands

Royal HaskoningDHV, 11 Newhall Street, Birmingham B3 3NY, UK

*Corresponding author. E-mail: andreas.giesen@rhdhv.com

Abstract Aerobic granular sludge is seen as the future standard for industrial and municipal wastewater treatment. Through a Dutch research and development program, a full-scale aerobic granular biomass technology has been developed – the Nereda® technology – which has been implemented to treat municipal and industrial wastewater. The Nereda® system is considered to be the first aerobic granular sludge technology applied at full-scale and more than 40 municipal and industrial plants are now in operation or under construction worldwide. Further plants are in the planning and design phase, including plants with capacities exceeding 1 million PE. Data from operational plants confirm the system’s advantages with regard to treatment performance, energy-efficiency and cost-effectiveness. In addition, a new possibility for extracting alginate-like exopolysaccharides (ALE) from aerobic granular sludge has emerged which could provide sustainable reuse opportunities. The case is therefore made for a shift away from the ‘activated sludge approach’ towards an ‘aerobic granular approach’, which would assist in addressing the challenges facing the wastewater treatment industry in Asia and beyond. Key words: aerobic granular sludge, biopolymer recovery, Nereda®, sustainable wastewater treatment

INTRODUCTION Aerobic granular sludge has been extensively researched over the last two decades as a part of the search for more sustainable wastewater treatment solutions. Conventional activated sludge (CAS) systems have key disadvantages such as slow settling flocculent biomass necessitating large clarifiers and low reactor biomass concentrations (typically 3–5 kgMLSS/m3), large treatment system footprints and relatively high system energy usage. It has been shown at the lab, pilot and the full scale that aerobic granular sludge has distinct advantages, when compared to CAS systems, including improved settling characteristics, which in turn allows for higher biomass concentrations and hence more compact treatment systems. A co-ordinated research partnership in the Netherlands led to the development of the Nereda® technology – a full-scale application of aerobic granular sludge. Currently, over 40 full scale Nereda® plants are operational or under design/construction across 5 continents. The operational full-scale plants have met effluent requirements whilst achieving more sustainable wastewater treatment with key advantages outlined below (compared to similarly loaded activated sludge systems): Page 319

988

Water Practice & Technology Vol 12 No 4 doi: 10.2166/wpt.2017.101

• 25–75% reduction in treatment system footprints as a result of higher reactor biomass concentrations and the non-use of secondary settling tanks; • 20–50% energy usage reduction and; • Associated capital and operational cost savings. This paper highlights the different Nereda® design configurations which have been developed to meet requirements at different sites across the world. Furthermore, results from several full-scale treatment plants are presented and the potential to extract a high-value reuse product (alginate) from Nereda® excess/waste sludge is discussed. AEROBIC GRANULAR BIOMASS AND THE NEREDA® TECHNOLOGY Starting with activated sludge, aerobic granular sludge can be formed by applying specific process conditions such as selectively wasting slow settling biomass and retaining faster settling sludge (de Kreuk et al. 2005). Furthermore, favouring slow growing bacteria such as Poly-phosphate Accumulating Organisms (PAOs) has been shown to enhance granulation (de Kreuk & van Loosdrecht 2006). Aerobic granular sludge consists of bio-granules, without carrier material, of sizes typically larger than 0.2 mm. The granular biomass can be used to biologically treat wastewater using similar processes to activated sludge system, however the granular sludge has a distinct advantage of faster settling velocities when compared to activated sludge, which allows for higher reactor biomass concentrations (e.g. 8–15 g/l) (de Kreuk et al. 2007). When aerated, an oxygen gradient forms within aerobic granules whereby the outer layers are aerobic and the inner core is anoxic or anaerobic (de Kreuk et al. 2007). Nitrifiers and heterotrophic bacteria proliferate in the aerobic outer layer of the granules, enabling the degradation of organics (COD removal) and nitrification (conversion of ammonia to nitrite/nitrate) respectively (de Kreuk et al. 2007). A simultaneous nitrification-denitrification process occurs whereby the formed nitrates (from nitrification) are denitrified (conversion of nitrate to nitrogen gas) in the anoxic core of the granules (Pronk et al. 2015). PAOs in the aerobic granules enable enhanced biological phosphorus removal whereby phosphate uptake occurs during aeration and phosphate rich waste sludge is subsequently removed from the system (de Kreuk et al. 2005). Aerobic granular sludge can therefore achieve biological nutrient removal in a single tank without the need for separate anaerobic and anoxic compartments or tanks. Comparatively, activated sludge systems capable of biological nitrogen and phosphorus removal require at least 3 tanks or zones (anaerobic, anoxic anaerobic) and multiple recycles between the zones or tanks (Wentzel et al. 2008). In the early 2000’s, lab-scale research at the Delft University of Technology (TU Delft), showed that aerobic granular sludge could be formed under a variety of conditions and that granular sludge could be used to achieve stable biological COD, phosphorus and nitrogen (de Kreuk et al. 2007). A collaborative public-private partnership was set up involving TU Delft, Royal HaskoningDHV, several Dutch District Water Authorities, STOWA (the Dutch Foundation for Applied Water Research). This partnership led to the development of the Nereda® wastewater treatment system, which is a full scale application of the aerobic granular sludge technology. Following initial pilot-scale research, the first full-scale Nereda® wastewater treatment plant was commissioned in 2006 at a cheese factory in the Netherlands (van der Roest et al. 2011). Subsequently, 18 full-scale Nereda® treatment plants have entered operation. Table 1 provides details of the operational plants as well as the full-scale plants under construction 11 plants) and in the final stages of design (11 plants). Nereda® operates a cyclical process with three cycle components or stages: simultaneous influent fill and effluent withdrawal; aeration/reaction and settling – all of which occur in a single reactor without partitions (Giesen et al. 2013). Granulation can be achieved via an incremental start-up process using activated sludge for seeding or alternatively granular seed sludge from other Nereda® Page 320

Water Practice & Technology Vol 12 No 4 doi: 10.2166/wpt.2017.101

989

Table 1 | List of full scale Nereda® treatment plants in operation, under construction and in the ﬁnal phases of design

Daily average 3

Operational plants

ﬂow (m /day)

Vika, Ede (NL)

Peak ﬂow 3

(m /h)

Person Equivalent (Calculated for p.e.

Greenﬁeld/Retroﬁt

a 54 g. BOD)

Start-up

CAS or SBR/Hybrid

50–250

1,500–5,000

2005

Retroﬁt

Cargill, Rotterdam (NL)

700

10,000–30,000

2006

Retroﬁt

Smilde, Oosterwolde (NL)

500

5,000

2009

Retroﬁt

STP Gansbaai (RSA)

5,000

400

63,000

2009

Greenﬁeld

STP Epe (NL)

8,000

1,500

41,000

2011

Greenﬁeld

STP Garmerwolde (NL)

30,000

4,200

140,000

2013

Greenﬁeld

STP Vroomshoop (NL)

1,500

200

12,000

2013

Greenﬁeld

STP Dinxperlo (NL)

3,100

570

11,000

2013

Greenﬁeld

STP Wemmershoek (RSA)

5,000

468

39,000

2013

Greenﬁeld

STP Frielas, Lisbon (PT)

12,000

1,850

44,000

2015

Retroﬁt

STP Ryki (PL)

5,320

465

38,600

2015

Greenﬁeld

Westfort Meatproducts, IJsselstein (NL)

1,250

330

43,000

2015

Greenﬁeld

STP Clonakilty (IRL)

4,896

622

23,000

2015

Greenﬁeld

STP Carrigtwohill (IRL)

6,750

844

41,000

2015

Greenﬁeld

STP Deodoro, Rio de Janeiro (BR)

Phase I - 64,800 Phase II - 86,400

4,590 6,120

360,000 480,000

2016 2025

Greenﬁeld

STP Kingaroy (AUS)

2,625

450

11,000

2016

Greenﬁeld

STP Simpelveld (NL)

3,668

945

10,000

2016

Greenﬁeld

STP Cork Lower Harbour (IRL)

18,280

1,830

72,000

2017

Greenﬁeld

Plants under construction

STP Highworth (UK)

1,719

197

10,000

2017

Greenﬁeld

STP Jardim Novo, Rio Claro (BR)

24,166

1,806

152,000

2018

Greenﬁeld

STP Hartebeestfontein (RSA)

5,000

208

52,000

2018

Greenﬁeld

STP Alpnach (CH)

14,000

1,872

48,000

2018

Greenﬁeld

STP Zutphen (NL)

10,128

550

237,000

2018

Greenﬁeld

STP Utrecht (NL)

55,000

13,200

343,000

2018

Greenﬁeld

STP Inverurie (UK)

10,871

544

47,204

2018

Retroﬁt

STP Kendal (UK)

26,000

1,749

103,000

2019

Greenﬁeld

STP Österröd, Strömstad (S)

3,730

360

13,000

2019

Greenﬁeld

STP Faro – Olhão (PT)

20,582

1,908

149,000

2019

Greenﬁeld

STP Ringsend, Dublin (IRL)

600,000

50,000

2,670,000

2021

Retroﬁt

Plants under design

STP Morecambe (UK)

17,000

2,088

33,000

2018

Greenﬁeld

STP Tatu, Limeira (BR)

57,024

3,492

322,000

2019

Greenﬁeld

STP Tijuco Preto, Sumaré (BR)

19.900

1.492

110.000

2019

Greenﬁeld

STP Breskens (NL)

3,500

1,000

31,300

2019

Greenﬁeld

STP Jardim São Paulo, Recife (BR)

Phase I – 22,792 Phase II – 67,764

1,871 5,577

109,000 325,000

2019 2025

Greenﬁeld

STP São Lourenço, Recife (BR)

Phase I – 18,842 Phase II - 25,123

1,287 1,715

105,000 140,000

2020 2024

Greenﬁeld

STP Jaboatão, Recife (BR)

Phase I - 109,683 Phase II - 154,483

8,536 12,037

609,000 858,000

2020 2025

Greenﬁeld

STP Kloten (CH)

26,000

2,850

125,000

2023

Retroﬁt

STP Barston (UK)

21,784

1,424

86,000

Tbd

Greenﬁeld

STP Walsall Wood (UK)

7,176

646

29,166

Tbd

Greenﬁeld

STP Radcliffe (UK)

5,324

463

24,722

Tbd

Greenﬁeld

Page 321

990

Water Practice & Technology Vol 12 No 4 doi: 10.2166/wpt.2017.101

plants can be used. The enhanced sludge settleability of aerobic granular sludge is evident from a comparison of typical full scale SVI (sludge volume index) values – for aerobic granular sludge the SVI5 (5 minutes) tends towards the SVI30 (30 minutes), with typical values at operational Nereda® plants around 30–60 ml/g (Giesen et al. 2013), whereas for activated sludge the SVI30 is typically in the range of 110–160 ml/g and the SVI5 is not measured because activated sludge exhibits minimal settling after 5 minutes (Tchobanoglous et al. 2004). Nereda® systems are preceded by conventional pre-treatment consisting of screening, grit removal and, depending on the application, FOG (fats, oils and greases) removal; whilst primary sedimentation is optional. Typical reactor depths range from 5.5 to 9 m, with lower and deeper depths possible; whilst secondary settling tanks and major sludge recycles are not required for the Nereda® system.

RESULTS FROM NEREDA® TREATMENT PLANTS New insights have emerged since implementing the first full-scale Nereda® installations allowing for further innovation, system development and design optimisation. Several system configurations have been developed to suit a variety of scenarios experienced from site to site. Two ‘greenfield’ or parallel extension approaches have been used, whilst two ‘brownfield’ approaches have also been developed – these configurations are detailed in Table 2 below. For ‘brown field’ Nereda applications, it is often possible to reuse existing infrastructure and implement a significant increase in biological treatment capacity against low investments. Examples of such applications in Table 1 are the retrofit of the existing SBR’s of Cargill’s wastewater treatment facility in Rotterdam (The Netherlands) and Irish Water’s Ringsend STP. The Nereda® at Lisbon’s Frielas STP is an example where conventional continuous activated sludge tanks were retrofitted. Detailed treatment performance of various industrial and municipal Nereda plants has been reported before (e.g. Giesen et al. 2013; Pronk et al. 2015) and below operation results of Ryki STP, Prototype Utrecht and hybrid Vroomshoop will be presented. Ryki STP – Poland

In the city of Ryki (Lublin Province, Poland) a new Nereda® wastewater treatment plants entered operation in February 2015. This is the first Nereda® installation located in the eastern part of Central Europe and also the first Nereda® plant that has to contend with low process temperatures during the winter period. The Ryki Nereda® plant is designed to treat 5,320 m3/d (dry weather), corresponding to 38,600 PE. In addition to the challenging winter temperatures, the plant has to treat a range of different incoming sewages (domestic, septic tanks and industrial) and has to handle extended industrial peak load periods. The combined pre-treated influent is fed to an influent buffer tank (500 m3) from where two Nereda® reactors (2,500 m3 each) are separately fed by three submersible pumps (‘1 buffer þ2 reactors configuration’). Biological treated wastewater is discharged to surface water via an existing pond. Table 3 shows the design loads for the plant, Figure 1 the wastewater temperatures experienced at the plant and lastly Table 4 shows the effluent performance compared to the effluent requirements. The Nereda® installation at Ryki has been operational for more than two years and continues to achieve effluent compliance, despite the low winter temperatures and highly variable seasonal loading. Vroomshoop STP – the Netherlands

A hybrid Nereda® conﬁguration was selected for the upgrade of the Vroomshoop STP (the Netherlands) and the new plant entered operation in 2013. The main feature of the hybrid conﬁguration Page 322

12 No 4

Cost-effective capacity and performance enhancement using existing infrastructure Use existing tanks or CAS reactors

Convert existing continuous activated sludge reactor, SBR or any suitable tank

4 Retroﬁt

Frielas STP (Portugal)

Vroomshoop STP (Netherlands) Enhance activated sludge system performance; Optimal use of existing infrastructure

Waste Nereda® sludge to activated sludge system

1 or more Nereda® reactors with excess sludge connection to activated sludge system

3 Hybrid

Wemmershoek STP (South Africa)

Optimised investments (2 versus 3 reactors)

Buffer stores inﬂuent between feeds to reactors

1 buffer þ2 reactors

2 Inﬂuent buffer followed by X reactors

Reference examples

Epe STP (Netherlands)

Advantages

Scalable for application to large (.100 ml/d) and mega (.500 ml/d) treatment plants

At least 1 reactor in feed phase at any given time

Conﬁguration characteristic

3 reactors

Typical Layout

1 Continuous feed, 3 þ reactors

Nereda® Conﬁguration

Table 2 | Nereda® conﬁgurations

‘Brownﬁeld sites’; Limited space or budget but require enhanced capacity and/or performance

‘Brownﬁeld sites’; Extension/ optimisation scenarios, utilising existing infrastructure

‘Greenﬁeld sites’; or extension to existing plants with parallel Nereda® system

Potential Applications

991 Water Practice & Technology Vol 12 No 4 doi: 10.2166/wpt.2017.101

Page 323

Water Practice & Technology Vol 12 No 4 doi: 10.2166/wpt.2017.101

992 Table 3 | Design loads for the Ryki Nereda® plant Design values Parameter

Domestic

Septic tankers

Industrial

Total

Daily dry weather ﬂow (m3/d)

2,400

120

2,800

5,320

Daily wet weather ﬂow (m3/d)

3,418

120

2,800

6,338

COD (kg/d)

1,680

384

2,500

4,564

BOD5 (kg/d)

960

156

1,200

2,316

TSS (kg/d)

1,200

144

400

1,744

Total N (kg/d)

192

112

326

Total P (kg/d)

Figure 1 | Temperatures at the Ryki WWTP.

Table 4 | Efﬂuent performance at the Ryki Nereda® plant Efﬂuent quality (average from April 2015 to February 2016) Parameter

Efﬂuent requirements

Reactor 1

Reactor 2

Pond Outlet

COD (mg/l)

125

BOD5 (mg/l)

5.5

6.3

4.4

TSS (mg/l)

4.5

Total N (mg/l)

5.7

5.5

5.0

Total P (mg/l)

0.9

0.8

(see Figure 2) is that the Nereda® waste sludge is fed into a parallel activated sludge system. The plant is designed with a dry weather hydraulic capacity of 156 m3/h and rain flow of 1,000 m3/h, whilst the design pollution load is 22,600 PE (population equivalents at 150 gTOD/PE). The discharge of the Nereda® waste or excess sludge into the activated sludge system has been found to significantly improve the sludge settleability of the activated sludge. Figure 3 shows how the SVI in the activated sludge system steadily decreased as a result of the addition of the Nereda® waste sludge, indicating improved sludge settleability. Improved settleability in an activated sludge system could allow for an increase in MLSS (mixed liquor suspended solids) concentrations in the activated sludge system and therefore increase the biological treatment capacity and/or; the possibility to allow higher hydraulic loading on the secondary settling tanks since the sludge settling rates are improved. Another potential advantage of this hybrid configuration is an improvement in biological phosphorus removal in the activated sludge system, since Nereda® waste sludge contains higher concentrations of PAOs when compared to activated sludge. Page 324

993

Water Practice & Technology Vol 12 No 4 doi: 10.2166/wpt.2017.101

Figure 2 | Schematic depiction of the Vroomshoop STP.

Figure 3 | Comparison of SVIs of the Nereda® and activated sludge systems at Vroomshoop STP (data from end-user: Waterschap Vechtstromen).

Between June and November 2014, energy usage monitoring at the Vroomshoop STP showed that the Nereda® side of the plant used on average 35% less energy than the activated sludge side. Furthermore, efﬂuent performance monitoring in 2014 showed the compliance of the plant under full loading conditions (see Table 5).

Prototype Nereda® Utrecht (PNU)

In 2013 a project speciﬁc Nereda® prototype (PNU) was installed at the existing order to investigate the potential of utilising Nereda® for the replacement 430,000 PE plant which is aging and utilises the non-optimal AB type activated The prototype consist of a single 1,000 m3 reactor which is designed to treat an

Utrecht STP in of the existing sludge process. average ﬂow of Page 325

Water Practice & Technology Vol 12 No 4 doi: 10.2166/wpt.2017.101

994

Table 5 | 2014 Efﬂuent performance at the Vroomshoop WWTP (data from end-user: Waterschap Vechtstromen)

Parameters

Average Inﬂuent (mg/l)

Average Efﬂuent (mg/l)

Requirement (mg/l)

Regulatory Compliance Criteria

Organics

COD BOD5

720 263

55 4

125 10

Limit (3� per year up to 250) Limit (3� per year up to 20)

Nitrogen

TN TKN NH4-N

– 66 –

NO2/NO3-N

–

7.2 5.2 Summer ¼ 1.4; Winter ¼ 3.0 2.0

10 – Summer ¼ 2 Winter ¼ 4 –

Yearly Average – Average (1 May - 1 Nov.) Average (1 Nov. - 1 May) –

8.9

0.9

PO4-P

–

0.6

–

Moving average of 10 successive samples –

TSS

317

Limit

Phosphorus

Suspended Solids

1,500 m3/day (9,000 PE), however the plant can be fed up to 600 m3/hr for test purposes. After successful demonstration and optimization of the design parameters for the Utrecht STP specific conditions, the PNU is operated by Royal HaskoningDHV as test and training facility. Whereas testing full-scale plant performance beyond the plant design conditions is often not possible because at operational plants effluent quality is a priority and the plant receives influent defined by the incoming sewer system, at the PNU facility it is possible for test purposes to operate well beyond the normal conditions. PNU is also used to validate usability and reliability of instrumentation and equipment design optimizations. Treated wastewater is decanted from Nereda® using a fixed overflow weir, similar to a conventional clarifier. In the design of the first municipal Nereda® plants, it was decided to discharge any particles that might lead to scum with the treated effluent as the obtained water quality fully meet the discharge requirements. To investigate the achievable effluent quality when – like in many clarifiers – scum forming particules are kept in the reactor, baffles were added to the PNU effluent launders in 2015. Figure 4 shows how the effluent suspended solids were reduced to below 10 mgTSS/l. Based on these results the optional use of scum baffles has been introduced in various full-scale designs where stringent requirements apply for suspended solids or total-P.

Figure 4 | Effluent suspended solids performance at the PNU facility with baffles (no primary clarification).

Page 326

995

Water Practice & Technology Vol 12 No 4 doi: 10.2166/wpt.2017.101

FURTHER DEVLOPEMENTS – ALE RECOVERY Research at TU Delft uncovered the ability to extract alginate-like exopolysaccharides (ALE) from aerobic granular sludge (Lin et al. 2010). Alginate is currently produced from seaweed at relatively high costs and is used in a variety of industries as a thickener or gel and as a basis for coatings. Aerobic granular sludge has been found to contain between 20 to 30% of ALE. Extracted ALE could potentially be used in the chemical sector, as a soil enhancer in agriculture or as a brick additive (van der Roest et al. 2015). The recovery of ALE from Nereda® excess sludge (aerobic granular sludge) is a potential re-use opportunity, whereby a waste stream could be converted into a product with a high resale value. Combining ALE extraction with the existing excess sludge treatment processes at wastewater treatment plants could also improve sludge treatment efﬁciency because ALE extraction reduces sludge volumes and the remaining (non-extracted) sludge has a higher digestibility and an improved dewaterability. The National Alginate Research Programme (NAOP) has been set up in the Netherlands to further research and develop this promising sustainable re-use concept. The NAOP is a public-private sector collaborative research initiative with the goal of developing sustainable and commercially viable ALE-extraction from Nereda® excess sludge (van der Roest et al. 2015). The NAOP is similar to the public-private collaborative partnership that successfully developed Nereda®. During the summer of 2017 a pilot study was carried out and based on the results two demo installations will be designed and realized in 2019.

DISCUSSION AND CONCLUSIONS Results from full-scale Nereda® treatment plants over the last decade have shown that Nereda® has numerous advantages when compared to similarly loaded activated sludge systems, including:

• 25–75% reduction in treatment system footprints as a result of higher reactor biomass concentrations and the non-use of secondary settling tanks;

• 20–50% energy usage reduction and; • Associated capital and operational cost savings. Nereda® treatment plants have been shown to achieve similar or improved enhanced biological nutrient (nitrogen and phosphorus) removal when compared to similarly loaded activated sludge systems. Furthermore, the possibility to recover ALE from Nereda® waste sludge has the potential to generate a reuse product with high commercial value. Four main Nereda® configurations have been developed for a wide range wastewater treatment scenarios ranging from ‘green-field’ systems to retrofits at ‘brown-field’ sites. The hybrid configuration (e.g. Vroomshoop STP) whereby Nereda® waste sludge is fed into a parallel activated sludge system has the potential to increase the loading capacity of the activated sludge system through improved sludge settleability. This configuration could therefore be applied advantageously for the extension of existing plants with an activated sludge line. The results achieved at full-scale Nereda® treatment plants show that aerobic granular sludge has clear and significant advantages over CAS systems. Currently sustainability requirements (including cost-effectiveness) are driving technological advancement and innovation. The advantages of Nereda® in comparison to activated sludge systems ultimately translate into more sustainable and cost-effective wastewater treatment. A shift away from the ‘activated sludge approach’ towards an ‘aerobic granular approach’ would assist in addressing the challenges facing the wastewater treatment industry in Asia and beyond. Page 327

996

Water Practice & Technology Vol 12 No 4 doi: 10.2166/wpt.2017.101

REFERENCES de Kreuk, M. K. & van Loosdrecht, M. C. M. 2006 Formation of aerobic granules with domestic sewage. Journal of Environmental Engineering 132(6), 694–697. de Kreuk, M. K., Heijnen, J. J. & van Loosdrecht, M. C. M. 2005 Simultaneous COD, nitrogen and phosphate removal by aerobic granular sludge. Biotechnology and Bioengineering 90(6), 761–769. de Kreuk, M. K., Kishida, N. & van Loosdrecht, M. C. M. 2007 Aerobic granular sludge – State of the art. Water Science Technology 55, 75–81. Giesen, A., de Bruin, L. M. M., Niermans, P. P. & van der Roest, H. F. 2013 Advancements in the application of aerobic granular biomass technology for sustainable treatment of wastewater. Water Practice & Technology 8(1), 47–54. Lin, Y., de Kreuk, M. K., van Loosdrecht, M. C. M. & Adin, A. 2010 Characterization of alginate-like exopolysaccharides isolated from aerobic granular sludge in pilot-plant. Water Research 44, 3355–3364. Pronk, M., de Kreuk, M. K., de Bruin, B., Kamminga, P., Kleerebezem, R. & van Loosdrecht, M. C. M. 2015 Full scale performance of the aerobic granular sludge process for sewage treatment. Water Research 84, 207–217. Tchobanoglous, G., Burton, F. L. & Stensel, H. D. 2004 Wastewater Engineering: Treatment and Reuse. Metcalf & Eddy, Mc Graw Hill, New York. van der Roest, H. F., de Bruin, L. M. M., Gademan, G. & Coelho, F. 2011 Towards sustainable waste water treatment with Dutch Nereda® technology. Water Practice & Technology 6(3). van der Roest, H. F., van Loosdrecht, M. C. M., Langkamp, E. J. & Uijterlinde, C. 2015 Recovery and reuse of alginate from granular Nereda sludge. Water 21 17(2), 48. Wentzel, M. C., Comeau, Y., Ekama, G. A., van Loosdrecht, M. C. M. & Brdjanovic, D. 2008 Chapter 7: phosphorus removal. In: Biological Wastewater Treatment: Principles, Modelling and Design (Henze, M., van Loosdrecht, M. C. M., Ekama, G. A. & Brdjanovic, D., eds). IWA, London.

Page 328

Water Quality Research Journal

ISSN 1201-3080 iwaponline.com/wqrj

Water Quality Research Journal is a a forum for original research dealing with the aquatic environment, and reports on new and significant findings that advance the understanding of the field. The journal’s scope includes: • Impact of current and emerging contaminants on aquatic ecosystems • Aquatic ecology (ecohydrology and ecohydraulics, invasive species, biodiversity, and aquatic species at risk) • Conservation and protection of aquatic environments • Responsible resource development and water quality (mining, forestry, hydropower, oil and gas) • Drinking water, wastewater and stormwater treatment technologies and strategies • Impacts and solutions of diffuse pollution (urban and agricultural run-off) on water quality • Industrial water quality • Used water: Reuse and resource recovery • Groundwater quality (management, remediation, fracking, legacy contaminants) • Assessment of surface and subsurface water quality • Regulations, economics, strategies and policies related to water quality • Social science issues in relation to water quality • Water quality in remote areas and cold climates For more details, visit iwaponline.com/wqrj

Page 332

178

52.3

2017

A novel cloud point extraction method for separation and preconcentration of cadmium and copper from natural waters and their determination by ﬂame atomic absorption spectrometry Cennet Karadaş

ABSTRACT A sensitive and simple cloud point extraction method was developed for simultaneous separation and preconcentration of cadmium and copper prior to their determination by ﬂame atomic absorption spectrometry. 4-phenyl-3-thiosemicarbazide was used as complexing agent and the cadmium and copper complexes were extracted from the aqueous phase by Triton X-114 surfactant.

Cennet Karadaş Department of Chemistry, Art and Science Faculty, Balıkesir University, Balıkesir 10100, Turkey E-mail: karadas@balikesir.edu.tr

The effects of parameters such as sample pH, ligand amount, concentration of surfactant, incubation temperature and time were optimized. For 20 mL of preconcentrated solution, the detection limits (3σ) were 0.20 and 0.49 μg L 1, and the enrichment factors were 20.7 and 19.9 for Cd(II) and Cu(II), respectively. In order to verify the accuracy of the developed method, certiﬁed reference materials (SLRS-5 river water and SPS-SW2 Batch 127 surface water) were analysed. Results obtained were in good agreement with the certiﬁed values. The proposed method was applied to tap water, river water and seawater samples with satisfactory results. Key words

| 4-phenyl-3-thiosemicarbazide, cadmium, cloud point extraction, copper, preconcentration

INTRODUCTION Heavy metals are considered to be one of the main sources

directly affect the liver and nervous system which can lead

of pollution in the environment since they have a signiﬁcant

to death (Baroumand et al. ). Excessive intake of

effect on its ecological quality (Tavallali et al. ). Cad-

copper would lead to accumulation of the metal in liver

mium is one of the most toxic elements among the heavy

cells and haemolytic crisis, jaundice and neurological dis-

metals (Ning et al. ). It causes different damage and

turbances (Wen et al. ). These heavy metals may enter

defects in lungs, kidneys and bones. Cadmium, with its

the food chain, accumulate in plants and animals, and

high half-life time from 10 to 33 years, can accumulate in

may cause damage to human health. For these reasons cad-

liver and kidneys (Ensafi et al. ). Its wide technological

mium and copper determination in water and biological

use as well as its production from burning oil and coal

matrices is a good tool for environmental and toxicological

and incineration of waste causes an extensive anthropogenic

monitoring.

contamination of soil, air and water (Tavallali et al. ).

Several techniques, including ﬂame atomic absorption

Although copper is one of the most essential elements in

spectrometry (FAAS) (Chen & Teo ; Tavallali et al.

the body and plays an important role in many body func-

; Ning et al. ; Baroumand et al. ; Naeemullah

tions, accumulation and an excess amount of it can

et al. ), electrothermal atomic absorption spectrometry

doi: 10.2166/wqrj.2017.004

Page 333

179

C. Karadaş

Water Quality Research Journal

A CPE method for preconcentration of Cd and Cu

52.3

2017

(ETAAS) (Lopez-Garcia et al. ; Falahnejad et al. ),

rich phase as well as its low cost, commercial availability

inductively coupled plasma optical emission spectrometry

and lower toxicity. Some parameters that inﬂuence the

(ICP-OES) (Silva et al. ; Zhao et al. ), spectropho-

extraction efﬁciency such as sample pH, ligand amount,

tometry (Wen et al. ; Yang et al. ), inductively

concentration of surfactant, effect of salt addition and incu-

coupled plasma mass spectrometry (ICP-MS) (Wang et al.

bation temperature and time were investigated and

) and voltammetry (Abbasi et al. ; Ensafi et al.

optimized. The proposed method was then applied to the

) have been used for the determination of Cu(II) and

determination of Cu(II) and Cd(II) in tap water, river

Cd(II) in various types of sample. Among these techniques,

water and seawater samples. The accuracy of the developed

FAAS has been widely used for the determination of heavy

method was veriﬁed by analysing certiﬁed reference

metals because of its low cost, ease of operation, high

materials (SLRS-5 river water and SPS-SW2 Batch 127

sample throughput and good selectivity (Wu et al. ).

surface water).

However, direct determination of heavy metal ions is generally difﬁcult due to various factors, in particular their low concentrations and matrix effects. In order to solve these

EXPERIMENTAL

problems, separation, preconcentration and matrix elimination

usually

required.

Various

preconcentration

Instruments

methods such as liquid–liquid extraction (LLE) (Karadaş & Kara ), solid phase extraction (SPE) (Moghadam

A Perkin Elmer model AAnalyst 200 (Shelton, CT, USA)

Zadeh et al. ; Sheikhshoaie et al. ), co-precipitation

ﬂame atomic absorption spectrometer equipped with deuter-

(Prasad et al. ), cloud point extraction (CPE) (Chen &

ium background correction and appropriate hollow cathode

Teo ; Tavallali et al. ; Ning et al. ; Naeemullah

lamps and an air–acetylene ﬂame (air and acetylene ﬂow

microextraction

rate 10 L min 1 and 2.3 L min 1, respectively) was used

(DLLME) (Wen et al. ; Lopez-Garcia et al. ), disper-

for determination of Cu(II) and Cd(II). The most sensitive

sive liquid–liquid microextraction based on solidiﬁcation of

wavelengths (nm) and lamp currents (mA) used for the

ﬂoating organic drop (DLLME-SFO) (Wu et al. ) and

determination of the analytes were as follows: Cu 324.75

ionic liquid-based single step microextraction (Khan et al.

and 30, and Cd 228.80 and 3, respectively. A Hanna Instru-

) have been used for the separation and preconcentra-

ments model 221 (Cluj-Napoca, Romania) digital pH-meter

tion of Cu(II) and Cd(II) from environmental matrices.

with a combined glass electrode was used for all pH

al.

),

dispersive

liquid–liquid

CPE is an attractive method that reduces the consump-

measurements. A Nuve ST 402 model thermostatic water

tion of and exposure to solvent, disposal costs and

bath (Ankara, Turkey) was used for controlling the tempera-

extraction time (Golbedaghi et al. ). This extraction

ture of the CPE experiments. A Hettich centrifuge model

method is based on the fact that most non-ionic surfactants

Rotaﬁx 32 A (Germany) was used to accelerate the phase

in aqueous solutions form micelles and become turbid when

separation.

heated to the cloud point temperature or in the presence of an electrolyte. Above the cloud point, the micellar solution

Reagents and solutions

separates into a surfactant-rich phase with a small volume and a diluted aqueous phase (Rezende et al. ; Zhao

All the reagents used were of analytical grade, and water

et al. ).

puriﬁed by a reverse osmosis system (AquaTurk Reverse

In this work, a new CPE method was developed for the

Osmosis System, HSC ARITIM, Istanbul, Turkey) was

preconcentration of Cu(II) and Cd(II) prior to FAAS deter-

used to prepare all the solutions. Nitric acid, hydrochloric

mination. The reagent 4-phenyl-3-thiosemicarbazide was

acid, sodium dihydrogen phosphate, sodium acetate, acetic

used as a chelating ligand. Triton X-114 was chosen as the

acid, ammonium acetate, boric acid and ethanol from

non-ionic surfactant for the work because of its low cloud

Sigma-Aldrich (St. Louis, MO, USA), Triton X-114 and

point temperature and the high density of the surfactant-

sodium hydroxide from Fluka (Gillingham, Dorset, UK),

Page 334

180

C. Karadaş

Water Quality Research Journal

A CPE method for preconcentration of Cd and Cu

and 4-phenyl-3-thiosemicarbazide and sodium tetraborate

52.3

2017

Sample preparation

from Aldrich (St. Louis, MO, USA) were used in the experiments. The laboratory glassware used was kept in 10% (v/v)

The proposed method was applied to natural water samples.

nitric acid overnight and rinsed with deionized water before

Tap water sample was collected directly at the laboratory

�1

) of the analytes

(Balıkesir University), river water was collected from

were prepared by dissolving appropriate amounts of

Küçük Bostancı River and seawater was collected from the

Cu(NO3)2.3H2O (Riedel-de haen from Sigma-Aldrich,

Aegean Sea close to the Edremit coast. The river water

St. Louis, MO, USA) and Cd(NO3)2.4H2O (Merck, Darm-

and seawater samples were ﬁltered through a cellulose

stadt, Germany) in 1% HNO3. Working standard solutions

membrane of 0.45 μm pore size, acidiﬁed to pH 2 using

were prepared daily from the stock standard solutions by

nitric acid and stored in pre-cleaned polyethylene bottles.

appropriate dilution with deionized water. Sodium dihydro-

The pH of the water samples (20 mL) was adjusted to pH

gen phosphate/phosphoric acid buffers for pH 2–3, sodium

8.0 using a few drops of 10% (w/v) sodium hydroxide sol-

acetate/acetic acid buffers for pH 4–5, ammonium acetate/

ution and then maintained using borate buffer solution.

acetic acid buffers for pH 6–7 and sodium tetraborate/

The proposed method was applied to the samples.

use. Stock standard solutions (1,000 mg L

boric acid buffers for pH 8–10 were used to adjust the pH of the solutions. The solution of the chelating ligand (2 × 10�2 M)

was prepared by dissolving appropriate

RESULTS AND DISCUSSION

amounts of the reagent in ethanol. The non-ionic surfactant, 1.0% (w/v) Triton X-114 was prepared by dissolving 1.0 g of

Optimization of the experimental variables

Triton X-114 in 100 mL of deionized water. The certiﬁed reference materials SPS-SW2 level 2 Batch 127 surface

The analytical parameters that affect the performance of

water (Spectrapure Standards AS, Oslo, Norway) and

CPE, such as sample pH, ligand amount, concentration of

SRLS-5 river water (National Research Council, Ottawa,

surfactant, incubation temperature and time and effect

Canada) were used for verifying the accuracy of the pro-

of salt addition were investigated and optimized. Standard

posed method.

solution (10 mL) containing 100 μg L�1 of Cu(II) and 50 μg L�1 Cd(II) was used in these experiments. A univari-

CPE procedure

ate optimization procedure was undertaken, i.e., varying one parameter at a time, keeping the others constant. All

An aliquot of the sample solution containing Cu(II) and Cd(II)

the experiments were carried out in triplicate.

ions was transferred to a 50 mL polyethylene centrifuge tube.

The extraction of metal ions by surfactant micelles gen-

Borate buffer (1.0 mL, pH 8.0), 0.5 mL of 2 × 10�2 mol L�1

erally occurs after the formation of a complex with sufﬁcient

ligand and 1.0 mL of 1.0% (w/v) Triton X-114 solution were

hydrophobicity (Silva et al. ). Since the pH is one of the

added. The mixture was diluted to 20 mL with deionized

main parameters for chelation reactions, the effect of pH on

water. The mixture was manually shaken for 5–6 sec and left

CPE procedure was investigated over the pH range of 2.0–

to stand for 10 min in a thermostated water bath set at 50 C.

10.0. The effect of the sample solution pH on the recovery

The resulting solution was centrifuged at 4,000 rpm for 10 min

of Cu(II) and Cd(II) is shown in Figure 1. Quantitative

to obtain phase separation. It was then cooled at þ4 C in a

recoveries were obtained at pH ranges 6.0–10.0 for Cu(II)

refrigerator for 10 min to increase the viscosity of the surfac-

and 8.0–10.0 for Cd(II). Therefore, all further experiments

tant-rich phase. The aqueous phase was carefully removed

were performed at pH 8.0 for the simultaneous extraction

with a Pasteur pipette and, to decrease its viscosity, the surfac-

of the Cu(II) and Cd(II).

tant-rich phase was diluted to 1.0 mL with 1.0 mol L�1 HNO3.

The concentration of the ligand has a direct effect on the

The ﬁnal solution was aspirated directly into the FAAS instru-

formation of metal–ligand complex as well as its extraction.

ment. The CPE procedure described above was also applied to

The effect of ligand amount on the recovery of Cu(II) and

the blank and calibration standards.

Cd(II) ions was examined over the ligand amount range Page 335

181

C. Karadaş

Water Quality Research Journal

A CPE method for preconcentration of Cd and Cu

Figure 3 Figure 1

mol L

52.3

2017

Effect of Triton X-114 concentration on the recovery of Cu(II) and Cd(II). Con ditions: sample volume 10 mL, sample pH 8.0, 0.5 mL of 2 × 10 2 mol L 1 ligand, incubation temperature 70 C, incubation time 30 min. W

Effect of sample pH on the recovery of Cu(II) and Cd(II). Conditions: sample volume 10 mL, 0.5 mL of 2 × 10

ligand solution, 0.5 mL of 1.0% (w/v)

Triton X-114, incubation temperature 70 C, incubation time 30 min. W

quantitative recoveries of the analytes were obtained at con0.0–4.2 mg. For this purpose, 0.5 mL of different concen 2

trations of ligand between 0.0 and 5 × 10

centrations between 0.025% and 0.1%. Above 0.1% (w/v),

was

the recovery of the analytes slowly decreased. Therefore, a

added to 10 mL of standard solution containing 100 μg L 1

concentration of 0.05% (w/v) Triton X-114 was selected

of Cu(II) and 50 μg L

mol L

Cd(II) at pH 8.0. The results are

for subsequent experiments.

given in Figure 2. The recovery of Cu(II) and Cd(II) was quan-

The effect of ionic strength on the extraction efﬁciency

titative in the ligand amount ranges 0.04–4.2 mg and 1.7–

of analytes was examined using NaCl at concentrations

4.2 mg, respectively. Therefore, 1.7 mg of ligand (0.5 mL of

from 0.0 to 0.5 mol L 1. The results are given in Figure 4.

2 × 10

mol L

) was used for further experiments.

According to the results obtained, salt addition has no sig-

The concentration of surfactant used in CPE is a critical

niﬁcant effect on the extraction efﬁciency of Cu(II) and

factor. In order to raise the efﬁciency of the extraction pro-

Cd(II). Therefore, all the extraction experiments were

cedure, the concentration of Triton X-114 in the sample

carried out without the addition of salt.

solution was optimized evaluating concentrations between

The shortest incubation time and the lowest possible

0.005% and 0.375% (w/v). As shown in Figure 3,

equilibration temperature are very important for completion

Figure 4 Figure 2

Effect of ligand amount on the recovery of Cu(II) and Cd(II). Conditions: sample volume 10 mL, sample pH 8.0, 0.5 mL of 1.0% (w/v) Triton X-114, incubation temperature 70 C, incubation time 30 min. W

Page 336

Effect of NaCl concentration on the recovery of Cu(II) and Cd(II). Conditions: 2

sample volume 10 mL, sample pH 8.0, 0.5 mL of 2 × 10

mol L

ligand,

0.5 mL of 1.0% (w/v) Triton X-114, incubation temperature 70 C, incubation time 30 min. W

182

C. Karadaş

Water Quality Research Journal

A CPE method for preconcentration of Cd and Cu

52.3

2017

of the reaction and efﬁcient separation of the phases (Naeemullah et al. ). The dependence of extraction efﬁciency upon incubation temperature and time was studied over the range of 30–70 C and 5–30 min, respectively. The results are W

given in Figures 5 and 6. Quantitative recoveries were obtained between 40 and 70 C. It was observed that an W

incubation time of 10 min is enough for quantitative extraction of the analytes. Therefore, an incubation temperature of 50 C and an incubation time of 10 min were selected as W

optimum. Effect of the volume of sample solutions

Figure 6

Effect of incubation time on the recovery of Cu(II) and Cd(II). Conditions: sample volume 10 mL, sample pH 8.0, 0.5 mL of 2 × 10

�2

mol L

�1

ligand,

0.5 mL of 1.0% (w/v) Triton X-114, incubation temperature 50 C. W

In order to obtain a high preconcentration factor, the sample volume is a key consideration. Under optimized conditions, the effect of sample volume on the extraction of Cu(II) and Cd(II) was examined. To evaluate the effect of sample volume on the recovery of Cu(II) and Cd(II), 10, 20 and 30 mL of sample solutions containing 200 ng Cu(II) and 100 ng Cd(II) were used. The results obtained are given in Table 1. Quantitative recoveries were obtained for the sample volumes studied. The preconcentration factors obtained were 10, 20 and 30 for a 10, 20 and 30 mL sample solution, respectively. Interference studies In order to demonstrate the selectivity of the developed method for determination of the analytes at trace levels, the

effects of common coexisting ions on the extraction of Cd(II) and Cu(II) were studied. In these experiments, 10 mL of solution containing 100 μg L�1 Cu(II) and 50 μg L�1 Cd(II) ions were added to interfering ions at a concentration of 10 mg L�1 and treated according to the recommended extraction procedure. The results are given in Table 2. The presence of the tested ions does not affect the recovery of Cu(II) and Cd(II) ions. In addition, the effect of the major matrix ions present in waters (Mg2þ, Ca2þ, Naþ, Kþ, SO2� 4 , � Cl�, CO2� 3 and NO3 ) on the recoveries of the analytes was

investigated using a synthetic seawater sample containing 1,270 mg L�1 Mg2þ, 400 mg L�1 Ca2þ, 10,800 mg L�1 Naþ, �1 CO2� 390 mg L�1 Kþ, 5,100 mg L�1 SO2� 4 , 600 mg L 3 ,

16,600 mg L�1 Cl�, and 620 mg L�1 NO� 3 . A 10 mL aliquot of the synthetic seawater solution spiked with 100 μg L�1 of Cu(II) and 50 μg L�1 Cd(II) was analysed according to the recommended procedure. The results are given in Table 2. It can be seen that the seawater matrix ions have no signiﬁcant effect on the recovery of Cu(II) and Cd(II) ions.

Table 1

The effect of sample volume on the recovery of Cu(II) and Cd(II) Analyte �1

concentration (μg L

)

Recovery (%)

Sample volume

Figure 5

Effect of incubation temperature on the recovery of Cu(II) and Cd(II). Con� � ditions: sample volume 10 mL, sample pH 8.0, 0.5 mL of 2 × 10 2 mol L 1 ligand, 0.5 mL of 1.0% (w/v) Triton X-114, incubation time 30 min.

(mL)

Cu(II)

Cd(II)

Cu(II)

Cd(II)

97.2 ± 3.5

102.6 ± 3.3

5.0

96.7 ± 1.7

97.1 ± 5.4

6.7

3.3

98.3 ± 5.4

100.5 ± 4.5

Page 337

183

Table 2

C. Karadaş

Water Quality Research Journal

A CPE method for preconcentration of Cd and Cu

52.3

2017

method. Table 3 summarizes the analytical characteristics of

Effect of interfering ions on the recovery of Cu(II) and Cd(II)

the optimized method, including linear working range, corre-

Recovery (%)

lation coefﬁcient (R), limit of detection (LOD), relative

Interfering iona

Added as

Cu(II)

Pb2þ

Pb(NO3)2

Cr3þ

Cr(NO3)3. 9H2O

Mn2þ

Mn(NO3)2. 4H2O

100.0 ± 3.3 98.0 ± 0.7

2þ

Cd(II)

standard deviation (RSD), equation of calibration curves, pre-

99.7 ± 5.3

105.8 ± 2.6

concentration factor (PF) and enrichment factor (EF). The

97.7 ± 0.8

96.8 ± 3.6

limits of detection, deﬁned as LOD ¼ 3 Sb/m and where Sb

Zn(NO3)2. 6H2O

97.9 ± 2.5

Fe3þ

Fe(NO3)3. 9H2O

100.0 ± 3.6 104.6 ± 6.2

centration, were found to be 0.49 μg L�1 for Cu(II) and

Ba2þ

Ba(NO3)2

96.9 ± 3.1

102.4 ±2.5

0.20 μg L�1 for Cd(II). The precision of the method was

Al3þ

Al(NO3)3. 9H2O

97.2 ± 2.0

101.0 ± 1.3

evaluated for a solution containing 5.0 μg L�1 of Cu(II) and

Sr(NO3)2

100.3 ± 0.4 98.5 ± 0.5

2.5 μg L�1 of Cd(II), and their RSD values were found to be

Ni2þ

Ni(NO3)2. 6H2O

98.3 ± 5.5

Synthetic seawater

KNO3, NaCl, MgSO4. 7H2O, CaCO3

101.8 ± 0.9 101.7 ± 2.8

3.1% for Cu(II) and 2.4% for Cd(II) (n ¼ 10). The preconcen-

2þ

Concentration of the interfering ions ¼ 10 mg L

�1

98.1 ± 4.7

is standard deviation of ten replicate blank signals and m is

96.6 ± 1.5

slope of the calibration curve obtained with 20-fold precon-

tration factor for the proposed method is calculated by the ratio of the sample volume (20 mL) to the ﬁnal volume (1 mL). Enrichment factors were calculated as the ratio of

the slopes of calibration graphs obtained using the preconcenTherefore, the proposed method can be applied to samples

tration method and direct aspiration. A comparison of the characteristic data obtained using

containing high amounts of salt.

the method developed with other reported preconcentration methods for Cu(II) and Cd(II) determination is summarized Analytical performance of the method and comparison

in Table 4. The limits of detection of the analytes are lower

with other methods

than or comparable to those obtained with other separation/ preconcentration methods.

The analytical performance of the proposed method was evaluated under the optimized conditions. Calibration graphs were constructed using 20 mL of the standard solutions buf-

Accuracy of the method

fered at pH 8.0 and containing the concentration ranges of 2.5–100 μg L�1 Cu(II) and 1.25–50 μg L�1 Cd(II). The stan-

In order to evaluate the accuracy of the proposed method,

dard solutions were processed by the optimized CPE

two certiﬁed reference materials, SLRS-5 river water and

Table 3

Analytical characteristics of the proposed CPE method

Parameters

Cu(II)

Sample volume (mL)

Calibration equation (with preconcentration) Working range (μg L

�1

)

A ¼ 4.05 × 10

Cd(II)

20 �3

2.5–100

�3

C þ 3.32 × 10

A ¼ 1.19 × 10�2 C þ 5.17 × 10�3

1.25–50

Correlation coefﬁcient (R)

0.9999

0.9998

Detection limit (LOD) (μg L�1)

0.49

0.20

RSD (%) (5.0 μg L�1 Cu(II) and 2.5 μg L�1 Cd(II))

3.1

Calibration equation (direct aspiration) Working range (μg L

�1

)

A ¼ 2.03 × 10 50–2,000

2.4 �4

�3

C þ 2.92 × 10

A ¼ 5.76 × 10�4 C þ 6.89 × 10�3

25–1,000

Preconcentration factor

Enrichment factor

19.9

20.7

Page 338

52.3

2017

SPS-SW2 Batch 127 surface water were analysed using the developed method. The analytical results are given in Table 5. The results obtained by the proposed method were in good agreement with the certified values and the precision was between 2.2 and 3.7% RSD. As seen in Table 5, the values obtained for the certified reference samples were between 95.0 and 108.0% of certified values. These results show the accuracy and repeatability of the developed method for the determination of Cu(II) and Cd(II) in water samples. In order to test the significance of differences between certified and obtained values, the Student’s t-test was applied to results of the proposed method and certified values. For 2 degrees of freedom at the 95% confidence level, critical t value is 4.30. As can be seen, USAE-SFODME: ultrasound-assisted emulsification solidified floating organic drop microextraction; SDSPE: suspension dispersive solid phase extraction.

CPE: cloud point extraction; IL-SSME: ionic liquid-based single step microextraction procedure; SPE: solid phase extraction; MDSPE: magnetic-dispersive solid phase extraction; DLLME: dispersive liquid–liquid microextraction;

Meng et al. () 8.5 40 20 1.5 (Cu) Co, Ni, Cu FAAS SDSPE

Farajzadeh et al. () 5 5 10 3.0 Cu FAAS DLLME

Khayatian & Hassanpoor () 28 6.7 13.4 4.1 (Cu) Fe, Cu FAAS USAE-SFODME

Mehdinia et al. ()

Mirabi et al. () 0

20 50

100 100

50 0.5

3.71 Cd

Cd FAAS

FAAS SPE

SPE

Ezoddin et al. () 5 50 25 0.16 (Cd) Cd, Pb FAAS MDSPE

Kara () 2 4.2 8.4 3.2 (Cu), 0.39 (Cd) Cd, Co, Cu, Mn, Ni, Pb, Zn FI-FAAS Micelle-mediated extraction

Khan et al. () Data not available 10 50 0.35 Cd FAAS (microinjection system) IL-SSME

Silva et al. ()

Chen & Teo () At least 30

65 15 10 1.2 (Cu), 1.0 (Cd)

0.27 (Cu), 0.099 (Cd) 64.3 (Cu), 57.7 (Cd) 50 Cu, Cd, Pb, Zn

Cu, Zn, Cd, Ni CPE

CPE

Method

Table 4

Water Quality Research Journal

A CPE method for preconcentration of Cd and Cu

ICP-OES

This work 30 20 20 0.49 (Cu), 0.20 (Cd) Cu, Cd FAAS

analysis (min) (mL)

Sample volume Preconcentration

factor 1)

Detection limit (LOD)

(μg L Analyte

Comparative data from recent studies on preconcentration–separation of copper and cadmium

Technique

Time of

Reference

C. Karadaş

FAAS

184

t values are smaller than the critical value of t at 95% conﬁdence level for all certiﬁed samples, indicating that there is no evidence of systematic error in the proposed method. Application of the method to real samples The proposed method was applied to tap water, river water and seawater samples. The applicability of the method was evaluated by spiking of these water samples with 10 μg L 1 of Cu(II) and 5 μg L 1 of Cd(II). The results are given in Table 6. The average percentage recovery values of Cu(II) and Cd(II) were 96 and 108 for tap water samples, 98 and 104 for river water samples, and 101 and 96 for seawater samples,

respectively.

These

results

demonstrate

the

reliability and accuracy of the method for the determination of Cu(II) and Cd(II) in natural water samples.

CONCLUSIONS The proposed method for simultaneous separation and preconcentration of Cu(II) and Cd(II) by CPE, using Triton X-114 as surfactant and 4-phenyl-3-thiosemicarbazide as complexing agent, has shown to be an efﬁcient, simple, accurate, precise, inexpensive, green and safe method. Triton X-114 is of relatively low cost and low toxicity. The surfactant-rich phase can easily be introduced into the nebulizer of the spectrometer after dilution with 1.0 mol L 1 HNO3 and determined directly by FAAS. The method requires nearly 30 min of sample preparation time per Page 339

185

Table 5

C. Karadaş

Water Quality Research Journal

A CPE method for preconcentration of Cd and Cu

52.3

2017

Results for the certiﬁed reference materials �1

Certiﬁed reference material

Analyte

Certiﬁed value (μg L

SPS-SW2 Batch 127 surface water

Cu(II) Cd(II)

100 ± 1 2.50 ± 0.02

SLRS-5 River water

Cu(II) Cd(II)

17.4 ± 1.3 0.0060 ± 0.0014

)

�1

Found valuea (μg L

Recovery (%)

RSDb (%)

95.0 ± 2.6 2.60 ± 0.13

95 104

2.2 5.0

3.3 1.3

18.8 ± 0.7 <LOD

108 –

3.7 –

3.5 –

)

Mean value ± standard deviation based on three replicate determinations.

RSD: relative standard deviation. pffiffiffiffi jμ � xj N , where t is statistical value (for 2 degrees of freedom, the critical value of t at the 95% confidence level is 4.30), s is the standard deviation, N is number of independent s determinations, x is the experimental mean value, and μ is the certified value. c

t¼

Table 6

Analytical results of water samples and the recovery of spiked analytes Founda

Added

Recovery

RSD

(%)

7.2 ± 0.2 16.8 ± 0.7 <LOD 5.4 ± 0.1

– 96 – 108

2.8 4.2 – 1.9

– 10 – 5

2.1 ± 0.2 11.9 ± 0.2 <LOD 5.2 ± 0.2

– 98 – 104

9.5 1.7 – 3.8

– 10 – 5

5.2 ± 0.2 15.3 ± 0.6 <LOD 4.8 ± 0.3

– 101 – 96

3.8 3.9 – 6.3

�1

Sample

Analyte

(μg L

Tap water

Cu(II)

– 10 – 5

Cd(II) River water

Cu(II) Cd(II)

Seawater

Cu(II) Cd(II)

)

�1

(μg L

)

Mean value ± standard deviation based on three replicate determinations.

sample. However eight samples can be prepared for analysis simultaneously. The developed method can be considered to be an alternative to other more sensitive analytical techniques, such as GFAAS, ICP-OES and ICP-MS of which the latter two instruments, especially ICP-MS, are not found in many laboratories due to their price.

REFERENCES Abbasi, S., Bahiraei, A. & Abbasai, F.  A highly sensitive method for simultaneous determination of ultra trace levels of copper and cadmium in food and water samples with luminol as a chelating agent by adsorptive stripping voltammetry. Food Chemistry 129 (3), 1274–1280. Baroumand, N., Akbari, A., Shirani, M. & Shokri, Z.  Homogeneous liquid–liquid microextraction via ﬂotation assistance with thiol group chelating reagents for rapid and

Page 340

efficient determination of cadmium(II) and copper(II) ions in water samples. Water, Air, & Soil Pollution 226 (2254), 1–8. Chen, J. & Teo, K. C.  Determination of cadmium, copper, lead and zinc in water samples by flame atomic absorption spectrometry after cloud point extraction. Analytica Chimica Acta 450 (1–2), 215–222. Ensafi, A., Allafchian, A. R. & Rezaei, B.  Polytetrafluorethylene membrane-based liquid three-phase micro extraction combined with in situ differential pulse anodic stripping voltammetry for the determination of cadmium ions using Au-nanoparticles sol-gel modified PtWire. Journal of the Brazilian Chemical Society 26 (7), 1482– 1490. Ezoddin, M., Majidi, B., Abdi, K. & Lamei, N.  Magnetic graphene-dispersive solid-phase extraction for preconcentration and determination of lead and cadmium in dairy products and water samples. Bulletin Environmental Contamination and Toxicology 95 (6), 830–835. Falahnejad, M., Zavvar Mousavi, H., Shirkhanloo, H. & Rashidi, A. M.  Preconcentration and separation of ultra-trace amounts of lead using ultrasound-assisted cloud point-micro solid phase extraction based on amine functionalized silica aerogel nanoadsorbent. Microchemical Journal 125, 236–241. Farajzadeh, M. A., Bahram, M., Mehr, B. G. & Jonsson, J. A.  Optimization of dispersive liquid–liquid microextraction of copper (II) by atomic absorption spectrometry as its oxinate chelate: application to determination of copper in different water samples. Talanta 75 (3), 832–840. Golbedaghi, R., Jafari, S., Yaftian, M. R., Azadbakht, R., Salehzadeh, S. & Jaleh, B.  Determination of cadmium(II) ion by atomic absorption spectrometry after cloud point extraction. Journal of the Iraian Chemical Society 9 (3), 251–256. Kara, D.  Preconcentration and determination of trace metals by flow injection micelle-mediated extraction using flame atomic absorption spectrometry. Talanta 2009, 79 (2), 429–435. Karadaş, C. & Kara, D.  Determination of copper(II) in natural waters by extraction using N-o-vanillidine-2-amino-pcresol and flame atomic absorption spectrometry. Instrumentation Science & Technology 42 (5), 548–561. Khan, S., Kazi, T. G. & Soylak, M.  A green and efficient insyringe ionic liquid-based single step microextraction

186

C. Karadaş

A CPE method for preconcentration of Cd and Cu

procedure for preconcentration and determination of cadmium in water samples. Journal of Industrial and Engineering Chemistry 27, 149–152. Khayatian, G. & Hassanpoor, S.  Development of ultrasoundassisted emulsification solidified floating organic drop microextraction for determination of trace amounts of iron and copper in water, food and rock samples. Journal of the Iranian Chemical Society 10 (1), 113–121. Lopez-Garcia, I., Vicente-Martinez, Y. & Hernandez-Cordoba, M.  Determination of cadmium and lead in edible oils by electrothermal atomic absorption spectrometry after reverse dispersive liquid–liquid microextraction. Talanta 124, 106–110. Mehdinia, A., Shegefti, S. & Shemirani, F.  A novel nanomagnetic task specific ionic liquid as a selective sorbent for the trace determination of cadmium in water and fruit samples. Talanta 144, 1266–1272. Meng, L., Chen, C. & Yang, Y.  Suspension dispersive solid phase extraction for preconcentration and determination of cobalt, copper, and nickel in environmental water by flame atomic absorption spectrometry. Analytical Letters 48 (3), 453–463. Mirabi, A., Dalirandeh, Z. & Rad, A. S.  Preparation of modified magnetic nanoparticles as a sorbent for the preconcentration and determination of cadmium ions in food and environmental water samples prior to flame atomic absorption spectrometry. Journal of Magnetism and Magnetic Materials 381, 138–144. Moghadam Zadeh, H. R., Ahmadvand, P., Behbahani, A., Amini, M. M. & Sayar, O.  Dithizone-modified graphene oxide nano-sheet as a sorbent for pre-concentration and determination of cadmium and lead ions in food. Food Additives & Contaminants: Part A 32 (11), 1851–1857. Naeemullah, Kazi, T. G. & Tuzen, M.  Development of novel simultaneous single step and multistep cloud point extraction method for silver, cadmium and nickel in water samples. Journal of Industrial and Engineering Chemistry 35, 93–98. Ning, J., Jiao, Y., Zhao, J., Meng, L. & Yang, Y.  Cloud point extraction–flame atomic absorption spectrometry method for preconcentration and determination of trace cadmium in water samples. Water Science and Technology 70 (4), 605–611. Prasad, K., Gopikrishna, P., Kala, R., Prasada Rao, T. & Naidu, G. R. K.  Solid phase extraction vis-a-vis coprecipitation preconcentration of cadmium and lead from soils onto 5,7dibromoquinoline-8-ol embedded benzophenone and determination by FAAS. Talanta 69 (4), 938–945.

Water Quality Research Journal

52.3

2017

Rezende, H. C., Nascentes, C. C. & Coelho, N. M. M.  Cloud point extraction for determination of cadmium in soft drinks by thermospray flame furnace atomic absorption spectrometry. Microchemical Journal 97 (2), 118–121. Sheikhshoaie, M., Shamspur, T., Mohammadi, S. Z. & Saheb, V.  Extraction of zinc, copper, and lead ions with a zeolite loaded by a multidentate schiff base ligand followed by flame atomic absorption spectrometric analysis. Separation Science and Technology 50 (17), 2680–2687. Silva, E. L., dos Santos Roldan, P. & Gine, M. F.  Simultaneous preconcentration of copper, zinc, cadmium, and nickel in water samples by cloud point extraction using 4-(2-pyridylazo)-resorcinol and their determination by inductively coupled plasma optic emission spectrometry. Journal of Hazardous Materials 171 (1–3), 1133–1138. Tavallali, H., Boustani, F., Yazdandoust, M., Aalaei, M. & Tabandeh, M.  Cloud point extraction–atomic absorption spectrometry for pre-concentration and determination of cadmium in cigarette samples. Environmental Monitoring and Assessment 185 (5), 4273–4279. Wang, X., Chen, J., Zhou, Y., Liu, X., Yao, H. & Ahmad, F.  Dispersive liquid–liquid microextraction and micro-solid phase extraction for the rapid determination of metals in food and environmental waters. Analytical Letters 48 (11), 1787–1801. Wen, X., Yang, Q., Yan, Z. & Deng, Q.  Determination of cadmium and copper in water and food samples by dispersive liquid–liquid microextraction combined with UV–vis spectrophotometry. Microchemical Journal 97 (2), 249–254. Wu, C. X., Wu, Q. H., Wang, C. & Wang, Z.  A novel method for the determination of trace copper in cereals by dispersive liquid–liquid microextraction based on solidification of floating organic drop coupled with flame atomic absorption spectrometry. Chinese Chemical Letters 22 (4), 473–476. Yang, S., Fang, X., Duan, L., Yang, S., Lei, Z. & Wen, X.  Comparison of ultrasound-assisted cloud point extraction and ultrasound-assisted dispersive liquid liquid microextraction for copper coupled with spectrophotometric determination. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 148 (5), 72–77. Zhao, L., Zhong, S., Fang, K., Qian, Z. & Chen, J.  Determination of cadmium(II), cobalt(II), nickel(II), lead(II), zinc(II), and copper(II) in water samples using dual-cloud point extraction and inductively coupled plasma emission spectrometry. Journal of Hazardous Materials 239–240, 206–212.

First received 30 January 2017; accepted in revised form 23 May 2017. Available online 8 July 2017

Page 341

52.2

2017

A computational ﬂuid dynamics analysis of placing UV reactors in series Patrick C. Young and Yuri A. Lawryshyn

ABSTRACT Ultraviolet (UV) light water treatment reactors are commonly used in both wastewater and drinking water disinfection. UV technology can effectively inactivate a large number of pathogens at low UV doses, however adenovirus requires a substantially higher dose than most pathogens of interest. In order to meet adenovirus inactivation requirements, UV reactors are often placed in series and the total inactivation is calculated as the sum of the reactors’ individual UV doses. In this paper, it is

Patrick C. Young (corresponding author) Yuri A. Lawryshyn Department of Chemical Engineering, University of Toronto, 200 College Street, Toronto, ON M5S 3E5, Canada E-mail: pat.young@utoronto.ca

shown that this simple summation treatment of UV dose may be acceptable. A parameter called the reactor additivity factor is introduced to properly characterize the interaction between UV reactors in series. Three types of UV reactors are modelled using computational ﬂuid dynamics, and their RAFs are computed. The validity of reactor additivity in practice in wastewater and drinking water systems is discussed. Key words

| computational ﬂuid dynamics, UV disinfection, UV reactor modelling, UV reactors in series

INTRODUCTION Ultraviolet (UV) light disinfection is a proven technology for

to achieve 4-log ‘virus’ inactivation credit. Since few vali-

wastewater and drinking water disinfection. In order to

dated UV disinfection reactors exist that are able to deliver

meet UV dosing requirements, several UV reactors, or

such a high UV dose, the National Water Research Institute

banks, are often placed in series. However, few studies

(NWRI) guidelines allow for UV drinking water and waste-

have addressed how putting UV reactors in series impacts

water reactors to be installed in series and the UV dose

their validation protocols. Health Canada and the United

delivered is calculated as the cumulative dose of the individ-

States Environmental Protection Agency recommend an

ual reactors (National Water Research Institute ). The

inactivation/removal of at least 4-log for enteric viruses,

reactors in series must be shown to be hydraulically inde-

i.e. adenovirus, for groundwater and surface water sources

pendent or the reactors in series must be validated in such

(United States Environmental Protection Agency ;

a way that the installed system is identical to the validated

Health Canada ). There is some regulatory concern

one. Therefore, the guidelines allow for a drinking water

that UV disinfection alone is not enough to effectively disin-

system to meet the requirement of 4-log adenovirus inacti-

fect adenovirus in drinking water in the absence of chlorine

vation by installing multiple UV disinfection units in

and ﬁltration processes.

series. The UVDGM speciﬁes that ‘good mixing should be

Although adenovirus is extremely resistant to UV disin-

conﬁrmed’ when placing UV reactors in series. However,

fection, it is not immune. The Ultraviolet Disinfection

the underlying assumption that UV doses are additive is

Guidance Manual (UVDGM) by Schmelling et al. ()

inconsistent in the literature and has not been thoroughly

speciﬁes a reactor equivalent dose (RED) of 186 mJ/cm

investigated.

doi: 10.2166/wqrj.2017.023

Page 343

P. C. Young & Y. A. Lawryshyn

Water Quality Research Journal

A CFD analysis of placing UV reactors in series

LITERATURE REVIEW

52.2

2017

than 1 from the perspective of the target organism. The author’s corollary is that if two identical reactors, each of

To date, there exist a few papers that consider the impact of

which can deliver a RED of 93 mJ/cm2 with MS2, are put

placing UV reactors in series on overall UV system perform-

in series, the resulting system will necessarily achieve a

ance. Tang et al. () investigated the interaction between

RED of at least 186 mJ/cm2 based on adenovirus.

multiple UV banks in series for an open channel conﬁgur-

It has been well established that computational ﬂuid

ation using MS2 as a test organism. The delivered UV

dynamics (CFD) analysis coupled with ﬂuence rate modelling

doses from multiple reactors were shown to be not exactly

is a reliable method for evaluating UV reactor performance.

additive, with the overall dose being greater than additive.

Furthermore, it is very difﬁcult to perform experiments to

However, no explanation for the result was given. Ferran

gain signiﬁcant technical insights into the effects of reactor

& Scheible () found that for two low pressure high

additivity. Speciﬁcally, there is no simple way to measure

output (LPHO) UV reactors in series in an open channel,

the dose received for each particle path per reactor for two

the RED was twice the RED of a single LPHO reactor.

reactors in series, in an effort to determine the overall dose

Ducoste & Alpert () numerically evaluated the RED

correlation between two reactors. Therefore, in this paper,

of UV reactors in series for both open channels and

the topic of reactor additivity will be explored from a numeri-

closed conduits. It was shown that additivity may only be

cal and CFD perspective. Two simple reactor conﬁgurations

assumed provided that there is sufﬁcient mixing between

will be considered to investigate both positive and negative

reactor banks. They also commented that the UV response

dose correlation (and will henceforth be referred to as the

kinetics of the target microorganism will impact the degree

‘correlated systems’). Additionally, two real-world reactor

of additivity. However, their results only looked at what

conﬁgurations will be investigated to demonstrate the

happens for the two cases of perfect mixing and no

phenomenon of reactor additivity in practice.

mixing between reactors.

As mentioned previously, the use of CFD for evaluating

Recently, Lawryshyn & Hofmann () looked at UV

UV reactor performance is an established practice. Unluturk

reactor additivity from a completely theoretical perspective.

et al. () coupled CFD velocity ﬁelds with ﬂuence rate

A reactor additivity factor (RAF) was introduced to quantify

models to compute the UV dose delivered to apple cider,

the degree of additivity. RAF was deﬁned as the RED of two

and found reasonable agreement between simulated and

reactors in series divided by twice the RED of a single reac-

experimental values. Lawryshyn & Cairns () showed

tor. Thus, an RAF of 1 means exact additivity, whereas an

that CFD UV reactor models can be carefully used in

RAF >1 means better than exact additivity and RAF <1

place of the bioassay testing of prototype reactors in order

means less than exact additivity – i.e. an RAF greater than

to speed up development and reduce associated prototyping

1 implies that the RED of the system is greater than the

costs. Sozzi & Taghipour () further demonstrated that

sum of the RED of the original reactors, and vice-versa for

CFD ﬂow ﬁelds were in agreement with particle image velo-

an RAF less than one. For two reactors in series with perfect

cimetry measurements and that the resulting microbial

mixing between the reactors, it was shown that the RAF will

inactivation was consistent with experimentally obtained

necessarily be one. Furthermore, it was shown that for sys-

biodosimetry results. Other studies also support the use of

tems with a negative correlation among the dose paths

CFD to predict reactor performance (Chiu et al. ; Lyn

between the two reactors, the RAF will be greater than or

et al. ; Pareek et al. ).

equal to one, whereas for a positive correlation, the RAF will be less than or equal to one. Additionally, it was shown that in the extreme case of perfect positive corre-

METHODOLOGY

lation, which is considered to be a worst case scenario, if the test organism is two or more times more sensitive than

In this section, theory is introduced that will allow us

the target organism, then the RAF will necessarily be greater

to discuss the performance of UV reactors placed in

Page 344

P. C. Young & Y. A. Lawryshyn

Water Quality Research Journal

A CFD analysis of placing UV reactors in series

52.2

2017

series. Additionally, the reactor models used will be

possible to construct such a reactor for demonstration pur-

introduced.

poses. Figures 1 and 2 depict two very simple cross ﬂow reactor conﬁgurations where one would expect positive

Theory

and negative dose path correlations, respectively. In the positively correlated case, the lamps in both R1 and R2

It has been well established that UV reactors deliver a distri-

are positioned to be near the lower wall so that particles

bution of doses to the microbes traversing the reactor

receiving a high dose in R1 will also receive a high dose in

(Wright & Lawryshyn ). The entry point as well as the

R2. In the negatively correlated case, the lamp in R1 is posi-

path that a microbe takes as it ﬂows through the reactor in

tioned near the bottom wall and the lamp in R2 is positioned

relation to the lamps results in its obtained UV dose.

near the top wall so that particles that receive a high dose in

Consider the two correlated systems depicted in Figures 1

R1 will receive a low dose in R2.

and 2. Suppose that there exist two identical UV reactors,

The performance of a UV reactor is characterized by a

R1 and R2, in series and that the dose distribution through

distribution of doses that microbes are expected to receive

each is identical. A microbe that receives a certain dose

as they traverse through the reactor. Therefore, it is generally

from R1 will not necessarily receive the same dose from

not useful to characterize a reactor’s inactivation potential

R2 due to mixing between reactors. The amount of mixing

by an average dose. UV reactor performance is often charac-

between R1 and R2 may be characterized by a correlation

terized by the RED or ‘reactor equivalent dose’. For ﬁrst

(represented mathematically by ρ) of the doses delivered

order microbial inactivation kinetics, RED is given by:

by R1 and subsequently R2 for a given microbial path. For the theoretical case of absolutely no mixing between reactors, one would expect perfect positive correlation in dose paths between R1 and R2, i.e. a given microbe would receive

�ð ∞ � 1 f(D)e�kD dD RED ¼ � ln k 0

(1)

exactly the same dose from each of the two reactors. For

where f (D) is the probability density function, i.e. the dose

positively correlated reactors, a microbe that receives a

distribution for a given reactor, and k is the microbe speciﬁc

high dose from R1 would be expected to receive a high

inactivation constant (see Wright & Lawryshyn () or

dose from R2. Similarly, particles that receive a low dose

Lawryshyn & Hofmann () for more details).

from R1 would be expected to receive a low dose from R2.

In the negatively correlated reactor case, although the

Perfect mixing between reactors implies zero correlation

reactors are effectively reversed in their alignment, the

between dose paths. With zero correlation, it is not possible

water layers on each side of the lamps of R1 and R2 are

to estimate the dose that a microbe receives from R2 given

the same. On an individual basis, it is expected that the

the dose that it receives from R1. There also exists an oppor-

RED for R1 and R2 will be very similar. However, across

tunity for the dose paths through R1 and R2 to be negatively

the entire system, the RED for the two reactors in series

correlated. In this case, a microbe that receives a high dose

will be different. Thus, the two reactor conﬁgurations, i.e.

in R1 is expected to receive a low dose in R2 and vice versa.

the positively and negatively correlated conﬁgurations, will

While this may be difﬁcult to achieve in practice, it is

have different additivities. Let us deﬁne a dimensionless

Figure 1

Positively correlated reactors in series.

Figure 2

Negatively correlated reactors in series.

Page 345

P. C. Young & Y. A. Lawryshyn

RED12 2RED1

52.2

2017

(k → ∞) the RAF will also approach 1 except for the case

additivity factor (RAF) given by:

RAF ¼

Water Quality Research Journal

A CFD analysis of placing UV reactors in series

of ρ ¼ –1, in which case the RAF will necessarily be greater (2ðaÞ)

or,

than 1. The effects of correlation and the inactivation constant on the RAF are summarized in Table 1. Numerical model

RAF ¼

RED12 RED1 þ RED2

(2ðbÞ)

CFD work was carried out using Workbench 14.5 (ANSYS ). All CFD models used the velocity-inlet boundary con-

where RED12 is the RED calculated across the entire system,

dition on the inlet, pressure-outlet on the outlet, and the no-

RED1 is the RED of the ﬁrst reactor, and RED2 is the RED

slip boundary condition (wall) was used on all wall and

of the second reactor. Equation (2(a)) is an idealization that

lamp surfaces. A pressure-based solver was used in Fluent,

assumes that the dose distributions for R1 and R2 are iden-

and turbulence was modelled with the realizable k�ε

tical. In practice, R1 will inﬂuence R2 and the dose

model. Particle tracking was done by Fluent and dose calcu-

distributions are not the same. Thus, Equation (2(b))

lation code was developed and executed using MATLAB

should generally be used. It should be noted that in the

R2008a. The process is summarized in Table 2 below.

case of different test versus target organisms, the RED of

The radial model (Blatchley ) was used for all ﬂu-

the system, i.e. RED12, should be calculated based on the

ence

target organism, whereas RED1 and RED2 should be

computational cost. Despite the fact that it is not as accurate

based on the test organism as is the case with a bioassay vali-

as other ﬂuence rate models, the objective of this study was

dation. As will be shown in the results, we will also present a

to show the relative trends in dose, and this is easily achiev-

RAF calculated with RED12 based on adenovirus and RED1

able with the radial model. It is deﬁned by:

rate

calculations

for

its

simplicity

and

low

and RED2 based on MS2. Although the proof has been omitted, it is intuitive that the negatively correlated reactor conﬁguration will have a

I(r) ¼

τ s ηL IL T α(r�R) 2πr

(3)

higher RED and superior inactivation performance to that of the positively correlated reactor conﬁguration. In the posi-

where I is the UV light intensity, r is the radial distance from

tively correlated reactor conﬁguration, there is extreme short

the centre of the lamp, R is the radius of the lamp sleeve, τs is

circuiting along the top of the reactor. Microbes ﬂowing along the bottom of the reactor will receive high UV doses from both R1 and R2, whereas particles ﬂowing along the top

Table 1

Trends in ρ and k on RAF

will receive very low doses through each of the two reactors. In the negatively correlated reactor configuration, the microbes flowing along the bottom of the reactor will first receive high doses from R1 and then low doses from R2

k!0

ρ ¼ �1

ρ<0

ρ¼0

ρ>0

AF ! 1

AF ¼ 1

AF ! 1

AF > 1

k!∞

AF ! 1

AF ¼ 1

AF ! 1

with the reverse happening along the bottom of the reactor. As corollary, it is clear that the RAF will be greater for the

Table 2

Software used in workﬂow

case of negative correlation than that of positive correlation. In fact, for any given value of k, the RAF is greater than 1 for

Operation

Vendor

Software

negative correlation, less than 1 for positive correlation, and

Geometry

ANSYS

DesignModeler

equal to 1 for zero correlation.

Meshing

ANSYS

Meshing

Physics and particle tracking

ANSYS

Fluent

Fluence rate modelling and dose calculations

Mathworks

MATLAB

The interactions with the inactivation constant are not as simple, however. In the case of a UV resistant organism (k → 0) the RAF will approach 1. For a sensitive organism Page 346

P. C. Young & Y. A. Lawryshyn

Water Quality Research Journal

A CFD analysis of placing UV reactors in series

52.2

2017

the lamp sleeve transmittance, ηL is the lamp efﬁciency, IL is

massless particles were injected uniformly across the inlet

the power intensity per unit length of the lamp, T is the UV

surface to achieve convergence. The number of particles

transmittance (UVT) of the water, and α is a non-dimensio-

used for the positively correlated and negatively correlated

nalizing coefﬁcient dependent on the UVT unit of

systems were 4158 and 3846 respectively. The number of

measurement. For a discrete dose distribution, RED is calcu-

particles was roughly doubled until the resultant RED

lated as:

between iterations differed by less than 1%.

N 1 1X e�kDi RED ¼ � ln k N i¼1

(4)

Wastewater reactor system The wastewater reactor was modelled to mimic a UV system

where N is the total number of particles, and Di is the dose

used in practice. Lamps were arranged in virtual (using a

of the ith particle (Wright & Lawryshyn ). In this paper,

symmetry plane) 4 × 4 banks, and included lamp supports.

D10 (dose required for a 1-log reduction of microbes) will

The ﬂow direction was parallel to the lamps. The lamps

primarily be used in place of k. D10 is deﬁned as follows:

were 1.5 m in length, had a diameter of 0.10 m, and formed a square grid with a spacing of 0.20 m between

ln (10) k

D10 ¼

(5)

lamp centres. Lamps were spaced such that there was a 0.05 m water layer between the sides, top and bottom, as can be seen in Figure 3 (dashed lines represent symmetry

Correlated systems The two correlated systems were modelled as described in Table 3. The reactors were modelled in three dimensions, with the lamps and lamp sleeves extending from wall to wall. A length of 10 hydraulic diameters (DH) was given for flow to develop before the lamps, and 5 DH was given after the lamps and before the outlet. A mesh was created of 7.0 × 105 hex elements. The minimum orthogonal quality was greater than 0.50 and grid convergence was achieved. Turbulent flow was achieved with a Reynolds number (Re) greater than 105 based on DH. A sufficient number of Table 3

Positively and negatively correlated reactor geometry Correlation

Feature

Positive

Negative

Reactor width

1.5 m

Lamp diameter

0.10 m

Inlet length

10 DH

Inter lamp length

10 DH

Lamp 1 top water layer

0.15 m

Lamp 1 bottom water layer

0.05 m

Lamp 2 top water layer

0.15 m

0.05 m

Lamp 2 bottom water layer

0.05 m

0.15 m

Figure 3

Wastewater reactor cross-section.

Page 347

P. C. Young & Y. A. Lawryshyn

Water Quality Research Journal

A CFD analysis of placing UV reactors in series

52.2

2017

planes). The spacing between banks was left as a variable between 0.5 and 6 m, the effect of which will be discussed later. The meshing algorithm used adaptive reﬁnement in order to give detail to the more complicated geometry features such as lamp supports (see Figure 4). Grid convergence was achieved with 3.9 × 106 tetra cells, and the minimum orthogonal quality was 0.18. Symmetry planes were used to model the open channel’s free surface and to create the virtual 4 × 4 lamp arrangement from a 2 × 4 lamp arrangement. The ﬂow was turbulent such that the Reynolds number based on hydraulic diameter was Figure 4

Wastewater reactor mesh.

greater than 105. Convergence was achieved by injecting 9900 massless particles using the same criteria as the correlated systems. Drinking water reactor A drinking water reactor was modelled after a generic residential system with two units placed in series. Each unit had an annular configuration, with the lamp placed in the centre and parallel to flow. The lamp length was 0.4 m, the sleeve radius was 0.01 m and the wall radius was 0.04 m, as can be seen in Figure 5. Meshing was done using adaptive refinement on proximity and curvature, and inflation was used to give more detail near the walls and lamps; as shown in Figure 6. Grid convergence was

Figure 5

Drinking water reactor cross-section.

Figure 6

Cross section of drinking water reactor mesh.

Page 348

achieved with 5.3 × 105 mixed tetra and hex cells, and the

P. C. Young & Y. A. Lawryshyn

Water Quality Research Journal

A CFD analysis of placing UV reactors in series

minimum orthogonal quality was 0.14. The ﬂow was turbu4

52.2

2017

RED for one reactor (or bank), followed by the individual

lent with a Reynolds number greater than 1.7 × 10 based on

RED of the subsequent reactor, and then the RED of the

the reactor’s diameter. Convergence was achieved by inject-

total system and using Equation (2(b)). Additionally, the cor-

ing 2066 massless particles using the same criteria as the

relation between the two dose distributions was calculated.

correlated systems.

In this section, the results obtained are compared to the theoretical results presented by Lawryshyn & Hofmann () with the intent of verifying them through numerical

RESULTS AND DISCUSSION

experiments.

CFD models were created for the correlated systems, the wastewater reactor, and the drinking water reactor. In all cases, the RAF was calculated by computing the individual

Correlated systems As discussed, two systems were modelled with conﬁgurations such that it was expected that one model would

Table 4

exhibit a positive correlation of dose paths between reac-

Calculated RED and RAF values of the correlated systems

tors, and the other model, negative. The predicted trends in correlation were observed, and the resultant REDs and

D10

RED1

RED2

RED12

RAF

Reactor

(mJ/

conﬁguration

cm2)

Positively correlated

1.00 10.00 20.00 30.00 40.00

12.15 20.00 25.45 29.63 33.02

11.85 19.31 24.51 28.52 31.76

22.87 34.22 42.50 49.11 54.71

0.94 0.86 0.84 0.84 0.85

0.54

1.00 10.00 20.00 30.00 40.00

11.94 20.00 25.41 29.46 32.69

11.54 19.46 24.95 29.10 32.39

26.57 47.46 60.81 69.74 76.13

1.13 1.21 1.20 1.19 1.17

–0.33

Negatively correlated

Figure 7

RAFs are summarized in Table 4. RED1 and RED2 refer to the REDs of the individual reactors 1 and 2 respectively, and RED12 refers to the total RED of the system. Lamp power was adjusted in each system such that the ﬁrst reactor in series would have a RED of 20 mJ/cm2 for a D10 of 10 mJ/cm2. It is important to note that in each case the dose histogram for the individual reactors is practically identical and it is difﬁcult to distinguish between the superimposed histograms, as can be seen in Figure 7. Thus, one would expect the numerical results to match the theory of

Correlated reactor normalized dose histograms.

Page 349

P. C. Young & Y. A. Lawryshyn

A CFD analysis of placing UV reactors in series

Lawryshyn & Hofmann (). With the exception of

Water Quality Research Journal

52.2

2017

Wastewater reactor system

where the D10 is 1 mJ/cm (due to numerical instabilities), the RAF tends towards 1 as the D10 increases, as is

An open channel reactor was modelled with a virtual 4× 4 par-

expected. For the positively and negatively correlated UV

allel ﬂow lamp conﬁguration. Two banks of lamps were placed

systems tested with MS2 (D10 ¼ 20 mJ/cm ), the RAF

in series as was discussed previously, and the distance

when targeting adenovirus (D10 ¼ 46.5 mJ/cm2) was calcu-

between the two banks was adjusted to be between 0.5 and

lated to be 1.16 and 1.57 respectively. This result is

6 m. Lamp power was adjusted such that the ﬁrst bank

consistent with the theory presented by Lawryshyn &

in series would have a RED of 20 mJ/cm2 for a D10 of

Hofmann ().

10 mJ/cm2. It was observed that the correlation between dose paths was positive and decreased almost to zero as the inter-bank length increased. The results of the RED and

Table 5

RAF analysis are summarized in Table 5. The dose distri-

Calculated RED and RAF values of the wastewater reactors

butions for reactor banks are virtually identical, as can be

Inter-lamp

D10

RED1

RED2

RED12

RAF

length (m)

(mJ/cm2)

0.50

1.00 10.00 20.00 30.00 40.00

9.54 20.00 23.09 24.60 25.51

9.19 19.69 22.83 24.38 25.32

17.58 34.79 42.25 46.11 48.50

0.92 0.87 0.91 0.94 0.95

0.39

1.00 10.00 20.00 30.00 40.00

9.77 20.00 23.48 25.20 26.26

9.52 19.56 22.99 24.69 25.72

17.59 35.94 43.87 48.03 50.57

0.91 0.91 0.94 0.96 0.97

0.17

1.00 10.00 20.00 30.00 40.00

9.80 20.00 23.11 24.62 25.54

9.64 19.84 22.88 24.38 25.29

18.11 37.75 44.80 48.26 50.33

0.92 0.94 0.97 0.98 0.99

0.07

3.00

6.00

Figure 8

Wastewater reactor normalized dose histograms.

Page 350

seen in Figure 8. Particle tracks coloured by velocity magnitude can be seen in Figure 9. The RAF tends towards 1 as the D10 increases. This convergence is more rapid as ρ approaches 0. When the reactor is validated for 1-log MS2, the RAFs on the 0.5, 3, and 6 m reactors are 1.08, 1.11, and 1.12 respectively, when adenovirus is assumed to be the target. Drinking water reactor system Two single-lamp drinking water reactors were placed in series, and the effect was analysed. Lamp power was adjusted such that the ﬁrst reactor in series would have a RED of 20 mJ/cm2 for a D10 of 10 mJ/cm2. The results of the RED and RAF computation can be seen below in

P. C. Young & Y. A. Lawryshyn

Figure 9

Water Quality Research Journal

A CFD analysis of placing UV reactors in series

52.2

2017

Particle tracks in the wastewater reactor.

Table 6. The resultant dose distributions for the two reactors are unimodal and positively skewed, and can be seen in Figure 10. Particle tracks of the microbes traversing the reactor are shown in Figure 11. As would be expected for a near zero correlation, the RAF very rapidly approaches 1 as the D10 increases. For the drinking water UV system tested with MS2, the RAF, when targeting adenovirus was calculated to be 1.07.

CONCLUSIONS Three fundamentally different UV reactor conďŹ gurations were analysed: the correlated systems, a wastewater reactor, Figure 10

Table 6

Drinking water reactor normalized dose histograms.

Calculated RED and RAF values of the drinking water reactors

and a drinking water reactor. For the correlated systems, it D10 (mJ/

RED1 (mJ/

RED2 (mJ/

RED12 (mJ/

RAF (mJ/

cm2)

1.00

14.57

10.86

25.53

1.00

0.04

10.00

20.00

17.71

37.28

0.99

20.00

21.34

19.98

41.07

0.99

30.00

21.95

21.00

42.78

1.00

40.00

22.29

21.57

43.75

1.00

was shown through the numerical experiments that a strong positive correlation leads to a RAF less than one and that a strong negative correlation leads to a RAF greater than one. For both positively and negatively correlated systems, the RAF was shown to converge to unity as the D10 increases. With the open-channel wastewater reactor it was shown that the correlation between dose paths is Page 351

P. C. Young & Y. A. Lawryshyn

Figure 11

A CFD analysis of placing UV reactors in series

Particle tracks in the drinking water reactor.

positive, and signiﬁcant with respect to the RAF. The correlation decreased as the second bank was placed further downstream. The implication of this is that the proximity between banks should be considered when placing reactors in series and assuming additivity. Reactor banks should be placed as far apart as reasonably possible. A residential scale drinking water reactor was also modelled, and the correlation between dose paths was found to be near zero. The very low correlation justiﬁes the additivity assumption for this reactor geometry. Finally, it was shown for all systems that additivity could be achieved when sizing was done based on MS2 validation and targeting adenovirus.

ACKNOWLEDGEMENTS The authors are very grateful for the generous support of Ontario Centres of Excellence, Trojan Technologies, and the University of Toronto.

REFERENCES ANSYS  FLUENT Theory Guide 14.0. ANSYS Inc., Canonsburg, Pennsylvania.

Page 352

Water Quality Research Journal

52.2

2017

Blatchley, E. R.  Numerical modelling of UV intensity: application to collimated-beam reactors and continuous-flow systems. Water Res. 31 (9), 2205–2218. Chiu, K., Lyn, D. A., Savoye, P. & Blatchley, E. R.  Integrated UV disinfection model based on particle tracking. J. Environ. Eng. 125 (1), 7–16. Ducoste, J. J. & Alpert, S.  Assessing the UV dose delivered from two UV reactors in series: can you always assume doubling the UV dose from individual reactor validations? In: Proceedings of 2nd North American Conference on Ozone, Ultraviolet & Advanced Oxidation Technologies, Toronto, Ontario, Canada. Ferran, B. & Scheible, O.  Biodosimetry of a full-scale UV disinfection system to achieve regulatory approval for wastewater reuse. In: Proceedings of the Water Environment Federation. World Conference on Ozone and Ultraviolet Technologies, Los Angeles, California, August 27-29, 2007, pp. 367–385. Health Canada  Guidelines for Canadian Drinking Water Quality. Summary 1, Table 1. Federal-Provincial-Territorial Committee on Drinking Water of the Federal-ProvincialTerritorial Committee on Health and the Environment August 2012, Health Canada. Lawryshyn, Y. A. & Cairns, B.  UV Disinfection of water: the need for UV reactor validation. Water Sci. Technol. Water Supply 3 (4), 293–300. Lawryshyn, Y. A. & Hofmann, R.  Theoretical evaluation of UV reactors in series. J. Environ. Eng 141 (10), 04015023-1– 04015023-15. Lyn, D. A., Chiu, K. & Blatchley, E. R.  Numerical modeling of flow and disinfection in UV disinfection channels. J. Environ. Eng 125 (1), 17–26. National Water Research Institute  Ultraviolet Disinfection Guidelines for Drinking Water and Water Reuse, 3rd edn. National Water Research Institute, Fountain Valley, California. Pareek, V. K., Cox, S. J., Brungs, M. P., Young, B. & Adesina, A. A.  Computational fluid dynamic (CFD) simulation of a pilot-scale annular bubble column photocatalytic reactor. Chem. Eng. Sci. 58 (3–6), 859–865. Schmelling, D., Cotton, C. & Mackey, E.  Ultraviolet Disinfection Guidance Manual for the Final Long Term 2 Enhanced Surface Water Treatment Rule. United States Environmental Protection Agency, USA. Sozzi, D. A. & Taghipour, F.  UV Reactor performance modeling by Eulerian and Lagrangian methods. Environ. Sci. Technol. 40 (5), 1609–1615. Tang, C., Kuo, J. & Huitric, S.  UV Systems for reclaimed water disinfection from equipment validation to operation. In: Proceedings of the Water Environment Federation. WEFTEC, Dallas, Texas, October 21–25, 2006, pp. 2930–2943. United States Environmental Protection Agency  National Primary Drinking Water Regulations; Giardia Lamblia, Viruses, and Legionella, Maximum Contaminant Levels, and Turbidity and Heterotrophic Bacteria (Surface Water

P. C. Young & Y. A. Lawryshyn

A CFD analysis of placing UV reactors in series

Treatment Rule). Final Rule, 43 FR 27486, 54(124). US EPA, Washington, DC, USA. Unluturk, S. K., Arastoopour, H. & Koutchma, T.  Modeling of UV dose distribution in a thin-ﬁlm UV reactor for processing of apple cider. J. Food Eng 65 (1), 125–136.

Water Quality Research Journal

52.2

2017

Wright, H. B. & Lawryshyn, Y. A.  An Assessment of the bioassay concept for UV reactor validation. In: Proceedings of the Water Environment Federation, 2000:(2). Disinfection 2000: Disinfection of Wastes in the New Millenium, New Orleans, Louisiana, March 15–18, 2000, pp. 378–400.

First received 11 June 2013; accepted in revised form 4 March 2017. Available online 21 March 2017

Page 353

Water Science & Technology

ISSN 0273-1223 iwaponline.com/wst

Water Science and Technology covers all aspects of the science and technology of wastewater treatment and water quality management worldwide. The journal’s scope encompasses five broad areas: • Wastewater treatment and transportation processes for stormwater and domestic, industrial and municipal effluents • Sources of pollution including hazardous wastes and source control • Effects and impacts of pollution on rivers, lakes, groundwater and marine waters • Water reuse and aquatic environmental restoration • Policy, strategy, control and management aspects of • water quality For more details, visit iwaponline.com/wst

Page 356

313

75.2

2017

Nitrite inhibition and limitation – the effect of nitrite spiking on anammox bioﬁlm, suspended and granular biomass Markus Raudkivi, Ivar Zekker, Ergo Rikmann, Priit Vabamäe, Kristel Kroon and Taavo Tenno

ABSTRACT Anaerobic ammonium oxidation (anammox) has been studied extensively while no widely accepted optimum values for nitrite (both a substance and inhibitor) has been determined. In the current �1 paper, nitrite spiking (abruptly increasing nitrite concentration in reactor over 20 mg NO� ) effect 2 -NL

on anammox process was studied on three systems: a moving bed bioﬁlm reactor (MBBR), a sequencing batch reactor (SBR) and an upﬂow anaerobic sludge blanket (UASB). The inhibition thresholds and concentrations causing 50% of biomass activity decrease (IC50) were determined in batch tests. The results showed spiked biomass to be less susceptible to nitrite inhibition. Although

Markus Raudkivi (corresponding author) Ivar Zekker Ergo Rikmann Priit Vabamäe Kristel Kroon Taavo Tenno Institute of Chemistry, University of Tartu, 14a Ravila St, 50411 Tartu, Estonia E-mail: markus.raudkivi@ut.ee

the values of inhibition threshold and IC50 concentrations were similar for non-spiked biomass (81 �1 and 98 mg NO� , respectively, for SBR), nitrite spiking increased IC50 considerably (83 and 2 -NL �1 240 mg NO� , respectively, for UASB). As the highest total nitrogen removal rate was also 2 -NL

measured at the aforementioned thresholds, there is basis to suggest stronger limiting effect of nitrite on anammox process than previously reported. The quantitative polymerase chain reaction analysis showed similar number of anammox 16S rRNA copies in all reactors, with the lowest quantity in SBR and the highest in MBBR (3.98 × 108 and 1.04 × 109 copies g�1 TSS, respectively). Key words

| anammox, deammoniﬁcation, nitrite spiking, reject water

INTRODUCTION Anaerobic ammonium oxidation (anammox) (Mulder et al. ) is a wastewater treatment process carried out by chemoautotrophic bacteria from the order Planctomycetales (Strous et al. a). The process uses NO� 2 as the electron into dinitrogen gas acceptor and oxidizes dissolved NHþ 4 in anoxic conditions (Van Hulle et al. ). It is a costeffective method that serves as an alternative to traditional nitrification-denitrification process (Van Hulle et al. ) due to significant saving on aeration energy, organic carbon consumption and biomass treatment costs (Lotti ). The most critical point for sustaining a stable anammox process with a high total nitrogen removal rate (TNRR) is maintaining the proper substrate NO� 2 concentrations. Nitrite has been recognized as an inhibiting compound for anammox organisms (Strous et al. b)

while the specific mechanism of nitrite inhibition is still unknown (Lotti et al. ). Other inhibiting/limiting parameters for the anammox process are free ammonia, dissolved oxygen (DO) and organic carbon/nitrogen (COD/N) ratio (Ali & Okabe ). Different inhibiting NO� 2 concentrations (IC50 ran�1 ging from 80 to 430 mg NO� 2 �NL ) have been observed for anammox process (Lotti et al. ) depending on the biomass type (Table 1). The highest IC50 values (concentration causing 50% inhibition) have been reported for the long-term tests either with gel biofilm carriers or with granular biomass (Kimura et al. ; Lotti et al. ). Lower IC50 values have been reported for suspended (flocculated) anammox (Strous et al. b; Bettazzi et al. ). In some cases, the reported high nitrite concentrations may instead be the

doi: 10.2166/wst.2016.456

Page 357

314

Table 1

M. Raudkivi et al.

Water Science & Technology

Nitrite inhibition/limitation for MBBR, SBR and UASB systems

75.2

IC50 values caused by nitrite concentrations and conditions for batch tests according to various authors �1

NO� 2 -N IC50 (mg L

350

)

�1

NO� 2 -N optimum (mg L

120

)

Temperature ( C)

Biomass type

Reference

Suspended

Dapena-Mora et al. ()

120

∼50

20–43

Suspended

Strous et al. (b)

Suspended

Bettazzi et al. ()

400

100a

Granular

Lotti et al. ()

Gel bioﬁlm carriers

Kimura et al. ()

>430 a

2017

100

Lower inhibiting nitrite concentrations than 100 mg

�1 NO� 2 -NL

were not determined.

result of biomass activity loss, not the main cause of it (Lotti et al. ). Balancing between optimal substrate concentrations and nitrite toxicity range is a challenge for the development of anammox technology and the wide range of observed nitrite inhibition levels can make it even more difﬁcult to design or operate anammox-based systems. Developing biomass with high tolerance to nitrite could make wider use and application of a more robust and cheaper N-removal technology possible. Our previous studies have suggested that tolerance to higher nitrite concentrations could be established by applying high nitrite concentration spikes in the reactor (Zekker et al. a). The tolerance to nitrite could also be linked to the anammox species present in the biomass, as Candidatus Brocadia anammoxidans is reportedly less tolerant to nitrite (IC 50 7 mmol nitrite) than both Candidatus Brocadia sinica (<16 mmol nitrite) and Candidatus Kuenia stuttgartiensis (13 or 25 mmol nitrite) (Oshiki et al. ). Though, as not all species of anammox bacteria have been thoroughly characterised and not all anammox studies have conducted microbial sequencing, this current article focuses more on the links between biomass type and nitrite spiking than microbiological differences. This study aimed to elucidate the reasons behind the wide variety of published results about the effect of nitrite inhibition on different types of anammox biomass. Moreover, the link between nitrite spiking (abruptly increasing the nitrite concentrations inside the reactor over 20 mg �1 NO� 2 -NL ) and biomass reaction to short-term exposure to extremely high nitrite concentrations was studied in batch tests for three different types of biomass (bioﬁlm, suspended, granular). Both the limiting and inhibitory effect of nitrite was researched in order to determine whether the biomass can adapt to high nitrite concentrations. Microbial research via the quantitative polymerase chain reaction (qPCR) method (with 16S rRNA primers Amx694F and Amx960R) was carried out in order the describe the most Page 358

common species of anammox bacteria present in used biomasses.

MATERIALS AND METHODS Continuous reactors setups, operation and inoculums Biomasses from three different laboratory-scale reactors – a moving bed biofilm reactor (MBBR (volume 20 L)), a sequencing batch reactor (SBR (10 L)) and an upflow anaerobic sludge blanket (UASB (2 L)) were used in this study. The specific details for all three biomasses during their continuous reactor operation period are presented in Table 2. All reactors were fed with reject water (composition in Zekker et al. ()) coming from the anaerobic tank of Tallinn wastewater treatment plant (Tallinn WWTP). The influent was fed using a peristaltic pump (Seko, Italy). The movement of biomass in all the reactors was ensured by mechanical stirring at approximately 100–200 rounds per minute (rpm) and additionally by coarse-bubble aeration. MBBR was an 20 L anammox reactor, operated under anoxic conditions (DO concentration <0.2 mg L�1) at 25.8 (±1.3) C. Around 10,000 ring-shaped carrier elements made of polyethylene (Bioflow 9) were used for microorganisms’ attachment material. The carriers occupied about 50% of the liquid volume of the reactor with total specific surface of 800 m2 m�3 (carriers’ interior protected specific surface of 500 m2 m�3). Hydraulic retention time (HRT) of 18–48 h was applied. Influent TN (total nitrogen) concentrations of up to W

Table 2

Speciﬁcations of the biomass used in batch tests

Reactor type

Biomass type

Inoculum’s origin

TNLR (g N m

MBBR

Bioﬁlm

Tallinn WWP

240–440

SBR

Suspended

Hannover WWP

20–50

UASB

Granular

Rotterdam WWP

400–600

�3

�1

)

315

M. Raudkivi et al.

Water Science & Technology

Nitrite inhibition/limitation for MBBR, SBR and UASB systems

� 900 mg N L�1 (NHþ were 4 -N=NO2 -N ratioabout 1=1:3) maintained by feeding the reactor with a mixture of reject water and NaNO2 solution. The biomass wet weight at the time of the batch tests was about 6 mg per carrier (Zekker et al. ). SBR was a 10 L deammonification reactor with intermittent aeration, inoculated with the biomass from Hannover anammox pilot plant (Germany) anammox reactor (Li et al. ). The plexiglass reactor was equipped with a water jacket and connected with a thermostat (Assistant 3180, Germany) operated at 26.0 (±0.5) C, DO was measured and controlled by DO controller (Elke Sensor, Estonia). Relatively short HRT of 15 h was applied. SBR was fed with a flow rate of ∼0.5 L h�1. The SBR was fed spikily fewer than 5% of the cycle time maintaining a high (50%) volumetric exchange ratio. SBR ran in cycles of 30 min aerobic (with DO up to 3 mg L�1) and 30 min anoxic phases. UASB was an 2 L anoxic anammox reactor inoculated with granular biomass from full-scale anammox UASB (Rotterdam, The Netherlands (van der Star et al. )). The plexiglass reactor was equipped with a water jacket and thermostated at 34 (±1) C. This reactor system was operated at higher temperatures than the other two in order to keep the biomass in similar conditions as in the Rotterdam WWTP (original temperature between 30 and 40 C (van der Star et al. )). The average HRT of the system was 22 h and the reactor was fed with the effluent of a 1.5 L nitritation reactor with TN concentrations up to 800 mg N L�1 � (NHþ 4 �N=NO2 �N ratio ∼ 1=1:3). In order to maintain a stable upflow rate for the granular sludge and help granules grow in size (maximum measured granule size >1 mm), upflow velocity of 4 L h�1 was used (Zekker et al. a). Firstly, continuous reactor operation was carried out to cultivate anammox biomass and to achieve sufficient TNRR. In order to cultivate higher tolerance to nitrite, the MBBR and UASB reactors were spiked with nitrite by abruptly raising the nitrite concentration inside the reactor to over 20 mg

75.2

2017

Table 3

Figure 1

Measured nitrite concentrations and TNRRs during both pre-batch and batch periods in (a) MBBR, (b) SBR and (c) UASB reactor.

�1 NO� (Figure 1; Table 3). In this study, the nitrite 2 -NL values over 20 mg N L�1 were considered as spikes, as the value is used commonly as a threshold of nitrite inhibition

Summary of reactor nitrite spiking and batch test results.

Reactor

Batch

Type

Abundancy (Anammox � 16S rRNA copies g 1 TSS)

Fraction of spikes (pre-batch)a

Fraction of spikes (batch)a

NO� 2 -N IC50 � (mg L 1)

NO� 2 -N optimum � (mg N L 1)

MBBR

1.04 × 109

9/24 (38%)

11/30 (37%)

∼85

SBR

3.98 × 108

No spiking

12/21 (57%)

15/30 (50%)

9/21 (43%)

240

UASB

4.72 × 10

Number of reactor samples with measured nitrite values over 20 mg N L�1 out of all measured samples.

Page 359

316

M. Raudkivi et al.

(Wett ). No nitrite spiking was applied in SBR as no nitrite was added externally due to the deammonifying biomass performing aerobic and anaerobic ammonium oxidation process was present in the reactor. As all reactors were operated with reject water from the Tallinn WWTP with inconsistent concentrations, fluctuations in reactor total nitrogen removal efficiency (TNRE) can be seen for all of the reactors (Figure 1). Some bigger decreases in reactor efficiencies could be the result of nitrite inhibition, which in some cases for the MBBR and UASB might have been unintentionally caused by the spiking. During the batch test period, drops in TNRR (up to 40%), can be the result of biomass being taken out and then returned to the reactor after batch testing. Calculated TNRR, TNRE values and maintained NO� 2 -N concentrations for both pre-batch (before batch tests) and batch periods are presented on Figure 1 for each of the systems. All figures, efficiency calculations and regression statistics were visualised and calculated with Origin. Batch tests The inhibiting/limiting effect of different nitrite concentrations on three different biomasses (MBBR biofilm, SBR sludge and UASB granular sludge) was studied in series of batch tests. The tests were performed in air-tight 800 mL, while the test bottles were mixed by magnetic stirrers (mixing rate ∼100 rpm). All batch tests were performed at 25 C in order to achieve comparability between different þ biomasses. The influent NO� 2 -N=NH4 -N ratio was prepared based on the anammox stoichiometric ratio 1.32/1 described by Strous et al. (). Acidic and alkaline solutions of micro- and macronutrients were added according to Zekker et al. () to maintain sufficient nutrient balance for the biomass. After the addition of substrates, the batch cell together with the biomass was deaerated for 15 min with argon, to ensure anoxic conditions for the test. Samples were taken after every 2 h during the 6-h period using the overpressure of argon. For the MBBR (biofilm) 200 carriers were used in each batch test, the average total suspended solids (TSS) concentration in the test cell was 2.2 g L�1. In the tests with suspended biomass the average TSS concentrations for SBR and UASB biomass were 2.4 g L�1 and 5.5 g L�1, respectively. The average TSS used for UASB was higher due to significant mineral part in the granules (40–60% of TSS weight), while both the extracted biofilm and the suspended biomass were almost fully composed of organic W

Page 360

Water Science & Technology

Nitrite inhibition/limitation for MBBR, SBR and UASB systems

75.2

2017

matter. The batch test TNRRs were calculated based on the linear regression of substrate concentrations in time for a gram of biomass TSS. The IC50 values have been calculated based on the slopes of the acquired regression equations. As all batch tests were performed at similar pH values (8 ± 0.6), free nitrous acid (FNA) concentrations were not considered (Puyol et al. ). Furthermore, whether nitrite inhibition is caused by NO� 2 or FNA was not determined in the current study. Chemical analysis Prior to analysis, the water samples were centrifuged at 4,000 rpm for 10 min to remove the solids. � � þ NH4 , NO2 , NO3 concentrations were measured spectrophotometrically according to Greenberg et al. (). The samples pH value was measured with a pH-meter connected with Jenway pH electrode (Germany) and DO was measured by Marvet Junior (Estonia) electrodes, respectively. The measurement of TSS was performed according to Greenberg et al. (). PCR methodology In case of biofilm, five biomass carriers were taken from the reactor and biomass was mechanically removed using a vortex mixer, followed by DNA extraction by MoBio Powersoil DNA isolation kit (USA) according to the manufacture’s instructions. For sludge samples, the same DNA isolation procedures were applied as for biofilm samples. The PCR products were purified with the JETquick Spin Column Kit (GENOMED GmbH) and then sequenced. 25–50 mg of biomass was applied for DNA extraction (Zekker et al. ). Pla46f /Amx368r primers were used for targeting anammox bacteria. Nested PCR was carried out according to the thermocycling parameters described by SànchezMelsió et al. (). PCR-denaturing gradient gel electrophoresis (PCR-DGGE) for detecting diversity of the most abundant microorganisms was conducted using the eubacterial primer set GC-BacV3f/907r (Koskinen et al. ). The sequencing was carried out according to Zekker et al. (). Quantitative polymerase chain reaction qPCR was conducted with primer sets Amx694F(GGGGAGAGTGGAACTTCTG) and Amx960R(GCTCCACCGCTT GTGCGAGC), which amplify about 285 bp fragments from most anammox bacteria 16S rDNA (Ni et al. ).

317

M. Raudkivi et al.

Water Science & Technology

Nitrite inhibition/limitation for MBBR, SBR and UASB systems

Cloning for the standard was performed using the Thermo Scientific InsTAclone PCR cloning kit according to the manufacturer’s instructions. JM109 cell line was used. Plasmid was purified from selected colonies using the GeneJET Plasmid minipreparation kit (Thermo Scientific). Dilutions of purified plasmid were used as standard in the qPCR reaction. PCR amplification and detection were performed in optical 96-well reaction plates. The PCR temperature programme was initiated during 12 min at 95 C, followed by 45 cycles of 10 s at 94 C, 20 s at 58 C, and 20 s at 72 C. Each PCR mixture (10 μL) was composed of 2 μL of 5x HOT FIREPol Eva Green qPCR Supermix (Solis BioDyne, Estonia), 0.25 μL of forward and reverse primers (100 μM) and 1 μL of template DNA.

75.2

2017

RESULTS AND DISCUSSION Nitrite inhibition on different biomass Batch tests were conducted during 340–490 (MBBR), 596– 738 (SBR) and 173–210, 244–284 (UASB) days of the reactor operation (Figure 1). The overall results of the batch tests and the regime of nitrite spiking during both prebatch and batch periods are presented in Table 3. The operational conditions such as elevated total nitrogen loading rate (TNLR) and high maintained nitrite values in the reactor could enhance the biomass’s tolerance towards high nitrite values. That was mostly seen for the biomass taken from UASB having both high TNLR (400–600 g N m�3 d�1) and 50% of the measured nitrite values over the threshold. The possible effects of spiking are discussed more thoroughly in the speciﬁc chapter. The ﬁrst set of batch tests were performed with biomass carriers taken from a MBBR (Figure 2). The error bars on the graphs represent standard deviation. The highest nitrite �1 concentration used in the batch tests was 73 mg NO� 2 -NL , �1 while the IC50 was calculated to be at 85 mg NO� 2 -NL . Due to technical problems with the MBBR, further tests with higher nitrite values could not be performed. Therefore, the calculated IC50 value is used to provide comparison with the other biomasses. The biomass from MBBR achieved the highest TNRR (5 mg N g�1 TSS h�1) compared to other systems and the value was measured at relatively low nitrite �1 concentration of 40 mg NO� 2 �NL . Similar results were described by Bettazzi et al. () (highest TNRR at 37 mg �1 NO� 2 -NL ) with suspended biomass, while Kimura et al.

Figure 2

The TNRR of the biomass from MBBR (bioﬁlm carriers) on different nitrite concentrations.

() have shown higher nitrite tolerance (highest TNRR at �1 100 mg NO� 2 -NL ) with biofilm. TNRR’s dependence on nitrite concentration for the suspended biomass from the SBR is shown on Figure 3. The overall resistance of suspended biomass to nitrite inhibition �1 measured as IC50 was 15% higher (98 mg NO� 2 -NL ) than for biofilm carriers. The highest TNRR (2.6 mg N g�1 TSS �1 h�1) was achieved at 81 mg NO� (p-value <0.05), 2 -NL which is two times higher than the respective concentration found for MBBR. The results differ greatly from the inhibitory nitrite values reported by Wett (), who found nitrite to be inhibitory in concentrations as low as 9 mg �1 NO� 2 -NL . Although a similar deammonification SBR system as ours was used in that study, no nitrite spiking was carried out. Furthermore, based on the results published

Figure 3

The TNRR of the biomass from SBR at different nitrite concentrations.

Page 361

318

M. Raudkivi et al.

Water Science & Technology

Nitrite inhibition/limitation for MBBR, SBR and UASB systems

in Wett (), the low nitrite values reported might be the result of anammox efﬁciency loss not the cause of it (Lotti et al. ). In the batch tests with granules from the UASB, the �1 highest TNRR was achieved at 83 mg NO� (p-value 2 -NL <0.05) (Figure 4), which was comparable to the respective value in the SBR. The IC50 value for the UASB biomass �1 was calculated to be at 240 mg NO� (p-value 2 -NL <0.05), which was by far the highest among three different biomasses used. Similarly, very high IC50 values were acquired by both Dapena-Mora et al. () (at 350 mg � �1 �1 NO� 2 -NL ) and Lotti et al. () (at 400 mg NO2 -NL , granular biomass). As the UASB was originally inoculated with biomass acquired from Lotti, the similar results were to be expected. Nitrite limitation Various other authors have achieved the highest TNRRs at �1 nitrite concentrations from 40 to 120 mg NO� 2 -NL (Strous et al. b; Dapena-Mora et al. ; Bettazzi et al. ). The fact that until ﬁrst signs of inhibition, the TNRR in the system rises with the increase of nitrite concentration (no plateau is usually reached), may refer to nitrite limitation in the anammox process. During the course of nitrite inhibition research, we also studied whether nitrite spiking has an effect on nitrite limitation. As shown in Table 3, no apparent effect from spiking on nitrite limitation was observed. Despite being spiked with different frequency, the batch tests with either SBR or UASB biomass showed similar results to low nitrite concentrations �1 (peak TNRRs at 81 and 83 mg NO� 2 -NL , respectively).

NO� 2 -N concentration (mg L

Page 362

2017

�1

)

TSS h

MBBR

SBR

UASB

2.50

0.60

1.05

1.18

1.97

1.78

2.90

5.03

3.84

Peak TNRR values.

�1

TNRR (mg N g

40 80 The TNRR of the biomass from UASB at different nitrite concentrations.

Calculated TNRRs at different NO2 concentrations in reactors �1

75.2

While MBBR and UASB were spiked with similar frequency, the nitrite concentrations at which the peak TNRR was achieved differed twofold. This indicates varying limiting concentrations for different types of reactors and biofilm. While in this research the peak TNRRs for suspended biomass were observed at around 80 mg �1 NO� 2 -NL , the peak TNRRs from other authors with suspended biomass have been at 37, 50 and 120 mg �1 NO� (Strous et al. (b), Dapena-Mora et al. 2 -NL () and Bettazzi et al. (), respectively). The results showed that while nitrite spiking does not have a significant effect on nitrite limitation, the type of biomass might be the most important factor concerning nitrite limitation. The calculated TNRRs for all reactors on different nitrite values is shown in Table 4. When MBBR was �1 operated within the threshold of 20 mg NO� 2 -NL , the achieved TNRR was 4 times higher than in SBR and 2.5 times higher than in UASB, while the peak TNRR values only differed 2 and 1.5 times, respectively. These results suggest that an anammox reactor with biofilm carriers would be the easiest to operate, as sufficient rates can be achieved in both under and over the strict threshold (Wett ). The effect of nitrite spiking on nitrite inhibition can be observed when comparing the TNRR at peak nitrite concentrations with IC50 values. The biomass from SBR, which was not spiked showed the peak TNRR values at 81 mg �1 �1 NO� and IC50 at 98 mg NO� 2 -NL 2 -NL , the slope of inhibition was steep (nitrite IC50 concentration only 1.2 times higher than the most efficient concentration). This means operating the reactor near the most efficient nitrite concentrations would be risky and difficult as even a slight increase in nitrite concentration could bring on significant inhibition. The inhibition slopes for both MBBR and UASB were gentler. MBBR was spiked 38% of the time and the difference between IC50 and peak TNRR nitrite concentrations was 2.1 times, while the respective values for UASB were 50% and 2.9 times. Table 4

Figure 4

2.79

2.37

)

3.83a

319

M. Raudkivi et al.

Water Science & Technology

Nitrite inhibition/limitation for MBBR, SBR and UASB systems

Based on the results of this study, nitrite spiking frequency did not have any visible effect on nitrite limitation. However, by adapting the anammox reactor to periodical high nitrite concentrations, the slope of inhibition can be decreased significantly, which makes the system more tolerable to even higher nitrite concentrations. This effect can be used to operate anammox reactors on higher stable nitrite concentrations in order to maximise TNRR without risking strong inhibition in the system. qPCR studies The results of the quantitative anammox 16S rRNA analysis showed that the highest abundance of anammox gene copies per a gram of TSS were determined in MBBR (1.04 × 109 copies g�1 TSS) (Figure 5). The amount of anammox gene copies in the biomass taken from SBR and UASB systems were similar (3.98 × 108 and 4.72 × 108 copies g�1 TSS, respectively, p-value < 0.05), which are, respectively, 2.6 and 2.2 times lower than in the biomass taken from MBBR. This result was expected as SBR was a deammonification system, a significant part of the biomass could belong to other bacteria, which do not carry the anammox gene. As the UASB granular biomass contains a considerable mineral part, the lower amount of gene copies per a gram of suspended solids was expected as well. In order to provide better comparability between the influent and biomass from the reactors, the quantitative anammox 16S rRNA analysis was also carried out from the reject water from Tallinn WWTP anaerobic tank. The amount of anammox 16S rRNA in the reject water was 2.26 × 106 copies g�1 TSS, which was a hundredfold lower than in the reactors.

75.2

2017

To compare the maximal TNRRs of different biomass, the qPCR results were used to calculate the maximum nitrogen removal rate per an anammox 16S rRNA gene copy (Figure 5). The biomass with the highest maximum TNRR per a gene copy was from the UASB (6.96 × 10�9 mg N anammox 16S rRNA copy�1 h�1), while the biomass from the SBR and MBBR achieved 6.52 × 10�12 and 4.86 × 10�12 mg N anammox 16S rRNA copy�1 h�1, respectively. Based on the results acquired by qPCR analysis, the higher limiting nitrite concentrations for anammox process could also mean higher Vmax values for well-functioning �1 anammox biomass. For the MBBR at 40 mg NO� 2 -NL the highest TNRR per a gene copy was about 1.4 times lower than for the UASB. The peak TNRR was achieved on similar nitrite concentrations for both UASB and SBR and the maximum TNRRs per gene copy were similar as well (6.96 × 10�9 mg N anammox 16S copy�1 h�1 and 6.52 × 10�9 mg N anammox 16S copy�1 h�1, respectively). In the biomass from the MBBR, Candidatus Brocadia fulgidia and Candidatus Kuenia stuttgartiensis (Zekker et al. ) and in the biomass taken from the SBR anammox clones closest to Candidatus Brocadia fulgidia (Zekker et al. b) were found as the most abundant ones. In the biomass from Rotterdam pilot plant, the inoculum for the UASB, Candidatus Brocadia anammoxidans was with relative abundance of 50–60% out of all anammox bacteria (van der Star et al. ). Although characterisation of different anammox cultures has been carried out in the recent years (Awata et al. ; Ali et al. ), no information was available for the characteristics and tolerance of Candidatus Brocadia fulgidia. For that reason, giving objective conclusions based on the microbiological data from the MBBR and SBR are difﬁcult. Although Candidatus Brocadia anammoxidans has been reported to be less tolerant to nitrite than other cultures (Oshiki et al. ), in current research this anammox bacteria showed the highest nitrite tolerance. This could either be due to the type of biomass (granular) or the effect of nitrite spiking, which could prove the importance of both in operating a stable anammox system.

CONCLUSION

Figure 5

Gene copies of anammox 16S rRNA and the maximum TNRR per a gene copy in reject water and in tested biomasses.

Nitrite inhibition and limitation for anammox process were studied with biomass taken from three different reactors: MBBR, SBR and UASB. Batch tests showed that while the response to nitrite inhibition can be lessened with nitrite spiking, nitrite limitation is primarily affected by the type Page 363

320

M. Raudkivi et al.

Nitrite inhibition/limitation for MBBR, SBR and UASB systems

of biomass. Nitrite spiking was carried out in the MBBR and UASB and both biomasses were less susceptible to nitrite inhibition – the difference between the highest TNRR (inhibition threshold) and IC50 values was two times in the �1 MBBR (40 and 85 mg NO� 2 -NL ) and three times in the � UASB (83 and 240 mg NO2 -NL�1 ). The SBR in which no nitrite spiking was carried out before batch tests, had a stronger and steeper response to nitrite inhibition (highest �1 TNRR at 81 and IC50 at 98 mg NO� 2 -NL ). The biomass from the MBBR achieved both the highest maximum TNRR (5.03 mg N g�1 TSS h�1) and highest abundancy in 16S rRNA gene copies (1.04 × 109 copies g�1 TSS). The reactor also showed the highest TNRRs at low nitrite concentrations, which could make operating a bioﬁlm reactor cheaper and technologically easier than suspended anammox biomass reactors. The highest TNRR per 16S rRNA anammox gene copies was achieved with biomass from the UASB at 83 mg �1 NO� 2 -NL . Contrary to earlier research on anammox reactors, reactors working on suspended or granular may function better at relatively high nitrite concentrations �1 around 60–80 mg NO� 2 -NL , indicating a strong limiting effect of nitrite on anammox process, which should be researched even further.

ACKNOWLEDGEMENTS This study was supported by projects (SLOKT11027T), and IUT20-16. Anne Paaver is acknowledged for chemical analyses of water samples and T. Lotti and M. Beier for supporting us with biomass.

REFERENCES Ali, M. & Okabe, S.  Anammox-based technologies for nitrogen removal: advances in process start-up and remaining issues. Chemosphere 141, 144–153. Ali, M., Oshiki, M., Awata, T., Isobe, K., Kimura, Z., Yoshikawa, H., Hira, D., Kindaichi, T., Satoh, H., Fujii, T. & Okabe, S.  Physiological characterization of anaerobic ammonium oxidizing bacterium ‘Candidatus Jettenia caeni’. Environ. Microbiol. 17, 2172–2189. Awata, T., Oshiki, M., Kindaichi, T., Ozaki, N., Ohashi, A. & Okabe, S.  Physiological characterization of an anaerobic ammonium-oxidizing bacterium belonging to the ‘Candidatus Scalindua’ group. Appl. Environ. Microbiol. 79, 4145–4148. Bettazzi, E., Caffaz, S., Vannini, C. & Lubello, C.  Nitrite inhibition and intermediates effects on anammox bacteria: a batch-scale experimental study. Process Biochem. 45, 573–580.

Page 364

Water Science & Technology

75.2

2017

Dapena-Mora, A., Fernandez, I., Campos, J. L., Mosquera-Corral, A., Mendez, R. & Jetten, M. S. M.  Evaluation of activity and inhibition effects on anammox process by batch tests based on the nitrogen gas production. Enzyme Microb. Technol. 40, 859–865. Greenberg, A. E., Clesceri, L. S. & Eaton, A. D.  Standard Methods for the Examination of Water and Wastewater. American Public Health Association, Washington, DC, USA. Kimura, Y., Isaka, K., Kazama, F. & Sumino, T.  Effects of nitrite inhibition on anaerobic ammonium oxidation. Appl. Microbiol. Biotechnol. 86, 359–365. Koskinen, P. E. P., Kaksonen, A. H. & Puhakka, J. A.  The relationship between instability of H2 production and compositions of bacterial communities within a dark fermentation fluidized-bed bioreactor. Biotechnol. Bioeng. 97, 742–758. Li, X., Zen, G., Rosenwinkel, K. H., Kunst, S., Weichgrebe, D., Cornelius, A. & Yang, Q.  Start up of deammonification process in one single SBR system. Water Sci. Technol. 50, 1–8. Lotti, T., van der Star, W. R. L., Kleerebezem, R., Lubello, C. & van Loosdrecht, M. C. M.  The effect of nitrite inhibition on the anammox process. Water Res. 46, 2559–2569. Mulder, A., Van de Graaf, A. A., Robertson, L. A. & Kuenen, J. G.  Anaerobic ammonium oxidation discovered in a denitrifying fluidized-bed reactor. Fems Microbiol. Ecol. 16, 177–183. Ni, B.-J. J., Hu, B.-L. L., Fang, F., Xie, W.-M. M., Kartal, B., Liu, X.W. W., Sheng, G.-P. P., Jetten, M., Zheng, P. & Yu, H.-Q. Q.  Microbial and physicochemical characteristics of compact anaerobic ammonium-oxidizing granules in an upflow anaerobic sludge blanket reactor. Appl. Environ. Microbiol. 76, 2652–2656. Oshiki, M., Shimokawa, M., Fujii, N., Satoh, H. & Okabe, S.  Physiological characteristics of the anaerobic ammoniumoxidizing bacterium ‘Candidatus Brocadia sinica’. Microbiology 157, 1706–1713. Puyol, D., Carvajal-Arroyo, J. M., Sierra-Alvarez, R. & Field, J. A.  Nitrite (not free nitrous acid) is the main inhibitor of the anammox process at common pH conditions. Biotechnol. Lett. 36, 547–551. Sànchez-Melsió, A., Cáliz, J., Balaguer, M. D., Colprim, J. & Vila, X.  Development of batch-culture enrichment coupled to molecular detection for screening of natural and man-made environments in search of anammox bacteria for N-removal bioreactors systems. Chemosphere 75, 169–179. Strous, M., Heijnen, J. J., Kuenen, J. G. & Jetten, M. S. M.  The sequencing batch reactor as a powerful tool for the study of slowly growing anaerobic ammonium-oxidizing microorganisms. Appl. Microbiol. Biotechnol. 50, 589–596. Strous, M., Fuerst, J. A., Kramer, E. H. M., Logemann, S., Muyzer, G., van de Pas-Schoonen, K. T., Webb, R., Kuenen, J. G. & Jetten, M. S. M. a Missing lithotroph identified as new planctomycete. Nature 400, 446–449. Strous, M., Kuenen, J. G. & Jetten, M. S. M. b Key physiology of anaerobic ammonium oxidation. Appl. Environ. Microbiol. 65, 3248–3250.

321

M. Raudkivi et al.

Nitrite inhibition/limitation for MBBR, SBR and UASB systems

van der Star, W. R. L., Abma, W. R., Blommers, D., Mulder, J. W., Tokutomi, T., Strous, M., Picioreanu, C. & Van Loosdrecht, M. C. M.  Startup of reactors for anoxic ammonium oxidation: experiences from the first full-scale anammox reactor in Rotterdam. Water Res. 41, 4149–4163. Van Hulle, S. W. H., Vandeweyer, H. J. P., Meesschaert, B. D., Vanrolleghem, P. A., Dejans, P. & Dumoulin, A.  Engineering aspects and practical application of autotrophic nitrogen removal from nitrogen rich streams. Chem. Eng. J. 162, 1–20. Wett, B.  Solved upscaling problems for implementing deammonification of rejection water. Water Sci. Technol. 53, 121–128. Wett, B.  Development and implementation of a robust deammonification process. Water Sci. Technol. 56, 81–88.

Water Science & Technology

75.2

2017

Zekker, I., Rikmann, E., Tenno, T., Lemmiksoo, V., Menert, A., Loorits, L., Vabamae, P. & Tomingas, M.  Anammox enrichment from reject water on blank bioﬁlm carriers and carriers containing nitrifying biomass: operation of two moving bed bioﬁlm reactors (MBBR). Biodegradation 23, 547–560. Zekker, I., Rikmann, E., Tenno, T., Kroon, K., Seiman, A., Loorits, L., Fritze, H., Tuomivirta, T., Vabamae, P., Raudkivi, M., Mandel, A. & Tenno, T. a Start-up of low-temperature anammox in UASB from mesophilic yeast factory anaerobic tank inoculum. Environ. Technol. 36, 214–225. Zekker, I., Rikmann, E., Tenno, T., Seiman, A., Loorits, L., Kroon, K., Tomingas, M., Vabamäe, P. & Tenno, T. b Nitritatinganammox biomass tolerant to high dissolved oxygen concentration and C/N ratio in treatment of yeast factory wastewater. Environ. Technol. 35, 1565–1576.

First received 20 June 2016; accepted in revised form 13 September 2016. Available online 9 November 2016

Page 365

2891

75.12

2017

Combined ultraﬁltration-electrodeionization technique for production of high purity water Anita Kusuma Wardani, Ahmad Nurul Hakim, Khoiruddin and I Gede Wenten

ABSTRACT Electrodeionization (EDI) is the most common method to produce high purity water used for boiler feed water, microelectronic, and pharmaceutical industries. Commonly, EDI is combined with reverse osmosis (RO) to meet the requirement of EDI feed water, with hardness less than 1 ppm. However, RO requires a relatively high operating pressure and ultraﬁltration (UF) as pretreatment which results in high energy consumption and high complexity in piping and instrumentation. In this work, UF was used as the sole pretreatment of EDI to produce high purity water. Tap water with

Anita Kusuma Wardani Ahmad Nurul Hakim Khoiruddin I Gede Wenten (corresponding author) Department of Chemical Engineering, Institut Teknologi Bandung, Jl. Ganesha 10, Bandung 40132, Indonesia E-mail: igw@che.itb.ac.id

conductivity 248 μS/cm was fed to UF-EDI system. The UF-EDI system showed good performance with ion removal more than 99.4% and produced water with low conductivity from 0.2 to 1 μS/cm and total organic compounds less than 0.3 ppm. Generally, product conductivity decreased with the increase of current density of EDI and the decrease of feed velocity and UF pressure. The energy consumption for UF-EDI system in this work was 0.89–2.36 kWh/m3. These results proved that UF-EDI system meets the standards of high purity water for pharmaceutical and boiler feed water with lower investment and energy consumption than RO-EDI system. Key words

| conductivity, electrodeionization, high purity water, ion removal, ultraﬁltration

INTRODUCTION High purity water is greatly important, especially for pharmaceutical and boiler feed water. Based on US and European pharmaceutical regulations, pharmaceutical industries require water with conductivity <1.3 μS/cm, total organic carbon (TOC) < 0.5 ppm as C, heavy metal <0.1 ppm as Pb, and aerobic bacteria <100 CFU/mL (Wang et al. ; Harfst ; Bennett ), while boiler feed water demands water with conductivity <1.1 μS/cm, TOC < 0.5 ppm as C, and silica <1 ppm (Scott ; Singh ). The conductivity of high purity water is less than 1 μS/cm at 25 C, with low quantities of TOC and total dissolved solids (TDS) (Bennett ; Bohus et al. ). Traditionally, ion-exchange system is used to produce high purity water in industry, but nowadays membrane processes are becoming popular as replacements for ion-exchange systems (Hernon et al. ). One of the membrane processes often used for high purity water production is electrodeionization (EDI) (Strathmann ). EDI has been applied for high purity water production at a large industrial scale (Khoiruddin et al. b). EDI W

combines ion-exchange resins and ion-selective membranes with direct current to remove ionized species from water (Hernon et al. ; Strathmann ; Alvarado et al. ; Lee & Choi ). It was developed to overcome the limitations of ion-exchange system, which needs regeneration of the resins and low quality product of electrodialysis (Bouhidel & Lakehal ; Nagarale et al. ; Wardani et al. ). Compared to conventional ionexchange system, EDI has the advantage of being a continuous process with stable product quality, which is able to produce high purity water without the need for acid or caustic regeneration (Helfferich ; Hernon et al. ; Lee et al. ). EDI is the most common method for producing high purity water that typically achieve more than 99.5% salt rejection with a resistivity of 1–18 MΩ cm and low quantities of TOC (Franken ; Grabowski et al. ; Wood et al. ). An EDI device contains alternating permselective anion-exchange membranes and cation-exchange membranes between two electrodes (Wood et al. ; Arar

doi: 10.2166/wst.2017.173

Page 367

2892

A. K. Wardani et al.

et al. ). The compartments in EDI stack consist of diluate or product compartments, concentrate compartments and electrode compartments. The compartments are filled with mixed-bed ion-exchange resins, which enhance the transport of ionic components from bulk solution toward the ion-exchange membranes under the force of a direct current (Helfferich ; Widiasa et al. ). When EDI is used for the production of high purity water, the ion-exchange resin beads enhance mass transfer, facilitate water splitting, and reduce stack resistance (Strathmann ). The direct current electrical field splits water into hydrogen and hydroxyl ions, which in turn continuously regenerate the ion-exchange resins. The exchanged ions are transferred through the membranes to the concentrate compartments and flushed from the system (Widiasa et al. ; Yeon et al. ). The quality of product water obtained by EDI process depends much on the characteristics of feed. It is usually required in present EDI technology that the hardness of feed water should be less than 1.0 ppm (as CaCO3) (Fu et al. ). Therefore, reverse osmosis (RO) is often compelled to be adopted for pretreatment of the EDI (Liang et al. ; Auerswald ; Wang et al. ; Song et al. ; Arar et al. ; Wenten & Khoiruddin ). Generally, RO needs ultrafiltration (UF) as pretreatment to filter out particles that may otherwise clog or damage the RO membrane. RO also requires a relatively high operating pressure, which results in high energy consumption and high complexity in piping and instrumentation (Liberman ; Fritzmann et al. ; Lee et al. ; Kucera ). This work aims to use UF as the sole pretreatment of EDI for high purity water production by varying UF pressure, EDI feed velocity, and current density. UF membrane was used to replace RO membrane due to its low operating pressure, less than 2 bar. UF can remove organic compounds and produce crystal clear water too (Aryanti et al. , ). In addition, the energy consumption and investment cost of UF-EDI system is much lower compared to RO-EDI system.

EXPERIMENTAL Experimental set-up UF-EDI system was used in this work to produce high purity water from tap water (Table 1). Polysulfone capillary UF membrane (GDP Filter, Indonesia) with pore size ±10 nm and effective area 2.51 m2 was used as pretreatment to Page 368

Water Science & Technology

Combined UF-EDI for high purity water production

Table 1

75.12

2017

Quality of feed water Feed water

TDS (ppm)

136 ± 0.8

Conductivity (μS/cm)

248 ± 1.0

Hardness (ppm)

6.1 ± 0.2

TOC (ppm)

4.5 ± 0.4

7.1 ± 0.2

produce EDI feed water. This pretreatment aimed to remove hardness and organic compounds before fed to EDI module. Meanwhile, the EDI stack consists of one diluate compartment, two concentrate compartments, and two electrode compartments (one anode and one cathode) (see Figure 1). Electrodes used in this work were stainless steel SS-304. Cation-exchange membrane (MC-3470) and anionexchange membrane (MA-3475) from Ionac Chemical Company (USA) were used as ionic selective barriers of the EDI stack. Properties of the membranes have been explained in the literature (Khoiruddin et al. a). Each membrane had effective area of 200 cm2 with the internal spacer for each concentrate and electrode compartments was 4 mm and for diluate compartment was 8 mm. Mixed ion-exchange resins with volume ratio 1:1 were ﬁlled to the diluate and concentrate compartments. The main characteristics of the ion-exchange resins used in this work are presented in Table 2. An adjustable power supply (homemade power supply) was used to produce direct current on EDI. It could supply voltage and direct current in the range of 0–100 V and 0–50 A, respectively. TDS and electrical conductivity of diluate and concentrate was measured every 10 minutes for 3 hours. Analytical method Total organic compounds and hardness Organic compounds of feed and product water were measured as TOC by TOC meter (Shimadzu TOC-VCPH, Mandel, Canada). The measurement was conducted at room temperature. Meanwhile, spectrophotometer UV-Vis (Spectronic 20D, ThermoFisher Scientiﬁc, USA) was used for measuring the hardness with Eriochrome Black T (EBT) solution as indicator. EBT solution was prepared by dissolving 50 mg of EBT powder into 50 mL of ethanol and placed in a dark and cool place. 10 mL of EBT solution was then diluted into 100 mL ethanol and named as EBT work solution. Optimum reaction between EBT and

2893

A. K. Wardani et al.

Figure 1

Water Science & Technology

Combined UF-EDI for high purity water production

2017

measured using conductivity meter (HI-98303, Hanna Instruments, Mauritius). Meanwhile, ion concentration was measured as TDS using TDS meter (TDS-3, HM Digital, Taiwan). Ion removal was calculated to show the decrease of ions in the diluate compartment using the following equation (Mulder ; Lu et al. ): R¼

Electrical conductivity and TDS One of the important characteristics of high purity water is the electrical conductivity. Electrical conductivity was |

75.12

Schematic diagram of UF-EDI system (CM: cation-exchange membrane and AM: anion-exchange membrane).

hardness ion occurred at pH 8–10 (Tachino et al. ), so buffer solution was prepared by dissolving 1 g of NH4Cl into 100 mL of NH4OH 12.50% solution. 2 mL EBT work solution and 2 mL buffer were added to 2 mL test solution, and diluted with deionized water until 10 mL, then put in spectrophotometer cuvette. Wavelength 530 nm was used to measure the absorbance value from each solution.

Table 2

Cin � Cout × 100 ð%Þ Cin

(1)

where R (%) is the removal of each ion, Cin (ppm) is the feed concentration, and Cout (ppm) is the product concentration.

Properties of ion-exchange resin (Dow Chemical Company) Amberlite™ IR120-Na

Amberlite™ IRA900-Cl

Type

Strong acid

Strong base

Matrix structure

Styrene divinylbenzene

Function group

Sulfonate

Trimethyl ammonium

Ion-exchange capacity

�2.00 eq./L

�1.00 eq./L

Moisture holding capacity

45–50%

58–64%

RESULTS AND DISCUSSION Determination of optimal feed conditions for EDI using UF membrane UF membrane was used to remove hardness and organic compounds as a pretreatment of EDI system to produce high purity water. Hardness was measured as concentration of CaCO3, while organic compounds were measured as TOC. In this work, some of hardness compounds as Page 369

2894

A. K. Wardani et al.

Combined UF-EDI for high purity water production

CaCO3 with molecule size 30–60 nm ( Jia et al. ) were rejected by UF membrane. UF rejection is determined mainly by the size and shape of solutes relative to the pore size in the membrane (Mulder ). The pore size of UF membrane used in this work was ±10 nm. Therefore, the molecules with size more than 10 nm were rejected by UF membrane. Hardness removal was also mentioned in the previous work (Tabatabai et al. ; Mika et al. ), where UF could remove more than 80% of hardness due to high physical–chemical interaction between hardness compounds and membrane surfaces. The other work also showed that UF can remove biodegradable organic compounds from feed solution, while the synthetic organic compounds can hardly be removed (Metcalf ). Figure 2 shows the hardness and TOC removal of UF permeates with pressure and feed velocity. When the pressure was increased, the amount of hardness and TOC in UF permeate also increased. Theoretically, when the pressure is too low, it will be difﬁcult to push the hardness ions and organic compounds through the membrane pores. Thus, the components drift to the UF retentate. Meanwhile, the amount of hardness and TOC decreased when the feed velocity was increased because contact time between feed solution and membrane surface decreased. The effect of velocity is important in the membrane ﬁltration process. A higher velocity can reduce membrane fouling by providing a shear force to sweep away deposited materials (Chen et al.

Figure 2

Effect of the UF pressure and feed velocity on the hardness and TOC of permeate.

Page 370

Water Science & Technology

75.12

2017

). This can slightly increase the retention of most components (Wei et al. ). Based on the results, the product with pressure up to 25 psi meets the standards of EDI feed water, with hardness less than 1 ppm. It implies that UF system was effective as pretreatment to produce EDI feed water. Furthermore, the optimal operating conditions were determined to be 1.25 m/s of feed velocity and 15 psi of UF pressure. These conditions gave the minimum value of hardness and TOC. Under these conditions, the permeate water quality was observed to be 0.921 ppm of the hardness and 0.659 ppm of the TOC. Voltage–current density characteristics The voltage–current density curves of EDI with different feed water are shown in Figure 3. The current density increased more rapidly than the voltage at higher voltage for UF permeate and lower voltage for brackish water. The reason is that even at low voltage, a signiﬁcant amount of Hþ and OH� ions are produced in the diluate compartment (Wang et al. ). Since more Hþ and OH� are produced at higher voltage, the resistance of the stack is decreased due to the higher conductivity of resin in Hþ and OH� form (Fu et al. ; Xing et al. ). Therefore, water with higher conductivity (brackish water) has a steeper curve due to its higher ion concentration.

2895

Figure 3

A. K. Wardani et al.

Combined UF-EDI for high purity water production

•

75.12

2017

Variation of current density with voltage in EDI process.

Similar to that reported by Song et al. (), the voltage–current density curve for this work could fall into two segments, as follows:

•

Water Science & Technology

First segment (0–50 V), the current density increased linearly as the voltage increased. The voltage–current density curve in this section follows Ohm’s law, where V ¼ IR. Second segment (50–80 V), the current density increased linearly as the voltage increased like the ﬁrst segment but more quickly. In this segment, water dissociation took place and a signiﬁcant amount of Hþ and OH� ions were produced. Consequently, more charge carriers are present and thus the current density is increased.

From voltage–current density curve, limiting current density can be determined. Limiting current density is the cross-point of the tangents drawn from ﬁrst segment and second segment (Doyen et al. ), and it is about 22.5 A/m2 according to Figure 3. In this work, the current density of 17.5 A/m2 (below the limiting current density), 22.5 A/m2 (limiting current density), and 27.5 A/m2 (above the limiting current density) were chosen to study the characteristics of ionic migration in different segments.

Effect of current density on quality of product water In this work, the UF permeate with conductivity 237 μS/cm was used as a feed solution. The EDI stack was operated with a constant feed velocity of 0.75 m/s. As shown in Figure 4(a), conductivity of diluate decreased with the increase of the current density. When the current density was increased from 17.5 A/m2 to 27.5 A/m2, the electrical conductivity of diluate decreased from 1 μS/cm (at 17.5 A/m2) to 0.3 μS/cm (at 27.5 A/m2). These results show that high purity water for pharmaceutical and boiler feed water can be obtained by using current density from 17.5 A/m2 up to 27.5 A/m2. The increase in current density also led to the increase in ion removal. At the end of the experiment of 180 minutes, the ion removal was 99.47%, 99.62%, and 99.92% for current density variation of 17.5, 22.5, and 27.5 A/m2, respectively. Theoretically, when the current density is too low, it will be difﬁcult to maintain a desired removal of the ions due to a lower strength of driving force (Arar et al. ). When the current density is raised, more electric potential is used for the transport of ions. Therefore, the conductivity of water decreased and ion removal increased, which was in agreement with the previous works (Meyer et al. ; Lu et al. ; Arar et al. , Page 371

2896

A. K. Wardani et al.

Water Science & Technology

Combined UF-EDI for high purity water production

75.12

2017

reduced stack resistance and participated in the ion transport through membranes. Effect of EDI feed velocity on quality of product water

Figure 4

Electrical conductivity of diluate and ion removal as a function of (a) current density and (b) feed velocity.

). At the limiting current density and above, water splitting occurs in the diluate compartment. At this condition, Hþ and OH� are formed by in situ water dissociation to regenerate the ion-exchange bed continuously (Zhang et al. ). When 100% and approximately 120% of the limiting current density (22.5 and 27.5 A/m2) were applied, the number of the generated Hþ and OH� ions from the water dissociation could be too high. Therefore, they not only helped to regenerate the ion-exchange resins, but also

Table 3

It was found that product conductivity and ion removal depended not only on the current density, but also on the feed velocity. In this work, the velocity of the UF permeate was varied from 0.5 to 1 m/s with a constant current density of 22.5 A/m2. When the feed velocity was increased, conductivity of the diluate increased and conductivity of the concentrate decreased. After 180 minutes, the electrical conductivity of diluate was 0.4, 0.7, and 0.9 μS/cm for feed velocity of 0.5, 0.75, and 1 m/s, respectively (Figure 4(b)). Increasing feed velocity also affected the ion removal. When the velocity was increased from 0.5 m/s to 1 m/s, the ion removal decreased from 99.77% (at 0.5 m/s) to 99.55% (at 1 m/s). When the velocity was increased, the residence time of the solution in the resin bed decreased. Thus, the diffusion kinetics of the ions from the solution to the ionexchange resin declined. This led to the decrease in the transport of ions to the concentrate compartments (Wen et al. ; Xing et al. ; Arar et al. ). All variations of feed velocity in this work produce high purity water suitable for pharmaceutical and boiler feed water. However, it is important to look for the optimum feed velocity since feed velocity is related to the product capacity. To achieve same product capacity, using higher feed velocity is more proﬁtable due to fewer numbers of modules needed, which leads to a reduction in investment cost. UF-EDI product characteristics and energy consumption The UF-EDI system showed good performance in producing high purity water. As shown in Table 3, water product has low conductivity from 0.2 to 1 μS/cm with TOC less than 0.3 ppm. These results showed that UF-EDI product for

Product quality and energy consumption of UF-EDI system Current density (A/m2)

EDI feed velocity (m/s)

Ion removal (%)

Feed water

–

248

2.452

–

Product

17.5 22.5 27.5 27.5 22.5 22.5 22.5

0.75 0.75 0.75 0.5 0.5 0.75 1

99.47 99.62 99.92 99.94 99.77 99.62 99.55

1.0 0.7 0.3 0.2 0.4 0.7 0.9

0.296 0.274 0.263 0.248 0.255 0.274 0.281

0.89 1.45 2.23 2.36 2.12 1.45 1.12

Page 372

Conductivity (μS/cm)

TOC (ppm)

Total energy (kWh/m3)

2897

Table 4

A. K. Wardani et al.

Water Science & Technology

Combined UF-EDI for high purity water production

75.12

2017

Comparison between RO-EDI and UF-EDI system Operating conditions Conductivity (μS/cm) RO/UF pressure (psi)

EDI feed ﬂow rate (m/s)

EDI voltage (V)

EDI current density (A/m2)

Ion removal

Feed

Product

Reference

RO-EDI

– – – – 40–200 180 112

– – – 5.5 2–20 0.11–0.28 0.3

– – – 20 – 20–40 30–70

– – – – 2–30 – –

>98% >98% >99% >99% >99% >99% >99%

– – 3.5–4.5 14–18 50–250 1,685 40–60

– – 0.05–0.06 0.005–0.01 0.006–0.01 1.1–5.9 0.3–0.4

Liang et al. () Auerswold () Prato & Gallagher () Wang et al. () Song et al. () Arar et al. () Wenten et al. ()

UF-EDI

10–30

10–20

40–60

17.5–27.5

>99%

248

0.2–1

This work

System

each operating condition meets the requirements of pharmaceutical and boiler feed water, with conductivity <1.1 μS/cm and TOC < 0.5 ppm (Scott ; Singh ). The optimum operating parameters including UF pressure, current density, and feed velocity were obtained according to the experimental results. In order to give an economic evaluation of the UF-EDI system in this work, a set of representative operating parameters was taken as follows: the UF pressure was 15 psi with feed velocity 1.25 m/s, the current density of EDI was 27.5 A/m2 and EDI feed velocity was 0.5 m/s. The energy consumption for UF process was evaluated by Equation (2) (Mulder ): EUF ¼

Q0 P η

(2)

where EUF is energy consumption of UF (kWh/m3), Q0 is UF feed ﬂow rate (m3/h), P is UF pressure (bar), and η is pump efﬁciency. Meanwhile, the energy consumption for EDI process was calculated using the following equation (Zuo et al. ; Lu et al. ): EEDI ¼

IVt L

(3)

where EEDI is energy consumption of EDI (kWh/m3), I is electrical current (ampere), V is voltage (volt), t is operating time (h), and L is water volume (m3). At the optimum conditions, the total energy consumption was 2.36 kWh/m3. Table 4 shows comparison of UF-EDI system and RO-EDI system from previous works. UF-EDI system had a performance as good as RO-EDI system, but with lower operating pressure. If RO-EDI was operated with same EDI module for UF-EDI system, energy consumption became much higher since operating pressure for RO is 20–30 times operating pressure for UF.

Generally, high operating pressure of RO not only leads to high energy consumption, but also high complexity in piping and instrumentation (Liberman ; Fritzmann et al. ; Lee et al. ; Kucera ). The materials of piping and instrumentation for RO must be able to withstand high pressure condition. It is usually necessary to use metals for the high pressure system (FILMTEC). This high complexity of RO leads to the increase of investment cost. Comparison of investment cost for RO-EDI and UF-EDI system is presented in Table 5. In general, the investment cost for the RO-EDI system is 2–4 times that for the UF-EDI system.

CONCLUSION This work used UF-EDI system as an alternative process for high purity water production. Such operating parameters as UF pressure, current density of EDI, and feed velocity were investigated in detail. The UF-EDI system showed good performance in producing high purity water, with product conductivity from 0.2 to 1 μS/cm and TOC less than 0.3 ppm. Ion removal of this system was more than 99.4%. Generally, product conductivity decreased with the increase Table 5

Comparison of investment cost of RO-EDI and UF-EDI system Capacity

Investment cost

System

(m3/hr)

(US$)

Reference

RO-EDI

58.14

925,608

120

1,120,000

Matzan et al. () Wenten et al. ()

120

515,400

UF-EDI

Estimated by GDP Filter

Page 373

2898

A. K. Wardani et al.

Combined UF-EDI for high purity water production

of current density of EDI and decrease of feed velocity and UF pressure. The optimum operating parameters for UF were obtained at pressure 15 psi and feed velocity 1.25 m/s. Meanwhile, current density 27.5 A/m2 and feed velocity 0.5 m/s were the optimum operating parameters for EDI. The energy consumptions for UF-EDI system in this work are 0.89–2.36 kWh/m3. These results proved that UF-EDI system meets the standards of high purity water for pharmaceutical and boiler feed water with lower investment and energy consumption than RO-EDI system.

ACKNOWLEDGEMENTS Financial assistance for this work has been provided by Lembaga Pengelola Dana Pendidikan (LPDP) Indonesia through Beasiswa Pendidikan Indonesia (BPI) and Program Penelitian, Pengabdian kepada Masyarakat, dan Inovasi (P3MI) Institut Teknologi Bandung. The authors would also like to thank GDP Filter Indonesia for the supporting data.

REFERENCES Alvarado, L., Ramírez, A. & Rodríguez-Torres, I.  Cr(VI) removal by continuous electrodeionization: study of its basic technologies. Desalination 249 (1), 423–428. Arar, O., Yuksel, U., Kabay, N. & Yuksel, M.  Removal of Cu2þ ions by a micro-flow electrodeionization (EDI) system. Desalination 277 (1–3), 296–300. Arar, O., Yuksel, O., Kabay, N. & Yuksel, M.  Demineralization of geothermal water reverse osmosis (RO) permeate by electrodeionization (EDI) with layered bed configuration. Desalination 317, 48–54. Arar, O., Yuksel, U., Kabay, N. & Yuksel, M.  Various applications of electrodeionization (EDI) method for water treatment – a short review. Desalination 342, 16–22. Aryanti, P. T. P., Yustiana, R., Purnama, R. E. D. & Wenten, I. G.  Performance and characterization of PEG400 modified PVC ultrafiltration membrane. Membrane Water Treatment 6 (5), 379–392. Aryanti, P., Joscarita, S. R., Wardani, A. K., Subagjo, S., Ariono, D. & Wenten, I. G.  The influence of PEG400 and acetone on polysulfone membrane morphology and fouling behaviour. Journal of Engineering and Technological Sciences 48 (2), 135–149. Auerswald, D.  Regeneration without chemicals. Ultrapure Water 11 (3), 52–61. Auerswold, D.  Optimizing the performance of a reverse osmosis/continuous electrodeionization system. Ultrapure Water 13 (4), 34–52. Bennett, A.  Water processes and production: high and ultrahigh purity water. Filtration & Separation 46 (2), 24–27.

Page 374

Water Science & Technology

75.12

2017

Bohus, V., Tóth, E. M., Székely, A. J., Makk, J., Baranyi, K., Patek, G., Schunk, J. & Márialigeti, K.  Microbiological investigation of an industrial ultra pure supply water plant using cultivation-based and cultivation-independent methods. Water Research 44 (20), 6124–6132. Bouhidel, K. E. & Lakehal, A.  Influence of voltage and flow rate on electrodeionization (EDI) process efficiency. Desalination 193 (1–3), 411–421. Chen, V., Fane, A. G., Madaeni, S. & Wenten, I. G.  Particle deposition during membrane filtration of colloids: transition between concentration polarization and cake formation. Journal of Membrane Science 125 (1), 109–122. Doyen, A., Roblet, C., L’Archevêque-Gaudet, A. & Bazinet, L.  Mathematical sigmoid-model approach for the determination of limiting and over-limiting current density values. Journal of Membrane Science 452, 453–459. FILMTEC Technical Manual Excerpt. System Design: Materials of Construction, Corrosion Control. DOW Chemical Company. Franken, T.  Ultrapure water: more than membrane technology alone. Membrane Technology 1999 (105), 9–12. Fritzmann, C., Löwenberg, J., Wintgens, T. & Melin, T.  Stateof-the-art of reverse osmosis desalination. Desalination 216 (1–3), 1–76. Fu, L., Wang, J. & Su, Y.  Removal of low concentrations of hardness ions from aqueous solutions using electrodeionization process. Separation and Purification Technology 68 (3), 390–396. Grabowski, A., Zhang, G., Strathmann, H. & Eigenberger, G.  The production of high purity water by continuous electrodeionization with bipolar membranes: influence of the anion-exchange membrane permselectivity. Journal of Membrane Science 281 (1–2), 297–306. Harfst, W.  Stringent water purity requirements for the pharmaceutical industry. Ultrapure Water 10 (7), 21–22. Helfferich, F. G.  Ion Exchange. Courier Corporation, New York. Hernon, B. P., Zanapalidou, R. H., Zhang, L., Siwak, L. R. & Schoepke, E.  Application of electrodeionization in ultrapure water production: performance and theory. In: Proceedings, 55th Annual Meeting International Water Conference, Pittsburg, PA, pp. 76–78. Jia, Z., Liu, Z. & He, F.  Synthesis of nanosized BaSO4 and CaCO3 particles with a membrane reactor: effects of additives on particles. Journal of Colloid and Interface Science 266 (2), 322–327. Khoiruddin, Widiasa I. N. & Wenten, I. G. a Removal of inorganic contaminants in sugar refining process using electrodeionization. Journal of Food Engineering 133, 40–45. Khoiruddin, K., Hakim, A. N. & Wenten, I. G. b Advances in electrodeionization technology for ionic separation - a review. Membrane Water Treatment 5 (2), 87–108. Kucera, J.  Reverse Osmosis: Industrial Processes and Applications. John Wiley & Sons, New Jersey. Lee, J. H. & Choi, J. H.  The production of ultrapure water by membrane capacitive deionization (MCDI) technology. Journal of Membrane Science 409–410, 251–256. Lee, J. W., Yeon, K. H., Song, J. H. & Moon, S. H.  Characterization of electroregeneration and determination of

2899

A. K. Wardani et al.

Combined UF-EDI for high purity water production

optimal current density in continuous electrodeionization. Desalination 207 (1–3), 276–285. Lee, K. P., Arnot, T. C. & Mattia, D.  A review of reverse osmosis membrane materials for desalination – development to date and future potential. Journal of Membrane Science 370 (1–2), 1–22. Liang, L.-S., Wood, J. & Haas, W.  Design and performance of electrodeionization systems in power plant applications. Ultrapure Water 9, 41. Liberman, B.  The importance of energy recovery devices in reverse osmosis desalination, In: The Future of Desalination in Texas Volume 2: Technical Papers, Case Studies and Desalination Technology Resources. Texas Water Development Board, Austin, Texas. Lu, H., Wang, J., Yan, B. & Bu, S.  Recovery of nickel ions from simulated electroplating rinse water by electrodeionization process. Water Science and Technology 61 (3), 729–735. Lu, H., Wang, Y. & Wang, J.  Recovery of Ni2þ and pure water from electroplating rinse wastewater by an integrated twostage electrodeionization process. Journal of Cleaner Production 92, 257–266. Matzan, E., Maitino, P. & Tate, J.  Deionization: cost reduction and operating results of an RO/EDI treatment system. Ultrapure Water 18 (8), 20–24. Metcalf, E.  Wastewater Engineering: Treatment and Reuse. McGraw-Hill, Boston. Meyer, N., Parker, W. J., Van Geel, P. J. & Adiga, M.  Development of an electrodeionization process for removal of nitrate from drinking water. Part 2: multi-species testing. Desalination 175 (2), 167–177. Mika, A. M., Childs, R. F. & Dickson, J. M.  Ultra-low pressure water softening: a new approach to membrane construction. Desalination 121 (2), 149–158. Mulder, M.  Basic Principles of Membrane Technology. Kluwer Academic Publishers, Dordrecht, The Netherlands. Nagarale, R. K., Gohil, G. S. & Shahi, V. K.  Recent developments on ion-exchange membranes and electromembrane processes. Advances in Colloid and Interface Science 119 (2–3), 97–130. Prato, T. & Gallagher, C.  Using EDI to meet the needs of pure water production. 61st Annual International Water Conference, Pittsburgh, PA. Scott, K.  Handbook of Industrial Membranes. Elsevier, Oxford, UK. Singh, R.  Production of high-purity water by membrane processes. Desalination and Water Treatment 3 (1–3), 99–110. Song, J. H., Song, M. C., Yeon, K. H., Kim, J. B., Lee, K. J. & Moon, S. H.  Puriﬁcation of a primary coolant in a nuclear power plant using a magnetic ﬁlter–electrodeionization hybrid separation system. Journal of Radioanalytical and Nuclear Chemistry 262 (3), 725–732. Song, J. H., Yeon, K. H., Cho, J. & Moon, S. H.  Effects of the operating parameters on the reverse osmosiselectrodeionization performance in the production of high purity water. Korean Journal of Chemical Engineering 22 (1), 108–114.

Water Science & Technology

75.12

2017

Strathmann, H.  Ion-Exchange Membrane Separation Processes. Elsevier, Amsterdam. Strathmann, H.  Electrodialysis, a mature technology with a multitude of new applications. Desalination 264 (3), 268– 288. Sun, X., Lu, H. & Wang, J.  Brackish water desalination using electrodeionization reversal. Chemical Engineering and Processing: Process Intensification 104, 262–270. Tabatabai, A., Scamehorn, J. F. & Christian, S. D.  Water softening using polyelectrolyte-enhanced ultrafiltration. Separation Science and Technology 30 (2), 211–224. Tachino, K., Tsubota, Y., Takechi, S., Nakajima, J., Yamashita, M., Isshiki, K., Fukumura, T. & Ukena, Y.  Hardness indicator. US Patent 6190611 B1. Wang, J., Wang, S. & Jin, M.  A study of the electrodeionization process high-purity water production with a RO/EDI system. Desalination 132 (1–3), 349–352. Wardani, A. K., Hakim, A. N., Khoiruddin, Destifen, W., Goenawan, A. & Wenten, I. G.  Study on the influence of applied voltage and feed concentration on the performance of electrodeionization in nickel recovery from electroplating wastewater. AIP Conference Proceedings 1805 (1), 030004. Wei, D. S., Hossain, M. & Saleh, Z. S.  Separation of polyphenolics and sugar by ultrafiltration: effects of operating conditions on fouling and diafiltration. International Journal of Chemical and Biomolecular Engineering 1, 10–17. Wen, R., Deng, S. & Zhang, Y.  The removal of silicon and boron from ultra-pure water by electrodeionization. Desalination 181 (1–3), 153–159. Wenten, I. G. & Khoiruddin  Reverse osmosis applications: prospect and challenges. Desalination 391, 112–125. Wenten, I. G., Khoiruddin, Arfianto, F. & Zudiharto  Bench scale electrodeionization for high pressure boiler feed water. Desalination 314, 109–114. Widiasa, I. N., Sutrisna, P. D. & Wenten, I. G.  Performance of a novel electrodeionization technique during citric acid recovery. Separation and Purification Technology 39 (1–2), 89–97. Wood, J., Gifford, J., Arba, J. & Shaw, M.  Production of ultrapure water by continuous electrodeionization. Desalination 250 (3), 973–976. Xing, Y., Chen, X. & Wang, D.  Variable effects on the performance of continuous electrodeionization for the removal of Cr(VI) from wastewater. Separation and Purification Technology 68 (3), 357–362. Yeon, K. H., Song, J. H. & Moon, S. H.  A study on stack configuration of continuous electrodeionization for removal of heavy metal ions from the primary coolant of a nuclear power plant. Water Research 38 (7), 1911–1921. Zhang, Z., Liba, D., Alvarado, L. & Chen, A.  Separation and recovery of Cr(III) and Cr(VI) using electrodeionization as an efficient approach. Separation and Purification Technology 137, 86–93. Zuo, W., Zhang, G., Meng, Q. & Zhang, H.  Characteristics and application of multiple membrane process in plating wastewater reutilization. Desalination 222 (1), 187–196.

First received 2 February 2017; accepted in revised form 8 March 2017. Available online 22 March 2017

Page 375

Water Science & Technology

Water Supply

ISSN 1606-9749 iwaponline.com/ws

Water Science and Technology: Water Supply covers all aspects of water supply. The journal’s scope includes: • Management of water resources (including reservoirs) • Water treatment technologies, including wastewater reuse and recycling • Water distribution systems • Drinking water quality • Water pollutants, characteristics and effects • Water utility management including economic and social aspects • Climate change effects on water • Norms and standards For more details, visit iwaponline.com/ws

Page 378

107

17.1

2017

Iron based sustainable greener technologies to treat cyanobacteria and microcystin-LR in water Virender K. Sharma, Long Chen, Blahoslav Marsalek, Radek Zboril, Kevin E. O’Shea and Dionysios D. Dionysiou

ABSTRACT The presence of the toxic cyanobacteria and cyanotoxin, microcystin-LR (MC-LR) and other cyanotoxins, in drinking water sources poses a serious risk to public health. Iron based technologies using magnetic zero-valent iron nanoparticles (nZVI) and ferrate ion (FeVIO2� 4 , Fe(VI)) represent greener approaches to remove cyanobacteria and degrade MC-LR in water. This paper reveals that nanoparticles of zero valent iron (nZVI) can destroy cyanobacteria in the source water and may play a preventive role in terms of the formation of cyanobacterial water blooms by removing nutrients like phosphate. Results on MC-LR showed that Fe(VI) was highly effective in removing MC-LR in water. Products studies on the oxidation of MC-LR by Fe(VI) demonstrated decomposition of the MC-LR structure. Signiﬁcantly, degradation byproducts of MC-LR did not contain signiﬁcant biological toxicity. Moreover, Fe(VI) was highly effective for the degradation of MC-LR in lake water samples. Mechanisms of removal and destruction of target contaminants by nZVI and Fe(VI) are discussed. Key words

| detoxiﬁcation, ferrate, harmful algal bloom, microcystin, oxidation, zero valent iron nanoparticles

Virender K. Sharma (corresponding author) Long Chen Department of Environmental and Occupational Health, School of Public Health, Texas A&M University, 1266 TAMU, College Station, TX, USA E-mail: vsharma@sph.tamhsc.edu Blahoslav Marsalek Academy of Sciences of the Czech Republic, Institute of Botany, Lidická 25/27, Brno 65720, Czech Republic Blahoslav Marsalek Radek Zboril Regional Centre of Advanced Technologies and Materials, Department of Physical Chemistry, Faculty of Science, Palacky University, Slechtitelu 11, Olomouc 78371, Czech Republic Kevin E. O’Shea Department of Chemistry and Biochemistry, Florida International University, Miami, FL, USA Dionysios D. Dionysiou Department of Biomedical, Chemical and Environmental Engineering (DBCEE), 705 Engineering Research Center, University of Cincinnati, Cincinnati, OH, USA

INTRODUCTION Cyanobacteria have critical functions in terrestrial and aquatic

produce heptatoxic MCs, which are stable in water. MCs

ecosystems, which include oxygen evolution, ﬁxation of nitro-

have been found in drinking waterbodies worldwide, causing

gen and carbon dioxide, and biomass production (Huo et al.

potential risk to human health (Li et al. ). Moreover,

). However, cyanobacteria are associated with many

MCs can easily accumulate in aquatic biota which has impli-

serious environmental problems, which have implications for

cations for human and environmental health. Among the

water quality and public health (Adamovsky et al. ). Cyano-

various MCs, MC-LR is the most common of the microcystins.

bacteria generate many toxins such as microcystins (MCs),

MC-LR is of great concern in water bodies due to its acute tox-

cylindrospermopsin, anatoxins nodularins, and saxitoxins,

icity (LD50 ¼ 50 μg kg�1 in mice) (Sharma et al. ). The

which can cause a signiﬁcant health hazard in drinking water (Sharma et al. ). Toxic effects include hepatotoxicity, cytotoxicity, neurotoxicity, embryotoxicity, dermatotoxicity or immunotoxicity.

Additionally,

cyanobacteria

commonly

World Health Organization has established a provisional guideline limit of 1 μg L�1 for MC-LR (Ibelings et al. ). In recent years, several technologies have been sought to

remove extracellular cyanobacteria and MC-LR in water

doi: 10.2166/ws.2016.115

Page 379

108

V. K. Sharma et al.

Water Science & Technology: Water Supply

Iron based sustainable greener technologies

17.1

2017

(Sharma et al. ; Jiang et al. ). Treatment may be classi-

capability to remove various contaminants from ground-

ﬁed into two categories: physical removal and oxidative

water and wastewater. In the past few years, emphasis has

transformation. Activated carbon, coagulation–ﬂocculation–

been placed on the nano ZVI nanoparticles (nZVI), which

sedimentation, and sand and membrane ﬁltration are examples

have shown remarkable reduction properties to remediate

of physical–chemical methods. Applications of physical pro-

numerous inorganic and organic contaminants (Crane &

cesses generally need replacement of materials (e.g. activated

Scott ; Yan et al. ; Jarošová et al. ). Small size,

carbon and membranes) and/or cleaning because of fouling.

large surface area, good transport properties and speciﬁc

Oxidation methods include UV-based advanced oxidation

mechanism of reaction with water under anaerobic con-

technologies, photocatalytic and chemical oxidation (Sharma

ditions are key properties for its effectiveness to remove

et al. , ). MC-LR is stable under natural sunlight and

contaminants (Klimkova et al. ; Mueller et al. ; Ray-

resistant to degradation by UV radiation (Westrick et al.

choudhury & Scheytt ; Yan et al. ; Filip et al. ;

). Photocatalytic degradation of MC-LR using titanium

Baikousi et al. ; Jarošová et al. ; Soukupova et al.

dioxide (TiO2) is promising (Sharma et al. ); however, it

). In recent years, developing composite materials con-

requires separation of the catalyst after removal of MC-LR

taining nZVI has been emphasized to enhance removing

and needs additional energy for the photoactivation of photo-

contaminants due to the combination of reduction/sorption

catalyst. Chlorine, chlorine dioxide, chloramine, ozone, and

or reduction/antimicrobial properties of hybrids (Marková

permanganate have been applied to remove MC-LR in water

et al. ; Petala et al. ; Baikousi et al. ). ZVI has

(Sharma et al. ). The reaction of chlorine with MC-LR

also shown inactivation of bacteria like Escherichia coli

resulted in chlorine substitution, which generates potentially

(Lee et al. ). In recent years, the role of nZVI in the

toxic chlorinated by-products (Acero et al. ; Huang et al.

destruction of cyanobacterial cells was explored (Marsalek

). In addition, the possibility of the reaction between chlor-

et al. ).

ine and bromide ion produces HOBr, which can produce toxic

Application of nZVI to remove cyanobacteria was car-

brominated by-products (Heeb et al. ). The degradation of

ried out using water inoculated with a Microcystis

MC-LR by chloramine was not signiﬁcant. Chlorine dioxide is

aeruginosa laboratory strain that remained in the colonial

capable of degrading MC-LR, but high doses are required. This

form (CCT12/2—8) (Marsalek et al. ). The average par-

limits the practical application due to the generation of chlorite

ticle size and surface area of applied nZVI were ∼70 nm

and chlorate as by-products after the use of chlorine dioxide

and ∼25 m2/g, respectively. Detailed chemical, microscopic,

(Kull et al. ). Applications of ozone and permanganate in

and microbiological analyses were performed (Marsalek

oxidizing MC-LR are promising (Sharma et al. ).

et al. ). The results on nZVI treatment of cyanobacteria

This paper deals with zero valent iron nanoparticles

showed multiple modes of action: (i) the removal of bioavail-

(nZVI) and high valent tetraoxy compound of iron (ferrate,

able phosphorus, (ii) the destruction of cyanobacterial cells,

O2� 4 ,

Fe(VI)) based technologies to treat cyanobacteria

and (iii) the immobilization of MCs (Marsalek et al. ).

and MC-LR in water (Marsalek et al. ; Jiang et al. ).

Release of cyanobacteria may thus be inﬂuenced by nZVI.

Both of these technologies are environmentally friendly and

Signiﬁcantly, the ecotoxicological study demonstrated that

can address some of the drawbacks of other treatment

nZVI was a highly selective agent (EC50 ¼ 50 mg/L against

methods. The effect of nZVI and Fe(VI) ions in treating cya-

cyanobacteria). This level of EC50 was 20–100 times lower

nobacteria and MC-LR under various environmental

than that the EC50 for ﬁsh, water plants, algae, and daph-

conditions are reviewed.

nids. Figure 1 shows the deformation of cells caused by the aggregated Fe(OH)3, which was generated as the major product from the nZVI treatment of cyanobacteria (Marsa-

ZVI NANOPARTICLES

lek et al. ). Furthermore Fe(OH)3, a nontoxic product, was capable of promoting ﬂocculation, resulting in gradual

ZVI has received tremendous interest in removing various

settling

contaminants (Bae & Hanna ). ZVI has a high

(Figure 1).

Page 380

the

decomposed

cyanobacterial

biomass

109

Figure 1

V. K. Sharma et al.

Water Science & Technology: Water Supply

Iron based sustainable greener technologies

17.1

2017

(a) Scanning electron microscope (SEM) images of cyanobacteria before treatment, (b) unused nZVI particles, (c) highly deformed cells after brief exposure to nZVI, and (d) completely destroyed cells surrounded by ferric oxide aggregates. (Adapted from Marsalek et al. (2012) with the permission of the American Chemical Society.)

FERRATE ION

oxide, generated from Fe(VI), acts as an efﬁcient coagulant to remove humic acids, radionuclides, metals, arsenic and non-

Fe(VI) ion in the aquatic environment has strong oxidation

metals (Horst et al. ; Prucek et al. , ). Fe(VI) as a

capability (Sharma ). For example, the redox potential of

pre-oxidant is able to decrease the concentration of disinfection

Fe(VI) in aqueous solution is the highest among other conven-

byproducts, formed during chlorination of water (Gan et al.

tional disinfectant and oxidants used in water and wastewater

; Yang et al. ). Recently, research in our laboratories

treatment (Jiang & Lloyd ). Numerous examples have dis-

has focused on the role of Fe(VI) in removing and oxidatively

played simultaneously disinfection, oxidation, and coagulation

transforming toxins such as MC-LR. Below is the summary of

properties of Fe(VI) (Eng et al. ; Sharma a; Filip et al.

results observed in studying the kinetics and oxidized products

; Jiang ; Prucek et al. ; Sharma et al. ). In a single

(OPs) and their toxicity in the oxidation of MC-LR by Fe(VI)

Fe(VI) dose treatment, inactivation of microorganisms, oxi-

(Jiang et al. ).

dative transformation of inorganic and organic contaminants, and toxins, as well as removal of toxic metals and phosphate

Kinetics

can be achieved (Sharma ; Jiang , ; Yates et al. ; Sharma et al. ). Fe(VI) as a disinfectant can inactivate

The oxidation of the MC-LR by Fe(VI) followed a second-

a wide range of microorganisms (Sharma b; Jiang ).

order kinetics (-d[Fe(VI)]/dt ¼ kapp[Fe(VI)][MC-LR]). The

The kinetics of the reactions with various pollutants with a var-

values of kapp showed a pH dependence with values ranged

iety of molecular and structural conﬁgurations (e.g. sulﬁde,

from

bisulﬁte, iodide, cyanides, ammonia, selenium, arsenic, azide,

0.08 mol�1 Ls�1 at pH 10.0. This indicates a rapid degra-

thiols, amines, amino acids) showed the feasibility of their

dation of MC-LR (Jiang et al. ). The comparison of the

removal by Fe(VI) (Lee et al. , ; Lee & von Gunten

rate constants for the oxidation of MC-LR with different oxi-

; Sharma , ; Zimmermann et al. ). The ferric

dants at neutral pH is presented in Table 1 (Kull et al. ;

1.3 ± 0.1 × 102 mol�1 Ls�1

7.5

8.1 ±

Page 381

110

V. K. Sharma et al.

Table 1

Water Science & Technology: Water Supply

Iron based sustainable greener technologies

Second-order rate constants and half-lives for oxidation of MC-LR by different oxidants at 22–25 C

17.1

2017

and trihydroxylation products. Hydroxylation of the

carbon–carbon double bond in the MHDA moiety also kapp M

k, Species

Ferrate(VI)a

HFeO� 4

(3.9 ± 0.2) × 102

(pK3 ¼ 7.23)

FeO2� 4

8.0 ± 2.0

Permanganateb

MnO� 4

Chlorinec (pKa ¼ 7.54)g

Chlorine dioxided e

Ozone a

�1 �1

Oxidant

�1 �1

t1/2

pH 7.0

2.5 × 102

155 s

3.6 × 102

107 s

HOCl

1.2 × 102

7.2 × 101

504 s

OCl�

6.8

ClO2

1.0

13.1 h

4.1 × 10 �1

4.1 × 10

[FeO2� 4 ]

0.08 s

�1

This study and half-life at dose [Fe] ¼ 1 mg L or ¼ 2.2 mg L . �1 From Rodríguez et al. (2007) and half-life at dose [Mn] ¼ 1 mg L�1 or [MnO� . 4 ] ¼ 2.2 mg L

Values at pH 7.2 taken from Acero et al. (2005) and half-life at [HOCl] ¼ 1 mg L�1.

d e

�1

From Kull et al. (2004) and half-life at [ClO2] ¼ 1 mg L

From Onstad et al. (2007) and half-life at [O3] ¼ 1.0 mg L�1. f At 25 C. W

tional group through the elimination of an H atom. Tautomerization of the enol group gave a chiral center at the alpha position. A pair of diastereotopic isomers consistent

b c

occurred, which resulted in the formation of an enol func-

At 25 C from Carrell Morris (1966). W

with m/z 1011.5510 (M þ 16) were thus obtained (Table 2). Fe(VI) also oxidized the diene group of the Adda moiety of

MC-LR via dihydroxylation to yield products with M þ 34 (Table 2). This corresponded to addition of two HO groups

without loss of H atoms. Hydroxylation yielded 1,2-, 3,4-, and 1,4-diol products (Table 2). Signiﬁcantly, the products seen from the attack on diene moiety were also reported in oxidation performed by photocatalytic and electrochemical process (Antoniou et al. ; Zhang et al. ; Zong et al. ; Liao et al. ). The cleavage of peptide bonds in MC-LR by Fe(VI) was also seen, which caused the hydrolysis

Acero et al. ; Onstad et al. ; Rodríguez et al. ).

of amide bonds of D-glu-MDHA and the

Ozone showed the highest value of kapp (Table 1). The efﬁ-

D-Asp

cient attack on the double bond of MC-LR by ozone may be

of the attack of Fe(VI) on the peptide bond was similar to

responsible for orders of magnitude faster reactivity in com-

the transformation of -NH ¼ C- amino acid functionality by

parison with other oxidants. The increasing order of the

L-Arg-Methyl

of the MC-LR by Fe(VI) ( Jiang et al. ). This step

Fe(V) (Bielski et al. ; Rush & Bielski ).

reactivity with MC-LR may be presented as chlorine dioxide < chlorine < Fe(VI) < Mn(VII) < O3 (Table 1). The half-life

Removal and biological toxicity assessment tests

(t1/2) for oxidizing MC-LR by O3 is less than a second whereas Fe(VI), Mn(VII), and chlorine oxidize MC-LR in seconds.

Removal of MC-LR by Fe(VI) was conﬁrmed by conducting

Ozone, chlorine, Mn(VII), and Fe(VI) are thus suitable oxi-

tests in water and lake water samples (Brno, Czech Republic)

dants to eliminate MC-LR in water treatment.

(Jiang et al. ). The lake water had total organic carbon of 7.9 mg L�1. In performing tests, the water samples were

OPs

spiked with MC-LR (25.0 μg L�1) and an addition of FeO2� 4 into the samples was 5.0 mg L�1. The removal of MC-LR in

Analysis of OPs of degradation of MC-LR was carried out by

deionized water was almost complete over the entire pH

high resolution liquid chromatography–mass spectrometry/

range of 6.0–8.0 at 20 C (Figure 3). At pH 7.0 and 8.0, the

mass spectrometry technique ( Jiang et al. ). The proposed

removal percentages were >99.0% (or <1 μg L�1), while a

structures for the OPs were based on the molecular formula

slight decrease at pH 6.0 (96.2%) was noticed. As shown

and are summarized in Table 2. Basically, four primary reac-

could in Figure 3, an Fe(VI) dose of 5.0 mg L�1 as FeO2� 4

tions occurred from the attacks of Fe(VI) on the aromatic

remove ∼ 75% in lake water at pH 7.0. This indicates that

ring, diene, enone, and amide functionalities of MC-LR by

the other components present in the lake water (e.g. dissolved

Fe(VI) (Figure 2). In hydroxylation of the aromatic ring,

organic matter) may also be reacting with Fe(VI) (Horst et al.

mono, di and trihydroxylation of the aromatic ring were

). Fe(VI) dose >5.0 mg L�1 would be required to comple-

obtained with corresponding m/z ¼ 1011.5510, 1027.5459,

tely remove MC-LR in the lake water ( Jiang et al. ).

and 1043.5408 (Table 2). Monohydroxylation involved the

The MC-LR is an inhibitor of protein phosphatase (PP1

loss of a hydrogen atom to yield a highly stabilized aromatic

and PP2A) enzymes, therefore, the PP1 inhibition was uti-

product (M þ 16). Further hydroxylation thus formed di-

lized to evaluate the biological activity of the Fe(VI)

Page 382

111

Table 2

V. K. Sharma et al.

Iron based sustainable greener technologies

Water Science & Technology: Water Supply

17.1

2017

OPs observed during hydroxylation of moieties of MC-LR by Fe(VI)

Moiety

OPs

Benzene ring

MDHA

Diene

treated solutions ( Jiang et al. ). When MC-LR was totally

and Fe(VI) ions in the system enhanced the photocatalytic oxi-

removed by Fe(VI) ion, the biological activity of OP was

dation of MC-LR (Figure 4). The effectiveness of Fe(VI) ion

almost completely eliminated. This demonstrated that the

was more than that of Fe(III) ion. Signiﬁcantly, complete

OPs of MC-LR were not biologically toxic ( Jiang et al. ).

removal of MC-LR was achieved by Fe(VI) in 30 min (Figure 4).

The removal of MC-LR by photocatalytic oxidation system

Formation of highly reactive intermediate Fe(V) species and

was also sought (Yuan et al. ; Sharma et al. ). Figure 4

also increasing amount of holes (i.e. oxidant) in iron species

shows the results in the TiO2-UV-MC-LR, Fe(III)-TiO2-UV-

containing TiO2 photocatalytical systems may have resulted

MC-LR, and Fe(VI)-TiO2-UV-MC-LR systems. Both Fe(III)

in enhanced removal of MC-LR (Sharma et al. ). Page 383

112

V. K. Sharma et al.

Water Science & Technology: Water Supply

Iron based sustainable greener technologies

17.1

2017

CONCLUSIONS

•

Both nZVI and Fe(VI) showed their potential as sustainable green materials to remove cyanobacteria and cyanotoxins in water.

•

nZVI was highly effective in destroying cyanobacteria via multiple modes of action.

•

The products of MC-LR oxidation by Fe(VI) were observed from the hydroxylation of benzene ring, diene, enone, and peptide bond of MC-LR, which did not have

Figure 2

any signiﬁcant toxicity.

Fe(VI) attacks on different moieties of MC-LR. (Adapted from Jiang et al. (2014) with the permission of the American Chemical Society.)

•

Fe(VI) could degrade MC-LR in water and lake water samples on a time scale of seconds.

•

Magnetic separation of generated iron oxides from nZVI and Fe(VI) treatment can be achieved using a cost effective low-gradient magnetic ﬁeld.

ACKNOWLEDGEMENTS The authors acknowledge the support of the United States National Science Foundation (CBET-1439314, 1236209, and 1235803) for this research. V. K. Sharma, R. Zboril, and B. Marsalek also acknowledge the support of the Operational Program Research and Development for Innovations–European

Regional

Development

Fund

(CZ.1.05/2.1.00/03.0058) and of the Technological Agency Figure 3

Removal of MC-LR in deionized water and lake water by Fe(VI) ([MC-LR] ¼ 25.0 μg L�1, [FeO24�] ¼ 5.0 mg L�1, and temperature 20 C). (Adapted from Jiang W

et al. (2014) with the permission of the American Chemical Society.)

of the Czech Republic–the project Environmental Friendly Nanotechnologies and Biotechnologies in Water and Soil Treatment (TE01020218).

REFERENCES

Figure 4

�1

The photocatalytic degradation of MCLR. Conditions: [ferrate(VI)] ¼ 0.08 mmol L � and Fe(III) ¼ 0.36 mmol L 1. (Adapted from Sharma et al. (2010) with the

permission of Springer Inc.)

Page 384

Acero, J. L., Rodriguez, E. & Meriluoto, J.  Kinetics of reactions between chlorine and the cyanobacterial toxins microcystins. Water Res. 39 (8), 1628–1638. Acero, J. L., Rodríguez, E., Majado, M. E., Sordo, A. & Meriluoto, J.  Oxidation of microcystin-LR with chlorine and permanganate during drinking water treatment. J. Water Supply: Res. Technol. –AQUA 57 (6), 371–380. Adamovsky, O., Moosova, Z., Pekarova, M., Basu, A., Babica, P., Svihalkova Sindlerova, L., Kubala, L. & Blaha, L.  Immunomodulatory potency of microcystin, an important water-polluting cyanobacterial toxin. Environ. Sci. Technol. 49 (20), 12457–12464.

113

V. K. Sharma et al.

Iron based sustainable greener technologies

Antoniou, M. G., Shoemaker, J. A., de la Cruz, A. A. & Dionysiou, D. D.  LC/MS/MS structure elucidation of reaction intermediates formed during the TiO2 photocatalysis of microcystin-LR. Toxicon 51 (6), 1103–1118. Bae, S. & Hanna, K.  Reactivity of nanoscale zero-valent iron in unbuffered systems: effect of pH and Fe(II) dissolution. Environ. Sci. Technol. 49, 1036–1045. Baikousi, M., Georgiou, Y., Daikopoulos, C., Bourlinos, A. B., Filip, J., Zboril, R., Deligiannakis, Y. & Karakassides, M. A.  Synthesis and characterization of robust zero valent iron/mesoporous carbon composites and their applications in arsenic removal. Carbon 93, 636–647. Bielski, B. H. J., Sharma, V. K. & Czapski, G.  Reactivity of ferrate(V) with carboxylic acids: a pre-mix pulse radiolysis study. Radiat. Phys. Chem. 44 (5), 479–484. Carrell Morris, J.  The acid ionization constant of HOCl from 5 to 35 C. J. Phys. Chem. 70, 3798–3805. Crane, R. A. & Scott, T. B.  Nanoscale zero-valent iron: future prospects for an emerging water treatment technology. J. Hazard. Mater. 211–212, 112–125. Eng, Y. Y., Sharma, V. K. & Ray, A. K.  Ferrate(VI): green chemistry oxidant for degradation of cationic surfactant. Chemosphere 63 (10), 1785–1790. Filip, J., Yngard, R. A., Siskova, K., Marusak, Z., Ettler, V., Sajdl, P., Sharma, V. K. & Zboril, R.  Mechanisms and efﬁciency of the simultaneous removal of metals and cyanides by using ferrate(VI): crucial roles of nanocrystalline iron(III) oxyhydroxides and metal carbonates. Chem. Eur. J. 17 (36), 10097–10105. Filip, J., Karlický, F., Marušák, Z., Lazar, P., Cerník, M., Otyepka, M. & Zboril, R.  Anaerobic reaction of nanoscale zerovalent iron with water: mechanism and kinetics. J. Phys. Chem. C 118 (25), 13817–13825. Gan, W., Sharma, V. K., Zhang, X., Yang, L. & Yang, X.  Investigation of disinfection byproducts formation in ferrate (VI) pre-oxidation of NOM and its model compounds followed by chlorination. J. Hazard. Mater. 292, 197–204. Heeb, M. B., Criquet, J., Zimmermann-Steffens, S. G. & Von Gunten, U.  Oxidative treatment of bromide-containing waters: formation of bromine and its reactions with inorganic and organic compounds – a critical review. Water Res. 48 (1), 15–42. Horst, C., Sharma, V. K., Clayton Baum, J. & Sohn, M.  Organic matter source discrimination by humic acid characterization: synchronous scan ﬂuorescence spectroscopy and ferrate(VI). Chemosphere 90 (6), 2013–2019. Huang, T., Zhao, J. & Chai, B.  Mechanism studies on chlorine and potassium permanganate degradation of microcystin-LR in water using high-performance liquid chromatography tandem mass spectrometry. Water Sci. Technol. 58 (5), 1079–1084. Huo, X., Chang, D. W., Tseng, J. H., Burch, M. D. & Lin, T. F.  Exposure of microcystis aeruginosa to hydrogen peroxide under light: kinetic modeling of cell rupture and simultaneous microcystin degradation. Environ. Sci. Technol. 49 (9), 5502–5510. W

Water Science & Technology: Water Supply

17.1

2017

Ibelings, B. W., Backer, L. C., Kardinaal, W. E. A. & Chorus, I.  Current approaches to cyanotoxin risk assessment and risk management around the globe. Harmful Algae 40, 63–74. Jarošová, B., Filip, J., Hilscherová, K., Tucek, J., Šimek, Z., Giesy, J. P., Zboril, R. & Bláha, L.  Can zero-valent iron nanoparticles remove waterborne estrogens? J. Environ. Manage. 150, 387–392. Jiang, J. Q.  Advances in the development and application of ferrate(VI) for water and wastewater treatment. J. Chem. Technol. Biotechnol. 89, 165–177. Jiang, J. Q.  The role of ferrate(VI) in the remediation of emerging micropollutants: a review. Desalin. Water Treat. 55 (3), 828–835. Jiang, J. Q. & Lloyd, B.  Progress in the development and use of ferrate(VI) salt as an oxidant and coagulant for water and wastewater treatment. Water Res. 36, 1397–1408. Jiang, W., Chen, L., Batchu, S. R., Gardinali, P. R., Jasa, L., Marsalek, B., Zboril, R., Dionysiou, D. D., O’Shea, K. E. & Sharma, V. K.  Oxidation of microcystin-LR by ferrate (VI): kinetics, degradation pathways, and toxicity assessment. Environ. Sci. Technol. 48, 12164–12172. Klimkova, S., Cernik, M., Lacinova, L., Filip, J., Jancik, D. & Zboril, R.  Zero-valent iron nanoparticles in treatment of acid mine water from in situ uranium leaching. Chemosphere 82 (8), 1178–1184. Kull, T. P. J., Backlund, P. H., Karlsson, K. M. & Meriluoto, J. A. O.  Oxidation of the cyanobacterial heptotoxin microcystin-LR by chlorine dioxide: reaction kinetics, characterization, and toxicity of reaction products. Environ. Sci. Technol. 38, 6025–6031. Lee, Y. & von Gunten, U.  Oxidative transformation of micropollutants during municipal wastewater treatment: comparison of kinetic aspects of selective (chlorine, chlorine dioxide, ferrateVI, and ozone) and non-selective oxidants (hydroxyl radical). Water Res. 44, 555–566. Lee, C., Jee, Y. K., Won, I. L., Nelson, K. L., Yoon, J. & Sedlak, D. L.  Bactericidal effect of zero-valent iron nanoparticles on Escherichia coli. Environ. Sci. Technol. 42 (13), 4927–4933. Lee, Y., Zimmermann, S. G., Kieu, A. T. & von Gunten, U.  Ferrate (Fe(VI)) application for municipal wastewater treatment: a novel process for simultaneous micropollutant oxidation and phosphate removal. Environ. Sci. Technol. 43, 3831–3838. Lee, Y., Kissner, Y. & von Gunten, U.  Reaction of ferrate(VI) with ABTS and self-decay of ferrate(VI): kinetics and mechanisms. Environ. Sci. Technol. 48, 5154–5162. Li, X., Zhao, Q., Zhou, W., Xu, L. & Wang, Y.  Effects of chronic exposure to microcystin-LR on hepatocyte mitochondrial DNA replication in mice. Environ. Sci. Technol. 49 (7), 4665–4672. Liao, W., Murugananthan, M. & Zhang, Y.  Electrochemical degradation and mechanistic analysis of microcystin-LR at boron-doped diamond electrode. Chem. Eng. J. 243, 117–126. Marková, Z., Šišková, K. M., Filip, J., Cuda, J., Kolár, M., Šafárová, K., Medrík, I. & Zboril, R.  Air stable magnetic bimetallic Fe-Ag nanoparticles for advanced antimicrobial treatment and phosphorus removal. Environ. Sci. Technol. 47 (10), 5285–5293.

Page 385

114

V. K. Sharma et al.

Iron based sustainable greener technologies

Marsalek, B., Jancula, D., Marsalkova, E., Mashlan, M., Safarova, K., Tucek, J. & Zboril, R.  Multimodal action and selective toxicity of zerovalent iron nanoparticles against cyanobacteria. Environ. Sci. Technol. 46 (4), 2316–2323. Mueller, N. C., Braun, J., Bruns, J., Cerník, M., Rissing, P., Rickerby, D. & Nowack, B.  Application of nanoscale zero valent iron (NZVI) for groundwater remediation in Europe. Environ. Sci. Pollut. Res. 19 (2), 550–558. Onstad, G. D., Strauch, S., Meriluoto, J., Codd, G. A. & Von Gunten, U.  Selective oxidation of key functional groups in cyanotoxins during drinking water ozonation. Environ. Sci. Technol. 41 (12), 4397–4404. Petala, E., Dimos, K., Douvalis, A., Bakas, T., Tucek, J., Zboril, R. & Karakassides, M. A.  Nanoscale zero-valent iron supported on mesoporous silica: characterization and reactivity for Cr(VI) removal from aqueous solution. J. Hazard. Mater. 261, 295–306. Prucek, R., Tuček, J., Kolarí̌ k, J., Filip, J., Marušák, Z., Sharma, V. K. & Zborǐ l, R.  Ferrate(VI)-induced arsenite and arsenate removal by in situ structural incorporation into magnetic iron(III) oxide nanoparticles. Environ. Sci. Technol. 47 (7), 3283–3292. Prucek, R., Tucek, J., Kolarik, J., Huskova, I., Filip, J., Varma, R. S., Sharma, V. K. & Zboril, R.  Ferrate(VI)-prompted removal of metals in aqueous media: mechanistic delineation of enhanced efﬁciency via metal entrenchment in magnetic oxides. Environ. Sci. Technol. 49, 2319–2327. Raychoudhury, T. & Scheytt, T.  Potential of zerovalent iron nanoparticles for remediation of environmental organic contaminants in water: a review. Water Sci. Technol. 68 (7), 1425–1439. Rodríguez, E., Onstad, G. D., Kull, T. P. J., Metcalf, J. S., Acero, J. L. & von Gunten, U.  Oxidative elimination of cyanotoxins: comparison of ozone, chlorine, chlorine dioxide and permanganate. Water Res. 41 (15), 3381–3393. Rush, J. D. & Bielski, B. H. J.  The oxidation of amino acid by ferrate(V). A pre-mix pulse radiolysis study. Free Rad. Res. 22, 571–579. Sharma, V. K.  Potassium ferrate(VI): environmental friendly oxidant. Adv. Environ. Res. 6, 143–156. Sharma, V. K. a Disinfection performance of Fe(VI) in water and wastewater: a review. Water Sci. Technol. 55 (1–2, Wastewater Reclamation and Reuse for Sustainability), 225–232. Sharma, V. K. b Ferrate studies for disinfection and treatment of drinking water. In: Advances in Control of Disinfection By-Products in Drinking Water Systems (A. Nikolaou, L. Rizzo & H. Selcuk, eds). Nova Science Publishers, Hauppauge, New York, pp. 373–380. Sharma, V. K.  Oxidation of nitrogen containing pollutants by novel ferrate(VI) technology: a review. J. Environ. Sci. Health, Part A: Toxic/Hazard. Subst. Environ. Eng. 45, 645–667. Sharma, V. K.  Oxidation of inorganic contaminants by ferrates (Fe(VI), Fe(V), and Fe(IV))- kinetics and mechanisms – a review. J. Environ. Manage. 92, 1051–1073.

Water Science & Technology: Water Supply

17.1

2017

Sharma, V. K.  Ferrate(VI) and ferrate(V) oxidation of organic compounds: kinetics and mechanism. Coord. Chem Rev. 257, 495–510. Sharma, V. K., Graham, N. J. D., Li, X. Z. & Yuan, B. L.  Ferrate (VI) enhanced photocatalytic oxidation of pollutants in aqueous TiO2 suspensions. Environ. Sci. Pollut. Res. 17 (2), 453–461. Sharma, V. K., Triantis, T. M., Antoniou, M. G., He, X., Pelaez, M., Han, C., Song, W., O’Shea, K. E., De La Cruz, A. A., Kaloudis, T., Hiskia, A. & Dionysiou, D. D.  Destruction of microcystins by conventional and advanced oxidation processes: a review. Sep. Purif. Technol. 91, 3–17. Sharma, V. K., Zhao, J. & Hidaka, H.  Mechanism of photocatalytic oxidation of amino acids: Hammett correlations. Catalysis Today 224, 263–268. Sharma, V. K., Zboril, R. & Varma, R. S.  Ferrates: greener oxidants with multimodal action in water treatment technologies. Acc. Chem. Res. 48, 182–191. Sharma, V. K., Chen, L. & Zboril, R.  A review on high valent FeVI (ferrate): a sustainable green oxidant in organic chemistry and transformation of pharmaceuticals. ACS Sustainable Chem. Eng. 4 (1), 18–34. Soukupova, J., Zboril, R., Medrik, I., Filip, J., Safarova, K., Ledl, R., Mashlan, M., Nosek, J. & Cernik, M.  Highly concentrated, reactive and stable dispersion of zero-valent iron nanoparticles: direct surface and site application. Chem. Eng. J. 262, 813–822. Westrick, J. A., Szlag, D. C., Southwell, B. J. & Sinclair, J.  A review of cyanobacteria and cyanotoxins removal/ inactivation in drinking water treatment. Anal. Bioanal. Chem. 397 (5), 1705–1714. Yan, W., Lien, H.-L., Koel, B. E. & Zhang, W.-X.  Iron nanoparticles for environmental clean-up: recent developments and future outlook. Environ. Sci. Process Impacts 15 (1), 63–77. Yang, X., Gan, W., Zhang, X., Huang, H. & Sharma, V. K.  Effect of pH on the formation of disinfection byproducts in ferrate(VI) pre-oxidation and subsequent chlorination. Sep. Purif. Technol. 156, 980–986. Yates, B. J., Zboril, R. & Sharma, V. K.  Engineering aspects of ferrate in water and wastewater treatment – a review. J. Environ. Sci. Health A 49, 1603–1604. Yuan, B., Li, Y., Huang, X., Liu, H. & Qu, J.  Fe(VI)-assisted photocatalytic degradating of microcystin-LR using titanium dioxide. J. Photochem. Photobiol. A 178 (1), 106–111. Zhang, Y., Zhang, Y., Yang, N., Liao, W. & Yoshihara, S.  Electrochemical degradation and mechanistic analysis of microcystin-LR. J. Chem. Technol. Biotechnol. 88 (8), 1529–1537. Zimmermann, S. G., Schmukat, A., Schulz, M., Benner, J., von Gunten, U. & Ternes, T. A.  Kinetic and mechanistic investigations of the oxidation of tramadol by ferrate and ozone. Environ. Sci. Technol. 46 (2), 876–884. Zong, W., Sun, F. & Sun, X.  Oxidation by-products formation of microcystin-LR exposed to UV/H2O2: toward the generative mechanism and biological toxicity. Water Res. 47 (9), 3211–3219.

First received 11 April 2016; accepted in revised form 21 June 2016. Available online 7 July 2016

Page 386

707

17.3

2017

In-stream detection of waterborne priority pollutants, and applications in drinking water contaminant warning systems Andrea G. Capodaglio

ABSTRACT Advancements in real-time water monitoring technologies permit rapid detection of in-stream, inpipe water quality, and alert of threats from waste loads. Legislation mandating the establishment of water resources monitoring, presence of hazardous contaminants in efﬂuents, and perception of the vulnerability of the water distribution system to attacks, have spurred technical and economic

Andrea G. Capodaglio DICAr, University of Pavia, Via Ferrata 3, Pavia 27100, Italy E-mail: capo@unipv.it

interest. Alternatively to traditional analyzers, chemosensors operate according to physical principles, without sample collection (online), and are capable of supplying parameter values continuously and in real-time. This review paper contains a comprehensive survey of existing and expected online monitoring technologies for measurement/detection of pollutants in water. The state-of-the-art in online water monitoring and contaminant warning systems is presented. Application examples are reported. Monitoring costs will become a lesser part of a water utility budget due to the fact that automation and technological simpliﬁcation will abate human cost factors, and reduce the complexity of laboratory procedures. Key words

| contaminant warning systems, dangerous pollutants, emerging pollutants, instrumentation, online monitoring, pollutants

INTRODUCTION Rapid and constant advancement of real-time water moni-

monitor network conditions, warning of potential contami-

toring and sensing technologies will make these an ever

nation events.

more important tool for the evaluation of online, in-pipe

Water-quality monitoring programs represent a balance

water quality, and the assessment of related life and health

between several factors: analytical capacity, collection, proces-

risks. Established technologies are now permitting rapid

sing, and maintenance of representative samples, and available

detection of water quality changes, health and environ-

resources, including technical, human and ﬁnancial. While

mental threats induced by waste loads, and other impacts.

monitoring has been traditionally driven by the development

Initially used in lieu of traditional monitoring mainly for

of increasingly more sophisticated analytical equipment allow-

reporting purposes, these instruments have recently been

ing lower detection limits and new constituent analyses, this

supplemented by speciﬁc software and interconnecting

increased capacity has often clashed against a number of limit-

networks to become true online quality monitoring of distri-

ations, including ﬁnancial investment availability for purchase,

bution systems, also known under the names of contaminant

and the capacity to collect uncontaminated and/or representa-

warning systems (CWS), or water quality event detection

tive samples suitable for the new technologies. An alternative

systems (EDS). These consist of an integrated system of sen-

to these traditional chemical cabinet analyzers is the use of

sors, supervisory tools and data acquisition, to continuously

chemosensors, which operate according to physical principles

doi: 10.2166/ws.2016.168

Page 387

708

A. G. Capodaglio

Water Science & Technology: Water Supply

In-stream detection of waterborne priority pollutants

17.3

2017

(e.g. light measurement and others), without sample collection

studying American water-supply systems in preparation for

(directly in-stream), supplying (true or surrogate) parameter

attacks (IonLife ). No such attacks have been ofﬁcially

values in real time. The other alternative, traditional manual

exposed (or disclosed) so far, however, accidental drinking

grab sampling followed by laboratory analysis, requires con-

water contamination events with variable (although usually

siderable manpower and only allows capture of small data

nonlethal) consequences occur almost every day in many

sets, mostly unrepresentative of the variance at the source,

parts of the world. The safety of water distribution systems

and allows potentially important events to occur undetected

has thus become of primary concern to governments, and

(Copetti et al. ). In addition, long-term environmental data-

research related to water quality monitoring has increased

bases often display signiﬁcant data shifts that exceed natural

signiﬁcantly in recent years.

variability, and which may be correlated with changes in

In view of these risks and the need for a safe and reliable

sampling and laboratory analysis techniques and methods

water supply, traditional monitoring routines can no longer

(Horowitz ). It has been shown, on the other hand, that

be considered satisfactory, especially since online, relatively

remotely acquired, continuously in situ monitored data can

cheap monitoring technologies, available for a larger

provide important early warning information about water

number of parameters than previously thought possible,

quality events (Glasgow et al. ). Comparative advantages

have become affordable (Capodaglio & Callegari ,

of online monitoring sensor technology versus traditional cabi-

; O’Halloran et al. ; Capodaglio et al. a).

net analyzers and manual sampling are summarized in Table 1.

Utilities around the world are now using some form of

During the last two decades, several studies have

online monitoring as warning systems for drinking water

revealed the presence of hazardous contaminants in

contamination,

anticipation

yet-to-be-speciﬁed

waters due to ‘common’ anthropic activities, including pesti-

regulations. In the USA, turbidity is currently the only indi-

cides (Öllers et al. ), natural and synthetic hormones

cator bearing a regulatory requirement for continuous

(Kolpin et al. ), plasticizers, personal care products

online monitoring (AWWA ); in Europe, current regu-

and pharmaceuticals (Daughton & Ternes ; Jones

lations (Council Directive 98/83/EC) do not specify the

et al. ). Since these may end up in water supplies, there

need for online measurements in drinking water systems,

is a clear need to be able to rapidly detect instances of acci-

although good practice suggests that, at least in critical situ-

dental (or deliberate) contamination in distribution systems,

ations, some basic continuous monitoring (e.g., turbidity)

due to the potential consequences for human health. These

should be implemented, given also the very affordable cost

data might not be measurable during routine ofﬂine monitor-

of last-generation sensors, today. Online CWS for water dis-

ing at drinking water treatment plants, or in various

tribution networks (WDN) have been studied in the last few

distribution system locations. Since existing laboratory

years, and are gradually being put into place. This paper con-

methods are too slow to develop operational responses,

tains a comprehensive review of existing and expected

they cannot provide a sufﬁcient level of public health protec-

online monitoring technologies for measurement/detection

tion in real time, therefore the need for better online

of pollutants in water. The state-of-the-art in online water

monitoring of water systems is clear (Storey et al. ).

monitoring and CWS is also presented, with some appli-

Water distribution systems are vital for the life and well-

cation examples.

being of cities and nations, but unlike other similarly vital installations, they are potentially accessible (also in the

Online water quality monitoring

‘unauthorized’ meaning) to almost everyone willing to do so. In the current geopolitical climate, water distribution sys-

AWWA () deﬁnes online water quality monitoring as the

tems may thus become relatively easy targets of terrorist

unattended sampling, analysis and reporting of parameters,

groups of any extraction, that could thus affect the fate of

producing data sequences at a greater frequency than that per-

large numbers of people with limited effort. In 2002 the

mitted by manual (grab) sampling, and allowing real-time

US FBI circulated a reserved warning to water industry man-

feedback for process control, water quality characterization

agers indicating that al-Qaida operatives may have been

for operational or regulatory purposes, and alert/alarm.

Page 388

709

Table 1

A. G. Capodaglio

Water Science & Technology: Water Supply

In-stream detection of waterborne priority pollutants

17.3

2017

Advantages of online monitoring sensor technology vs manual and ﬁxed cabinet sampling

Issue

Online sensors

Cabinet analyzers

Manual sampling

Installation costs

One sensor can detect several parameters at once. Extremely simple in-stream/in-pipe installation (protection from vandalism/theft needed). Sensor costs are low compared to analyzers. Can run on battery for long times.

Installation requires proper operatoraccessible housing (protected from vandalism/ theft) with service lines. Instrumentation is expensive and needs automatic samplers.

Very small installation costs. May require trucking a boat to the river or building sampling ports in pipe network.

O&M costs

Extremely low. Can work unassisted for long periods. Usually, visual inspection suggested bi-weekly or monthly.

High. Frequent personnel control, calibration check, reagent costs.

Highest. A team is engaged for each campaign. Cost of laboratory procedures and sample handling, and possible errors must be considered.

Site accessibility

Site must be accessed at installation and in case of maintenance/ repositioning. No need to access if working properly. Minimal disturbance.

Site must be accessed often. Medium– high disturbance.

Requires team working on-site during campaigns. Maximum disturbance. Access can affect measurement.

Sampling frequency

Sample-less. Measurement is instantaneous, can be set from fractions of second on.

Depends on technical times for analysis. Limited capacity (e.g. of automatic sampler). Need for sample handling.

Even during ‘continuous sampling’ events the frequency is limited by the operators’ training and technology used.

Data availability and uncertainty

Instantaneous, can be transmitted wirelessly to a receiving station, and/or stored locally. Uncertainty due to missing data is highly unlikely. Systematic data error due to calibration deviation possible but retraceable.

After analysis, same as online sensor, however, the delay due to the analytical process cannot be eliminated. Uncertainty of missing data due to failed procedure possible. Systematic data error due to miscalibration possible.

Unless simple parameters are measured locally by hand-held devices, samples have to be transported and worked up in the laboratory. Uncertainty due to missing data depending on monitoring protocols, but quite possible. Systematic and random data errors due to manual handling of samples and human interference highly possible.

Water quality dynamics

Can fully capture water quality dynamics at the short- and longterm ranges, due to continuous, virtually unlimited data collection.

Can capture some trends, depending on proper preliminary setting and sampler’s capacity. Usually ﬁnite data collection capacity.

Almost undetectable within a single campaign.

Health protection

Early online detection of contaminants may allow prompt response. Used on CWS/EDS systems.

Detection of contaminants may not be quick enough for adequate intervention. Automatic determination of hazardous pollutants not always possible.

Only for slow-moving contaminants in far-off locations (e.g. groundwater).

Online monitoring of pollutants and dangerous sub-

(a) variability, in space and time (in general very low for

stances is important for different purposes, including plant

groundwater, low for lakes, high for rivers, very high

management, pollution control and reduction, and limiting

for discharge channels and urban or industrial drainage,

the environmental impact of discharges. Online instrumen-

or in-plant, in-pipe monitoring);

tation must be placed at selected, representative locations

(b) vulnerability, including type and location of possible

in water system networks, and must be periodically main-

contaminating activities, time-of-travel of contaminants

tained. Monitoring requirements can be deﬁned according

to intake/point of use, natural/technological barriers’

to monitored water type, considering its:

effectiveness, control options after alarm. Page 389

710

A. G. Capodaglio

Water Science & Technology: Water Supply

In-stream detection of waterborne priority pollutants

17.3

2017

The ‘ideal’ location for control of contaminants is as close

referred to as supervisory control and data acquisition

to their potential source as possible. Source water with low

(SCADA). These consist of individual online instruments, con-

vulnerability is characterized by few potential contaminant

nected to programmable logic controllers or remote telemetry

activities, transit times longer than those required for labora-

units, that convert output signals to the desired units, compare

tory analysis, and the presence of multiple physical barriers

them to criteria set by users, and generate signals for alarm or

between contaminating activities and point of intake. In a

control to process equipment. A host computer is used to visu-

source with moderate vulnerability, online monitoring of sur-

alize, store, or to further utilize data for speciﬁc purposes

rogate parameters (such as total organic carbon (TOC),

(Figure 1). A few cities around the world have already adopted

dissolved organic carbon (DOC), UV254, pH and conduc-

such systems, as illustrated by Allen et al. ().

tivity) should be considered. In high-vulnerability water sources, online monitoring of chemical–physical–biological

Online monitoring technology overview

parameters (turbidity, pH, conductivity, redox, ﬁsh toxicity) and surrogate parameters, in addition to speciﬁc indicators

Table 3 summarizes the ﬁve main classes of online monitor-

(e.g. volatile organic compounds (VOCs), phenols and

ing instrumentation. In this review, just the ﬁrst four classes

speciﬁc toxicity tests) may be preferred. In water/wastewater

will be examined, with discussion of basic operating

treatment applications monitoring must consider possible

principles, state-of-the-art, and evaluation of technology for

process optimization options, response times, signiﬁcative

online applications in water and wastewater monitoring.

sampling frequency, and allow adequate process-control lead time. For drinking water protection, multi-barrier approaches

Physical monitors

based on the concept that contaminants must be subject to as many points of control/treatment (barriers) as possible, prior

Well-established technologies used for monitoring physical

to tap, are usually adopted (O’Halloran et al. ).

parameters include: light scattering/blocking (turbidity,

Table 2 summarizes monitoring requirements and

particles, suspended solids (SS)), light absorbance (color),

objectives for various types of activities in the speciﬁc case

electrochemical (conductivity, reduction–oxidation poten-

of water distribution system applications.

tial (Redox)), electrophoretic (streaming current), and

Availability of real-time information is one of the key

other (radioactivity, temperature). Most of these have been

beneﬁts of online monitoring: the information must be con-

commercially available as online instrumentation for some

veyed to the appropriate user by means of systems often

time (Table 4).

Table 2

Online monitoring objectives and strategies for water distribution systems (modiﬁed from AWWA 2002)

Activity

Monitoring strategy

Objectives

Contaminant source identiﬁcation

Surrogate parameters (TOC, DOC, UV254, pH, conductivity); Speciﬁc parameters (related to known sources of contamination); Biotests and toxicity tests

Deﬁne potential contamination in relation to vulnerability of source water

Monitoring of discharges into the source water

Speciﬁc organic/inorganic contaminants

Identify water pollution accidents

Best management practices/protection of water source

Hydrological parameters; Environmental parameters (solar radiation, O2, chloride)

Prevent source deterioration; Environmental management

Drinking water quality protection

Speciﬁc organic/inorganic contaminants; Treatmentrelated parameters (ﬂow, turbidity, pH, TOC, DOC, etc.); Biotests/toxicity

Allow appropriate responses to contaminant presence (intake shut-up, additional treatment, treatment adjustment)

Emergency response

Speciﬁc organic/inorganic contaminants; Biotests/ toxicity

Drinking-water pollution control; Risk management; Treatment modiﬁcation

Page 390

711

Figure 1

A. G. Capodaglio

Water Science & Technology: Water Supply

In-stream detection of waterborne priority pollutants

17.3

2017

SCADA system for environmental monitoring.

Inorganic monitors

stripping voltammetry (ASV, CSV) methods was developed and launched on the market quite recently (Kissinger &

Inorganic monitors are used in online mode to detect inﬂuent and efﬂuent water quality, and/or treatment process control; applicable technologies are listed in Table 5. Online monitoring of inorganic constituents, with the exception of chemical titration technology (alkalinity, acidity, hardness), is still in the early phases for many elements of interest to

Heineman ; Jothimuthu et al. ; Yue et al. ; Table 3

Online monitoring instrumentation classes

Type of monitors

Application examples

Physical

Turbidity, particles, color, conductivity, total dissolved solids (TDS), streaming current, radioactivity, temperature, redox potential

Inorganic

pH, dissolved oxygen (DO), hardness, acidity, alkalinity, disinfectants such as chlorines and ozone, metals, ﬂuoride, nutrients, cyanide

Organic

carbon (BOD, COD or TOC), hydrocarbons, UV adsorption, VOCs, pesticides, DBPs

Biological

nonspeciﬁc, algae, protozoa, pathogens

Hydraulic

ﬂow, level and pressure

drinking water applications. For metals, typical available technologies (non-existent until very recently for many metals of interest, like As, Cd, Pb, Hg, Se, Zn) are adaptations to automatic mode of complex colorimetric methods originally developed for laboratory applications, and therefore expensive and/or complex to operate, and still not suitable for installation in remote or unmanned sites. Online instrumentation based on anodic or cathodic

Page 391

712

Table 4

A. G. Capodaglio

Water Science & Technology: Water Supply

In-stream detection of waterborne priority pollutants

17.3

Most appropriate technology

Other technologies

Low turbidity raw water

Single beam (tungsten or LED) turbidimeter

Particle counters; Particle monitors

Clariﬁed water; Filter efﬂuent

Modulated four-beam turbidimeter

High turbidity raw water

Ratio turbidimeter; Modulated four-beam turbidimeter

Filter backwash

Transmittance turbidimeter; Surface scatter; Ratio turbidimeter; Modulated four-beam turbidimeter

Color

Online colorimeter; Spectrophotometer

TDS

Two-electrode conductivity probe; Electrode-less (toroidal) probes

2017

Physical online monitor technology (modiﬁed from AWWA 2002)

Application

Table 5

Laser light source (660 nm) and improved optics turbidimeters

Online inorganic monitor technology (modiﬁed from AWWA 2002)

Parameter

Currently applied technology

Other technologies

DO, pH

Ion-selective electrodes

Fiber-optic chemical sensors (FOCS)

Hardness

EDTA tritration online; Ion-speciﬁc electrodes (ISE)

(FOCSs or optodes) for pH, DO; Iodometric DO measurements

Alkalinity

Online alkalinity titrator

ClO2: Iodometry, Amperometric meth. I, DPD, amaranth, chlorophenol red, LGB dye, ion chromatography

Iron, manganese, metals

X-ray ﬂuorescence (complex), colorimetry

CSV, ASV stripping voltammetry; Graphene-based EC-sensors

Ammonia, nitrite

Colorimetric, FOCS (ammonia)

Nitrate

Ion sensitive gas membrane electrodes, UV spectrometry

Phosphorus, cyanide

Colorimetric, FOCS (cyanide)

Bullough et al. ; Nunes et al. ), with detection limits

Some promise for future applications comes from develop-

down to 0.5–10 μg/l (clean water), depending on sample

ments in optode technology, coupled with miniaturized

type and actual analyte (ModernWater ).

spectrophotometry, due to their low-cost, low power require-

Developments in miniaturization technology and new

ments and long-term stability. Optodes (or optrodes) are

materials, (i.e. carbon nanotubes) recently allowed design of

optical sensors formed by a polymeric matrix coated onto the

fully automated, online metal monitors able to provide continu-

tip of an optical ﬁber, capable of (optically) measuring a

ous monitoring in liquid streams (Hanrahan et al. ).

speciﬁc substance, with the aid of a chemical transducer,

Graphene has recently attracted strong scientiﬁc and tech-

applying various measurement methods, such as reﬂection,

nological interest, showing great promise in many diverse

absorption, evanescent wave, chemiluminescence, surface

applications, from electronics, energy storage, and fuel cells,

plasmon resonance (SPR), and, by far the most popular, lumi-

to biotechnologies because of its unique properties (Shao

nescence (ﬂuorescence and phosphorescence). Optodes may

et al. ). Graphene-based electrochemical sensors have

provide viable alternatives to electrode-based sensors, or

been developed for the detection of heavy metal ions. However,

more complicated analytical instrumentation (Tengberg et al.

no commercial graphene-based products for environmental

; Xie et al. ), although they often still do not have resol-

monitoring applications are available as of now.

ution comparable to the most recent cathodic microsensors.

Page 392

713

Table 6

A. G. Capodaglio

In-stream detection of waterborne priority pollutants

Carbon fractions measured by organic carbon analyzers (modiﬁed from AWWA 2002)

Water Science & Technology: Water Supply

17.3

2017

For this reason, in addition to its use in mandatory monitoring, and notwithstanding a lack of speciﬁc regulations,

Carbon fraction

Abbr.

Deﬁnition

Total carbon

Sum of organically and inorganically bound carbon (incl. elemental C) in water

Total inorganic carbon

TIC

Sum of elemental carbon, CO2, CO, CN, CS, CCl4, etc.

Total organic carbon

TOC

Organic carbon bound to particles <100 μm (TOC ¼ TC– TIC)

Dissolved organic carbon

DOC

Organic carbon in water bound to particles <45 μm

Nonpurgeable organic carbon

NPOC

OC present after sample scrubbing to eliminate inorg. C and VOCsa

Volatile organic carbon

VOC

TOC fraction removed from the sample by gas stripping

many water utilities already routinely use online organics monitoring to some degree. Table 6 shows different fractions measured by an organic carbon analyzer. Most organic compounds in water absorb UV radiation: their concentration can thus be estimated using spectrometry. Originally, a single UV source with wavelength of 254 nm was used for such measures. However, recently, instrumentation reading the entire UV–VIS (Ultraviolet–visible spectroscopy) adsorption spectrum (200–750 nm) was introduced (S-can ), and UV absorption is now a commonly used methodology. To quantify organic contamination, due to a multitude of substances, cumulative parameters such as chemical oxygen demand (COD), biochemical oxygen demand (BOD) or spectral absorption coefﬁcient are often used. Evidence shows strong

correlation between organic carbon measured with UV and

Organic monitors

it was shown that several other parameters can be inferred

Most commercial TOC analyzers actually measure NPOC.

that measured with standard methods (Figure 2). In addition, by correlating their concentration values to UV full spectrum This class of monitors includes TOC analyzers, UV

absorption (Figure 3). Furthermore, several common organic

absorption and differential spectroscopes, chip-based micro-

compounds have absorption spectra that make their identiﬁ-

machined devices and chromatographs. Although not all of

cation quite easy with appropriate instrumentation (Figure 4).

these are suitable for online, on-site applications, this speciﬁc

Fluorescence spectroscopy has also been indicated

technology is much more developed than that for inorganics.

recently as a promising tool for online monitoring of organic

Figure 2

Correlation between BOD5 and COD laboratory results and the results measured with a spectrometric probe (S::can website, 2015).

Page 393

714

A. G. Capodaglio

Water Science & Technology: Water Supply

In-stream detection of waterborne priority pollutants

Figure 3

Correspondence between spectral absorption areas and quality parameters (S::can website, 2015).

Figure 4

Spectral absorption of benzene, with the typical ﬁve-peak shape (S::can website, 2015).

17.3

2017

matter, although no commercial products exist, yet (Shutova

phase detection (after volatilization), and resistance-based

et al. ).

sensors; some methods, however, give merely an indication

In addition to organic matter, hydrocarbons are probably

of the presence/absence of oil.

the main class of contaminants found in surface and ground-

VOCs, including aromatic compounds, halogenates and

water. Methods for online detection include: ﬂuorometry,

trihalomethanes, evaporate when exposed to air, and can be

reﬂectivity, light scattering and turbidity measurement, ultra-

of health concern when found in water (trihalomethanes are

sonic methods, electrical conductivity, spectroscopy, gas-

disinfection byproducts (DBPs) – possible precursors to

Page 394

715

A. G. Capodaglio

Water Science & Technology: Water Supply

In-stream detection of waterborne priority pollutants

17.3

2017

formation of suspected carcinogens). Current monitoring

close correlations with regulatory microbial measures

technologies for VOCs include purge-and-trap gas chrom-

(Bridgeman et al. ).

atography (GC) with ﬂame ionization (FID), electron

LED UV ﬂuorescence sensors adopting a special combi-

capture (ECD) or photoionization detectors or mass spec-

nation of UV and ﬂuorescence measurements have been

trometry (MS). Most of these methods require skilled

tested for online monitoring of dissolved organic matter

operators, puriﬁcation and pre-concentration, sample injec-

and to predict DBP formation potential during water treat-

tion and results analysis. Detection limits for different

ment. This application has demonstrated the potential

substances vary according to the detector method (Yongtao

applicability of LED UV/ﬂuorescence sensors for online

et al. ; Capodaglio & Callegari ).

water monitoring (Li et al. ).

Pesticides, including insecticides, fungicides and herbicides, comprise triazines and phenylurea compounds; they

Biological monitors

are monitored in surface waters in order to detect accidental pollution. Online monitoring of pesticides can be carried out

Biosensors, deﬁned as devices incorporating a biological, bio-

using composite techniques, such as:

logically derived, or biomimicking material, integrated within a physicochemical transducer, offer some advantages for

• • •

high-pressure

liquid

chromatography

(HPLC)/diode

environmental analysis, compared to conventional methods,

array (DA) detection;

since they are cheap and simple to use, and are frequently

GC separation and mass spectrometer (MS) detection;

able to evaluate complex matrices with minimal sample prep-

liquid chromatography/MS.

aration. Biosensors should be distinguished from bioassays or bioanalytical systems, which require additional sample pro-

Each technique is capable of optimally detecting a group

cessing

(e.g.

reagent

addition).

Advantages

include

of compounds, for example, HPLC/DA can be used for atra-

miniaturization and portability possibilities, permitting their

zine, chlortoluron, cyanazine, desethylkatrazine, diuron,

use as on-site devices. In addition to the identiﬁcation of

hexazinone, isoproruton, linuron, metazachlor, metha-

speciﬁc chemicals, some biosensors offer the possibility of

benzthiazuron, metobrorumon, metolachlor, metoxuron,

measuring biological effects, such as toxicity, cytotoxicity,

monolinuron, sebutylazine, simazine and terbutylazine

genotoxicity, or endocrine disrupting effects.

(AWWA ).

This information could be, in some cases, more relevant

In theory, any analytical laboratory method can be

than speciﬁc chemical composition. Online biological moni-

adapted for online use, provided that requirements for con-

tors are an active area of R&D due to increasing regulatory

sumables and manual intervention can be minimized: for

and public demand. At this time, many biological monitors

this reason, current online systems are often a ‘robotized’

are relatively new and can still be considered experimental/

adaptation of ofﬂine procedures, however, this solution is

unique laboratory-based applications, although commercial

not always the most efﬁcient. Novel technologies, such as

tests have started in water monitoring, for BOD, nitrate and

optochemical sensors, biosensors, and microbiological

pesticide assessment (Bahadır & Sezgintürk ).

sensors, are being tested for organics and hydrocarbon

Table 7 shows an overview of the most common types

analysis. Advances already in use include differential UV

of online biological monitors. Table 8 summarizes the com-

spectroscopy for DBP detection and microphase solid-

parative features of biosensors versus current online LC–MS

phase extraction (SPE) for analysis of semivolatile organics

methods (Rodriguez-Mozaz et al. ).

(Yongtao et al. ).

At the moment, bacterial-based systems (Kim & Gu )

A novel LED-based prototype instrument, detecting ﬂu-

show poor sensitivity and low ease of operation. Develop-

orescence peaks C and T (surrogate parameters for organic

ments will likely derive from improved ﬁngerprinting of

and microbial matter, respectively), was recently developed

organisms, and cost reduction. Signiﬁcant advances can be

and tested. Although correlating well with regulatory

expected from protozoan monitor technology, with UV

organic surrogate measures, the device did not provide

absorption/scattering analytical techniques that may soon Page 395

716

Table 7

A. G. Capodaglio

Water Science & Technology: Water Supply

In-stream detection of waterborne priority pollutants

17.3

2017

Common online biological monitors

Technology

Measurement

Comments

Fish tests

Swimming pattern Ventilation rate Bioelectric ﬁeld Avoidance patterns

Low sensitivity Sophisticated requirements Requires exotic ‘electric ﬁsh’ species Interpretation complex

Daphnid tests

Swimming activity Behavior

Good performance, no determination of causes

Mussel tests

Shell positions/opening

Can concentrate pollutants to levels many times greater than found in the water. Long organism lifespan. Similar results over different species

Algae tests

Fluorescence (photosynthesis)

Commercial monitors available

Bacteria tests

Luminescence Respiration of nitriﬁers

Commercially available, toxicity data for over 1,000 compounds

Chlorophyll-a

Fluorometry

Interference with pigments, diss. organics, sensitive to environmental variables

Chlorophyll-a and algal absorption

Reﬂectance radiometry

Commercial systems available

Protozoan monitors

Measurement requires preliminary concentration/centrifugation of sample Laser scanning cytometry Particle characterization

By ﬁltration on membrane cartridge

UV spectroscopy Multi-angle light scattering Nucleic acid molecules and magnetized microbeads

W/modiﬁed blood cell separators, minimal operation time Analysis possible within 3 min, particles must be conﬁrmed by trained operator Measure particle size/distribution, high number of false positive and negative results Online system, unlabeled parasites, differentiation problems Successfully tested in laboratory Oocysts detected within 20 min, not fully automated

allow automated detection of Cryptosporidium and Giardia.

underlying principle is that the current generated by an

Molecular techniques initially applied to the recognition of

MFC directly relates to the metabolic activity of the electro-

organisms’ genomic sequence in clinical applications (Bej

active bioﬁlm at the anode surface, thus any disturbances of

) have shown great potential for detection of pathogens

their metabolic pathways are translated into a change in elec-

in water, and are producing interesting results that could

tricity production (Molognoni et al. ). Their application,

soon lead to widespread online use. Very recently, a prototype

supported by interpretation software, would not be limited

automated biosensor for fast (8 hours) identiﬁcation and

to organic carbon, but also to water toxicity and speciﬁc com-

quantiﬁcation of Escherichia coli contaminations in ground,

pounds (Chouler & Di Lorenzo ; Yang et al. ). For

surface and drinking water was proposed and tested. The

interpretation of these data, the use of artiﬁcial neural net-

instrument is based on a three-electrode potentiostat using

works, often adopted in wastewater-related modelling

electrochemical assays to detect E. coli using their β-galactosi-

(Raduly et al. ), has been proposed.

dase activity (Ettenauer et al. ).

Molecular methods for detection of microbial pathogens

Microbial fuel cells (MFCs), biological systems capable of

have in fact been established, however, most of these have

degrading organic matter with direct generation of electrical

important limitations, associated with the time necessary to

energy, intensively investigated as an alternative to traditional

isolate and/or identify the pathogen and detection accuracy.

wastewater treatment processes (Capodaglio et al. ,

Research towards their improvement relies on methods of

b), have recently been highlighted as a technology with

culture on selective media, immunological approaches,

potential for rapid and simple testing of water quality. The

nucleic

Page 396

acid-based

assays

and

DNA

microarrays

717

Table 8

A. G. Capodaglio

Water Science & Technology: Water Supply

In-stream detection of waterborne priority pollutants

17.3

2017

Comparative features of online SPE–LC–MS methods vs biosensors for environmental analysis (modiﬁed from Rodriguez-Mozaz et al. 2007)

Online SPE–LC–MS

Biosensors

Comparatively higher sample volumes of water are necessary

Small sample volumes are sufﬁcient to obtain enough sensitivity

Matrix effect; ionic suppression or enhancement in MS spectrometry

Matrix effects. Variable depending on biorecognition principle and transduction element

Preconcentration of the sample necessary (SPE)

Direct analysis of the sample. Minimal sample preparation

Multi-residue analysis

Limited multi-analyte determination

Automatization and minimal sampling handling

Possible automatization of the system

Direct and fast elution of the sample after preconcentration. Minimal degradation

Direct analysis after sampling is possible. Minimal degradation

No biological stability restrictions

Low biological material stability

Determination of chemical composition

Determination of biological effect and of bioavailable pollutant content

Compound selectivity by using speciﬁc sorbents (MIPs and immunosorbents)

Compound selectivity by using speciﬁc biological recognition element

Minimal consumption of organic solvents (elution with the LC mobile phase)

Consumption of organic solvents avoided. Direct analysis of contaminant in water

Generation of organic solvent waste

Minimal and non-contaminating waste

Short analysis time and high throughput

Faster analysis. Real-time detection and high throughput

Limited portability. Laboratory conﬁned

Availability of portable biosensor systems

Applicability to early-warning and on-site monitoring

Qualiﬁed personnel required

Qualiﬁed personnel not required. User friendly

Expensive equipment

Cost-effective equipment

(Lemarchand et al. ). Molecular ﬁngerprinting was

Figure 5 shows, as an example, spectral ﬁngerprinting of a

recently demonstrated as an effective monitoring tool for

municipal wastewater, with three spectral readings, in the

detection of cyanobacteria in surface waters (Loza et al. ).

wavelength range 230–730 nm, recorded at the same point within a short interval. Individual spectra show clearly

Fingerprinting

different features, indicating a pronounced water-quality change occurring in the 18 minutes elapsed since the ﬁrst

Fingerprinting methods describe the use of a unique chemical

reading. Although this indication alone, in general, will

signature, isotopic ratio, mineral species, or pattern analysis

not individuate the compound(s) responsible, it can never-

to identify different chemicals. Optical ﬁngerprinting by

theless trigger an alert to the operator, indicating deviation

UV, VIS, and near-infrared (NIR) absorption spectroscopy

from routine conditions. Fingerprinting is used, in conjunc-

can be effectively achieved by low-cost and compact devices

tion with sophisticated algorithms and statistical software,

that can be linked to an online diagnostic system, to directly

in CWS or EDS, described below.

identify some compounds (e.g. benzene, Figure 4) present in the water, or to indicate the possibility of their presence.

Spectral photometric (spectrometric) methods are probably the most interesting, currently available mature

A ﬁngerprint contains much more information about

technology to cover most online monitoring needs, and

water quality than a single-wavelength instrument can

speciﬁcally ﬁngerprinting. They are recommended by the

provide, allowing more accurate and comprehensive assess-

US-EPA (EPA ) for online monitoring over traditional

ments. In optical ﬁngerprinting, a wide portion of the UV,

analytical techniques, having been tested for online drink-

VIS and NIR spectrum is monitored simultaneously at

ing-water quality

high measurement frequencies (minutes or fractions);

traditional reagent-based analyzers (EPA ).

monitoring applications, instead of

Page 397

718

Figure 5

A. G. Capodaglio

Water Science & Technology: Water Supply

In-stream detection of waterborne priority pollutants

17.3

2017

Optical ﬁngerprinting in a pipe, indicating rapid water-quality changes (S::can website, 2015).

The main features that have contributed to the wide

about a water system not otherwise available. As shown in

acceptance of spectrometric methods, in comparison to

Figure 6, nitrate proﬁles measured continuously are com-

photometric ones, are:

pared to calculate travel times between monitoring sites,

•

determine water age, and verify the network’s hydraulic

cost efﬁciency: the continuous UV/VIS spectrum enables simultaneous measurement of, e.g., organic carbon, nitrate and turbidity, for which only one spectrometer is required, instead of three photometers;

•

lower cross-sensitivity to turbidity, coloration, surface growth, etc.: potential interferences, not detectable by single/dual wavelength measurement, are nearly always compensated using spectral information;

•

greater precision, higher selectivity and reproducibility: since cross-sensitivity is substantially reduced, heterodyning of signals due to interference/noise is significantly less than with photometers; furthermore, individual substances and/or groups can be allocated to specific spectral features, resulting in very high reproducibility, without the absolute necessity of specific calibration;

•

qualitative evaluation: in addition to calibrated parameters, qualitative spectral information contained in the ‘ﬁngerprint’ can be directly applied for alarm and control systems.

model (Thompson & Kadiyala ). Integrated CWS Following the ‘9–11’ events in the USA and the completion and review of a risk assessment procedure for public water systems serving populations greater than 3,300, as mandated by the US Bioterrorism Act of 2002, water distribution systems were identiﬁed as one of the most vulnerable areas of attack from potential terrorist or extremist groups. In consideration of deliberate hostile actions on water supplies, although none has been reported to date, in January 2002 the FBI circulated a reserved bulletin warning water industry managers that al-Qaida may have been studying American water-supply systems in preparation for attacks (IonLife ). Homeland Security Presidential Directive 9 required the US Environmental Protection Agency (US-EPA) to develop a program for utilities to improve protection of their water distribution systems.

Continuous, consistent online data obtained by such

Online quality monitoring of distribution systems has

instruments can be used to extract useful information

been investigated extensively for some time (Grayman

Page 398

719

Figure 6

A. G. Capodaglio

In-stream detection of waterborne priority pollutants

Water Science & Technology: Water Supply

17.3

2017

Comparison of continuous nitrate proﬁles in different sections of a distribution network (S::can website, 2015).

et al. ; Hasan et al. ; CHMHill ), generating so-

Regulatory compliance beneﬁts include the ability to

called CWS, or Water Quality EDS. These consist of inte-

maintain proper chlorine residuals and pH control in the

grated in situ sensors, SCADA systems, designed to

network (to avoid Pb and Cu leaching from pipes). Warning

continuously monitor network conditions and warn of any

of intentional or unintentional contamination in distribution

potential contamination events. In addition to security issue

systems is somewhat more complex. Specialized analyzers

detection, their beneﬁts may be categorized as operational

are available, including GCs that may detect speciﬁc con-

enhancements, regulatory compliance, and contamination

taminants and toxicity monitors that can provide general

warning. Operational enhancements include continuous indi-

warnings. Due to the large number of potential contami-

cation of water quality in the distribution system beyond that

nants, however, it is more practical to monitor for

possible through routine regulatory sampling. Early indi-

indications of contamination through changes in the same

cations of water quality problems may consist of unusually

water quality parameters, or surrogates, often used for oper-

low residual chlorine, impending nitriﬁcation (elevated

ational monitoring (Table 9).

ammonia), turbidity excursions caused by mains breaks,

Real-time monitoring strategies are the answer to detect-

and other unusual quality changes. Monitoring is achieved

ing low probability/high impact events at an early stage,

through measurement of parameters already familiar to utili-

while chronic or long-term risks should still be monitored

ties (e.g., chlorine residual), and/or other parameters

with traditional sampling. Considering the safety of drink-

relatively new to these applications (e.g., TOC). The US-

ing-water supplies, realistic detection limits of available

EPA recommends monitoring four key parameters, namely:

online instrumentation must be taken into account

TOC, pH, conductivity, and chlorine (EPA ), while a

(Figure 7). These could be enhanced by combining tech-

later study, considering all instruments and technologies

nologies, e.g. spectrometry with online toxicity tests

available, suggested the following parameters for online

(Weingartner ), but often their sensitivity will not sufﬁce

water-quality monitoring systems: conductivity, chlorine

to warn about possible long-term contaminant effects.

(combined), pH, oxidation/reduction potential (ORP), temperature, turbidity, and UV absorption (CHMHill ).

An ideal EDS will react to most types of threat agents, at concentrations far below the LD50 lethal dose (in turn, Page 399

720

Table 9

A. G. Capodaglio

Water Science & Technology: Water Supply

In-stream detection of waterborne priority pollutants

17.3

2017

Parameters typically included in operational water quality monitoring systems

Parameter

Signiﬁcance

TOC

Total organic carbon. Elevated turbidity excursions can be associated with a breakthrough at the water treatment plant or scouring and release of bioﬁlm within the distribution system.

Residual chlorine

A sudden loss in residual could promote bioﬁlm growth and potential violation of the Total Coliform Rule.

Conductivity

Its measurement provides an easy method for identifying mixing or different water sources, which can have a signiﬁcant impact on many industrial operations.

Controlled for disinfection and corrosion control. The formation of some disinfection byproducts is pH dependent.

Turbidity

Provides warning of a system disruption created by a surge or reversal in ﬂow that scours the pipeline. This could be caused by a pipeline break, hydrant knockover, or other problems that will impact chlorine residual and customer satisfaction.

Note: Utilities that use chloramines for disinfection should also measure ammonia, nitrates, and DOC to provide early warning of nitrification in the distribution system. The first water quality indicator of nitrification will be the increase of ammonia, which will occur before nitrites and nitrates begin to increase.

Figure 7

An example of realistic detection limits of online sensors (redrawn from Weingartner 2013).

et al. ; Hall et al. ). Surrogate parameters can therefore provide valid information on the presence of contaminants within a distribution system. The challenge then is to analyze surrogate parameter signals to identify changes that are signiﬁcantly beyond the range of natural

much higher than drinking limits), provide distinct signals to each threat, not respond to harmless substances or operational ﬂuctuations (i.e. ‘false alarms’), and have a ‘fast’ (real-time) response. The sensoristic component can consist of various sensing platforms, including contaminantspeciﬁc sensors, or quality sensors (e.g., pH, Cl, electrical conductivity, etc.) currently installed in many municipal water distribution systems to provide ‘surrogate’ data to the CWSs. Table 10 lists in-pipe physical and chemical parameters

ambient variability of the background water quality, or to establish a detection baseline. Detection of events by simply using upper and lower thresholds of parameter concentration is virtually impossible; hence complex patternrecognition algorithms are indispensable. These can be implemented in advanced, user-friendly event-detection software, maintaining use of any already-installed sensors for event detection and water protection, resulting at the moment in the most economical and effective solution to distribution system security (Weingartner ).

that can be reliably measured along with current technologies. The most difﬁcult issue is to distinguish actual

EDS implementation examples

contamination from natural ﬂuctuations of the water matrix. Figure 8 summarizes capability, reliability and

The CANARY EDS software (Hart et al. ) is an open-

O&M requirements for existing physico-chemical in-pipe

source software platform developed by the US-EPA. It gath-

sensors.

ers water quality inputs from SCADA systems and processes

Laboratory experiments and pipe system tests have

the data using event detection algorithms and statistical

proved that a majority of potential contaminants will

models to determine the probability of an anomalous

change the value of at least one surrogate parameter from

event occurring within the distribution network (McKenna

normal background levels (Byer & Carlson ; Cook

& Hart ).

Page 400

721

Table 10

A. G. Capodaglio

Water Science & Technology: Water Supply

In-stream detection of waterborne priority pollutants

In-pipe measurable physical and chemical parameters with currently available technology

17.3

2017

of large water-security systems, based on distributed spectral (and other) sensors, centralized data collecting stations, and

Available technology

on several types of event detection software. The ﬁrst implementation of a working water security

General parameters Pressure

On-chip

Temperature

On-chip PPT

On-chip or electrolyte

ORP

On-chip or electrolyte

Conductivity

On-chip

Dissolved oxygen (on-chip)

On-chip

Chlorine (free, total)

On-chip amperometric; use of ORP

Chloramines: calculated from total free chlorine

On-chip amperometric; use of ORP

General optical parameters

system in the USA was in Glendale (Arizona) with about ten monitoring stations fully integrated into a central database (Thompson ). In New York City, all existing manual sampling stations are being converted into monitoring stations with the use of online spectral monitors: over 30 have been already installed, and the proprietary software Moni::Tool was selected as the event detection software (EPA ). Philadelphia’s water utility developed a comprehensive CWS for its drinking water system under an EPA-WSI grant, where selection of available instrumentation was made by ﬁeld comparison between products from ﬁve different manufacturers

Turbidity (optical)

Optical

(PWD & CHMHill ). The City of Dallas (Texas) devel-

Color

Optical

oped a CWS consisting of four spectral UV-VIS monitors

UV254–simple surrogate organics indicator

Optical

based at the water treatment facilities, giving a ‘ﬁngerprint’

Spectral TOC/DOC–broad organics detector

Optical

UV spectral alarms

Optical

in the system. Monitoring is supported by an EDS constantly

NH4 (chloraminating systems): ISE

ISE

of expanded water quality monitoring capabilities are a 24/7

NO2 (chloraminating systems): spectral hi-resolution UV-VIS

Optical

view of water quality available to staff, including parameters

NO3 (groundwater under agricultural inﬂuence): spectral UV-VIS

Optical

UV spectral absorbance, DOC, pH, and free ammonia. All

Hydrocarbon alarm: UV-VIS or ﬂuorescence

Optical

Spectral parameters for special purpose

of the water leaving each plant and 32 distribution monitors providing continuous water quality analysis at 16 checkpoints checking for anomalies in the background. Reported beneﬁts

such as nitrate, total chlorine turbidity, TOC, conductivity, are web-accessible to the city’s water operators and constitute

Other important parameters that no sensors exist for Arsenic

None

Endocrine disruptors

None

Pesticides/Herbicides

None

valuable information for the detection of water quality changes from intentional or unintentional actions, natural phenomena and/or problems at treatment plants (Sanchez & Brashear ). Bratislava Water Company (BVS) is responsible for the operation of the water and wastewater systems of the capital of Slovakia, supplying a population of over 600,000. Drink-

CANARY was tested along with four other proprietary

ing water is produced in seven treatment facilities from more

EDS software tools under real-life conditions by the

than 150 groundwater sources. Given the high quality of raw

US-EPA (EPA ). The results of this evaluation were

water the only treatment is chlorination, to prevent micro-

encouraging, as the conclusions regarding EDS performance

biological growth during distribution. To ensure that

showed that event detection is possible, but the ability to

contamination of a source would not compromise overall

detect anomalous conditions strongly depends on EDS con-

high quality, BVS implemented a monitoring system that

ﬁguration, baseline variability of the monitoring location,

oversees all sources, coupled with an EDS sending an

and the nature of the change, with overall positive response.

alarm in case of an unexpected event. Measured parameters

The US-EPA runs the Water Security Initiative (WSI), a nationwide project to support investigation and deployment

include

TSS,

turbidity,

NO3-N,

COD,

BOD,

TOC,

DOC, UV254, color, benzene, toluene, xylene (BTX), O3, Page 401

722

Figure 8

A. G. Capodaglio

In-stream detection of waterborne priority pollutants

Water Science & Technology: Water Supply

17.3

2017

Synthetic assessment of capability, reliability and O&M requirements for existing physico-chemical in-pipe sensors: for the most common online quality parameters, a score representing an assessment of current technology in terms of selectivity, reliability and maintenance requirements is given. Higher scores generally indicate ‘better’ performance (except in the case of maintenance, where a lower score indicates lower requirements) (redrawn from Weingartner 2013).

H2S, assimilable organic carbon (AOC), temperature and

of-the-art application examples examined. It is clear that

pressure. Monitoring occurs by means of submersible online

technological development in this ﬁeld is very rapid, and

spectrometric probes combined with a centralized system,

that astonishing advances are anticipated in several areas

continuously analyzing four spectral alarm parameters from

(ﬁngerprinting, opto-chemical sensors, biosensors, molecular

each site. An evaluation of the EDS showed that the spectral

techniques). Some of the technologies mentioned, although

alarm system is able to detect contamination events down to

promising, are not at commercial or at online, standalone

100 μg/L TOC, 25 μg/L carbendazim, 100 μg/L benzene and

application stage. Software applications, together with new-

50 μg/L saxitoxin (BVS ).

generation sensors, are also contributing to the identiﬁcation

Under an EU-funded project, SMaRT-OnlineWDN, a

of otherwise difﬁcult-to-monitor parameters. In some cases,

group of European research and industrial partners is cur-

the presence of contaminants not directly observable online

rently investigating the development of an online security

can be inferred by water property (e.g. absorbance) changes

management system for WDN based on sensor measure-

or by indirect indicators (with statistical analysis software),

ments of water quality and quantity, with planned

giving rise to the possibility of water quality ‘alerts’, pending

applications ranging from the detection of deliberate con-

more detailed analysis with traditional methods. Examples

tamination, to improved operation and control of a WDN

of CWS applications, arisen from perceived threats to the

under normal and stress conditions. An online running

safety of water supply networks, have also been illustrated.

model, automatically calibrated to the measured sensor

CWS is perhaps the sector in which more rapid development

data, will give detailed information on contamination sources

is expected in the coming years, possibly creating a drive for

(localisation and intensity) (SMART-Online WDN ).

further technological breakthroughs. In spite of high technology instrumentation being developed, monitoring costs are bound to become a lesser and

DISCUSSION AND CONCLUSIONS

lesser part of water utility budgets due to the fact that automation and technological simpliﬁcation will abate human

In this paper, an overview of existing instrumentation appli-

cost factors (maintenance and other labor forms) and signiﬁ-

cable to water and wastewater online monitoring and

cantly reduce the complexity of procedures (sample

forthcoming developments has been given, and a few state-

preparation, reagent requirements, etc.).

Page 402

723

A. G. Capodaglio

In-stream detection of waterborne priority pollutants

Proper interpretation and use of the growing mass of water quality data that will become available through new monitoring and information technologies will allow better management of water resources, and of water/wastewater treatment facilities.

REFERENCES Allen, M., Preis, A., Iqbal, M., Srirangarajan, S., Lim, H. B., Girod, L. & Whittle, A. J.  Real-time in-network distribution system monitoring to improve operational efficiency. J. AWWA 103 (7), 63–75. AWWA  Online monitoring for drinking water utilities. In: Cooperative Research Report (E. Hargesheimer, O. Conio & J. Popovicova, eds). AWWA Research Foundation, Denver, CO, USA. Bahadır, E. B. & Sezgintürk, M. K.  Applications of commercial biosensors in clinical, food, environmental, and biothreat/biowarfare analyses. Analytical Biochemistry 478, 107–120. Bej, A. K.  Molecular based methods for the detection of microbial pathogens in the environment. Journal of Microbiological Methods 53, 139–140. Bridgeman, J., Baker, A., Brown, D. & Boxall, J. B.  Portable LED fluorescence instrumentation for the rapid assessment of potable water quality. Sci. Total Environ. 524–525, 338–346. Bullough, F., Fenech, C. & Bridle, H.  Advances in water quality monitoring of inorganics: current trends. Journal of Water Resource and Protection 5, 40–48. BVS  Early Warning System for Monitoring of Bratislava’s Water Supply. Internal Report, BVS, Bratislava, Slovakia. Byer, D. & Carlson, K. H.  Real-time detection of intentional chemical contamination in the distribution system. American Water Works Association Journal 97 (7), 130–133. Capodaglio, A. G. & Callegari, A.  Online Monitoring Technologies for Drinking Water Systems Security. Risk Management of Water Supply and Sanitation Systems, NATO Science for Peace and Security Series. Springer, Dordrecht, The Netherlands. Capodaglio, A. G. & Callegari, A.  Water supply systems security: novel technologies for the online monitoring of unforeseeable events. WIT Transactions on The Built Environment, Safety and Security Engineering VI 51, 251–263. Capodaglio, A. G., Molognoni, D., Dallago, E., Liberale, A., Cella, R., Longoni, P. & Pantaleoni, P.  Microbial fuel cells for direct electrical energy recovery from urban wastewaters. The Scientific World Journal 2013, 634738. Capodaglio, A. G., Callegari, A. & Molognoni, D. a Online monitoring of priority and dangerous pollutants in natural and urban waters: a state-of-the-art review. Management of Environmental Quality: An International Journal 27 (5), 507–536.

Water Science & Technology: Water Supply

17.3

2017

Capodaglio, A. G., Molognoni, D. & Vilajeliu Pons, A. b A multi-perspective review of microbial fuel-cells for wastewater treatment: bio-electro-chemical, microbiologic and modeling aspects. AIP Conference Proceedings 1758 (1), 030032. CH2M HILL  Contamination Warning System Demonstration Pilot Project: Implementation and Assessment. Internal Report, CH2MHill, Philadelphia, PA, USA. Chouler, J. & Di Lorenzo, M.  Water quality monitoring in developing countries; can microbial fuel cells be the answer? Biosensors 5, 450–470. Cook, J., Roehl, E., Daamen, R., Carlson, K. & Byer, D.  Decision support system for water distribution system monitoring for homeland security. In: Proceedings of the AWWA 2005 Water Security Conference, Denver, CO, USA, WSC61381. Copetti, D., Valsecchi, L., Sanfilippo, L., Capodaglio, A. G., Moscetta, P. & Tartari, G.  Real-time evaluation of nutrient concentrations in surface waters using a portable probe. In: Proceedings, IWA XIII Specialist Conference on Watershed and River Basin Management, 9–14 September, San Francisco, USA. Daughton, C. G. & Ternes, T. A.  Pharmaceuticals and personal care products in the environment: agents of subtle change? Environmental Health Perspectives 107 (6), 907–915. EPA  Distribution System Water Quality Monitoring: Sensor Technology Evaluation Methodology and Results. A Guide for Sensor Manufacturers and Water Utilities, EPA 600/R09/076, US Environmental Protection Agency Office of Research and Development, Water Infrastructure Protection Division, National Homeland Security Research Center, Cincinnati, OH, USA. EPA  Water Quality Event Detection System Challenge: Methodology and Findings. Office of Water (MC-140) EPA 817-R-13-002, Cincinnati, OH, USA. Ettenauer, J., Zuser, K., Kellner, K., Posnicek, T. & Brandl, M.  Development of an automated biosensor for rapid detection and quantification of E. coli in water. Procedia Engineering 120, 376–379. Glasgow, H. B., Burkholder, J. M., Reed, R. E., Lewitus, A. J. & Kleinman, J. E.  Real-time remote monitoring of water quality: a review of current applications, and advancements in sensor, telemetry, and computing technologies. Journal of Experimental Marine Biology and Ecology 300, 409–448. Grayman, W. M., Deininger, R. A. & Males, R. M.  Design of Early Warning and Predictive Source Water Monitoring Systems. AWWA Research Foundation, Denver, CO, USA. Hall, J., Zaffiro, A. D., Marx, R. B., Kefauver, P. C., Krishnan, E. R., Haught, R. C. & Herrmann, J. C.  On-line water quality parameters as indicators of distribution system contamination. American Water Works Association Journal 99 (1), 66–72. Hanrahan, H., Patila, D. G. & Wang, J.  Electrochemical sensors for environmental monitoring: design, development

Page 403

724

A. G. Capodaglio

In-stream detection of waterborne priority pollutants

and applications. Journal of Environmental Monitoring 6, 657–664. Hart, D. B., McKenna, S. A., Klise, K. A., Cruz, V. & Wilson, M.  CANARY: A water quality event detection algorithm development tool. In: Proceedings ASCE World Environmental and Water Resources Congress 2005 (K. C. Kabbes, ed.). ASCE, Reston, VA, USA. Hasan, J., States, S. & Deininger, R.  Safeguarding the security of public water supplies using early warning systems: a brief review. Journal of Contemporary Water Research and Education 129, 27–33. Horowitz, A. J.  A review of selected inorganic surface water quality-monitoring practices: are we really measuring what we think, and if so, are we doing it right? Environ. Sci. Technol. 47, 2471–2486. IonLife  How safe is our water? The threat of terrorism. Available at IonLife website http://www.ionlife.org (accessed July 2008). Jones, O. A. H., Voulvoulis, N. & Lester, J. N.  Human pharmaceuticals in the aquatic environment. Environ. Technol. 22 (12), 1383. Jothimuthu, P., Wilson, R. A., Herren, J., Haynes, E. N., Heineman, W. R. & Papautsky, I.  Lab-on-a-chip sensor for detection of highly electronegative heavy metals by anodic stripping voltammetry. Biomedical Microdevices 13 (4), 695–703. Kim, B. C. & Gu, M. B.  A multi-channel continuous water toxicity monitoring system: its evaluation and application to water discharged from a power plant. Environmental Monitoring and Assessment 109 (1–3), 123–133. Kissinger, P. T. & Heineman, W. R. (eds)  Laboratory Techniques in Electroanalytical Chemistry 2. Marcel Dekker, New York. Kolpin, D. W., Furlong, E. T., Meyer, M. T., Thurman, E. M., Zaugg, S. D., Barber, L. B. & Buxton, H.  Pharmaceuticals, hormones, and other organic wastewater contaminants in U.S. streams, 1999–2000: a national reconnaissance. Environmental Science & Technology 36, 1202–1211. Lemarchand, K., Masson, L. & Brousseau, R.  Molecular biology and DNA microarray technology for microbial quality monitoring of water. Critical Reviews in Microbiology 30 (2), 145–172. Li, W., Jin, J., Li, Q., Wu, C., Lu, H., Zhou, Q. & Li, A.  Developing LED UV fluorescence sensors for online monitoring DOM and predicting DBPs formation potential during water treatment. Water Research 93, 1–9. Loza, V., Perona, E. & Mateo, P.  Molecular fingerprinting of cyanobacteria from river biofilms as a water quality monitoring tool. Applied Environmental Microbiology 79 (5), 1459–1472. McKenna, S. A. & Hart, D.  On-line identification of adverse water quality events from monitoring of surrogate data: CANARY software. In: Proceedings, Singapore International Water Week Conference.

Page 404

Water Science & Technology: Water Supply

17.3

2017

ModernWater  Modern Water company website http://www. modernwater.com/monitoring/trace-metals/online-ova. Molognoni, D., Puig, S., Balaguer, M., Capodaglio, A. G., Callegari, A. & Colprim, J.  Multiparametric control for enhanced biofilm selection in microbial fuel cells. J. Chem. Technol. Biotechnol. 91 (6), 1720–1727. Nunes, C. N., Pauluk, L. E., dos Anjos, V. E., Chierici Lopes, M. & Quináia, S. P.  New approach to the determination of contaminants of emerging concern in natural water: study of alprazolam employing adsorptive cathodic stripping voltammetry. Analytical and Bioanalytical Chemistry 5, 1–9. O’Halloran, R., Fogelman, S. & Zhao, H.  Current Online Water Quality Monitoring Methods and Their Suitability for the Western Corridor Purified Recycled Water Scheme. Urban Water Security Research Alliance Technical Report No. 10, UWSRA, City East, Queensland, Australia. Öllers, S., Singer, H. P., Fässler, P. & Müller, S. R.  Simultaneous quantification of neutral and acidic pharmaceuticals and pesticides at the low-ng/l level in surface waters. Journal Chromatography A 911 (2), 225–231. PWD & Ch2MHill  Philadelphia Water Department Contamination Warning System Demonstration Pilot Project: Selection of Online Water Quality Monitoring Technologies and Station Design. White Paper submitted to EPA as part of the Water Security Initiative Grant awarded to Philadelphia Water Department, Philadelphia, PA, USA. Raduly, B., Gernaey, K. V., Capodaglio, A. G., Mikkelsen, P. S. & Henze, M.  Artificial neural networks for rapid WWTP performance evaluation: methodology and case study. Environ. Model. Software 22, 1208–1216. Rodriguez-Mozaz, S., Lopez de Alda, M. J. & Barcelo, D.  Advantages and limitations of on-line solid phase extraction coupled to liquid chromatography–mass spectrometry technologies versus biosensors for monitoring of emerging contaminants in water. Journal Chromatography A 1152, 97–115. Sanchez, C. & Brashear, K.  City of Dallas Water Utilities Contamination Warning System. In City of Dallas Dept. of Public Works website http://dallascounty.org/department/ pubworks/media/InfrastructureForum_2011 (accessed June 2015). S-can  S::can company website http://www.s-can.at (accessed May–July 2015). Shao, Y., Wang, J., Wu, H., Liu, J., Aksay, I. A. & Lina, Y.  Graphene based electrochemical sensors and biosensors: a review. Electroanalysis 22 (10), 1027–1036. Shutova, Y., Baker, A., Bridgeman, J. & Henderson, R. K.  Spectroscopic characterization of dissolved organic matter changes in drinking water treatment: from PARAFAC analysis to online monitoring wavelengths. Water Research 54, 159–169. SmartOnlineWDN  Online Security Management and Reliability Toolkit for Water Distribution Networks. Project website. Accessible on http://www.smart-onlinewdn.eu/ (accessed 4 May 2016).

725

A. G. Capodaglio

In-stream detection of waterborne priority pollutants

Storey, M. V., van der Gaag, B. & Burns, B. P.  Advances in online drinking water quality monitoring and early warning systems. Water Research 45 (2), 741–747. Tengberg, A., Hovdenes, J., Andersson, H., Brocandel, O., Diaz, R., Hebert, D., Arnerich, T., Huber, C., Kortzinger, A., Khripounoff, A., Rey, F., Ronning, C., Schimanski, J., Sommer, S. & Stangelmayer, A.  Evaluation of a lifetimebased optode to measure oxygen in aquatic systems. Limnology and Oceanography Methods 4, 7–17. Thompson, K. A.  Operational enhancements resulting from the development and implementation of a contaminant warning system for Glendale (AZ). In: Proceedings 2008 AWWA Water Security Congress. Denver, CO, USA. Thompson, K. & Kadiyala, R.  Protecting water quality and public health using a smart grid. Computational Water, Energy, and Environmental Engineering 2, 73–80. Weingartner, A.  Real-time technologies for early contamination warning in water security systems.

Water Science & Technology: Water Supply

17.3

2017

Presentation given at ERNCIP Operator’s Workshop, 11–12 September, Brussels, Belgium. Xie, X., Zhai, J. & Bakker, E.  pH independent nano-optode sensors based on exhaustive ion-selective nanospheres. Analytical Chemistry 86 (6), 2853–2856. Yang, W., Wei, X. & Choi, S.  A two-channel bacteria-based biosensor for water quality monitoring. In: 2015 IEEE Sensors, pp. 1–4. Yongtao, L., George, E. J., Hansen, E. M., Toma, J. J., Werner, M. & Smith, P. D.  SPE Twin-PAL. Online in situ monitoring of semivolatile organic compounds in water. In: Proceedings, AWWA Water Quality Technology Conference, Salt Lake City, UT, USA. Yue, W., Bange, A., Riehl, B. L., Riehl, B. D., Johnson, J. M., Papautsky, I. & Heineman, W. R.  Manganese detection with a metal catalyst free carbon nanotube electrode: anodic versus cathodic stripping voltammetry. Electroanalysis 24 (10), 10–18.

First received 3 June 2016; accepted in revised form 20 September 2016. Available online 7 October 2016

Page 405

Journal of Water Supply: Research and Technology

AQUA

Journal of

H2Open Journal

ISSN 0003-7214 iwaponline.com/aqua

Hydroinformatics

ISSN 2616-6518 iwaponline.com/h2open

ISSN 1464-7141 iwaponline.com/jh

Journal of

Water & Climate Change

Journal of

ISSN 2040-2244 iwaponline.com/jwcc

Water Quality Research Journal

ISSN 1201-3080 iwaponline.com/wqrj

Water Policy

ISSN 1477-8920 iwaponline.com/jwh

Journal of

ISSN Online 1751-231X iwaponline.com/wpt

ISSN 1366-7017 iwaponline.com/wp

Water Science & Technology

Water Reuse & Desalination

ISSN 0029-1277 iwaponline.com/hr

Water Practice & Technology

Oï¬&#x192;cial Journal of the World Water Council

Water & Health

Hydrology Research

ISSN 2220-1319 iwaponline.com/jwrd

Water Science & Technology

Water Supply

ISSN 0273-1223 iwaponline.com/wst

Journal of

Water, Sanitation & Hygiene for Development

ISSN 2043-9083 iwaponline.com/washdev

ISSN 1606-9749 iwaponline.com/ws

IWA Publishing key contacts Managing Director: Rod Cookson - rod.cookson@iwap.co.uk Sales Manager: Ian Morgan - imorgan@iwap.co.uk Marketing & Digital Development Manager: Margarita Lygizou - mlygizou@iwap.co.uk Marketing & Communications Assistant: Sarah Cooper - scooper@iwap.co.uk Insights & Bibliometic Analyst: Jim Eisenberg - jeisenberg@iwap.co.uk Books Commissioning Editor: Mark Hammond - mhammond@iwap.co.uk Production Editor: Niall Cunniffe - ncunniffe@iwap.co.uk Publishing Manager: Emma Gulseven - egulseven@iwap.co.uk Managing Editor: Michelle Herbert - eherbert@iwap.co.uk Editoral Assistant: Emma Buckingham - ebuckingham@iwap.co.uk Open Access Publisher: Sara Bosshart - sbosshart@iwap.co.uk

IWA Publishing Alliance House 12 Caxton Street London SW1H 0QS, UK Tel: +44 (0)20 7654 5500 Fax: +44 (0)20 7654 5555 Email: publications@iwap.co.uk

2017 Top Cited Papers from IWA Publishing Journals

Articles inside

Water Supply - Top Cited Papers 2017