Crojfe 2 2014 kb by zpandur

2014

Editorial

FORMEC 2015 in Austria 4th – 8th October 2015 in Linz

Austria will be hosting the 48th International Symposium on Forestry Mechanisation (FORMEC, http:// formec.boku.ac.at/) from 4th October to 8th October 2015 in Linz. In keeping with FORMEC tradition, the Conference will start with an Ice Breaker on Sunday evening, followed by two days of scientific conference and an optional one- or two-day visit of the AUSTROFOMA exhibition at Hochficht (http://www.austrofoma.at/news.html). The FORMEC Conference is organised by the Institute of Forest Engineering, University of Natural Resources and Life Sciences, Vienna (BOKU). The core issues targeted at the FORMEC Conference are: Þ Forest harvesting systems; Þ Wood transportation;

Croat. j. for. eng. 35(2014)2

Þ Forest road network planning and construction; Þ Environmental effects of forest operations and forest work sciences. AUSTROFOMA has established itself as an international gathering for the forestry and forest equipment sector. In the distinctive uplands of the Bohemian Forest, the practical and forward looking use of forestry machinery will demonstrate that managing forests in the interests of nature need not be incompatible with modern forestry technology. We are looking forward to welcoming many participants and to your interesting contributions! Karl Stampfer and Gernot Erber

109

Original scientific paper

Comparison of Cost Efficiency of Mechanized Fuel Wood Thinning Systems for Hardwood Plantations on Farmland Raffaele Spinelli, Natascia Magagnotti, Fulvio Di Fulvio, Dan BergstrĂśm, Matteo Danelon, Giorgio Alberti Abstract A harwarder is a machine used for both wood harvesting and extraction. A small and a large harwarder (SH and LH) were time studied whilst thinning hardwood plantations established on agricultural land in Italy. Two treatments were studied: whole tree sections (WT) or firewood logs integrated with tree tops (IH) were harvested and forwarded to the roadside. The selective thinning yielded 45 tonnes of fresh biomass (t) per hectare. The average productivity of the SH and LH with the WT harvesting treatment were 3.46 and 2.77 t per gross productive work hour, respectively. The SH was more efficient for felling and loading, while the LH was more efficient in the terrain transport work. The productivity of both machines was about 15% lower for IH treatment. The harwarder based thinning operation gave a harvesting cost between 18 and 34 â&#x201A;Ź/t under the conditions studied. Thus, the operational cost per t of the SH was less than for the LH. The harvesting cost decreased with increasing size of harvested trees for both machines. The level of stand damage caused by both harwarders was almost as low as the levels recorded in the literature for motor-manual thinning. The LH was able to handle larger trees than the SH in the studied conditions. The LH gives higher flexibility, since it can be used more efficiently in thinning of larger trees and in larger plantations than in the present study. Keywords: harvesting, harwarder, firewood, biomass, agricultural land

1. Introduction In the early 1990s, the EU launched a new afforestation programme, with the intention of controlling agricultural production, reducing the shortage of wood products and achieving a number of social and environmental benefits (Tassone et al. 2004). This programme was bolstered by a number of ambitious funding schemes, offering specific grants for the afforestation of arable land (Kassioumis et al. 2004). By the end of the 1990s, approximately 900,000 ha of land across the EU had been afforested with both softwood and hardwood (Du Breil 2000, Cogliastro et al. 2007). More hectares were planted in the following years, causing a significant increase in the European forest area. 148,000 ha of new plantations had been established in Italy by the year 2000, of which ca. 60% were planted with hardwood species (Magnani et al. 2005). The most popular species were walnut (Juglans regia Croat. j. for. eng. 35(2014)2

L.) and cherry (Prunus avium L.), which offered both fast growth and high-quality timber (Gold and Hanover 1987, Dupraz 1994). In general, the plantations were established with two or more main crop species, often combined with nurse tree species, which provide side protection and improve bole quality (Bohanek and Groninger 2003, Cutter et al. 2004). Currently, the success of these plantations depends on their capacity to produce good quality stem wood within a relatively short time, estimated to be between 30 and 40 years (Mary et al. 1999). Timely thinning has a crucial role to play in the selection of the best trees, maintaining the right density and preventing fast growing nurse trees from overtopping the main crop (Bohanek and Groninger 2005). Usually, early thinnings generate poor financial returns due to the handling of small trees, which have

111

R. Spinelli et al.

Comparison of Cost Efficiency of Mechanized Fuel Wood Thinning Systems for Hardwood ... (111–123)

relatively low financial value but have a relatively high operational cost per unit harvested (Kärhä et al. 2003, Spinelli and Magagnotti 2010). Traditionally, forest owners can balance the cost of thinning young forests with revenues drawn from harvest of older forests. That is not the case with the new plantations, which were all planted in the same period. Furthermore, tree plantations on farmland are often small and scattered, which makes machine relocations a critical cost issue (Spinelli et al. 2009). Given the large area planted in the last two decades, Europe is facing a serious forest management problem and risk to miss an opportunity to obtain much needed biomass. In thinning stands, where a large share of the cut trees are undersized for e.g. pulp production, a fuel wood harvest, or a combined fuel wood and pulp wood harvest, may yield higher profits compared to a pure roundwood harvest (cf. Di Fulvio and Bergström 2013). Therefore, the key issue is to develop effective work systems that can make thinning of the new hardwood plantations established on ex-arable land financially viable. A possible solution can be the use of a single machine system that combines the harvester and forwarder work, a harwarder. The main advantages of this machine system are that: Þ an entrepreneur owns and relocates only one machine at a time (Asikainen 2004), Þ its productivity, compared to a dual machine system, can also achieve a similar level for short hauling distances. For this reason, the harwarder may offer a lower harvesting cost compared to the standard harvester and forwarder system when the thinning intensity is below 55 m3 ha–1 (Kärhä 2006) and/or stand size is below a biomass removal <250 m3 (Väätäinen et al. 2006a). Furthermore, some harwarders can be driven on public roads and can therefore be independently relocated over short distances, without needing a truck and trailer unit. The objective of this work was to study the productivity and cost efficiency of harwarder systems used for fuel wood thinning of hardwood plantations. The following factors were considered: two machine sizes (small or large) and two products harvested (parts of whole trees or firewood logs and tree tops). The impact of the machines on the ground and remaining trees was also measured and compared.

2. Materials and methods Two different thinning systems were considered: Whole Tree (WT); harvesting and forwarding of unde-

112

limbed whole-tree sections to the roadside and Integrated Harvesting (IH); integrated harvesting of firewood logs and tree tops, with the two assortments separately forwarded to the roadside. Each thinning system was combined with two sizes of harwarder, »small« (SH) and »large« (LH), giving four unique treatments studied in the field.

2.1 Study sites Two sites in Italy were used for the trials, one located in San Daniele (46°08’N, 13°00’E) and one in Persereano (45°57’N, 13°17’E) (Table 1). The stand in San Daniele was 15 years old and was planted in rows spaced 4x2 m, with cherry (Prunus avium L.), European walnut (Juglans regia L.) and common ash (Fraxinus excelsior L.), the latter acting as a nurse tree. The stand in Persereano was 17 years old and was planted in rows spaced 4x3 m, with sessile oak (Quercus robur L.), cherry, sycamore maple (Acer pseudoplatanus L.) and common ash, the latter two species acting as nurse trees. 12 plots were located in San Daniele and 4 in Persereano. Each plot measured 50x40 m and was randomly assigned to each treatment, in order to allow evenly spread conditions for comparison. Each treatment was repeated four times, giving a total of 16 repetitions. Table 1 Stand characteristics of study sites: average values (standard deviation in parentheses) Stand

San Daniele

Persereano

12.5 (1.2)

13.8 (1.0)

1,187 (23)

853 (16)

14.8 (2.2)

12.8 (1.0)

Dry mass, odt/ha

58.4 (11.2)

52.5 (4.1)

Dry mass per tree, odkgb

49.3 (10.2)

61.6 (4.6)

Fresh mass, t/ha

94.2 (18.1)

84.7 (6.6)

Fresh mass per tree, kgbc

79.5 (16.5)

99.4 (7.4)

Plots, n a

DBH, cm

Density, trees/ha 2

Basal area, m /ha b

Mean diameter weighted by basal area Estimated by means of dendrometric parameters c Estimated fresh mass, for wood with a moisture content of 38% b

2.2 Harvesting and chipping systems The SH was a Vimek Biocombi 610 (Vimek AB, Sweden, www.vimek.se), with 6 wheels, a 44 kW engine and a weight of 4.9 t. It was equipped with a crane with a reach of 5.2 m which, in turn, was equipped Croat. j. for. eng. 35(2014)2

Comparison of Cost Efficiency of Mechanized Fuel Wood Thinning Systems for Hardwood ... (111–123)

R. Spinelli et al.

with a Hypro grapple-saw fitted with accumulating arms. The machine had a loading bunk with compacting stakes for load-compression. The LH was a Pfanzelt Felix 206 (Pfanzelt Maschinenbau GmbH, www.pfanzelt-maschinenbau.de), with 4 wheels and a 130 kW engine. It was equipped with a crane with a reach of 8.5 m. The boom tip was equipped with a quick connection device for time-efficient changing of heads. The heads used on the machine were a Logmax 5000 harvester head and a timber grapple for roundwood. The machine had an extendable load bunk and the complete unit weighed 14 t. All machines were operated by experienced operators, who had run their machines for several years but had no experience of these types of plantations.

tree cuts per crane cycle and loading while moving), driving loaded/unloaded, unloading work time and the delay times were all separately measured (cf. Magagnotti and Spinelli 2012). The sum of the net worktime elements was used as the Productive Machine work time (PM0). PM0 was converted to gross productive work time (PMH15), which includes delays shorter than 15 minutes. Delay time is, however, typically erratic and its magnitude may vary greatly over time. Thus, it is most probably incorrectly measured when measured over short periods, as in this study (c.f. Spinelli and Visser 2008). For this reason a delay factor of 0.835 was used to convert PMH0 to PMH15 for the two harwarders according to Kuitto et al. (1994).

2.3 Work methods

2.5 Field measurements

Both sites were thinned selectively, in order to create enough space around good quality crop trees, while removing the trees that were defective in some way, or direct competitors. The thinning intensity ranged between 40% and 55% of the initial tree density per ha. At the time of harvest, the trees to be cut had already been selected and marked. The SH thinned 2 rows of plantation per swath. In the WT treatment, the trees were cut and directly loaded onto the load bunk; the tree tops were first cut on the standing trees, then the butt log was felled and loaded. In the IH treatment, ca. 3–4 m long tree tops were cut off from the standing tree and directly loaded onto the bunk. Subsequently, in a separate load, the firewood butt logs were cut and directly loaded. The LH carried out cutting and forwarding as separate operations. It cut 2–4 rows per swath. In the WT treatment, trees were cut and bucked in ca. 5 m lengths. Tree branches were compressed with the harvester head while feeding the material through the head. Bunched tree sections were piled along strip-roads. In the IH treatment, separate piles were produced for firewood logs and treetops. Subsequently, the harvesting head was switched for the timber grapple and the load space was extended for forwarding work.

The harvested mass per plot and load were weighed using a portable plate scale. Different products were weighed separately. At the end of the trial, Table 2 Cost assumptions and machine hourly rates Small harwarder

Large harwarder

Investment , €*

145,000

350,000

Resale (20%), €

29,000

70,000

1,400

19,333

35,000

Interests, € year

3,867

9,100

Insurance, € year–1

2,500

9,450

16,800

945

1,680

9,667

17,500

45,762

82,580

32.7

59.0

10.5

15.8

63.2

94.8

Machine

Service life, years –1

Utilization, PMH15 year Interest rate, % Depreciation, € year–1 –1

–1

Diesel, € year

–1

Lubricant, € year

Maintenance, € year–1 –1

2.4 Time study

Total, € year

The time study was carried out between the 21st and 25th of October 2013. It was carried out during daylight; there was some rainfall during the study. The time consumption per plot and machine configuration was measured using a Husky Hunter™ field computer running Siwork 3 software. Recording was carried out at 0.6 s intervals (100 per minute). The net work-time elements cutting and loading (i.e. including crane movement time from first to last

Total, € PMH15–1

Croat. j. for. eng. 35(2014)2

Crew, n. –1 15

Labour, € PMH

Overheads (20%), € PMH15–1 –1 15

Machine rate, € PMH

* The investment cost was obtained from the manufacturers and was based on 2013 price lists

113

R. Spinelli et al.

Comparison of Cost Efficiency of Mechanized Fuel Wood Thinning Systems for Hardwood ... (111–123)

all the biomass was chipped and transported for weighing at a certified weighbridge. Weighbridge figures were then used to correct the plate scale figures. The moisture content (MC, wet basis) was determined in total on 9 wood chips samples (each sample=1 L). Samples were collected and weighed fresh in the field and then dried in a ventilated oven at 70˚C until a constant weight was reached. The fuel consumption of each machine was measured by starting with a full tank and refilling it at the end of the working day.

2.6 Machines cost Machine costs were calculated using the method described by Eliasson (2013) (Table 2). Machine service life estimates as well as the costs of insurance, repairs and services were obtained directly from the machine owners. The labour cost was set to 20 € per PMH15, inclusive of indirect salary costs. The calculated operational cost of all machines was increased by 20% to account for overhead costs (cf. Hartsough 2003).

2.7 Soil compaction and tree damage measurements In the San Daniele study site, stand damage and soil compaction measurements were carried out. Stand damage was determined by inspecting all standing trees left in each plot after harvest. Wounds with an exposed area smaller than 10 cm2 were not recorded, as they had little impact on a tree’s health or wood quality (cf. Whitney 1991). Soil compaction was determined by sampling 10 cores per plot: 5 on inter-rows driven over by the machines and 5 on inter-rows that had not been driven over by machinery representing undisturbed soil conditions. Cores were collected in rings made of thin walled stainless steel tubing, with an internal diameter of 8 cm and a height of 5 cm, corresponding to a volume of 250 cm3. Rings were pushed into the soil, down to a depth of 5 cm, after removing the litter layer. These rings were then removed from the soil and the samples were trimmed and placed into sealed plastic bags (one sample per bag). Samples were weighed before and after being oven-dried until they reached a constant weight at 70˚C. These data were used to calculate the bulk density and the moisture content of each sample. Once in the laboratory, the soil was passed through a 2 mm sieve in order to calculate total porosity. The ground pressure applied by the loaded axles of both harwarders was calculated as described by Komandi (1990), by using the maximum axle loads obtained from the portable scales as input data.

114

2.8 Analysis and statistics Analysis of variance (ANOVA) was used in order to analyse initial stand and removal properties. A general linear model (GLM) was used for analysis of time consumptions, productivities and costs. Differences between the methods were examined using Tukey’s post-hoc HSD pair-wise tests of means. The properties of the initial stand, the removal characteristics, time consumption, productivities, costs, damage and soil impact in each plot were compared using the model: g ij = m +ai + bj + ai ´ bj + eij

(1)

Where: m overall mean, a harwarder: »small« (SH) vs. »large« (LH), b product: »whole trees (WT)« vs. »integrated (IH)«, e random error. Removal per ha and tree mass were tested as covariates in the GLM. Harvesting costs for the different harwarders and treatments were then modelled as a function of harvested tree mass and annual usage of the equipment. Statistics were carried out using Minitab® (Minitab Inc.).

3. Results 3.1 Properties of removal In total, 140 fresh t (88 odt) were harvested during the experiment. The MC of the biomass ranged from 34% to 42%, with a mean value of 38%. The thinning intensity reached almost 50% of trees/ha, which corresponded to an average removal of 45 fresh t/ha (28 odt/ha) (Table 3). The mean tree mass harvested varied between 57 and 114 fresh kg (from 35 to 71 odkg) (Table 3). The average DBH (weighted by basal area) before thinning was 13 cm, the average DBH of the remaining trees was 14 cm after the thinning operation. The percentage of removed trees and the removal mass per ha were similar for the four different treatments, while the average harvested tree size was slightly larger in treatments with the small harwarder, with this difference being close to significant (Table 3). The large harwarder extracted 9% more firewood biomass compared to the small machine (Table 3) and this difference was significant. The forwarding distance was, on average, 217 m for all treatments (Table 3). Croat. j. for. eng. 35(2014)2

Comparison of Cost Efficiency of Mechanized Fuel Wood Thinning Systems for Hardwood ... (111–123)

R. Spinelli et al.

Table 3 Properties of removals Harwarder

Product

Removal fresh mass, tha-1

Removal fresh mass per tree, kg

Extraction distance, m

Firewood, % total fresh mass

ANOVA p-value*

Harwarder p-value

Product p-value

Mean

0.815

0.771

Sd.

–

Min.

–

Max.

–

Mean

0.262

0.790

Sd.

–

Min.

–

Max.

–

Mean

0.055

0.0531

Sd.

–

Min.

–

Max.

108

115

–

Mean

231

210

218

211

0.860

Sd.

101

–

Min.

140

125

100

150

–

Max.

360

311

332

282

–

Mean

–

<0.001

–

Sd.

–

Min.

–

Max.

–

Harvested plots, n

Removal intensity, % number of trees

Note: Fresh mass, for wood with a moisture content of 38%. *The bold p-values indicate statisticalyl significant difference ( p£0.05). The p-value for interactions harwarder × product are >0.8 in all cases (not shown in this table)

3.2 Work efficiency The total study time for the two harwarders (including machine delays) was 45.9 hours, of which delay time represented 3.7 hours (8.1%). The incidences of delays on the PM0 time for the SH and LH was 10.1% and 7.3%, respectively. A full load for the SH contained, on average, 3.2 t (2.0 odt) (Sd=0.7 t) of whole-tree biomass, 2.7 t (1.7 odt) (Sd=0.5 t) of firewood logs and 1.7 t (1.1 odt) (Sd=0.3 t) of tree tops, corresponding respectively to 70%, 54% and 34% of the load capacity (5 t) for this machine. A full load of the LH contained 4.2 t (2.6 odt) (Sd=1.0 t) of whole-tree biomass, 4.2 t (2.6 odt) (Sd=1.2 t) of fireCroat. j. for. eng. 35(2014)2

wood and 3.7 t (2.3 odt) (Sd=1.0 t) of tree tops corresponding respectively to 42%, 42% and 37% of the load capacity (10 t). The average driving speed was 1.1 m/s (Sd=0.2 m/s) and 1.5 m/s (Sd=0.4 m/s), respectively, for the SH and LH. Of the total recorded PM0 time, the felling and loading work represented 78%, extraction work 12% and unloading work 10%. The tree mass had a significant effect on the total PMH0/t, due to the fact that felling/loading efficiency was significantly dependent on tree size (Table 4). However, the tree size had no significant effect on driving time and it was not significant for unloading work. The removal (t/ha) was

115

R. Spinelli et al.

Comparison of Cost Efficiency of Mechanized Fuel Wood Thinning Systems for Hardwood ... (111–123)

Table 4 Analysis of variance table for the GLM harvesting time consumption, productivity and cost

Felling and loading, PMH0/t

Driving, PMH0/t

Unloading, PMH0/t

Total net time, PMH0/t

Gross productivity, t/PMH15

Harvesting cost, €/t

Variables

Adj SS

Adj MS

p-value*

Tree mass, kg

0.73565

27.88

0.000

Harwarder

0.32407

12.28

0.006

Harwarder×Tree mass, kg

0.14028

5.32

0.044

Product

0.47226

17.90

0.002

Harwarder×Product

0.01313

0.50

0.497

Residual error

0.26386

0.02639

–

Tree mass, kg

0.009277

4.01

0.073

Harwarder

0.024931

10.79

0.008

Harwarder×Tree mass, kg

0.006282

2.72

0.130

Product

0.040255

17.42

0.002

Harwarder×Product

0.025437

11.01

0.008

Residual error

0.023109

0.002311

–

Tree mass, kg

0.000130

0.09

0.768

Harwarder

0.008940

6.34

0.031

Harwarder×Tree mass, kg

0.000310

0.22

0.649

Product

0.000479

0.34

0.573

Harwarder×Product

0.001261

0.89

0.367

Residual error

0.014109

0.001411

–

Tree mass, kg

0.88851

23.97

0.001

Harwarder

0.10038

2.71

0.131

Harwarder×Tree mass, kg

0.07710

2.08

0.180

Product

0.74988

20.23

0.001

Harwarder×Product

0.00009

0.00

0.962

Residual error

0.37063

0.03706

–

Tree mass, kg

1.138

22.25

0.001

Harwarder

0.031

0.61

0.453

Harwarder×Tree mass, kg

0.014

0.27

0.612

Product

1.097

21.45

0.001

Harwarder×Product

0.050

0.98

0.346

Residual error

0.512

0.051

–

Tree mass, kg

102.32

18.10

0.002

Harwarder

72.47

12.82

0.005

Harwarder×Tree mass, kg

18.31

3.24

0.102

Product

88.99

15.74

0.003

Harwarder×Product

3.39

0.60

0.456

Residual error

56.53

5.65

–

Note: Fresh mass, for wood with a moisture content of 38%. *The bold p-values indicate statistically significant difference (p 0.05)

116

Croat. j. for. eng. 35(2014)2

Comparison of Cost Efficiency of Mechanized Fuel Wood Thinning Systems for Hardwood ... (111–123)

strongly correlated (p<0.001) to the harvested tree mass, which explained 60 % of its variability. For this reason, only the tree mass was used as a covariate in the analyses. The interaction of tree mass and harwarder type was significant for the felling and loading work-time; the SH was more sensitive to an increase in tree size than the LH. The type of harwarder had a significant effect on all time elements: the SH was more efficient for the felling and loading work, while the LH was more efficient for the driving and unloading work. The difference in total PM0 time consumption per t was below 5% in the average studied conditions and it was not significant (Tables 4 and 5). The type of product had a significant effect on the PM0 time consumption: the WT treatment had, on avTable 5 Corrected averages according to the GLM for the average tree mass studied (82 kg), minimum and maximum time consumptions, productivities and costs Harwarder Product

LH IH

Felling and loading, PMH0/t

0.167a

0.200ab

0.225b

0.271c

Min.

0.139

0.158

0.231

0.217

Max.

0.183

0.220

0.279

0.323

Driving, PMH0/t

0.036

0.057

0.019

0.022c

Min.

0.026

0.044

0.013

0.019

Max.

0.041

0.070

0.028

0.024

Unloading, PMH0/t

0.042a

0.039a

0.015b

0.016b

Min.

0.035

0.011

0.013

Max.

0.047

0.041

0.018

0.017

Total net time, PMH0/t

0.245a

0.295b

0.259ab

0.308b

Min.

0.208

0.244

0.256

0.252

Max.

0.271

0.325

0.318

0.360

Gross productivity, t/PMH15

3.46a

2.81b

3.19ab

2.77b

Min.

3.08

2.57

2.63

2.32

Max.

4.02

3.43

3.26

3.3 Productivity and cost On average for all treatments, the gross harvesting productivity (including felling, extraction and delay time) varied from 2.3 to 4.0 t/PMH15 (1.4–2.5 odt/PMH15). The average fuel consumption was 4.5 and 8.0 l/PMH15 for the SH and LH, respectively. The harvesting cost per t was 34–39% significantly lower for the SH (Table 5). The IH treatment cost from 4.2 to 4.5 €/t more than WT, and the combination of IH treatment and the large harwarder cost significantly more (Table 5). Two cost models (Eqs. 2 and 3) for the SH (CSH) and LH (CLH) harwarder, based on the results in Table 4 (all significant variables included) were obtained: CSH = 29.170 – 0.105 ´ Tm – 1.932 ´ P [€/t]

(2)

CLH = 53.152 – 0.252 ´ Tm – 2.402 ´ P [€/t]

(3)

R2 adj. = 0.93, F = 50.1, p = 0.001 Where: Tm tree mass, kg, P product type, dummy variable (0=IH, 1=WT).

3.32

Harvesting cost, €/t

18.27

22.50

29.71

34.22c

Min.

15.75

18.46

29.03

28.56

Max.

20.52

24.58

36.09

40.89

Note: weights are fresh, for wood with a moisture content of 38% Different superscript letters in individual rows indicate a significant difference (p 0.05) between treatments according to Tukey’s HSD tests of means

Croat. j. for. eng. 35(2014)2

erage, a 16% lower time consumption per t than the IH treatment and this reduction was significant in the case of the SH (Tables 4 and 5).

R2 adj. = 0.90, F = 34.3, p = 0.001

Harvested plots, n

R. Spinelli et al.

Fig. 1 Harvesting cost as function of tree mass for the four different treatments (calculated using Eqs. 2 – 3 with the average removal fresh mass set to 45 t/ha)

117

R. Spinelli et al.

Comparison of Cost Efficiency of Mechanized Fuel Wood Thinning Systems for Hardwood ... (111–123)

The harvesting cost at the average stem mass (82 kg) varied between 18 and 34 €/t (30 and 55 €/odt). It was one third lower for the SH system (Table 5). Of the total harvesting cost, felling and extraction accounted for 82% to 85%, and delays accounted for 15% to 18% (Fig. 2).

3.4 Soil compaction and frequency of tree damage The initial soil densities were similar for the different plots (Fig. 3).The average soil bulk density increased from 1.04 to 1.25 gcm–3 for the SH and from 1.08 to 1.47 gcm–3 for the LH (Fig. 3). The effect was significant for the LH (p=0.004) (c.f. Fig. 3). The soil porosity was, on average, 40 % and decreased to 30% for the SH and 20% for the LH. The effect was significant for the LH (p=0.01).

Fig. 2 Average harvesting cost by treatment and activity in the field study. (Calculated with the average removal fresh mass set to 45 t/ha and average removal tree fresh mass set to 82 kg) If the tree size is increased from 50 kg to 120 kg, the harvesting cost for the SH and LH is reduced by 32% and 45%, respectively (Fig. 1).

Damage to residual trees varied between 1% and 4% of the total number of remaining trees. Mean damage frequency was slightly higher for the LH and the WT treatment, but the differences were not significant.

4. Discussion The SH was found to be more time-efficient for the felling and loading work compared to the LH. This can be explained by the fact that the SH direct-loaded the cut trees while the LH was unable to direct-load. The inter-row space in the test plantations was too narrow for direct-loading of the LH. Many studies indicate that direct-loading is more efficient than separate loading (Andersson and Eliasson 2003, Talbot et al. 2003, Ringdahl et al. 2012, Wester and Eliasson 2003). The LH felling and loading work efficiency per t increased rapidly with harvested tree size, while the SH efficiency was less sensitive to different tree sizes. This could suggest that the SH was already operating close to the limit of its size capacity and, therefore, increases in tree size were balanced by proportional increases in working time. The LH was more efficient for the transportation work due to both higher driving speed and larger load capacity. In addition, the LH had higher efficiency for the unloading work since it was equipped with a timber-grapple, which had a larger handling capacity than the grapple-saw used on the SH.

Fig. 3 Average soil bulk density before and after machine activity for the four analysed treatments (n=3 per column, the error bars represent the standard deviations). Columns that do not share any letter are significantly different (p 0.05) according to Tukey’s HSD tests of means

118

A different share of harvested firewood was obtained for the two machines, which could be due to the different ways trees were handled by the operators. The main instruction given to them was to produce firewood logs up to the lowest large branch. In each case, the operators selected the bucking point. The inCroat. j. for. eng. 35(2014)2

Comparison of Cost Efficiency of Mechanized Fuel Wood Thinning Systems for Hardwood ... (111–123)

R. Spinelli et al.

tegrated harvesting of firewood logs and tree tops reduced the felling/loading and extraction work efficiency, compared to whole-tree sections. This is explained as follows: Þ in the IH treatment, the wood needed to be sorted into separate loads into the stand, while in the WT no sorting took place, Þ in the IH treatment, a firewood load contained a similar mass to a whole-tree section load, but a full load of tree crowns contained 12–47% less mass than loads of whole-tree sections. The productivities of the two harwarders were similar. The productivity levels for the LH are very close to those reported in the literature for similar machines in early thinning operations in boreal forests in Scandinavia (c.f. Laitila and Asikainen 2006, Di Fulvio and Bergström 2013). The productivity of the small Vimek harwarder used in this study was almost twice as high compared to a previous model of the machine studied when used in the early thinning of Pinus contorta (Nordin 2011). However, in Nordin´s (2011) study, the ground conditions were much more challenging than in this study. This may help explain the large difference in productivities especially when coupled with the fact that the engine power has increased 2.4 times and the machine load capacity has increased by 35% on this new model, resulting in a 29% larger load for this study. The main asset of an LH is the flexibility of the machine. For example, it can be used for a variety of operations, i.e. from early thinning to final felling work of trees with stem volumes up to 0.3 m3. The SH used in this study is designed for early thinning and the purchase of an SH depends on the ability to secure enough thinning work on a yearly basis. The LH may, however, complement the thinning of a small number of plantations with other, more conventional, forest operations. Furthermore, the harvester head on the LH is better suited for processing logs than the simple grapple-saw on the SH, which provides a higher potential for value recovery, as saw, pulp and firewood logs can attract a higher price than energy chips (Spinelli et al. 2013). Thus, it seems that the LH is likely to have a higher annual utilization than the SH. A sensitivity analysis was carried out under the assumption that both machines were exclusively used for the early thinning of plantations on farmland. The annual harvested mass varied from 1,000 to 6,500 fresh t. The annual utilization (PMH15/year) of each harwarder was then re-calculated, by assuming the mean harwarder productivity (t/PMH15) as in Table 5. The analysis showed that the harvesting cost is below 35 €/t for the LH, if the machine harvested at least Croat. j. for. eng. 35(2014)2

Fig. 4 Harvesting cost as a function of annual harvested biomass. (Calculated with the average removal fresh mass set to 45 t/ha and average removal tree fresh mass set to 82 kg) 4,000 t per year (i.e. at least 1,200 PMH15/year) (Fig. 4 and 5); this figure corresponds to about 90 ha per year given the average removal of 45 t/ha. Assuming a maximum utilization of 2,000 PMH15 per year, one single machine could not harvest more than 150 ha per year.

Fig. 5 Harvesting cost as a function of annual usage of harwarders for four treatments. (Calculated with the average removal fresh mass set to 45 t/ha and average removal tree fresh mass set to 82 kg)

119

R. Spinelli et al.

Comparison of Cost Efficiency of Mechanized Fuel Wood Thinning Systems for Hardwood ... (111–123)

The two harwarders cost the same per t for WT when the LH harvests 6,500 t/year and the SH harvests 3,000 t/year (Fig. 4). These figures correspond to an annual utilization of 900 and 2,000 PMH15/year, respectively, for the SH and LH (Fig. 5). For the IH harvesting, they cost the same per tonne if the SH is used for 1,000 PMH15/year and the LH for 2,000 PMH15/year (Fig. 5). At any annual utilization rate, both harvesting chains presented in this study resulted in a lower harvesting cost than any of the motor-manual alternatives studied in similar conditions by Magagnotti et al. (2011, 2012). If the current labour cost and fuel prices are taken into account, the motor-manual thinning costs between 25 and 40 €/t, while the mechanized chains tested at the time cost between 20 and 22 €/t. The actual mechanized harvesting system studied cost between 18 and 34 €/t. Therefore, the SH might offer cost savings compared to the other mechanized options tested before. In contrast, the LH offers no cost benefits over conventional dual-machine systems. The harwarder options offer additional savings on relocation cost, especially considering that both harwarders tested are road-legal, which means that they can relocate independently over short distances. This may offer some benefits over the other mechanized chains such as e.g. excavator based feller bunchers that need dedicated transportation for relocation (Väätäinen et al. 2006b). This advantage of the harwarder becomes especially important with farmland forests in Italy, which are typically fragmented and may have an average area below 1 ha (Alberti et al. 2005). The impact on both stands and soil was minor in this study and was near to the levels recorded for motor-manual operations, where the frequency of residual stand damage was 3.4% and the increase in soil bulk density 17% (Magagnotti et al. 2011). In particular, soil bulk density after machine activity in this study was still below the 1.7–1.8 gcm–3 range considered as the threshold value for optimal root elongation (Heilman 1981). The significant soil compaction in the case of the LH could be related to the higher ground pressure produced, which was calculated to 270 kPa, compared to a ground pressure of 220 kPa obtained for the SH using Komandi’s (1990) formula. The compaction found can also be related to the high moisture content recorded at the time of harvest (21%), which was similar on all harvested plots. The depth at which the observations for evaluation of compaction and porosity were made seemed to be appropriate, as the main impacts of wood extraction are generally concentrated within the first 10 cm layer (Ampoorter et al. 2010), especially in Mediterranean and sub-Mediterranean soils (Makineci et al. 2007).

120

4.1 Strengths and weaknesses of the study The two harwarder systems were studied under similar thinning conditions with no significant differences in thinning intensity and biomass removal per ha. Somewhat larger trees were, however, harvested with the SH, and this could have had some effect on our results; however this effect would be small as the tree size was used as a covariate when analysing time consumption and costs. Removal intensity, biomass yield and tree mass are similar to the ones reported by previous studies of the same type of operations in similar plantations (Magagnotti et al. 2011, Magagnotti et al. 2012), suggesting that the test sites are representative of average work conditions. A plot length of 50 m was too short for accumulating a full load when the stand stocking was low. For this reason, the extraction times were corrected for partial loads, occasionally derived from study design. In such a case, machine travel time was corrected using the ratio between the actual scaled load and the reference full load. The latter was assumed to be the maximum load actually carried during the whole study for each machine and assortment, which was visually assessed as the »optimum« load size, when the bunk was clearly unable to accommodate any more wood. In order to remove the effect of different forwarding distances, the extraction time for each plot was corrected in each single plot by using the speed of the machine in the plot and the average forwarding distance recorded over the whole study. Both operators were professionals and had worked with their respective machines for several years in thinning operations. However, productivity levels between harvester operators have been noted to vary significantly by up to 40% in thinning (Ovaskainen et al. 2004). The LH needed to change its configuration from a harvester to a forwarder and this extra time was not accounted for in this study (ca. 20 minutes for each configuration change c.f. Di Fulvio and Bergström 2013). This time can be significant in small stands and it will be less relevant as the stand area or removal volumes per site increase.

4.2 Future work The results of this and other studies (Magagnotti et al. 2011, Magagnotti et al. 2012) provide information on harvesting systems for selective thinning of hardwood plantations on farmlands. More detailed analyses of the whole supply system from the plantation to the end users are needed, where the processing and Croat. j. for. eng. 35(2014)2

Comparison of Cost Efficiency of Mechanized Fuel Wood Thinning Systems for Hardwood ... (111–123)

transportation of firewood and whole tree sections to their respective customers are also included. These kinds of analyses can be carried out using simulations of machines working in different plantation environments (i.e. in terms of tree sizes, growing stocks and plantation sizes) with the introduction of factors such as market demands and biomass prices as stochastic variables.

5. Conclusions The main objective of this work was to study the productivity and cost efficiency of harwarder systems in fuel wood thinning of hardwood plantations considering: two machine sizes (small and large) and two products harvested (1: tree-parts of whole trees and 2: firewood logs and tree tops). The productivity of the harvesting work was, on average, 14–24% higher for removal of whole-tree parts in comparison to the extraction of firewood logs and tree tops. This difference was significant for the SH and close to significant for the large harwarder. The SH was more efficient for felling and loading, while the LH was more efficient in the extraction work. Therefore, the productivity was similar for the two harwarders in the studied conditions. The harvesting cost was, on average, 18.3–29.7 €/t when harvesting whole-tree parts and was 22.5–34.2 €/t for the harvesting of firewood and tree tops. The SH cost significantly less for both treatments (WT and IH). The system with the SH was more cost-effective due to the lower hourly operational cost of the harwarder, which was mainly as a result of the lower machine investment cost. For an annual utilization time of 2,000 hours for the LH, the SH must be used for at least 900 hours/year when harvesting whole-tree parts to equal the cost-efficiency of the LH and 1,000 hours/year for an integrated harvest. The main drawback of the SH is the fact that it is limited to thinning works. Therefore it requires securing of enough thinning work during the year to reach sufficient utilization and a reasonable hourly cost compared to a larger machine. This study indicates that thinning of farmland plantations may offer good working conditions (e.g. flat areas, absence of roughness, schematic work pattern) for a small harwarder, but they remain challenging for a large machine due to the relatively small trees handled and the limited manoeuvrability in the stands. The LH is, thus, more efficient in stands with long forwarding distances (e.g. large plantation areas) and in the removal of large tree sizes. Croat. j. for. eng. 35(2014)2

R. Spinelli et al.

Acknowledgements The research leading to these results has received funding from the Regione Autonoma Friuli Venezia Giulia under the LR 26/2005 art. 16 funding scheme (»Arboplan Project«) and from the European Union Seventh Framework Programme (FP7/2012-2015] under grant agreement n°311881 (»INFRES Project«). The authors gratefully acknowledge the support of Mr. Diego Chiabà and Ms. Giulia Olivotto with field data collection and laboratory analyses.

6. References Alberti, G., Candido, P., Peressotti, A., Turco, S., Piussi, P., Zerbi, G., 2005: Aboveground biomass relationships for mixed ash (Fraxinus excelsior L. and Ulmus glabra Hudson) stands in Eastern Prealps of Friuli Venezia Giulia (Italy). Annals of Forest Science 62(8): 831–836. Ampoorter, E., Van Nevel, L., De Vos, B., Hermy, M., Verheyen, K., 2010: Assessing the effects of initial soil characteristics, machine mass and traffic intensity on forest soil compaction. For. Ecol. Manag. 260(10): 1664–1676. Andersson, J., Eliasson, L., 2004: Effects of three harvesting work methods on Harwarder productivity in final felling. Silva Fennica 38(2): 195–202. Asikainen, A., 2004: Integration of work tasks and supply chains in wood harvesting – cost savings or complex solutions? International Journal of Forest Engineering 15(2): 11–17. Bohanek, J., Groninger J., 2003: Impacts of intensive management on black walnut growth and bole quality at mid-rotation. Forest Science 49(4): 522–529. Bohanek, J., Groninger, J., 2005: Productivity of European black alder (Alnus glutinosa) interplanted with black walnut (Juglans nigra) in Illinois, U.S.A. Agrofor. Syst. 64(2): 99–106. Cogliastro, A., Gagnon, D., Bouchard, A., 1997: Experimental determination of soil characteristics optimal for the growth of ten hardwoods planted on abandoned farmland. For. Ecol. Manage. 96 (1–2): 49–63. Cutter, B., Coggeshall, M., Phelps, J., Stokke, D., 2004: Impacts of forest management activities on selected hardwood wood quality attributes: a review. Wood Fiber Sci. 36(1): 84–97. Du, Breil, de Pontbriand, J., 2000: European experiences with regulation 2080/92 and the new afforestation policy under Agenda 2000. Proceedings of the Scientific Symposium »New forests for Europe: afforestation at the turn of the century«. Freiburg, Germany 16–17 February: 23–50. Dupraz, C., 1994: Prospects for easing land tenure conflicts with agroforestry in Mediterranean France: a research approach for intercropped timber orchards. Agrofor. Syst. 25(3): 181–192.

121

R. Spinelli et al.

Comparison of Cost Efficiency of Mechanized Fuel Wood Thinning Systems for Hardwood ... (111–123)

Eliasson, L., 2013: Machine Cost Calculation Model, downloaded on Sept. 14th, from: http://www.forestenergy.org/ pages/costing-model---machine-cost-calculation/

ni, O., Rossi, F., 2005: What role for afforestation in Italian strategies towards the Kyoto Protocol? Hints from a Kyoto forest in the Po Valley (Northern Italy). Forest@ 2: 333–344.

Di Fulvio, F., Bergström, D., 2013: Analyses of a single-machine system for harvesting pulpwood and/or energy-wood in early thinnings, International Journal of Forest Engineering 24(1): 2–15.

Mary, F., Dupraz, C., Delannoy, E., Liagre, F., 1999: Incorporating agroforestry practices in the management of walnut plantations in Dauphiné, France: an analysis of farmers’ motivations. Agrofor. Syst. 43(1–3): 243–256.

Gold, M., Hanover, J., 1987: Agroforestry systems for the temperate zones. Agrofor. Syst. 5(2): 109–121.

Nordin, L., 2011: Produktivitet och lönsamhet vid skogsbränsleuttag i klena gallringar – En tidsstudie av Vimek 608 BioCombi i contortabestånd. Examensarbete. (Productivity and profitability in early bioenergy –thinnings – A time study of Vimek 608 BioCombi in stands of Lodgepole pine) SLU, Institutionen för skoglig resurshushållning. Arbetsrapport 315. http://stud.epsilon.slu.se/2469/

Hartsough, B., 2003: Economics of harvesting to maintain high structural diversity and resulting damage to residual trees. Western Journal of Applied Forestry 18(2): 133–142. Heilman, P., 1981: Root penetration of Douglas-fir seedlings into compacted soil. For. Sci. 27(4): 660–666. Kärhä, K., Jouhiaho, A., Mutikainen, A., Mattila, S., 2003: Mechanized energy wood harvesting from early thinnings. International Journal of Forest Engineering 16(1): 23–36. Kärhä, K., 2006: Whole-tree harvesting in young stands in Finland. Forestry Studies 45: 118–134. Kassioumis, K., Papageorgiu, K., Christodoulou, A., Blioumis, V., Stamou, N., Karameris, A., 2004: Rural development by afforestation in predominantly agricultural areas: issues and challenges from two areas in Greece. Forest Policy and Economics 6(5): 483–449. Komandi, G., 1990: Establishment of soil-mechanical parameters which determine traction on deforming soil. Journal of Terramechanics 27(2): 115–124. Kuitto, P.J., Keskinen, S., Lindroos, J., Oijala, T., Rajamäki, J., Räsinen T., Terävä, J., 1994: Mechanized cutting and forest haulage. Mätsäteho Tiedotus, Report 410, Helsinki, Finland ISBN 951-673-139-2. 38 p. Laitila, J., Asikainen, A., 2006: Energy Wood Logging from Early Thinnings by Harwarder Method. Baltic Forestry 12(1): 94–102. Makineci, E., Demir, M., Comez, A., Yilmaz, E., 2007: Chemical characteristics of the surface soil, herbaceous cover and organic layer of a compacted skid road in a fir (Abies bornmulleriana Mattf.) forest. Transport. Res. Part D, 12(7): 453– 459. Magagnotti, N., Nati, C., Picchi, G., Spinelli, R., 2011: Mechanized thinning of walnut plantations established on exarable land. Agroforestry Systems 82(1): 77–86. Magagnotti, N., Pari, L., Picchi, G., Spinelli, R., 2012: Energy biomass from the low-investment fully mechanized thinning of hardwood plantations. Biomass and Bioenergy 47: 195200. Magagnotti, N., Spinelli, R., 2012: Good practice guidelines for biomass production studies. COST Action FP-0902. Florence: CNR IVALSA. 50 p. Magnani, F., Grassi, G., Tonon, G., Cantoni, L., Ponti, F., Vicinelli, E., Boldreghini, P., Nardino, M., Georgiadis, T., Faci-

122

Ovaskainen, H., Uusitalo, J., Väätäinen, K., 2004: Characteristics and Significance of Harvester Operators’ Working Technique in Thinnings. Journal of Forest Engineering 15(2): 67–77. Ringdahl, O., Hellström, T., Lindroos, O., 2012: Potentials of possible machine systems for directly loading logs in cut-tolength harvesting. Canadian Journal of Forest Research 42(5): 970–985. Spinelli, R., Magagnotti, N., 2010: Comparison of two harvesting systems for the production of forest biomass from the thinning of Picea abies plantations. Scandinavian Journal of Forest Research 25(1): 69–77. Spinelli, R., Visser, R., 2008: Analyzing and estimating delays in harvester operations. International Journal of Forest Engineering 19(1): 35–40. Spinelli, R., Magagnotti, N., Picchi, G., 2009: Deploying mechanized Cut-to-Length technology in Italy: fleet size, annual usage and costs. International Journal of Forest Engineering 21(2): 23–31. Spinelli, R., Ebone, A., Gianella, M., 2013: Biomass production from traditional coppice management in northern Italy. Biomass and Bioenergy. Manuscript submitted in June 2013. Talbot, B., Nordfjell, T., Suadicani, K., 2003: Assessing the utility of two integrated harvester-forwarder machine concepts through stand-level simulation. International Journal of Forest Engineering 14(2): 31–44. Tassone, V., Wesseler, J., Nesci, F., 2004: Diverging incentives for afforestation from carbon sequestration: an economic analysis of the EU afforestation program in the south of Italy. Forest Policy and Economics 6(6): 567–578. Väätäinen, K., Liiri, H., Röser, D., 2006a: Cost competitiveness of harwarders in CTL-logging conditions in Finland – a discrete-event simulation study at the contractor level. Proceedings of the International Precision Forestry Symposium, Stellenbosch University, South Africa, 5–10 March. Econo Print, SA. 451–463 p. Väätäinen, K., Asikainen, A., Sikanen, L., Ala-Fossi, A., 2006b: The cost effect of forest machine relocations on logging costs in Finland. Forestry Studies 45: 135–41. Croat. j. for. eng. 35(2014)2

Comparison of Cost Efficiency of Mechanized Fuel Wood Thinning Systems for Hardwood ... (111–123) Wester, F., Eliasson, L., 2003: Productivity in Final Felling and Thinning for a Combined Harvester-Forwarder (Harwarder). International Journal of Forest Engineering 14(2): 45–51.

R. Spinelli et al.

Whitney, R., 1991: Quality of eastern white pine 10 years after damage by logging. For. Chronicle 67: 23–26.

Authors’ address: Raffaele Spinelli, PhD. e-mail: spinelli@ivalsa.cnr.it Natascia Magagnotti, PhD. e-mail: magagnotti@ivalsa.cnr.it CNR IVALSA Via Madonna del Piano 10 50019 Sesto Fiorentino (FI) ITALY Fulvio Di Fulvio, PhD.* e-mail: fulvio.di.fulvio@slu.se Dan Bergström, PhD. e-mail: dan.bergstrom@slu.se Department of Forest Biomaterials and Technology Swedish University of Agricultural Sciences Skogsmarksgränd, 90183, Umeå SWEDEN

Received: January 21, 2014 Accepted: March 11, 2014 Croat. j. for. eng. 35(2014)2

Matteo Danelon, MSc. e-mail: matteo.danelon@uniud.it Giorgio Alberti, PhD. e-mail: giorgio.alberti@uniud.it Department of Agricultural and environmental Sciences University of Udine, via delle Scienze 206 33100 Udine ITALY * Corresponding author

123

Original scientific paper

Estimating Annual Available Amounts of Forest Biomass Resources with Total Revenues and Costs during the 60-Year Rotation in a Mountainous Region in Japan Kazuhiro Aruga, Ayami Murakami, Chikara Nakahata, Reiko Yamaguchi, Masashi Saito, Takuyuki Yoshioka Abstract This study extracted production forests and estimated the annual available amounts of forest biomass resources under profitable forest management. Production forests were extracted as sub-compartments where expected revenues surpassed all costs, from planting to final harvesting, for a 60-year rotation. These revenues and costs were estimated for two types of timber harvesting systems (a conventional operation system using a chainsaw and mini-forwarder, and a mechanized operation system using a processor and forwarder) and three types of forest biomass harvesting systems (normal extraction, landing sales, and no biomass extraction) in each sub-compartment using a geographic information system. Then, annual available amounts of forest biomass resources were estimated on the basis of annual supply potentials from production forests. The model was then applied to Nasushiobara City and the Kanuma area in Tochigi Prefecture, Japan. As a result, the number of profitable sub-compartments was estimated as 2,814 out of a total of 5,756 in Nasushiobara City, and 22,872 out of a total of 32,851 in the Kanuma area. The annual amounts of available forest biomass resources were estimated as 11,849 m3 y–1 and 115,213 m3 y–1 in Nasushiobara City and the Kanuma area, respectively. These amounts largely exceed the annual demands of a 500 kW woody biomass power generation plant planned in Nasushiobara City (6,000 m3 y–1) and a chip production factory located in the Kanuma area (12,000 m3 y–1), respectively. €1 = 143 yen on March 13, 2011 Keywords: economic balance, geographic information system, harvesting system, production forest, supply potential

1. Introduction Japan is dependent on the imports of oil, coal, and natural gas for the majority of its energy supply. The energy self-sufficiency rate in 2010 was just 5% (Japan Forestry Agency 2013). In order to secure a stable supply of energy, alternatives to fossil fuel, for example »renewable energy« such as solar, wind, rivers, geothermal heat, and biomass will need to be developed. Among various biomass resources in Japan, woody biomass in particular attracts attention. This is not just because it is abundant, but also because its energy use Croat. j. for. eng. 35(2014)2

is expected to contribute to revitalizing forests and forestry product industries, which have been depressed for the last 30 years. Maintaining the relevant ecological, economic, and social functions of manmade forests, which are behind in tending, is also important. In July 2011, the »Feed-in Tariff (FIT) Scheme for Renewable Energy Use« was introduced in accordance with new legislation entitled the »Act on Purchase of Renewable Energy Sourced Electricity by Electric Utilities«. Under the FIT program, electricity generated

125

K. Aruga et al.

Estimating Annual Available Amounts of Forest Biomass Resources with Total Revenues ... (125â&#x20AC;&#x201C;138)

from woody biomass is to be procured for 20 years at a fixed price (without tax) for unused materials such as the logging residue: 32 yen/kWh, general materials such as sawmill residue: 24 yen/kWh, and recycled materials such as construction waste wood: 13 yen/kWh (Agency for Natural Resources and Energy 2012). Power generated from unused materials is offered incentives. Therefore, use of logging residue will be promoted in the near future. Numerous studies have examined the availability of woody biomass resources using GIS. Iuchi (2004) and Kamimura et al. (2009) developed techniques for estimating the supply potential of woody biomass, including logging residues, sawmill residues, and construction waste woods, in terms of regional energy in units of cities and towns. In addition to supply potentials, Yoshioka and Sakai (2005) and Kinoshita et al. (2009) devised techniques for estimating the regional harvesting volumes and costs of logging residues in units of sub-compartments corresponding to conventional forest management units in Japan, whereas Yagi and Nakata (2007) and Yamamoto et al. (2010) developed techniques that expressed them in units of kilometer-scale grids of cities and towns. Furthermore, Nord-Larsen and Talbot (2004), Aruga et al. (2006a), RĂ¸rstad et al. (2010), and Aruga et al. (2011) discussed the long-term feasibility of timber and forest biomass resources by predicting future forest resources using growth models while optimizing the allocation of fuel wood using linear programming or random search. Moreover, Aruga et al. (2006b) and Panichelli and Gnansounou (2008) discussed the scales and locations of bio-energy facilities based on the relationship between the annual available amounts and the procurement costs of forest biomass resources, whereas Ranta (2005), MĂśller and Nielsen (2007), and Viana et al. (2010) devised a technique for expressing it at a national level. In addition to these methods for the estimation of volumes and costs, Yamaguchi et al. (2010) and Nakahata et al. (2014) developed a technique for estimating the available amount of logging residues in consideration of the economic balances estimated from regional revenues and costs of both timber and logging residues in units of sub-compartments, whereas Kinoshita et al. (2010) and Kamimura et al. (2012) established a technique to express them in units of cities and towns. However, these studies have not considered regeneration expenses, which are important for conducting sustainable forest management. In contrast, Aruga et al. (2013) developed a method for extracting production forests based on economic balances by considering regeneration expenses after

126

final felling operations. Aruga et al. (2013) defines production forests as sub-compartments where expected revenues surpassed all costs of thinning and final felling operations. Then, Aruga et al. (2013) estimated available amounts of timber and forest biomass resources from profitable sub-compartments. Regeneration expenses include site preparation, planting, weeding, vine cutting, pruning, and forest inventory. About 2,500 seedlings/ha are assumed to be planted and weeding operations are assumed to be conducted once a year for ten years after planting in this study. Then, vine cutting operations are assumed to be conducted 10 and 12 years after planting and pruning operations are assumed to be conducted 15 and 25 years after planting. However, Aruga et al. (2013) estimated only the supply potentials and available amounts of timber and forest biomass resources based on the current situation. In order to plan power generation plants considering available amounts of forest biomass resources, future supply potentials and available amounts of forest biomass resources should be projected. Therefore, this study first estimated revenues and costs of precommercial and commercial thinning operations as well as final felling operations with two types of timber harvesting systems (a conventional operation system using a chainsaw and mini-forwarder, and a mechanized operation system using a processor and forwarder) and three types of forest biomass harvesting systems (normal extraction, landing sales, and no biomass extraction) in each sub-compartment using a geographic information system. Then, production forests were extracted as sub-compartments where expected revenues surpassed all costs, from planting to final harvesting, for a 60-year rotation. Finally, the most economical timber and forest biomass harvesting system for each sub-compartment was determined and annual available amounts of forest biomass resources were estimated on the basis of annual supply potentials from production forests.

2. Study sites and data The study sites were Nasushiobara City (Aruga et al. 2013) and the Kanuma area (Aruga et al. 2011) in Tochigi Prefecture, Japan. The gross area of Nasushiobara City is 59,280 ha, and the forest area is 38,689 ha (65% of the gross area). The area of national forests is 24,981 ha and that of private and local government forests is 13,708 ha. In this study, major plantation species such as Japanese cedar (Cryptomeria japonica) and Japanese cypress (Chamaecyparis obtusa), owned by private individuals and organizations as well as local Croat. j. for. eng. 35(2014)2

Estimating Annual Available Amounts of Forest Biomass Resources with Total Revenues ... (125–138)

K. Aruga et al.

Fig. 1 Stand species of Nasushiobara City governments, were analyzed. Private individuals and organizations along with local governments own 7,340 sub-compartments of Japanese cedar comprising 2,850 ha, and 2,521 sub-compartments of Japanese cypress comprising 1,103 ha (Fig. 1). The average slope angle is relatively low (10°) and the road network density is relatively high (27 m/ha). An agrarian organization in the Nasunogahara area in Tochigi Prefecture is willing to conduct thinning operations and extract thinned woods for woody biomass power generation in cooperation with a Forest Owners’ Co-operative in Nasushiobara City in order to nurture river resources as well as maintain forests for soil and water conservation. Croat. j. for. eng. 35(2014)2

The Kanuma area consists of Kanuma City and the town of Nishikata. The gross area is 52,200 ha and the forest area is 35,593 ha (68% of the gross area). The area of national forests is only 1,642 ha and that of private and local government forests is 33,951 ha. Private individuals and organizations as well as local governments own 30,104 sub-compartments of Japanese cedar comprising 17,341 ha and 19,957 sub-compartments of Japanese cypress comprising 9,950 ha (Fig. 2). The average slope angle is relatively high (22°) and the road network density is relatively low (18 m/ha). The Kanuma area was one of the famous forestry areas in Tochigi Prefecture (the site index that ranks the order of stand production capacity was higher than that for

127

K. Aruga et al.

Estimating Annual Available Amounts of Forest Biomass Resources with Total Revenues ... (125â&#x20AC;&#x201C;138)

Fig. 2 Stand species of the Kanuma area Nasushiobara City). A chip production factory, aimed at supplying a portion of the chips to a woody biomass power generation plant, is located in the Kanuma area. Forest-registration data (stand ages, tree species, and site indices) and GIS data (information on roads and sub-compartment layers) from the Tochigi Prefectural Government were used in the study, as were 50 m grid digital elevation models (DEMs) from the Geographical Survey Institute. Private individuals and organizations as well as local governments of Nasushiobara City own 4,456 sub-compartments of Japanese

128

cedar comprising 2,761 ha and 1,300 sub-compartments of Japanese cypress comprising 918 ha based on 50 m meshes. Those of the Kanuma area owned 22,735 sub-compartments of Japanese cedar comprising 17,247 ha and 10,116 sub-compartments of Japanese cypress comprising 7,262 ha. The sub-compartments in both areas were significantly fewer than the actual numbers because there were many small sub-compartments of less 0.25 ha that could not be recognized with 50-m meshes. However, the areas were not significantly smaller than the actual areas. Croat. j. for. eng. 35(2014)2

Estimating Annual Available Amounts of Forest Biomass Resources with Total Revenues ... (125–138)

K. Aruga et al.

Table 1 Direct operating expenses Operation

Expense, yen m–3

System

Reference

Chainsaw

Felling

53 / Vn + 65

Both

Nakahata et al. 2011

Chainsaw

Processing

39 / Vl + 329

Conventional

Nakahata et al. 2011

Processor

Processing

207 / Vl + 161

Mechanized

Nakahata et al. 2011

Mini grapple-loader

Bunching

1,999

Conventional

Nakahata et al. 2011

Grapple-loader

Bunching

1,199

Mechanized

Nakahata et al. 2011

Mini forwarder

Forwarding

(769 + 0.508 Lf) / Rf

Conventional

Nakahata et al. 2011

Forwarder

Forwarding

(378 + 0.301 Lf) / Rf

Mechanized

Nakahata et al. 2011

Truck

Transporting

(778 + 0.031 Lt) / Rt

Both

Sawaguchi 1996

Machine

Vn – stem volume, m3 stem–1; Vl – extracted volume per stem, m3 stem–1; Lf – forwarding distance, m; Lt – transporting distance, m; Rf and Rt – loading capacity rates

3. Methods Production forests were extracted and annual available amounts of forest biomass resources were estimated in the following order: 1) estimation of supply potentials of timber and logging residues based on the cutting and extraction rates in thinning and final felling operations during a 60-year rotation; 2) estimation of total expenses from planting to final felling operations during a 60-year rotation; 3) estimation of revenues from thinning and final felling operations during a 60-year rotation; 4) estimation of economic balances during a 60-year rotation; 5) extraction of profitable sub-compartments as production forests; and 6) estimation of annual available amounts of forest biomass resources on the basis of annual supply potentials from profitable sub-compartments. A simple version of »Methods« will be described below. Full technical details will be found in an earlier paper (Aruga et al. 2013). Thinning and final felling operations were assumed to be conducted based on stand ages. First (precommercial), second (commercial) thinning and final felling operations were assumed to be conducted at 25, 40, and 60 years old, respectively. Supply potentials of timber and forest biomass resources on each subcompartment were estimated from the cutting, extraction, timber, and forest biomass rates. The rate of forest biomass to whole tree for pre-commercial thinning operations was 100% whereas those for commercial thinning operations and final felling operations were 55% and 26%, respectively. This study investigated a conventional operation system and a mechanized operation system. This Croat. j. for. eng. 35(2014)2

study also examined three types of forest biomass harvesting systems: 1) normal extraction, 2) landing sales, and 3) no biomass extraction. All costs, including the direct and indirect operating expenses associated with each machine, strip-road and landing establishment expenses, and regeneration expenses, were estimated (Eq. 1). A=D+S+L+I+R

(1)

Here, A, D, S, L, I, and R denote all costs, direct operating expenses, strip-road establishment expenses, landing establishment expenses, indirect operating expenses, and regeneration expenses, respectively. Direct operating expenses, given in Table 1 (Nakahata et al. 2011, Sawaguchi 1996), included labor and machinery expenses (maintenance, management, depreciation, fuel, and oil expenses). The stem volume Vn (m3 stem-1) of each sub-compartment was estimated using yield tables (Forest Experiment Station 1955, 1961) with stand species, ages, and site indices in the forest registration, and the extracted volume per stem, Vl (m3 stem–1) was estimated by multiplying Vn with the stem extraction rate (80% for the first thinning and final felling operations and 50% for the second thinning operation). Forwarding and transporting costs were changed according to the loading capacity rates for timber and forest biomass resources upon the first thinning, second thinning, and final felling operations. Forwarding distances were estimated by the average distances from the landings to all grids within the sub-compartments. Landings were set within grids so as to minimize their distances from roads, centers of gravity in the sub-compartment, and the power gen-

129

K. Aruga et al.

Estimating Annual Available Amounts of Forest Biomass Resources with Total Revenues ... (125–138)

Table 2 Cost estimation Normal extraction 1 T

Landing sale

Felling

Processing

Bunching

Forwarding

Transporting

Landing establishment

Machine transportation

Garage maintenance

Incidental personnel

Overhead costs

X X

1 F

Handling fees

No biomass extraction

1 – first thinning; 2 – second thinning; 3 – final felling T – timber; F – forest biomass resources X – entire cost consideration; Y – cost consideration according to timber and forest biomass volumes

eration plant in Nasushiobara City or chip production factory in Kanuma City. Transporting distances from the landings to log markets and to the power generation plant in Nasushiobara City or chip production factory in Kanuma City were calculated with the Dijkstra method (Dijkstra 1959). If sub-compartments were unconnected to existing roads, forest road net-

works were planned to connect landings by the Dijkstra method. The expenses of forest road establishment, which should be paid for by public budgets, were not considered in this study. Strip roads were assumed to be established for forwarding operations. The strip-road density d (m ha–1) was assumed to be related to the average slope angle of each sub-compartment, q (°), according to the following equation (Fig. 3): d = 956.72 q–0.52

(2)

The strip-road cost S (yen) was then estimated by multiplying d by the following unit strip-road cost s (yen m–1): s = 67 e0.116 q for the conventional operation system (Sawaguchi 1996) (3) 0.117 q s = 220 e for the mechanized operation system (Sawaguchi 1996) (4) The expenses of the landing-establishment L (yen) were estimated by the following equation: L = 187.63 V

(5)

Here, V is the extracted volume per hectare (m ha–1). Machine transportation expenses, garage-maintenance expenses, incidental personnel expenses, overhead costs, and handling fees associated with the log market were considered as indirect operating expenses (Zenkoku Ringyo Kairyo Fukyu Kyokai 2001). 3

Fig. 3 Average slope angle and strip road density

130

Croat. j. for. eng. 35(2014)2

Estimating Annual Available Amounts of Forest Biomass Resources with Total Revenues ... (125–138)

In addition to these timber and forest biomass extraction costs, regeneration expenses included those for site preparation, planting, weeding, vine cutting, pruning, and forest inventory. The regeneration expenses were estimated as 2,512,376 yen ha–1 for Japanese cedar and 2,892,365 yen ha–1 for Japanese cypress (Okawabata 2003). For normal extraction of the first thinning operation, only thinned woods left in the forest after precommercial thinning were extracted as forest biomass resources. Therefore, all costs were related to forest biomass extraction (Table 2). Logging residues upon the second thinning and final felling operations were considered as a by-product of timber harvesting. Therefore, operations for forest biomass extraction started after processing, and all costs, excluding those for forest biomass extraction as well as associated indirect costs, were considered as timber extraction costs. For landing sales of the first thinning operation, only felling and processing costs as well as associated indirect costs were considered. For the case of no biomass extraction of the first thinning operation, only felling costs and associated indirect costs were considered whereas processing and extraction costs as well as associated indirect costs were not. For the first thinning operation of landing sales, only felling and processing operations were assumed and timber extraction was not considered. For no timber extraction, use of the processor was an unrealistic assumption even for the processing operation of the mechanized operation system. Therefore, chainsaw felling and processing operations were assumed. Therefore, small-sized strip roads for the conventional operation system were assumed to be constructed upon the first thinning operation of the mechanized operation system, and those for the mechanized operation system were assumed to be constructed upon the second thinning operation of the mechanized operation system. For the first thinning operation of no biomass extraction, no timber or forest biomass extraction was assumed. Therefore, strip roads were assumed to be constructed for the second thinning operation, unlike normal extraction and landing sales, for which they were assumed to be constructed upon the first thinning operation. Incomes were estimated using supply potentials and log prices: 11,000 yen m–3 for Japanese cedar, 22,000 yen m–3 for Japanese cypress, and 3,400 or 1,000 yen m–3 for normal extraction or landing sales, respectively, of forest biomass resources. For thinning operations, the subsidy was estimated by the standard unit cost, area, assessment coefficient, and subsidy rate (Tochigi prefectural government 2010). Subsidies on Croat. j. for. eng. 35(2014)2

K. Aruga et al.

the first thinning operation for normal extraction, the first thinning operation for landing sales or no biomass extraction, and the second thinning operation for all types of operations were 235,504 yen ha–1, 64,291 yen ha–1, and 338,188 yen ha–1, respectively. For strip-road establishment on sub-compartments where thinning operations were conducted with subsidies, the subsidy (Table 3) was also estimated with the standard unit cost, length, assessment coefficient, and subsidy rate (Tochigi prefectural government 2010). In this study, the subsidy for regeneration was also considered. Subsidies were estimated as 1,227,400 yen ha–1 for Japanese cedar and 1,219,240 yen ha–1 for Japanese cypress. Table 3 Subsidies for strip-road establishment (yen m–1) Average slope angle, °

Conventional

Mechanized

159

191

230

137

276

147

477

183

850

4. Results and discussion 4.1 Normal extraction Upon the first thinning operation with the conventional operation system, only a few sub-compartments on large areas and gentle slopes were profitable (Table 4). However, upon the first thinning operation with the mechanized operation system, no sub-compartments were profitable. With the mechanized operation system, machine transportation expenses were high and the extracted volumes of the first thinning operation were small. Therefore, machine transportation expenses per extracted volume were high. This was the reason why no sub-compartments were profitable upon the first thinning operation. Upon the second thinning and final felling operations, there were more profitable sub-compartments with the mechanized operation system than with the conventional operation system, because although machine transportation expenses with the former were higher, productivity was also higher and subsequent costs were lower. In particular, upon the final felling operation, almost all sub-compartments were profitable with the mechanized operation system.

131

K. Aruga et al.

Estimating Annual Available Amounts of Forest Biomass Resources with Total Revenues ... (125â&#x20AC;&#x201C;138)

Table 4 Profitable sub-compartments with normal extraction Nasushiobara City

Kanuma area

Rate, % Area, No. of subcompartments ha First thinning

No.

Area

Average slope Area, No. of subangle, compartments ha Â°

Rate, % No.

Area

Average slope angle, Â°

0.3

2.2

4.9

0.0

0.1

11.2

1,634

1,752

28.4

47.6

9.6

10,399

8,135

31.7

33.2

21.6

1,812

1,839

31.5

50.0

9.4

12,355

10,268

37.6

41.9

21.4

1,530

1,452

26.6

39.5

8.8

10,682

8,250

32.5

33.7

21.5

Total with R*

658

737

11.4

20.0

9.7

7,543

6,492

23.0

26.5

22.2

First thinning

0.0

2,161

2,609

37.5

70.9

11.9

16,155

19,139

49.2

78.1

22.7

5,550

3,628

96.4

98.6

9.5

32,822

24,502

99.9

100.0

21.5

3,218

2,968

55.9

80.7

9.6

21,058

20,920

64.1

85.4

21.6

1,022

1,352

17.8

36.8

9.3

7,927

8,589

24.1

35.0

21.2

Second thinning Conventional Operation Final felling System Total without R*

Second thinning Mechanized Operation Final felling System Total without R* Total with R* *Regeneration expenses

The proportions of profitable sub-compartments in Nasushiobara City and the Kanuma area, without considering regeneration costs, were 27% and 33%, respectively, with the conventional operation system and 56% and 64%, respectively, with the mechanized operation system. Similarly to the second thinning and final felling operations, there were more profitable sub-compartments with the mechanized operation system than with the conventional operation system. However, the profitable sub-compartments in Nasushiobara City and the Kanuma area in consideration of the regeneration costs decreased to 11% and 23%, respectively, with the conventional operation system and 18% and 24%, respectively, with the mechanized operation system. This shows the current state of affairs of Japanese forestry, in which many forest owners are unwilling to conduct regeneration operations after final felling operations.

4.2 Landing sales Upon the first thinning operation, the proportions of profitable sub-compartments were 36% and 14% in Nasushiobara City and the Kanuma area, respectively (Table 5). There were more profitable sub-compartments than with normal extraction (Table 4). However, upon the second thinning operation of the mechanized operation system, the numbers were similar to those for normal extraction in Nasushiobara City and lower

132

in the Kanuma area because strip roads for the conventional operation system were assumed to be constructed upon the first thinning operation and those for the mechanized operation system were assumed to be constructed upon the second thinning operation. On the other hand, profitable sub-compartments also increased significantly relative to those with normal extraction upon the second thinning operation of the conventional operation system. Upon the final felling operation, almost all sub-compartments were profitable with the conventional operation system and all sub-compartments were profitable with the mechanized operation system. Relative to normal extraction, profitable sub-compartments in Nasushiobara City and the Kanuma area, without considering regeneration costs, increased significantly to 100% and 97%, respectively, with the conventional operation system and 94% and 96%, respectively, with the mechanized operation system. However, unlike normal extraction, the proportions with the mechanized operation system were smaller than those with the conventional operation system owing to strip-road construction upon the second thinning operation of the mechanized operation system. Similarly to normal extraction, profitable subcompartments in Nasushiobara City and the Kanuma area in consideration of the regeneration costs decreased to 31% and 49%, respectively, with the convenCroat. j. for. eng. 35(2014)2

Estimating Annual Available Amounts of Forest Biomass Resources with Total Revenues ... (125â&#x20AC;&#x201C;138)

K. Aruga et al.

Table 5 Profitable sub-compartments with landing sales Nasushiobara City No. of Area, subha compartments First thinning

Rate, % No.

Area

Average No. of Area, slope subha angle, compartments Â°

Rate, % No.

Area

Average slope angle, Â°

2,095

1,328

36.4

36.1

5.9

4,419

1,894

13.5

7.7

9.9

4,760

3,430

82.7

93.2

9.6

32,532

24,430

99.0

99.7

21.6

5,645

3,652

98.1

99.3

9.6

32,713

24,475

99.6

99.9

21.4

5,755

3,679

100.0

9.7

31,941

24,281

97.2

99.1

21.1

Total with R*

1,779

1,855

30.9

50.4

9.5

16,211

17,664

49.3

72.1

21.9

First thinning

2,095

1,328

36.4

36.1

5.9

4,419

1,894

13.5

7.7

9.9

Second thinning

2,335

2,322

40.6

63.1

7.6

11,009

12,416

33.5

50.7

17.7

Final felling

5,756

3,679

100.0

9.7

32,851

24,509

100.0

21.5

Total without R*

5,423

3,570

94.2

97.0

8.9

31,680

24,191

96.4

98.7

21.0

Total with R*

2,648

2,484

46.0

67.5

8.9

21,029

20,328

64.0

82.9

20.9

Second thinning Conventional Operation Final felling System Total without R*

Mechanized Operation System

Kanuma area

*Regeneration expenses

tional operation system and 46% and 64%, respectively, with the mechanized operation system, although the proportions were higher than those with normal extraction. Therefore, landing sales were effective as an economical timber and forest biomass harvesting system at these research sites.

4.3 No biomass extraction The first thinning operations were assumed to be pre-commercial and all costs were covered by subsidies (Table 6). Strip roads were assumed to be constructed upon the second thinning operation, unlike normal extraction and landing sales in which strip roads were assumed to be constructed upon the first thinning operation. For the conventional operation system, normal biomass extraction expenses were higher than strip-road establishment expenses. Therefore, there were more profitable sub-compartments without biomass extraction than with normal extraction upon the second thinning operation of the conventional operation system. On the other hand, there were fewer profitable sub-compartments without biomass extraction than with normal extraction upon the second thinning operation of the mechanized operation system, owing to the high costs of strip-road establishment for the latter case. In both operation systems and areas, the profitable sub-compartments without biomass extraction were fewer than those Croat. j. for. eng. 35(2014)2

with landing sales upon the second thinning operation. Upon the final felling operation, the number of profitable sub-compartments without biomass extraction was higher than that with normal extraction and equal to or lower than that with landing sales. The number of profitable sub-compartments without considering regeneration costs was higher than that with normal extraction and lower than that with landing sales. Profitable sub-compartments decreased when regeneration costs were considered. However, similar to the case when regeneration costs are not considered, the number was higher than that with normal extraction and lower than that with landing sales. Therefore, forest biomass harvesting contributed to economic balances under sustainable forest management with a certain biomass harvesting system (e.g., landing sales) at these research sites.

4.4 Most economical timber and forest biomass harvesting system during a 60-year rotation The most economical timber and forest biomass harvesting system for each sub-compartment could be classified as being one of the seven types of systems listed in Table 7. The most common type for the final felling operation was the mechanized operation system with landing sales. For the second thinning operation, the most common type was the conventional or mechanized operation system with landing sales.

133

K. Aruga et al.

Estimating Annual Available Amounts of Forest Biomass Resources with Total Revenues ... (125–138)

Table 6 Profitable sub-compartments with no biomass extraction Nasushiobara City

Kanuma area

Rate, % Area, ha

No.

Area

5,756

3,679

100.0

9.7

Conven- Second thinning tional Final felling Operation System Total without R* Total with R*

3,302

3,001

57.4

81.6

4,827

3,447

83.9

3,786

3,186

65.8

1,492

1,411

First thinning Mecha- Second thinning nized Final felling Operation System Total without R* Total with R*

5,756 1,844

First thinning

Rate, %

Average No. of slope subangle, compartments °

No. of subcompartments

Average slope angle, °

Area, ha

No.

Area

32,851

24,509

100.0

21.5

9.8

21,495

21,259

65.4

86.7

21.5

93.7

9.4

32,713

24,475

99.6

99.9

21.4

86.6

9.9

29,941

23,781

91.1

97.0

21.0

25.9

38.4

9.3

13,421

14,010

40.9

57.2

21.7

3,679

100.0

9.7

32,851

24,509

100.0

21.5

1,983

32.0

53.9

7.2

8,195

9,546

24.9

38.9

17.0

5,756

3,679

100.0

9.7

32,851

24,509

100.0

21.5

4,915

3,448

85.4

93.7

9.2

31,303

24,107

95.3

98.4

21.0

2,213

2,201

38.4

59.8

8.7

19,230

19,557

58.5

79.8

21.4

*Regeneration expenses

Table 7 Most economical timber and forest biomass harvesting system for each sub-compartment Nasushiobara City First thinning

Second thinning

Final felling

Kanuma area Area, ha

Average area, ha

Average slope angle, °

12,841

12,691

0.99

25.2

25.0

334

1,177

3.52

25.6

3.7

2.43

12.8

1.30

6.2

1,044

869

0.83

11.8

No. of subcompartments

Area, ha

Average area, ha

Average No. of subslope angle, compartments °

734

713

0.97

18.0

336

3.46

2.57

153

199

974

395

0.41

13.6

9,082

4,980

0.55

24.8

3,639

1,424

0.39

6.5

9,209

3,032

0.33

13.8

132

544

4.12

17.5

312

1,691

5.42

24.2

5,756

3,680

0.64

9.7

32,851

24,509

0.75

21.5

Total

C – conventional operation system; M – mechanized operation system 1 – normal extraction; 2 – landing sale; 3 – no biomass extraction

In contrast, the most common type for the first thinning operation was the conventional operation system with normal extraction or landing sales. The most economical timber and forest biomass harvesting system for a sub-compartment was influenced by the area and average slope angle of the subcompartment (Table 7). In terms of area, b, c, and g were classified as having large areas; a and d were classified as having medium areas; and e and f were classified as having small areas. The biomass harvesting

134

systems of b, c, and g were classified as normal extraction because the areas were large, the extracted volumes were relatively large, and indirect costs were subsequently reduced despite the first thinning operation. The biomass harvesting systems of a, d, e, and f on smaller areas were classified as landing sales or pre-commercial thinning upon the first thinning operation. In terms of average slope angles, a, b, e, and g were classified as being on steep slopes while c, d, and f were Croat. j. for. eng. 35(2014)2

Estimating Annual Available Amounts of Forest Biomass Resources with Total Revenues ... (125–138)

K. Aruga et al.

Table 8 Profitable sub-compartments with the most economical timber and forest harvesting system for each sub-compartment Nasushiobara City

First

Second

Final

thinning

felling

Profitable

Kanuma area

Rate, %

No. of subArea, No. of subcompartments ha compartments 538

588

73.3

Profitable

Rate, %

Area

No. of subcompartments

Area, ha

No. of subcompartments

Area

82.5

9,674

11,616

75.3

91.5

311

92.8

92.6

330

1,148

98.8

97.6

100.0

153

199

100.0

1,044

869

100.0

704

270

72.3

68.4

7,603

4,397

83.7

88.3

1,175

715

32.3

50.2

3,882

1,647

42.2

54.3

Total

127

520

96.2

95.6

310

1,676

99.4

99.1

2,814

2,672

48.9

72.6

22,872

21,423

69.6

87.4

C – conventional operation system; M – mechanized operation system 1 – normal extraction; 2 – landing sale; 3 – no biomass extraction

classified as being on gentle slopes. The timber harvesting systems of a, b, e, and g on steep slopes were categorized under only the conventional or mechanized operation system. Although the road density was reduced on steep terrain, unit roading costs increased, and hence, subsequent roading costs increased. Therefore, the timber harvesting systems on steep slopes were categorized exclusively under either the conventional or mechanized operation system in order to reduce roading costs for a 60-year rotation. In contrast, the timber harvesting systems of c, d, and f on gentle slopes were categorized under mixed timber operation systems consisting of both the conventional and mechanized operation systems. Therefore, roading costs of both the conventional and mechanized operation systems were included, although they were low because of the gentle slopes. All sub-compartments of c and d were profitable because the areas were relatively large and the slopes were gentle (Table 8). However, only 36% and 45% of the sub-compartments of f on gentle slopes were profitable in Nasushiobara City and the Kanuma area, respectively, because their areas were small. The proportions of profitable sub-compartments of g and b on large areas and steep slopes were about 90%, and those of a and e on smaller areas and steeper slopes were about 80%. The proportion of profitable subcompartments on large areas was thus higher. The numbers of profitable and deficit sub-compartments were 2,814 and 2,942 among a total of 5,756 subcompartments in Nasushiobara City. In contrast, the corresponding numbers were 22,872 and 9,979 among Croat. j. for. eng. 35(2014)2

a total of 32,851 sub-compartments in the Kanuma area. The numbers of profitable sub-compartments, when considering regeneration costs, were higher than those before applying the most economical timber and forest biomass harvesting system to each subcompartment.

4.5 Annual available amounts of forest biomass resources The supply potentials of timber and forest biomass resources were 1,572,395 m3 and 972,672 m3, respectively, in Nasushiobara City (Table 9) and 12,587,431 m3 and 7,868,852 m3, respectively, in the Kanuma area for a 60-year period (Table 10). Therefore, the annual supply potentials of timber and forest biomass resources were 26,207 m3 y–1 and 16,211 m3 y–1, respectively, in Nasushiobara City and 209,791 m3 y–1 and 131,148 m3 y–1, respectively, in the Kanuma area. The annual available amounts of timber and forest biomass resources were estimated on the basis of the annual supply potentials from profitable sub-compartments with the most economical timber and forest biomass harvesting system. As a result, the annual available amounts of timber and forest biomass resources were 19,084 m3 y–1 and 11,849 m3 y–1, respectively, in Nasushiobara City (Table 9) and 184,019 m3 y–1 and 115,213 m3 y–1, respectively, in the Kanuma area (Table 10). These amounts largely exceed the annual demands of a 500 kW woody biomass power generation plant scheduled to be built in Nasushiobara City (6,000 m3 y–1) and a chip production factory in the Kanuma area (12,000 m3 y–1), respectively.

135

K. Aruga et al.

Estimating Annual Available Amounts of Forest Biomass Resources with Total Revenues ... (125â&#x20AC;&#x201C;138)

Table 9 Supply potentials and available amounts of timber and forest biomass resources in Nasushiobara City, m3 Supply Potential

First Thinning Second Thinning

Available Amounts

Timber

Biomass

Total

Timber

Biomass

Total

182,561

134,549

211,657

281,187

492,844

155,163

206,134

361,297

Final Felling

1,360,737

508,925

1,869,662

989,899

370,229

1,360,128

Total

1,572,395

972,672

2,545,067

1,145,062

710,912

1,855,975

26,207

16,211

42,418

19,084

11,849

30,933

Annual

Table 10 Supply potentials and available amounts of timber and forest biomass resources in the Kanuma area, m3 Supply Potential

Available Amounts

Timber

Biomass

Total

Timber

Biomass

Total

1,526,346

1,345,625

Second Thinning

1,712,658

2,275,270

3,987,928

1,506,232

2,001,033

3,507,264

Final Felling

10,874,773

4,067,236

14,942,010

9,534,935

3,566,128

13,101,063

Total

12,587,431

7,868,852

20,456,283

11,041,167

6,912,785

17,953,952

209,791

131,148

340,938

184,019

115,213

299,233

First Thinning

Annual

5. Conclusions This study extracted production forests and estimated the annual available amount of forest biomass resources under profitable forest management. Production forests were extracted as sub-compartments where expected revenues surpassed all costs, from planting to final harvesting, for a 60-year rotation. Iuchi (2004), Yoshioka and Sakai (2005), Yagi and Nakata (2007), Kamimura et al. (2009), Yamamoto et al. (2010), Yamaguchi et al. (2010), Kamimura et al. (2012), Aruga et al. (2013) and Nakahata et al. (2014) estimated only the supply potentials and available amounts of timber and forest biomass resources based on the current situation. This study could project future supply potentials and available amounts of forest biomass resources in order to plan power generation plants considering available amounts of forest biomass resources. Furthermore, Aruga et al. (2006ab), Kinoshita et al. (2009), Kinoshita et al. (2010), and Aruga et al. (2011) projected future supply potentials and available amounts of forest biomass resources, but did not consider regeneration expenses for the sustainability of forest management which this study included. Moreover, this study examined the most economical timber and forest biomass harvesting system temporally and spatially. The model was then applied to Nasushiobara City and the Kanuma area in Tochigi Prefecture, Japan. As

136

a result, the proportions of profitable sub-compartments without considering regeneration costs were higher than those considering regeneration costs. This is indicative of the current state of affairs of Japanese forestry, where many forest owners are unwilling to perform regeneration operations after final felling operations. Therefore, it is important to develop low-cost regeneration operations or to extend rotation ages by reducing the number of regeneration operations for sustainable forest management and hence sustainable use of forest biomass resources amid the current forestry situation in Japan. According to forest biomass harvesting, the number of profitable sub-compartments without biomass extraction was higher than that with normal extraction and lower than that with landing sales. Therefore, forest biomass harvesting contributed to economic balances under sustainable forest management with a certain system of biomass harvesting. The most economical timber and forest biomass harvesting system for each sub-compartment was also determined in this study. Profitable sub-compartments increased by applying the most economical timber and forest biomass harvesting system to each sub-compartment. The annual amounts of available forest biomass resources largely exceed the annual demands of a 500 kW woody biomass power generation plant scheduled to be built in Nasushiobara City and a chip production factory in the Kanuma area, respectively. Croat. j. for. eng. 35(2014)2

Estimating Annual Available Amounts of Forest Biomass Resources with Total Revenues ... (125–138)

This model could help forest planners consider biomass harvesting when establishing forest plans. The proportion of profitable sub-compartments on large areas was higher than on small areas. Therefore, costs are expected to decrease by extending forestryoperation sites while merging sub-compartments. The Japan Forest Agency has implemented measures on »coordination and consolidation of forestry practices«. Such measures will ensure coordination among a number of small forest owners to conduct forestry practices on a large scale. In future studies, we intend to expand forestry-operation sites while merging subcompartments in order to reduce the costs. The Japan Forest Agency has also implemented measures for long-term rotation management because revenues from final felling operations cannot cover regeneration costs under the current conditions. Extending the cutting ages is expected to increase revenues owing to an improvement in log prices. However, in the present study, the effects of changing log prices were not considered. Future studies will also address log prices along with log quality.

Acknowledgement We are grateful to the Tochigi Prefectural Government for providing the required data. A part of this study was supported by Nissei Zaidan.

6. References Agency for Natural Resources and Energy, 2012: Settlement of the details of the Feed-in Tariff scheme for renewable energy, including purchase price and surcharge rates. http:// www.meti.go.jp/english/press/2012/0618_01.html (Accessed on June 18, 2012). Aruga, K., Yoshioka, T., Sakurai, R., 2006a: Long-term feasibility of timber and forest biomass resources at an intermediate and mountainous area: balance of harvesting volumes using random search. Journal of Japan Forest Engineering Society 21: 49–59. Aruga, K., Tasaka, T., Yoshioka, T., Sakurai, R., Kobayashi, H., 2006b: Long-term feasibility of timber and forest biomass resources at an intermediate and mountainous area: (2). Examining the optimum scale of an energy plant. Journal of Japan Forest Engineering Society 21: 185–192. Aruga, K., Murakami, A., Nakahata, C., Yamaguchi, R., Yoshioka, T., 2011: Discussion on economic and energy balances of forest biomass utilization for small-scale power generation in Kanuma, Tochigi prefecture, Japan. Croatian Journal of Forest Engineering 32(2): 571–586. Aruga, K., Murakami, A., Nakahata, C., Yamaguchi, R., Saito, M., Kanetsuki, K., 2013: A model to estimate available timber Croat. j. for. eng. 35(2014)2

K. Aruga et al.

and forest biomass and reforestation expenses in a mountainous region in Japan. Journal of Forestry Research 24: 345–356. Dijkstra, E.W., 1959: A note on two problems in connexion with graphs. Numerische Mathematik 1: 269–271. Forest Experiment Station, Forest Bureau, Ministry of Agriculture, 1955: Japanese Cedar Yield Table in Abukuma Area of the North Kanto Region. Tokyo. Forest Experiment Station, Forest Bureau, Ministry of Agriculture, 1961: Japanese Cypress Yield Table in the Kanto Region. Tokyo. Forestry Agency, The ministry of Agriculture, Forestry and Fisheries of Japan, 2013: Annual report on forest and forestry in Japan. Fiscal Year 2012 (summary), Tokyo. Iuchi, M., 2004: Development of the support system for biomass energy business plans: the data base and evaluation models to simulate the collection cost. Socio-economic Research Center, Y03023, 1–26. Kamimura, K., Kuboyama, H., Yamamoto, K., 2009: Estimation of spatial distribution on wood biomass supply potential for three prefectures in the northern Tohoku Region. Journal of the Japan Institute of Energy 88: 877–883. Kamimura, K., Kuboyama, H., Yamamoto, K., 2012: Wood biomass supply costs and potential for biomass energy plants in Japan. Biomass and Bioenergy 36: 107-115. Kinoshita, T., Inoue, K., Iwao, K., Kagemoto, H., Yamagata, Y., 2009: A spatial evaluation of forest biomass usage using GIS. Applied Energy 86: 1–8. Kinoshita, T., Ohki, T., Yamagata, Y., 2010: Woody biomass supply potential for thermal power plants in Japan. Applied Energy 87: 2923–2927. Möller, B., Nielsen, P.S., 2007: Analyzing transport costs of Danish forest wood chip resources by means of continuous cost surfaces. Biomass and Bioenergy 31: 291–298. Nakahata, C., Aruga, K., Takei, Y., Yamaguchi, R., Ito, K., Murakami, A., Saito, M., Tasaka, T., Kanetsuki, K., 2011: Improvement on operational efficiencies and costs of extracting thinned woods using a processor and a forwarder in Nasunogahara area: (II) based on comparative analyses of current operations and mechanized operations. Bulletin of Utsunomiya University Forest 47: 27–34. Nakahata, C., Uemura, R., Saito, M., Kanetsuki, K., Aruga, K., 2014: Estimating harvesting costs and projecting available amounts of logging residues with small-scale forestry in Nasushiobara, Tochigi Prefecture, Japan. Journal of Forestry Research: DOI 10.1007/s11676-014-0482-x. Nord-Larsen, T., Talbot, B., 2004: Assessment of forest-fuel resources in Denmark: Technical and economic availability. Biomass and Bioenergy 27: 97–109. Okawabata, O., 2003: An estimation of necessary manpower and cost of reforestation on cedar and cypress. Journal of Japan Forest Engineering Society 18: 195–200. Panichelli, L., Gnansounou, E., 2008: GIS-based approach for defining bioenergy facilities location: A case study in Northern Spain based on marginal delivery costs and resources competition between facilities. Biomass and Bioenergy 32: 289–300.

137

K. Aruga et al.

Estimating Annual Available Amounts of Forest Biomass Resources with Total Revenues ... (125–138)

Ranta, T., 2005: Logging residues from regeneration fellings for biofuel production-a GIS-based availability analysis in Finland. Biomass and Bioenergy 28: 171–182. Rørstad, P., Trømborg, E., Bergseng, E., Solberg, B., 2010: Combining GIS and forest modelling in estimating regional supply of harvest residues in Norway. Silva Fennica 44: 435–451. Sawaguchi, I., 1996: Studies on forest-road evaluation and forest-road standards in mountain forests: (I) characteristics of parameters for forest-road evaluation. Bulletin of the Forestry and Forest Products Research Institute, 372: 1–110. Tochigi Prefectural Government, 2010: Forestation program standard unit cost table of Fiscal year 2010. Tochigi. Viana, H., Cohen, W.B., Lopes, D., Aranha, J., 2010: Assessment of forest biomass for use as energy. GIS-based analysis of geographical availability and locations of wood-fired power plants in Portugal. Applied Energy 87: 2551–2560. Yagi, K., Nakata, T., 2007: Economic analysis on small-scale forest biomass gasification considering regional resource dis-

tribution and technical characteristics. Journal of the Japan Institute of Energy 86: 109–118. Yamamoto, H., Nakata, T., Yabe, K., 2010: Design of biomass co-firing system considering resource distribution and transportation optimization. Journal of the Japan Institute of Energy 89: 42–52. Yamaguchi, R., Aruga, K., Murakami, A., Saito, M., Ito, K., 2010: Development of the model to estimate the harvesting volumes and costs of logging residues considering economic balances of timber and logging residue harvesting in Sano city, Tochigi Prefecture. Journal of the Japan Institute of Energy 89: 982–995. Yoshioka, T., Sakai, H., 2005: Amount and availability of forest biomass as an energy resource in a mountain region in Japan: a GIS-based analysis. Croatian Journal of Forest Engineering 26(2): 59–70. Zenkoku Ringyo Kairyo Fukyu Kyokai, 2001: Management of Forestry Mechanization (in Japanese). Tokyo. 1–239.

Authors’ address: Assoc. Prof. Kazuhiro Aruga, PhD.* e-mail: aruga@cc.utsunomiya-u.ac.jp Ayami Murakami, MSc. e-mail: ayamy_mumu@yahoo.co.jp Chikara Nakahata, MSc. e-mail: c.nakahata0927@gmail.com Reiko Yamaguchi, MSc. e-mail: r.yamaguchi0210@gmail.com Utsunomiya University Faculty of Agriculture Department of Forest Science 350 Mine Utsunomiya 321-8505 JAPAN Assist. Prof. Masashi Saito, PhD. e-mail: m_saito@shinshu-u.ac.jp Shinshu University Faculty of Agriculture Department of Forest Science 8304 Minamiminowa Kamiina 399-4598 JAPAN

Received: March 13, 2014 Accepted: May 29, 2014

138

Assoc. Prof. Takuyuki Yoshioka, PhD. e-mail: yoshioka.takuyuki@nihon-u.ac.jp Nihon University College of Bioresource Sciences Department of Forest Science and Resources 1866 Kameino Fujisawa 252-0880 JAPAN * Corresponding author Croat. j. for. eng. 35(2014)2

Original scientific paper

Productivity Study of WoodPac Bundling of Logging Residues and Small Stems Iwan Wästerlund, Anders Öhlund Abstract A new approach for procuring logging residues has been introduced, in which the residues are compacted into cylindrical bales known as composite residue logs (CRLs). Some large-scale productivity studies have been undertaken on different bundling machines, and preliminary calculations on the possible benefits associated with bundling have been based on limited material or on prototypes. The aims of the study presented here were to measure the effects of concentrating forest fuel on bundling productivity with a WoodPac machine and to test if there was a difference between bundling logging residues and small stems. The results show that the WoodPac machine produced 19.3 bundles per effective hour (E0), equivalent to 28.5 MWh, in a clear-cut spruce stand. Productivity was not influenced by the amount of green mass as long as there was more than 5 Mg per 100 m driving distance and the material was collected in heaps. About 20% of the handled material was shaved off, mainly as fine material, so 24% more was handled than appeared in the bundles. Productivity could be raised to 24 bundles per E0 with logging residues, but with young stems the productivity may be 50% lower. Forwarding bundles was at least 2.5 times more productive than forwarding loose logging residues. In conclusion, bundling could be of interest when economical and environmental aspects of the whole chain are considered. Keywords: wood energy, shave-off, forwarding

1. Introduction A formerly widely used system for procuring fuel wood in Sweden and Finland incorporates fuel adapted harvesting at clear cuts, forwarder extraction of logging residues and their piling at landings as storage areas (Andersson et al. 2002). This system has several drawbacks. Three basic chains for bio fuel handling can be used; (i) Maintain pieces whole and chip them just before use, or (ii) compact the material as bales or bundles, or (iii) chip as soon as possible. All chains have some drawbacks. The second chain needs extra machinery from beginning to collect plus compact the pieces to better handling units, and the third chain needs a chipper at least from the landing, so storing losses may occur. In the second case long distance transport costs may be reduced compared to not treatCroat. j. for. eng. 35(2014)2

ed slash (Johansson et al. 2006, Engblom 2007, Spinelli et al. 2012), but the extra cost for specialised machines is not favoured in the practise although transport distances of more than 60 km may need some compaction of the logging residues to be of economic interest (Johansson et al. 2006, Engblom 2007). Both time and financial costs are high when collecting logging residues after clear-felling, or small stems following a first commercial thinning (Brunberg 1991, Hakkila and Nurmi 1997, Eliasson 1998, Cuchet et al. 2004, Kärhä and Vartiamäki 2006, Jylhä and Laitila 2007). A possible way to reduce the cost of harvesting of small stems is to utilise whole trees (Hakkila 1989), especially in combination with multi-tree handling machines (Brunberg et al. 1990). At extraction of slash, the maximum load on forwarders and lorries is only 35–50% (7–8 tonnes on a 16 tonne forwarder or

139

I. Wästerlund and A. Öhlund

Productivity Study of WoodPac Bundling of Logging Residues and Small Stems (139–151)

14 tonnes on a 40 tonne truck) of the vehicles bearing capacities and the transportation costs are high if logging residues or small trees are not compressed (Carlsson et al. 1983, Brunberg et al. 1998, Johansson et al. 2006). A possible solution is to compress the material into bundles at an early stage in the logistic chain. The advantage of this is that a small bunch of trees or logging residues can be kept together from the forest to the factory where the bunch can be processed (Schiess and Yonaka 1982, Arola et al. 1985, Flinkman and Thörnqvist 1986, Johansson et al. 2006, Spinelli et al. 2012). These bundles may also be used as storing units (Pettersson and Nordfjell 2007). Small bunches can be handled during transportation with the same equipment as industrial cut-to-length round wood (Schiess and Yonaka 1982, Liss 1995, Johansson et al. 2006). The bundles can be compressed using a number of methods, and the forces needed to compress the residues and small trees into fairly dense bundles are actually quite small (cf. Nordfjell and Liss 2000). Besides logging residues, forest fuels from powerline corridors comprise a potential resource that has been largely neglected to date. Lanes wider than 5 m cover approx. 140,000 hectares of productive forestland in Sweden (Anon.1989), and every year approx. 13,000 hectares of lanes are cleared (Jonsson et al. 1992). Silviculture in these lanes is aimed at maintaining the power lines reliability and to facilitate their maintenance. For example, the power company Skellefteå Kraft in Sweden has 12,000 km of lanes, including 1,000 km of 40–metre broad 130 kV power lines. These lanes are coppiced every 12–20 years when the trees are 3–5 m tall, and their diameter at stump height is 4–8 cm. Coppicing can be carried out motor-manually by brush-saw during the snow-free season at a cost of 160–300 €/ha (Larsson 1998). Furthermore, a 3-metre broad inspection track and 2-m radius circles around each pole are kept completely free from brushwood. The cost of all these activities may account for more than half of the total maintenance costs for the power lines. A rather new procurement approach that has been introduced, and may still be under development, involves compacting logging residues into cylindrical bales known as composite residue logs (CRL, cf. Andersson et al. 2002). There are currently six major types of equipment on the market for compressing logging residues. Three of them (Fiberpac, Fixteri and Pika RS 2000) compress the material through a funnel, tie strings around the bundle and finally cut the bundle to the desired length, e.g. 3–4 m (Brunberg et al. 1998, Kärhä and Vartiamäki 2006). The fourth type (WoodPac) tumbles the material within a compartment with

140

spiked rollers, which shave off some of the fine material (mainly needles and twigs) and produces 3.4 m long bundles. Finally, strings are tied around the bundle, which is then released from the compartment and falls to the ground. The fifth, older type is baling (Bala Press AB), which heavily compresses logging residues into 1.2×1.2 m cylindrical bales (Andersson and Hudson 1997), is hard to chip due to the size. The baling and bundling is done at the harvesting site to facilitate handling for transporting the forest fuel. The sixth type is a lorry mounted compression unit, which compresses with high forces up to 4.8 m long bundles (Lindroos et al. 2010). Productivity may be quite high but the applied forces are far above those needed for the material and need terrain transport of loose material. Preliminary calculations indicate that the transport benefits could well justify the costs involved in bundling the logging residues (cf. Johansson et al. 2006, Engblom 2007, Spinelli et al. 2012). The WoodPac bundling system seems to have been studied least intensively of these six approaches. It is also of interest since it releases some of the fine material back to the forest. This shaving off of needles and twigs is likely to have both advantages and disadvantages. Advantages include the facts that disproportionately high amounts of nutrients will be left in the forest, since the fine material contains relatively high levels of nutrients, and the heating plant will receive less troublesome minerals in the ash (cf. Orjala et al. 2000, Aho and Silvennionen 2004). The disadvantage is that the shaving will most likely reduce the productivity of the machinery. WoodPac is a hydraulically driven machine that is placed on and powered by vehicles, such as a medium-sized forwarder. The mass of the WoodPac machine is about 7.6 Mg and the hydraulic power unit should be able to pump 2 – 2.3 l s–1 generating 28–32 MPa pressure. The compression space contains eight cylinders, two of which are equipped with spikes. The bundles produced are 3.4 m long and have a radius of about 0.7 m. Some few large-scale productivity studies have been undertaken on these bundling machines (Kärhä and Vartiamäki 2006) and preliminary calculations on the benefits associated with bundling have been based on limited material. One issue to address is how concentrating the fuel material in heaps on the clear-cut area affects the bundling. Further parameters that need to be investigated are the amounts of fine material that are released by the WoodPac machine, and the benefits (if any) of this material for the forest soil. An additional question to consider is whether the bundling equipment can bunch small trees that are longer Croat. j. for. eng. 35(2014)2

Productivity Study of WoodPac Bundling of Logging Residues and Small Stems (139–151)

than the compartment. Finally, studies to date seem to have concentrated on the production of raw material, and little information has been gathered on the amount of energy contained in the bundles produced.

2. Aims The aims of the study were to measure the productivity of bundling with a WoodPac machine, to assess the effects (if any) of the concentration of forest fuel on productivity and to evaluate possible differences in productivity between bundling logging residues and small stems. The amount of material shaved off and the energy content of the bundles were also considered in the productivity analyses.

3. Material and Methods 3.1 Description and preparation of experimental sites Material from three locations was used: two clearfelled area and a power line corridor. Logging residues were collected at a final felling stand 15 km west of Umeå (lat. 65.51 N, long. 20.17 E) and a final felling mid-south of Sweden (Sävsjö, Småland, 57° 25´N; 14° 40´O). The first stand (ca. five ha in area) was harvested at the end of May, and 221 m³ solid volume under bark/ha (sub ha–1), comprising 66% Norway spruce (Picea abies (L.) Karst.), 32% Scots pine (Pinus silvestris L.), and 2% deciduous trees, was extracted according to the log list from the harvester. The minimum top diameter for pulpwood was 5 cm. The stand was 130 years old and naturally regenerated. The sec-

I. Wästerlund and A. Öhlund

ond stand was a 0.58 ha 120 yr-old spruce stand, cut in autumn. The tree harvester operators were instructed to perform fuel-adapted harvesting, that is trees were to be delimbed at the side of the machine and the residues were not to be driven on. The two forwarder operators were also instructed not to drive on the piles of residues when extracting timber. After harvesting, the piles were left to dry for a month. The average basal area of the 552 piles produced was 4.06 m² (s. dev. 1.4), assuming that the basal area formed a circle, and the average height was 0.76 m (s. dev. 0.18), so the average volume was about 1.5 m3. The third location was a power line corridor, 100 km N Umeå (lat 64.28 N, long 21.17 E), which was examined during a cleaning operation. About 500 m of the 40 m wide corridor was used for the study. All trees (0.5 to 6 m tall) were felled motor-manually on the 22nd and 23rd of May 2002, and the coppicing crew were instructed to fell trees perpendicular to the power line to allow them to be easily picked up by the bundling machine. The regeneration was a mixed birch (60%) and pine (30%) stand with 8,000 stems ha–1 and, according to calliper measurements, an average diameter at breast height (Dbh) of 1.7 cm (four circular plots). Some larger pines with Dbh 6.5 cm were also present. Bundling was done along a route in the area selected to reflect the different densities of young trees.

3.2 Bundling Bundling was done with a 2001 version WoodPac unit mounted on a Rottne SMV Rapid 13.5 Mg forwarder with a Rottne RK 90 crane equipped with a

Fig. 1 Woodpac during release of the compression chamber. (A) Fabricated CRL can be seen in the foreground. From the power line; (B) Front frame of WoodPac. (Photo: I. Wästerlund) Croat. j. for. eng. 35(2014)2

141

I. Wästerlund and A. Öhlund

Productivity Study of WoodPac Bundling of Logging Residues and Small Stems (139–151)

Hultdin 28 slash grapple. The operator had had several years of experience with this machinery. The eight WoodPac rollers are driven and connected with a transmission chain to ensure they all rotate at the same speed (Fig. 1A and 1B). Material is fed in by crane into a slot between the two top rollers. The width of the feeding slot restricts the diameter of the material to less than 30 cm. The top rollers are spiked and rotate in opposite directions to allow the material to be fed into the compression chamber. As material is fed into chamber, the rolling binds the material together, forming a CRL. Once the operator judges the chamber to be full, a twine dispenser feeds polypropylene agricultural twine between the rollers on three different places on the bundle. When the tying process is complete, the twine is automatically cut. The process is then manually stopped to allow the chamber to open and the CRL to be ejected. The mass of each CRL was determined when loading the CRLs onto the forwarder with a TB 3000 cranemounted scale, calibrated each day with a concrete weight (mass 588 kg). CRLs were carefully gripped in the middle and the mass was read when the CRL was in balance and the display was steady. Following this procedure, the accuracy should be within ±5 kg, according to the specifications. A 5% sample of the 552 piles of logging residues was randomly selected. From each of the 28 piles, an 8–10 litre sample of loose logging residues was collected in the middle of June. The piles were chipped into a container with an Erjo 765 chipper. Samples were transported the same day to an accredited laboratory to analyse their moisture content at 105° C 24 h, according to Swedish stan-

Table 1 Moisture content measured in 28 heaps of logging residues before bundling, five CRLs after production and five CRLs composed of young stems after production. S. dev. = standard deviation Log. residues before bundling

CRL log residues after bundling

CRL young stem, after bundling

Av. moisture cont, %

28.6

24.3

S. dev.

4.4

5.2

2.5

dard method SS 18 71 70 (Anon. 1997). To test if the moisture content varied after bundling, samples were also taken from five bundles and measurements were done after chipping the whole bundles (Table 1). Moisture contents are given on a green weight basis.

3.3 Time studies The operator of the bundling machine decided his own route through the clear-felled area, but was instructed to drive as normally as possible, since he had long experience in bundling. The time study was done as a correlation study with snap-back timing (Anon. 1978), using a Husky Hunter computer running SIWORK3 software (Rolew 1988). The observation unit was one bundle i.e. from the time the bundler chamber was completely closed until it was closed again. Work with the bundling machine was split into eight work elements (Table 2). If more than one work element was performed simultaneously, the time for the work element with the highest priority was recorded. All element times were measured as effective times (E0, Anon. 1978). Delay times were measured, but not included in the analysis.

Table 2 Description of time elements used to analyse bundling work Priority

Work element

Description

Crane cycle

From when crane starts to move from the machine to grip material until the grip drops material into the bundler (or stops moving for other work to be done on the machine)

Feeding

Time when the grip is used to press down material into the bundler or the operator disperses material for the bundler

Tying

Time from when twine dispenser starts until the bundling chamber opens to eject a CRL

Unloading

From the time the chamber opens until it is closed again

Driving

From the time the wheels start to move until they stop or a higher priority element starts

Miscellaneous

Productive time used for other essential tasks

Bundling

Time when logging residues are being compressed and no higher priority work is being done

Bundler delays

Non-productive time caused by problems with bundling machine, not included in the study

142

Croat. j. for. eng. 35(2014)2

Productivity Study of WoodPac Bundling of Logging Residues and Small Stems (139–151)

I. Wästerlund and A. Öhlund

Table 3 Work elements and priorities for the time study on forwarding bundles Priority

Work element

Description

Crane out

From the time the crane starts to move from rest point to grip material until the grip grasps the material (or the crane is stopped for driving)

Crane in

From when the grip has grasped the material, and lifts it, until it drops it into the machine (or the crane is stopped, before the machine is moved to collect more material)

Re-gripping

Time when the crane moves to grip again, or carry more material within a crane cycle

Driving to area

Time from when the wheels start to move at the landing until they stop for material to be collected, or a higher priority element starts

Driving within area

From when the wheels start to move until the wheels stop or a higher priority element starts. Only includes driving within harvesting area

Driving to landing

From when the wheels start to move after the last crane cycle on the bundling area until the wheels stop at the landing, or a higher priority element starts. Only includes driving after leaving loading area, when machine is full

Miscellaneous

Time required for other essential tasks

Delays

Non-productive time, not included in the study

In the time studies, 250 bundles of logging residues and 22 bundles of young stems were analysed. After fabrication, the diameter of each CRL was measured at three places, 0.2–0.3 m from both ends and in the middle. The mean rear, mid-point and front-end diameters (s. dev. within brackets) of CRLs made of logging residues were 0.73 m (0.020), 0.74 m (0.020) and 0.73 m (0.019), respectively. For CRLs made of young trees, the corresponding figures were 0.75 m (0.046), 0.76 m (0.026) and 0.77 m (0.051). Each CRL was tagged at both ends and was marked with GPS coordinates obtained using a Magellan 320 receiver.

3.4 Forwarding The terrain transport of the bundles was done with a 14 Mg Hemek Ciceron TD 81 forwarder with a Fiskars 71 crane within a few days after bundling. The grapple was a conventional Cranab 028. The forwarder loading area was 4.8 m2. The observation unit for forwarding was a full load of bundles, i.e. from the time the wheels started to move from the landing until a new load of bundles was unloaded. The type and number of bundles taken for each load, and their order of loading, were noted to provide records of the distances travelled and the mass of the loads. The time elements used in the analysis are shown in Table 3. Four loads were analysed in this way.

3.5 Complementary studies Småland To measure the amount of logging slash handled for each bundle, a separate study was done in Småland on the amounts of material that were shaved off. For Croat. j. for. eng. 35(2014)2

this, the machine was placed on a tarpaulin while it produced one bundle. All biomass was then cleaned from the machine, which was moved away from the tarpaulin and the content was weighed. The prepared bundle was also weighed, numbered and positioned as usual. In addition, the number of crane cycles required to produce the bundle was counted as well as the time taken for tying and unloading it. The last two tasks were done as controls, since they were partially masked during normal operation. Eight bundles were analysed in this way (Bohm Larsson 2004). To get a rough estimation of fuel consumption, the machines started with a full tank and the amounts required to refill them during or at the end of the study were recorded (Umeå study).

4. Results Production data are based on the production of 1 Mg (green weight) CRLs on the ground. Moisture contents at the time of bundling were interpolated between moisture contents in the logging residue heaps and newly produced bundles (Table 1). The bundler produced 19.3 bundles per effective hour (E0) from logging residues, with an average mass of 368 kg (Table 4). At a moisture content of 27%, the production was 5.2 Mg dry matter per hour (DM, dry mass per bundle, 270 kg). When bundling rather poorly dispersed young stems in the power line corridor, the productivity and production fell to only 10 bundles and 2.6 Mg dry matter per hour. Differences in productivity measured in terms of min per Mg and bun-

143

I. Wästerlund and A. Öhlund

Productivity Study of WoodPac Bundling of Logging Residues and Small Stems (139–151)

Table 4 Productivity when bundling logging residues and young stems with WoodPac (MC=moisture content, DM=dry matter) Green mass per bundle, kg

Min. per Mg, green weight

Bundles per E0 hour

Mg DM prod per E0 hour

Logging residues, MC=27%

Table 5 Average time consumption for each work element when bundling logging residues (n=250 bundles) with 27% moisture content, and young stems (n=22) with 26% moisture content. Standard deviations shown in brackets, and percentage of effective time in italic Cmin per 1 Mg green log. residues, S. dev. and %

Cmin per 1 Mg green young st., S. dev. and %

Crane cycle

4.49 (1.02) 50.8

11.23 (2.80) 63.6

Feeding

0.95 (0.52) 10.7

1.18 (0.39) 6.6

Work element

Average (n=250)

368

8.83

19.3

5.2

S. dev.

38.2

1.89

3.63

1.06

Young stems, MC=26% Average (n=22)

Driving

1.03 (0.85) 11.7

2.47 (1.53) 14.0

350

17.63

10.1

2.6

Compression

0.75 (0.37) 8.5

1.04 (0.40) 5.9

S. dev.

37.6

3.90

1.60

0.52

Tying

1.00 (0.57) 11.3

0.98 (0.31) 5.6

Unloading

0.44 (0.09) 5.0

0.53 (0.11) 3.0

Miscellaneous

0.18 (0.30) 2.0

0.21 (0.24) 1.2

Effective time

8.83 (1.89)

17.63 (3.90)

Delay bundler

0.4 (1.67)

9.23

17.63

dles per hour depend on the mass of the bundles, since values were based on bundles. The average effective time for slash bundling per Mg green material was 8.8 min (Table 5). For lifting the material in with the crane, it took about 50% of the time for bundling logging residues, and about 64% of the time for young stems. Compared to logging residues, the proportion of time consumed by driving during bundling was high for young stems, indicating that the concentration of the material could have great influence on this variable. Most other time elements seem to be quite consistent between the two types of

Total time

material used, indicating that they could be machinedependent. However, the substantial standard deviation obtained for tying time (which reflects delays that occurred in the bundler) indicates that the system could be improved.

Fig. 2 Effective time required to produce 1 Mg of bundles plotted against distance driven to collect the material for the bundles produced with logging residues

144

Croat. j. for. eng. 35(2014)2

Productivity Study of WoodPac Bundling of Logging Residues and Small Stems (139–151)

I. Wästerlund and A. Öhlund

Fig. 3 Effective time required to produce 1 Mg of bundles plotted against the distance driven to collect the material to produce bundles with young stems (r2=0.52)

Fig. 4 Time for the work elements loading and driving during loading/bundling required to produce 1 Mg of bundles plotted against distance driven to collect the material for logging residues (r2=0.025) Although the time study indicated that there might be a weak correlation between the time required to make 1 Mg of bundles and the distance driven to collect the material, there seemed to be no such effect when collecting logging residues (Fig. 2). For young stems, on the other hand, production declines when the concentration of biomass is low (Fig. 3). Croat. j. for. eng. 35(2014)2

A variable that could be related to bundling productivity was concentration of the material (as indicated by the amounts produced per 100 m driven). The effective times for the measured work elements for bundling logging residues were randomly or nonsignificantly correlated to this concentration measure (Fig. 4). However, for young stems, two elements –

145

I. Wästerlund and A. Öhlund

Productivity Study of WoodPac Bundling of Logging Residues and Small Stems (139–151)

Fig. 5 Time required for the work elements loading and driving during loading to produce 1 Mg of bundles compared with distance driven to collect the young stems (r2=0.71) crane work (loading) and driving during loading – were strongly related to production, indicating that productivity would be significantly decreased at concentrations less than about 5 Mg material produced per 100 m driven (cf. Figure 5). In the shaving study, the amount of fine material shaved off while bundling logging residues was 96 kg (s. dev. 19), and the green mass of the prepared bundles was 402 kg (s. dev. 25). Thus, the machine handled 1.24 times more material than the amount bundled, implying that to produce 370 kg bundles, as in this part of the investigation, 456 kg of material had to be handled on average (Fig. 6).

Forwarding the bundles was investigated only as a case study. The average load for the four trips examined was 12 bundles, and the forwarding distance was on average 280 m per trip. Since the bundles were quite dry, the average mass per load was only 5 Mg but the average volume was 16 m3 (full forwarder load = 16.3 m3). The average effective time for forwarding was 3.9 cmin per Mg green mass, equivalent to 15.6 Mg per E0 (Table 6). These productivity figures are not high compared to timber forwarding in terms of mass, but in volume terms they are very high (cf. Fig. 7). The amount of fuel consumed was 292 litres when bundling logging residues, giving an average con-

Fig. 6 Composition of material flow (kg DM) when producing one bundle including the peeled off material (Smland material)

146

Croat. j. for. eng. 35(2014)2

Productivity Study of WoodPac Bundling of Logging Residues and Small Stems (139–151)

Table 6 Average time consumption in cmin per Mg green mass and per unit volume for each work element when forwarding CRLs with 27% MC. Standard deviations shown in brackets Cmin per 1 Mg green mass

Cmin per m3

Crane out

0.60 (0.02)

0.18 (0.005)

Crane in

0.68 (0.01)

0.21 (0.003)

Driving during loading

0.95 (0.40)

0.29 (0.123)

Driving loaded

0.48 (0.16)

0.15 (0.048)

Gripping

0.10 (0.21)

0.03 (0.065)

Re-gripping

0.01 (0.01)

Driving empty

0.89 (0.26)

0.27 (0.079)

Miscellaneous

0.19 (0.09)

0.06 (0.029)

Effective time

3.89 (0.51)

1.19 (0.168)

Work element

Delays Total time

0 3.89

1.19

sumption of about 23 litres per hour. During forwarding, the fuel consumption was 13 litres. This means that the fuel consumption was about 23/19.3=1.2 litre/bundle during bundling residues and 13/(4*12)=13/48=0.3 l forwarding per bundle. This means that (1.21x9.7)=11.7 kWh diesel were consumed to produce a bundle with logging residues worth (19.7x0.27)/3.6=~1.5 MWh and 65 litres were consumed to make 22 bundles (65/22=2.95 l/bundle),

Fig. 7 Productivity when forwarding CRLs in terms of transported volume for a single driving distance Croat. j. for. eng. 35(2014)2

I. Wästerlund and A. Öhlund

with the young stems with an energy content of about 19.7x260x1/3600=1.42 MW, and (0.3x9.7)/1.42=2 kWh diesel consumed to forward a bundle and per produced MWh bundle. Thus from the energy point of view, bundling of young stems in unorganised collection cannot be motivated. The energy content in each bundle (19.7 MJ/kg DM) is based on calorific values from Pettersson and Nordfjell (2007). Energy content in diesel is set to ~9.7 kWh/litre for environment class 1 diesel, which is commonly used for forestry machines in Sweden.

5. Discussion One drawback of this study was that the material had unusually low moisture contents because the weather was warm and dry (so both the logging residues and young stems dried out rapidly) in the delay between harvesting and bundling. In colder, moist conditions, the green residues may have moisture content around 50% (cf. Hakkila 1989). In such cases the bundles could have a green mass of about 540 kg and a dry mass about the same as found here. The working practices in the study on young stems were not ideal, since the stems were scattered on the ground, and the productivity figures presumably reflect this. The figures obtained may indicate the productivity for motor-manually cleaned areas, but the young stems should preferably be bunched before bundling (cf. Johansson and Gullberg 2002), which would probably considerably improve the production. However, the experiment showed that it was possible to feed in stems up to about twice the length of the chamber. If the stems were just a little longer than this, there was no need for the operator to do anything more than place the root end to one side of the chamber. Some long trees (5–6 m) had to be broken up with the grapple after some time causing increased bundling time (cf. Table 4). Production data were based on times recorded per bundle, and the locations of the bundles produced. The amounts of the basic material of logging residues in the heaps, and their locations, were not recorded, so the productivity could not be related to the heap size. On the other hand, the distance travelled to produce a bundle should indicate the available amounts of material reasonably well. The GPS marking of the bundles may have had an accuracy of ±15 m. The accuracy is greater if the GPS receiver (Magellan 320) is placed on the same spot for a while than if a very rapid reading is taken. In this

147

I. Wästerlund and A. Öhlund

Productivity Study of WoodPac Bundling of Logging Residues and Small Stems (139–151)

case, the receiver was left on the bundles for 30–35 s to ensure that the spot determinations were as accurate as possible, at that time. The positions from the receiver were transferred in NMEA and then converted to RT90 using the program GPS Pathfinder Office 2.70. A constant height of 60 m above the ellipsoid WGS84 was assumed at conversion. The productivity of the WoodPac machine was 19.3 bundles with logging residues per E0hour, but there was substantial variation in time expenditure between the best and worst bundles. The average effective time for the 10 best bundles was about 6 min per Mg (25 bundles per E0) and 14 min for the 12 worst, indicating that improvements can be made to both the work method and the machine (cf. discussion in Kärhä and Vartiamäki 2006). Differences between the 10 best and 10 worst bundles did not depend on the available amounts of material per 100 m, but there was a threefold spread in time consumption for each of the elements feeding, driving during bundling, tying and miscellaneous. More detailed studies on these work elements and conditions are needed to identify the reasons for the large differences observed. Delays occurred in the work element tying for several reasons, e.g. sometimes the twine broke, sometimes the tying was disrupted by branches sticking out from the bunch and sometimes the string did not attach to the bundle. Improvements to the machine are already being made to reduce the time requirements for this work element from 30 s to about 19–20 s per bundle. The average bundle in the present study contained about 270 kg DM biomass. With an average energy content of 19.7 MJ kg-1 (Pettersson and Nordfjell 2007), each bundle would contain 1.48 MWh and the production per hour could amount to 28.9 MWh. Andersson and Hudson (1997) reported that the Bala Press baling machine could produce 17–18 bales per E0 hour when baling material, at stump site, with moisture contents similar to those in the present study. The energy content was 1.6–1.7 MWh per bale, indicating productivity per hour of about 29 MWh, similar to that found for the WoodPac. Andersson et al. (2000) and Kärhä and Vartiamäki (2006) describe studies in which both Fiberpac and WoodPacmachines were examined, and the WoodPac data they provide are about consistent with the results obtained in the present study. The Fiberpac machine was found to produce 18–35 bundles per effective hour in the cited studies. The bundles produced by Fiberpac and WoodPac machines are about the same size, but Fiberpac bundles may be less compressed (0.98 MWh per bundle according to Kärhä and Vartiamäki 2006). Assuming that the DM content in the Fiberpac bundle

148

is 250 kg for similar material as above, Fiberpac would then produce 27–47 MWh per E0 (productivity figures of 10–30 bundles per E15are indicated in the Timberjack brochures, Anon. 2003). However, Kärhä and Vartiamäki (2006) state that the productivity is greatly influenced by the quality of residue piles and windrows, as was also found in the present studys. Spinelli et al. (2008) report a production of 25 green ton/h when bundling poplar trees on a poplar plantation (11 DM ton/h at 55% moisture content). The Fixteri bundler produced 2.8–3.7 m3 with mainly small stems (17–89 litres/sten) at first thinning (2.56 m long, around 50% in moisture content and roughly 325x0.5x19.7/3600 = 0.89 MWh), which is much less than WoodPac but, on the other hand, it cut the stems (Jylhä and Laitila 2007). Rogbico produced 25–38 MWh per machine hour, which is much more but, on the other hand, all material was already transported to the landing (forwarding logging residues took all the profit from the bundling according to Lindroos et al. 2010) However, the WoodPac machines handle 24% more material than the amount produced as bundles. This decreases the difference in the amount of logging residues handled per hour between the two bundling systems. WoodPac tumbling and shaving procedure also appears to take about 15–20% more time than the Fiberpac compaction process. On the other hand, 20% of the fine material is left in the forest when using WoodPac, which could be advantageous (see below). The low production with young stems may be caused, to some extent, by the fact that some trees had to be bent in back to the chamber, but the major problem in this case was definitely the low biomass concentration on the ground. The trees were simply spread out, albeit in one direction, after the motormanual cleaning and bundle production of only 2.5–3 Mg DM per E0 hour may be significantly lower than potential figures. The only clear effect of biomass concentration was found in this part of the study: at more than 5 Mg green material produced per 100 m driven, time consumption stabilised at around 14 min Mg–1. This indicates that a production rate of 4.3–4.4 Mg per hour could easily be achieved with better concentration of young stems, which should be confirmed in further studies, which is similar to production reported for mountain forests and maritime pine (Cuchet et al. 2004, Spinelli et al. 2012). The forwarding productivity (15.6 Mg per E0) was only moderate (but comparable to reported figures, Johansson et al. 2006, Kärhä and Vartiamäki 2006), relative to handling logs, if only the transported mass is considered. However, the CRLs produced filled up the forwarder bunk. For handling loose logging residues with single travel distances of 280 m (the average Croat. j. for. eng. 35(2014)2

Productivity Study of WoodPac Bundling of Logging Residues and Small Stems (139–151)

distance in the present study), normal productivity would be about 15 m3 per hour (Hakkila and Nurmi 1997). According to Brunberg et al. (1998) forwarding productivity may be equal to about 8.5 green Mg per E0 at distances of 280 m (adjusted for the moisture content found here). In the present case, the volume of the loads was about 16 m3, indicating terrain transport productivity of about 50 m3 per E0. Bundling logging residues would thus improve the efficiency of terrain transport more than 2.5 fold, as indicated by Andersson et al. (2000).

5.1 Suggested improvements On average, seven crane cycles were required to produce one bundle with a 0.26 m2 grapple. An interesting question to address is whether a bigger grapple on the base machine would result in shorter feeding times, in spite of the rather small inlet slot. With a bigger grapple, the concentration of the material could have more effect. The best figures obtained suggest that, with improved working techniques, it should be possible to attain the time consumption levels indicated in Table 7. Feeding of young stems would, of course, be improved using material collected with a feller-buncher, but long stems would undoubtedly need more handling compared to logging residues. Production of 24 bundles per E0 hour would be attainable, indicating an E15 production of about 20 bundles, but only half this number with young stems.

5.2 Environmental aspects The machine consumed 292 litres of diesel while producing 250 bundles, equivalent to roughly 1.2 l per Table 7 Possible effective times for production of 1 Mg green material with 50% moisture content using the WoodPac bundler Cmin per 1 Mg

Cmin per 1 Mg

Logging res

Young stems

Crane cycle

2.43

5.55

Feeding

0.52

1.01

Driving + bundling

0.42

1.04

Compression

0.45

0.69

Tying

0.42

Work element

Unloading

0.28

0.31

Miscellaneous

0.13

0.14

Effective time

4.65

9.16

Bundles per E0

Croat. j. for. eng. 35(2014)2

I. Wästerlund and A. Öhlund

bundle, when bundling logging residues. Thus, the bundling costs in terms of energy expenditure were equivalent to about 0.8% of the energy content of the prepared bundles. The fuel consumption of forwarding the bundles was estimated to be 13 litres, equivalent to 0.27 l per bundle and 0.18% of transported energy. If transporting loose logging residues, the machine would spend three times more time, and the energy expenditure could amount to 0.5–0.6%. Thus, more than half of the energy consumed for bundling is recouped solely by the reductions in energy costs for forwarding. Based on data from Bohm Larsson (2004) about 25.3 kg/ha of nitrogen, 3 kg of phosphorous, and 10.6 kg/ha of potassium were left on the spruce clear-cut area thanks to the shave off, which is worth quite a lot if replaced with the bought fertilizer. Nevertheless, the tumbling and shaving action will inevitably take some time, placing the WoodPac system at a disadvantage relative to both the Fiberpac and Bala Press processes. The work elements »driving+bundling« and »compression« account for 20% of the time. As discussed above, this is roughly equivalent to the time difference between the WoodPac and Fiberpac production techniques. According to Rheén (2004) branches of young spruce trees may have an ash content of 2.12%, and according to Nurmi (1993), the foliage contents may be as high as 2.2–8.7%. The high levels of inorganic substances in the needles may cause problems during combustion (Orjala et al. 2000, Aho and Silvennionen 2004) and it may be better to leave them in the forest. Thus, the lower bundling productivity obtained with WoodPac could be compensated by producing bundles of higher quality due to lower ash contents, and perhaps better drying properties (Jirjis and Nordén 2002, Pettersson and Nordfjell 2007), but these possibilities are still to be proven and evaluated (cf. Asikainen et al. 2002). Risks related to reductions in long-term site productivity and costs of ash recycling may also be reduced by using the WoodPac system (Burger 2002, Hakkila 2002). The catchment of material in the forest would be reduced but, on the other hand, the losses accrued during storage handling and transports could also be reduced with the shaving technique. An important question to consider is how much the potential reduction in transport costs, reductions in the amounts of nutrients taken from the site, better drying capacities and better fuel are worth. If a high price-weighting is attributed to these factors (or possibly any price), it could even be worth increasing the proportion of material shaved off, either by prolonging the tumbling or by equipping the rollers with sharper or longer spikes.

149

I. Wästerlund and A. Öhlund

Productivity Study of WoodPac Bundling of Logging Residues and Small Stems (139–151)

6. Conclusions The studies have shown the following results: Þ The WoodPac machine produced 19.3 bundles with logging residues per E0, equivalent to 28.5 MWh. Þ The productivity was not influenced by the amount of green mass, as long as there was more than 5 Mg per 100 m driving distance and the material was collected from heaps. Þ Productivity can be improved by about 20% for logging residues. Þ Young stems longer than the compartment can be bundled, but their bundling may take twice as long as bundling logging residues. Bunching stems before using the bundler may considerably improve the production, and this possibility needs to be studied. Þ About 20% of the handled material was shaved off, mainly as fine material, and left in the forest, leaving some NPK back to the forest. Þ Forwarding productivity was improved at least 2.5 fold with bundles compared to forwarding loose residues. Þ Energy expenditure for bundling and for forwarding the bundles at a distance of 280 m was equivalent to about 0.8% and 0.2%, respectively, of the energy content of the bundles produced.

Acknowledgements The authors would like to thank Magnus Pettersson and MalinBohm Larsson for their help during the study, plus Erland Josefsson, the constructor of WoodPac, for his support and help during the experiments. The study was financed via the NIFES project Semimobile pelletiser and Efokus, Sollefteå.

7. References Aho, M., Silvennionen, J., 2004: Preventing chlorine deposition on heat transfer surfaces with aluminium – silicon rich biomass residue and additive. Fuel, Vol 83. Anon., 1978: Forest work study nomenclature. The Nordic Work Study Council, NISK, Boks 61, 1432 Ås, Norway, 81–99. Anon., 1989: Kraftledningar i fysisk planering [Power lines in physical planning]. Statens Energiverk/Statens Naturvårdsverk/Boverket/Vattenfall, PBL/NRL Nr 27. (In Swedish). Anon., 1997: SS 18 71 70, Biofuels and peat – Determination of total moisture. Swedish Standard Institution, Stockholm, 1–65 (In Swedish). Anon., 2003: Timberjack 1490D Slash bundler. www.timberjack.com/products/forest-energy/1490D.htm, Timberjack Forestry Group 2003.

150

Andersson, G., Asikainen, A., Björheden, R., Hall, P.W., Jirjis, R., Mead, D.J., Nurmi, J.,Weetman, G. F. 2002: Production of forest energy. In: Richardson, J., Björheden, R., Hakkila, P., Lowe, A. T., Smith, C.T. (eds.). Bioenergy from sustainable forestry, guiding principles and practice, Kluwer Academic Publisher,Dordrecht, 49–123. Andersson, G., Hudson, B., 1997: Baling of forest residues – a system analysis. In: Hakkila, P., Heino, M.,Puranen, E. (eds.),Forest management for bioenergy, The Finnish Forest Research Institute, Vantaa, Research Paper (640): 102–110. Andersson, G., Nordén, B., Jirjis, R., Åstrand, C., 2000: Composite residue logs cut fuel costs. The Forestry Research Institute of Sweden, Uppsala,Resultat No 4. (In Swedish with English summary). Arola, R.A., Radcliffe, R. C., Winsauer, S.A., 1985: Chunking bundled small-diameter stems. Forest Prod. J. 35(4): 40–42. Asikainen, A., Björheden, R., Nousianen, I., 2002: Cost of wood energy. In: Richardson, J., Björheden, R., Hakkila, P., Lowe, A. T., Smith, C.T. (eds.), Bioenergy from sustainable forestry, guiding principles and practice, Kluwer Academic Publisher, Dordrecht, 125–157. Bohm Larsson, M., 2004: Fractions and nutrient removal when bungling green logging residues with WoodPac. SLU, Dept. Silviculture, Umeå, Report No16 (In Swedish with English summary). Brunberg, B., 1991: Productivity norms for stand-operating single-grip harvesters in thinnings – A study of the literature. The Forest Operations Institute of Sweden, Stockholm, Report 3. (ISSN 0346-6671) (In Swedish with English summary). Brunberg, B., Hedenberg, Ö., Jonsson, T., 1990: Multitree technology – Ist impacts on logging costs and pulpmill raw materials. The Forestry Research Institute of Sweden, Stockholm, Report No 3. (ISSN 0346-6671) (In Swedish with English summary). Brunberg, B., Frohm, S., Nordén, B., Thor, M., 1998: Forest bioenergy fuel – final report of commissioned projects. The Forest Operations Institute of Sweden, Stockholm, Redogörelse 5. (ISSN 1103-4580) (In Swedish with English summary). Burger, J. A., 2002: Soil and long-term site productivity values. In: Richardson, J., Björheden, R., Hakkila, P., Lowe, A.T., Smith, C.T. (eds.). Bioenergy from sustainable forestry, guiding principles and practice, Kluwer Academic Publisher, Dordrecht, 165–189. Carlsson, T., Larsson, M., Nordén, B., 1983: Lastbilstransporter av träddelar – studier 1981/82 [Lorry transports of tree sections – studies 1981/1982]. Forskningsstiftelsen Skogsarbeten, Stockholm,Resultat No 9. (ISSN 0280-1884) (In Swedish). Cuchet, E., Roux, P., Spinelli, R., 2004: Performance of a logging residue bundler in the temperate forests of France. Biomass & Bioenergy 27: 31–39. Engblom, G., 2007: System analyses of Wood fuel transports. SLU, dept. Forest Resource Management, Umeå, Report No 175. (In Swedish with English summary). Eliasson, L., 1998: Analyses of single-grip harvester productivity. ActaUniversitatis Agriculturae Suecicae, Silvestria 80: 1–24. Croat. j. for. eng. 35(2014)2

Productivity Study of WoodPac Bundling of Logging Residues and Small Stems (139–151) Finkman, M., Thörnqvist, T., 1986: Storage of bundled undelimbed pulpwood and logging residues. Dept. of For. Prod., Swed. Univ. of Agric. Sciences, Uppsala, Report 180. (ISSN 91-576-2756-8) (In Swedish with English summary). Hakkila, P., 1989: Utilization of residual forest biomass. Springer Series in Wood Science, Berlin, ISBN 3-540-50299-8. Hakkila, P., Nurmi, J., 1997: Logging residues as a source of energy in Finland. In: Hakkila, P., Heino, M., Puranen, E. (eds.). Forest management for bioenergy, The Finnish Forest Research Institute, Vantaa, Research Paper 640: 90–101. Hakkila, P., 2002: Operations with reduced environmental impact. In: Richardson, J., Björheden, R., Hakkila, P., Lowe, A.T., Smith, C.T. (eds.). Bioenergy from sustainable forestry, guiding principles and practice, Kluwer Academic Publisher, Dordrecht, 244–261. Kärhä, K., Vartiamäki, T., 2006: Productivity and costs of slash bundling in Nordic conditions. Biomass&Bioenergy 30: 1043–1052. Jirjis, R., Nordén, B., 2002: Stock piling of composite residue logs (CRLs), small biomass losses and no health problems. The Forestry Research Institute of Sweden, Uppsala, Resultat No 12. (In Swedish with English summary). Jonsson, M., Kjellberg, M., Lindholm, D., 1992: Utilization of non commercial wood from operations for energy forestry. Vattenfall Research Bioenergi, Vällingby U(B) 1992/34. Projekt Skogskraft Rapport No 11. (ISSN 11100-5130) (In Swedish with English summary). Johansson, J., Gullberg, T., 2002: Multiple handling in the selective felling and bunching of small trees in dense stands. Int. J. For. Eng. 13(2): 25–34. Johansson, J., Liss, J-E., Gullberg, T., Björheden, R., 2006:Transport and handling of forest energy bundles – advantages and problems. Biomass&Bioenergy, 30: 334–341. Jylhä, P., Laitila, J., 2007: Energy wood and pulpwood harvesting from young stands using a prototype whole-tree bundler. Silva Fennica 41(4): 763–779. Larsson, U., 1998: Power-line corridors – a resource of forest fuel production. SLU, Forest Technology, Umeå, Students reports No17. (In Swedish, English summary).

I. Wästerlund and A. Öhlund

Liss, J-E., 1995: Bunthantering av klena träd från först gallringar [bunching of small trees in first thinning]. SLU Info/Skog, Sveriges lantbruksuniversitet, Garpenberg, Småskogsnytt, 5:12–15 (In Swedish). Lindroos, O., Matisons, M., Johansson, P., Nordfjell, T., 2010: Productivity of a prototype truck-mounted logging residue bundler and road-side bundling system. Silva Fennica 44(3): 547–559. Nordfjell, T., Liss, J-E., 2000: Compressing and drying of bunched trees from a commercial thinning. Scand. J. For. Res. 15: 284–290. Nurmi, J., 1993: Heating values of the above ground biomass of small-sized trees. Acta Forestalia Fennica 236: 1–30. Orjala, M., Ingalsuo, R., Patrikainen, T., Hämäläinen, J., 2000: Combusting of wood chips produced by different harvesting methods in fluidised bed boilers. The 1st World Conference and Exhibition on Biomass for Energy and Industry, Sevilla, 6p. Pettersson, M., Nordfjell, T., 2007: Fuel quality changes during seasonal storage of compacted logging residues and young trees. Biomass & Bioenergy 31(11–12):782–792. Rhen, C., 2004: Chemical composition and gross calorific value of the above-ground biomass components of young Picea abies. Scand. J. For. Res. 19: 72–81. Rolew, A. M., 1988: Siwork 3 version 1.1. Work study and field data collection system based on Husky Hunter handheld computer. Danish Forest and Landscape Research Institute, Lyngby, Denmark,1–35. Schiess, P, Yonaka, K., 1982: Evaluation of a new concept in biomass fiber field processing and transportation. In: Sarkanen, K., Tillman, D., Jahn, E. (eds.), Progress in biomass conversion, Academic Press, New York, Vol(3): 183–214. Spinelli, R., Nati, C., Magagnoti, N., 2008: Harvesting shortrotation poplar plantations for biomass production. Croat. J. For. Eng. 29(2): 129–139. Spinelli, R., Magagnoti, N., Picchi, G., 2012: A supply chain evaluation of slash bundling under the conditions of mountain forestry. Biomass&Bioenergy 36: 339–345.

Authors’ address:

Received: April 17, 2013 Accepted: April 28, 2014 Croat. j. for. eng. 35(2014)2

Prof. Iwan Wästerlund, PhD. * e-mail: iwanolasgarden@telia.com Olasgarden forest and roads Solvägen 9 918 32 Sävar SWEDEN Anders Öhlund (Ringbjer), MSc. e-mail: anders.ohlund@sca.com SCA Forest AB Måsvägen 20 94153 Piteå SWEDEN * Corresponding author

151

Original scientific paper

Stakeholders’ Perspectives on Utilization of Logging Residues for Bioenergy in Turkey Hasan Alkan, Mehmet Korkmaz, Mehmet Eker Abstract Although using logging residues for bioenergy is not a new issue for countries such as Sweden, Finland, Austria, Germany, etc. that are developed in terms of forestry, it is a new issue that requires studying for countries such as Turkey. This study investigates the views of forest engineers working in forest enterprises, researchers working in forestry research institutes and academicians working at universities concerning the use of logging residues for bioenergy that are not currently used in energy production. Within the framework of the study, a questionnaire was sent out to 181 forest engineers, 77 academic staff members and 29 research institute employees, a total of 287 respondents. According to the results of the study, logging residues that are either left in the forest floor or collected by forest villagers for the purpose of fire wood have a favorable potential for energy and forestry if they are used in bioenergy production. Thus, the issue is substantial and of primary importance for Turkey. On the other hand, there are barriers in developing bioenergy sector and using logging residues for this purpose. In order to remove these barriers, first of all, forestry administration should clarify its strategies and policies related to the issue. Keywords: Bioenergy, logging residues, stakeholder’s perspectives, Turkey

1. Introduction In parallel with population growth, requirements for energy increase every single day while energy sources are rapidly decreasing. The speed of energy consumption in the world is 300 thousand times higher than the speed of fossil fuels formation. In other words, fossil fuel formed in a thousand years is consumed only in one day (Yılmaz et al. 2003). For this reason, the fact that fossil fuel reserves such as petrol, coal and natural gas, will come to an end at most between the years 2030 and 2050 should be taken into consideration (Akyüz 2010). On the other hand, consumption of world energy chiefly from fossil fuels can cause environmental problems such as air, water and soil pollution to a dangerous extent (Ertürk et al. 2006). Another matter related to the issue is the possibility of employment that can emerge based on using new energy resources and types of production (Gökcöl et al. 2009, Halder et al. 2012). Croat. j. for. eng. 35(2014)2

Thus bioenergy has become one of the most dynamic and rapidly changing sectors of the global energy market. In general, three main categories of bioenergy resources are used globally: forest biomass, agricultural biomass and waste biomass (Resch et al. 2008, Halder et al. 2012). Forest biomass has the potential to be one of the most convenient energy sources in the future as it was in the past (Ladanai and Vinterbäck 2009). In USA, one of the major sources of woody biomass comes from conventional forests, particularly logging residues (Gan and Smith 2006). 90% of biomass on earth consists of the animals and microorganisms together with stems, branches, needles and leafs and debris (Saraçoğlu 2006a). Logs that are left in the forest and not evaluated due to low additional value, their main stems, roots, stumps, stem end, top of stem together with timber, cone, bark, needles and leafs of small diameter trees that are damaged in felling or transport can be called as logging residues (Röser et al. 2008, Eker et al. 2009). Although logging residues

153

H. Alkan et al.

Stakeholders’ Perspectives on Utilization of Logging Residues for Bioenergy in Turkey (153–165)

have the potential to be used directly or indirectly like other forest biomass, which is used as clean and renewable energy source, it is known that it not used as much as possible (Ateş et al. 2007, Aruga et al. 2011, Saraçoğlu 2010, Eker 2011). Due to rising costs based on factors such as: topographic structure of Turkey, actual timber harvesting system, work force, lack of technology, etc., logging residues can be left in the forest without undergoing a process (Acar et al. 2001, Saraçoğlu 2006b). As a result of not evaluating the logging residues and leaving them in the forest, fuel material density and risk of fire in the forest base increase and the problem of bark beetle damage and rejuvenation obstacles can emerge. Not collecting the logging residues and leaving them in the harvesting area can cause the loss of a potentially energy supplying material for the local population and loss of possible employment opportunities. On the other hand, the literature on forest biomass also addresses various environmental, ecological, economical, and logistical issues associated with harvesting and transporting logging residues (Dirkswager et al. 2011, Kühmaier and Stampfer 2012). For example in low site indexed areas, removing logging residues together with nutrients will impoverish the growing environment (Sterba 2003, Hacker 2005). In this respect, in Finland it is regulated by law that 30% of logging residues must be left in the area (Hakkila 2006). Turkey has an important potential in terms of renewable energy resources. The forest is the most important of these resources. The government owns approximately 99.9% of Turkish forest areas. Thus, 21.7 million hectares of forests are controlled, protected and managed through The General Directorate of Forestry (GDF), which has a large rural organization expanding along the whole country, founded by the state (Öztürk et al. 2010). The ratio of these forest areas are approximately 27.8% of the total area of the country. Tree volume in our forests is about 1.5 billion m3 every year, and regular and legal production of 14–16 million m3 of wood raw material is carried out. Forest villagers are commonly employed in the production activities (GDF 2012). It is impossible to say that necessary attention is paid and required investment is made to provide renewable energy resources in Turkey. Based on this, 80% of energy requirements is still supplied by importing and the environment is polluted rapidly by intense fossil fuel use. Besides, researches related to renewable energy are not adequate. Thus, the perception and attitude of stakeholders to the issue are different, and consistent strategies and policies for making use of forests to produce bioenergy cannot be

154

carried out. To remove these handicaps and start using harvest residues as raw material for energy sector is closely related with the approach, attitude and behavior of related organizations and institutions and other stakeholders.

2. Aim of the study The aim of this study is to determine the views of forest administration employees, forest research institute employees and academicians who are considered the most important stakeholders of the issue and the ones that have the role in forming the regulation to use biomass, bioenergy and harvest residues in bioenergy. Based on stakeholders’ views it has been found out that the biomass and its components are not well determined and such issues are brought into conformity with the now current literature.

3. Materials and methods The study was designed as a two-step process. In the first phase, the researches related to the issue in Turkey and globally were supplied by general survey and they were evaluated. In the second phase, researchers (research and development – R&D) and executers that are the stakeholders in the issue were determined. Research and development group was composed of academic staff working in various forestry faculties in Turkey that have or might have the potential to carry out theoretical or practical researches in the field and researchers working in forestry research institutes. Strategies and policies determined about forestry were applied by forest engineers who work within the framework of forestry general directorate. For this reason, the group of executers in the study is composed of engineers working in the General Directorate of Forestry.

3.1 Data collection Primary data were collected in the course of the field survey by using a questionnaire. Existing literature and secondary data such as reports, plans, etc. were also investigated in this study, as well as other material. In order to determine the views of group members about bioenergy, an electronic questionnaire form was produced. The questions included in the forms were prepared according to criteria of questionnaire forming. In the questionnaire, based on the specifications of the subject; open ended, multiple choice questions with questions including two or more answers were used. In order to try the questions and eliminate the shortcomings, ideas of academicians and executers Croat. j. for. eng. 35(2014)2

Stakeholders’ Perspectives on Utilization of Logging Residues for Bioenergy in Turkey (153–165)

were consulted and pre-tests were carried out. As a result, some of the questions were excluded from the questionnaire, while some other questions that were considered important were added. The final form was filled in by each member of R&D and execution group. In Turkey, there are 217 Forestry Operation Directorates under 27 Regional Forestry Directorates and there are 1,308 Forest Sub-District Directorates empoying 4,799 forest engineers. In order to determine the view of the group of executers with 10% sampling error, the sampling size was calculated by the formula below (Karasar 2005): n= Where: n Z N p, q

Z2 ´ N ´ p ´ q N ´ D2 + Z2 ´ p ´ q

(1)

sampling size; safety coefficient (for 95% safety level Z=1.96); population size (4,799); probability of availability of the mass to be measured in the main mass; D sampling error accepted (10%). Based on the above formula, it was calculated that it was necessary to provide 94 questionnaires. However, in this study, 181 questionnaire forms were filled and analyzed to increase the reliability of the study. Actually, 205 forest engineers participated in the questionnaire activity on behalf of the group of executers. Since there were significant deficiencies in the questionnaire received and in transferring the data into MS Excel data base, 24 questionnaires were excluded from the study. In the study conducted about R&D group, the sample size was calculated. Academicians working in different 9 forest faculties and researches working in 8 research institutes was selected as a target R&D group, to fill the questionnaire from in electronic media. In this respect, the number of participants who were thought to be related to the subject and asked to participate in the study was approximately 150. 29 researchers from research institute and 77 academic staff members from universities participated in the questionnaire activity by R&D group. The questionnaire studies conducted on-line showed various feedback levels, namely between 6% and 73%. In general, feedback level between 20% and 40% is considered acceptable (Tekin and Zerenler 2005, Derinalp 2007). In this study, the feedback level was 70.6%. Additionally, in order to carry out situation analyses related to the issue, interviews were made with forest villagers, cooperative administrators, raw material supCroat. j. for. eng. 35(2014)2

H. Alkan et al.

pliers, forestry operation directorates, investors and other stakeholders and significant data were obtained.

3.2 Analysis of data For the statistical analyses of the questionnaire, SPSS 15.0 package program was used. In the analyses, data about questions were installed at first, and frequencies and percentages were used according to the specification of the questions. Statistical relationship between answers given by executers and R&D group members were analyzed by using Mann-Whitney U test (Kalaycı 2010, Özdamar 1999, Nachar 2008). In order to analyze the overall-case and summarize it effectively, SWOT analysis was used. SWOT is an abbreviation for strength, weakness, opportunity and threat. SWOT analysis aims at identifying the strengths, weaknesses, opportunities and threats in the environment. Having identified these factors, strategies are developed. These strategies may develop the strengths; they can eliminate the weaknesses, exploit the opportunities or counter the threats. While the strengths and weaknesses are identified by an internal evaluation, the opportunities and threats are identified by an external evaluation (Dyson 2004).

4. Result and discussion 4.1 Profile of stakeholders Upon investigating the profile of stakeholders, it was found out that the majority of the practitioners (83.3%) were in middle age group (26 to 45). The other age groups were: 9.94% for 18 to 25 age group and 6.63% for 46 to 65 age group. This case bears importance in terms of obtaining the expected data in the study. Besides, 57.46% of the participants had 6 or more years of vocational experience. When R&D group was investigated, it was understood that: 69.73% of the academicians were working in Forest Engineering, 25% were working in Forest Industry Engineering and 5.27% were working in other departments. The academicians with the highest rate of participation were Assistant Professors (41.56%). It was followed by Associate Professors (19.48%) and Professors (18.18%). The units with the highest amount of participation in R&D group were: forest economics (15.38%), silviculture and afforestation (11.54%), forestry harvesting and transportation (11.54%) and forest product chemistry (9.62%) departments.

4.2 Definition and components of logging residues There is not yet an adequate term for identifying residual materials such as thin branches, bark, needles

155

H. Alkan et al.

Stakeholders’ Perspectives on Utilization of Logging Residues for Bioenergy in Turkey (153–165)

Table 1 Forest biomass components to be used for the purpose of energy production* Components

R&D group

Execution group

Frequency

Branch woods

72.64

50.83

Cutting residues

80.19

149

82.32

Energy forest

43.40

14.36

Small-diameter trees

53.77

19.89

Industrial residues

39.62

33.15

Shrubs

58.49

38.67

Needles, leaves

33.96

26.52

Cones

47.17

40.88

Bark

56.60

106

58.56

Other

12.26

4.97

*Respondents had the chance to choose more than one option

and leaves, logs, etc., materials that emerge upon harvesting and not evaluated because they are unmerchantable in our country. They are identified as harvesting residue in GDF (2009), felling residue in Ateş et al. (2007), and logging residue in Eker et al. (2009) and Saraçoğlu (2010). Views of the stakeholders related to the issue are given in Fig. 1. As seen in Fig. 1, harvesting residue, logging residue and cutting residue are the three most commonly offered terms for both stakeholder groups. When evaluations by Eker et al. (2013), international literature, production process, quality of the outcome, and other

issues are taken into consideration, it can be said that logging residue is more suitable than other terms. Thus, in order to identify this substance, the term logging residue was used. There are differences as to which components of forest biomass to use for the purpose of producing bioenergy and which components to use as logging residues (Table 1). It can be seen that according to stakeholders, felling residues and branches and wood are considered to be the most important components to form logging residues. Besides, with species whose bark can be removed, the bark is also considered as an important component.

4.3 Current use of logging residues in Turkey The fact that renewable natural resources are consumed rapidly according to the current literature, globTable 2 Current evaluation method of logging residues Evaluation method

Fig. 1 The term offered to define the woody residual material

156

Frequency

It is left in stand

53.04

It is stored in stand for mulching and protection of seedling

35.36

It is collected by forest villagers with reasonable unit prices

130

71.82

It is collected by supplies because of standing tree selling

22.10

It is burnt in stand

9.39

Other

4.42

Croat. j. for. eng. 35(2014)2

Stakeholders’ Perspectives on Utilization of Logging Residues for Bioenergy in Turkey (153–165)

H. Alkan et al.

Table 3 Views about adequacy of the current activities Current activities are necessary and adequate in terms of constituting and developing wood based bioenergy sector in our country Stakeholders

I agree

I am Neutral

I disagree

Number

R&D group

5.66

16.04

78.30

Execution group

7.19

23.20

126

69.61

al climate change scenarios, approaches to reduce the dependence on fossil fuels and petroleum products by the reasons of carbon cycle and sustainable development made bioenergy production and use of logging residues for this purpose important (Ertürk et al. 2006, Halder et al. 2012, Resch et al. 2008, Ladanai and Vinterback 2009, Yılmaz et al. 2003). However, using logging residues in producing bioenergy is still not a common issue (Aruga et al. 2011, Ateş et al. 2007, Eker et al. 2009, Eker 2011, Röser et al. 2008, Saraçoğlu 2010). In countries where logging residues are used as a source of bioenergy, the logging residues after the production of wood raw material is chipped and made ready for various purposes (Spinelli and Hartsough 2001, Yoshioka et al. 2002, Röser et al. 2008, Laitila et al. 2013). For example, 20% of production in Finland (Malinen et al. 2001), 18% in Sweden and 14% in Austria are supplied from bioenergy made of wood and plant wastes (Saraçoğlu 2008). In USA, the total amount of recoverable logging residues reaches 36.2 million dry tonne (Gan and Smith 2006). It is stated by Demirbaş (2001) that the potential of forest residues as bioenergy source is 18 million ton and their energy value is 5.5 Mtep (million ton petrol equivalent). It is stated by Saraçoğlu (2008) that the annual biomass potential of Turkey is 17.2 Mtep. Still, the use of logging residues in bioenergy utilization is not yet applicable and the activities mentioned are not used in Turkey. According to the data obtained by the group of executers, the most common use of logging residues by forest villagers is for cooking and/or heating. These logging residues are collected by forest villages by a moderate price paid as a subvention. In times when supply costs are high, and heating and cooking needs are supplied by replacement products, forest villagers are not very interested in this method of use. In such conditions, the residues stay in their places without undergoing a process and they are left to rot. In places where logging residues are considered as an obstacle in reforestation activities, forestry organizations move these residues to a certain place and burn them. By the help of standing tree selling which is commonly used in the recent years, the utilization of logging residues has also started (Table 2). Croat. j. for. eng. 35(2014)2

4.4 Need for using logging residues for bioenergy The necessity of using logging residues for bioenergy is becoming obvious. Furthermore, the answer to the question »Do you think the issue of logging residues-bioenergy is one deserving attention, research and focus on?« was Yes by 94.34% of R&D group members and 93.37% of execution group members. When comparing the answers to the question, no statistical difference between groups was established (p>0.05). On the other hand, there was a significant difference between views of academic staff members and researchers working in forestry research institutes that form R&D group (U=991.500, p=0.027) according to Mann-Whitney U Test (Table 81 – Test number 1). Academic staff members believe that it is necessary to increase the use of logging residues for bioenergy. Necessary scientific, R&D quality or applied researches about logging residues have not been carried out so far. Views of the stakeholders concerning the premise »Current activities are necessary and adequate in terms of constituting and developing wood based bioenergy sector in our country« are shown in Table 3. According to the findings of Mann Whitney-U test (Table 8 – Test number 2), there is no statistically significant difference between the views of the two groups (U=8,775.000, p=0.121). The most important reason for the stakeholders view that the current studies are inadequate and that more researches should be carried out is that they believe that in Turkey bioenergy should be produced by logging residues in times when there is enough raw material and its supply does not cause any disagreement in terms of economic, social and ecologic concerns (Table 4). One of the reasons of the stakeholders view that current studies are inadequate and that more research

Within the framework of the study, in order to compare the groups, all Mann-Whitney U tests conducted are given in Table 8 as Test number 1, 2, 3, 4 and 5 respectively.

157

H. Alkan et al.

Stakeholders’ Perspectives on Utilization of Logging Residues for Bioenergy in Turkey (153–165)

Table 4 Views concerning the necessity of using forest biomass (logging residues) to produce energy* Stakeholders

R&D

Executers

I agree

Premises

I am neutral

I disagree

Number

It is chiefly necessary for Turkey to produce energy (heat, power and fuel) from wood biomass under current conditions

72.64

12.26

15.09

It is a luxury for Turkey to produce heat and power from wood biomass under current conditions. However, it has a potential to form a market by the support of energy forestry and other activities

60.38

17.92

21.70

In times when there is enough raw material and its supply does not cause any disagreement in terms of economic, social and ecologic concerns, it is necessary to use wood biomass in bioenergy production

85.85

5.66

8.49

It is chiefly necessary for Turkey to produce energy (heat, power and fuel) from wood biomass under current conditions

130

71.83

20.44

7.73

It is a luxury for Turkey to produce heat and power from wood biomass under current conditions. However, it has a potential to form a market by the support of energy forestry and other activities

12.71

19.89

122

67.40

150

82.87

11.05

6.08

*The respondents had the chance to choose more than one option

Table 5 State of acknowledgement of the importance of logging residues in bioenergy sector R&D group

Level of knowledge

Execution group

Number

I have no information

5.66

6.07

I have some information

44.34

112

61.88

I have a stock of knowledge which I obtained from different sources

37.74

28.18

I have made theoretical and practical researches. I have full knowledge of the issue

11.32

3.87

Other

0.94

Total

106

100.00

181

100.00

should be carried out is that parties concerned have no enough information. Table 5 shows the stakeholders’ state of acknowledgement in using logging residues for bioenergy production. As shown in Table 5, in Turkey there is still no adequate stock of knowledge concerning using forest biomass for the purpose of producing bioenergy. According to the results of Mann-Whitney U test (Table 8 – Test number 3), there is a statistically significant (U=7,753.000, p=0.002) difference between the views of researchers (R&D group) and executer groups. These are two separate groups whose views are consulted

158

within the framework of knowledge level related to the subject. R&D group members state that they have more information than the employees concerning the issue. The most important reason for this is that they study global literature more than the employees of the General Directorate of Forestry.

4.5 Potential gains by using logging residues for the purpose of energy production Stakeholders’ views related to potential gains by using logging residues for the purpose of energy production are shown in Table 6. As shown in Table 6, 98.0% Croat. j. for. eng. 35(2014)2

Stakeholdersâ&#x20AC;&#x2122; Perspectives on Utilization of Logging Residues for Bioenergy in Turkey (153â&#x20AC;&#x201C;165)

H. Alkan et al.

Table 6 Views related to potential gains by using logging residues for the purpose of energy production * Potential gains

R&D group

Execution group

Number

Supports local energy production

72.64

132

72.93

Supports rural development

70.75

116

64.09

It helps these activities to be carried out regularly by obtaining material from tending and thinning

42.45

30.39

Helps to reduce fuel material deposited in forest and decrease fires

67.92

142

78.45

Create employment opportunities in energy industry

62.26

109

60.22

Within the framework of optimization of sustainability of forest utilization, it can produce new vocational opportunities to forest engineers, private forestry companies and technical staff

53.77

102

56.35

It can contribute to improve wild life habitats

8.49

9.39

It can supply additional source of income for forestry management

52.83

107

59.12

It will not have any gain

1.89

1.10

Other

5.66

3.87

*The respondents had the chance to choose more than one option

of R&D group and approximately 99.0% of executer group believe that using logging residues for the purpose of energy production could possibly have a gain. In the new sector to be formed by using logging residues for the purpose of energy production, there will be potential opportunities such as supporting local energy production, providing new working opportunities in terms of energy sector, supporting rural development, bringing professional and attractive gains for forest engineers, supplying forestry managements with new sources of income, giving positive effects to forest ecosystem when conducted appropriately, etc. The group with the highest level of gain was R&D with supporting local energy production, while the gain for the executer group was the reduction of fuel material deposition in forest and risk of fires. Furthermore, especially executer group members believe that after final cutting removing logging residues from forest will give positive effects in terms of the health and reforestation of the area. There is statistically significant difference between groups (U=8,329.000, p=0.030). The views of R&D group are more pessimistic (Table 8 â&#x20AC;&#x201C; Test number 4).

4.6 Obstacles in using logging residues for the purpose of energy production Views of stakeholders about the establishment of the bioenergy sector and obstacles in using logging Croat. j. for. eng. 35(2014)2

residues for the purpose of energy production are shown in Table 7. As shown in Table 7, more than 50% of the obstacles considered by R&D group are: lack of necessary efforts by competent authorities, difficulty in sustainable supply of logging residues, high supply costs (production, transportation) of logging residues and ecologic damages caused while removing logging residues from the forest area. More than 50% of the obstacles considered by execution group are: difficulty in providing adequate number and quality of working force, high supply costs (production, transportation) of logging residues, lack of necessary efforts by competent authorities. The first condition for considering bioenergy as an alternative or additional source of energy and using logging residues as a raw material or product for energy sector is that stakeholders (Stidham and SimonBrown 2011) and competent authorities make the necessary efforts by developing their strategies and policies, and do sample applications and divert the parties in this field. The stakeholders who participated in the study conducted by Dwivedi and Alavalapati in USA in the year 2009 also addressed this issue. In Turkey, the highest authority for this issue is the Ministry of Energy and Natural Resources. The ministry has made some important steps related to the issue. For example, in order to generalize the use of renewable

159

H. Alkan et al.

Stakeholders’ Perspectives on Utilization of Logging Residues for Bioenergy in Turkey (153–165)

Table 7 Views about the factors to prevent establishment and development of energy sector* R&D group

Factors

Execution group

Number

Lack of necessary efforts by competent authorities (lack of policies and even strategies)

66.04

53.04

Difficulty in providing adequate number and quality of working force

30.19

123

67.96

High supply costs (production, transportation) of logging residues

51.89

109

60.22

Raw material (for example fiber-chip sector) competition in forest products industry

35.85

30.39

High distribution costs of energy obtained from logging residues

28.30

23.20

Difficulty in sustainable supply of logging residues

59.43

40.88

Lack of methods and techniques necessary for the supply of logging residues (difficulty in supplying and operating chopping equipment in forest operations)

49.06

34.25

Ecologic damages caused while removing logging residues from the forest area

50.94

29.83

Potential negative effects of the use of logging residues on wild life and habitat

36.79

23.76

Other

1.89

4.97

*The respondents had the chance to choose more than one option

energy resources, they have issued the law No. 5,346 concerning »Using Renewable Resources to Produce Electricity« in 2005. With this low, biomass sources were included in the renewable energy resources. Also by the change in the law (Dec. 29, 2010), energy purchase price, which was 5.5 Euro cent/kWh, was changed to USD and this purchase price was determined as 13.3 USD – the highest for companies that carry out production based on biomass with the guarantee that their production would be purchased. Besides, on condition that local machinery and equipment is used in the facilities, there is an incentive between 0.4–2 USD cent/kWh. However, the General Directorate of Forestry that is the institution with the highest authority on the issue has not determined a strategy or policy about the issue, yet. In the recent days, The General Directorate of Forestry started a study to promote foreign investments on the South West of Turkey (for example Muğla Region), but considering the fact that there might be difficulty in the supply of raw material by the investment in some sectors, such as chip wood sector, and the pressures directed upon the institution, the General Directorate of Forestry interrupted the study. Due to the attitude of the General Directorate of Forestry, investors think that there will be problems in the sustainable supply of logging residues as raw material. Stidham and Simon-Brown (2011) state that distrust between investors, suppliers of raw material and other parties, is a significant obstacle in the development of the sector.

160

In Turkey, an adequate amount of trust has not yet been formed between the parties. Not taking steps in this issue, which is important for a sustainable implementation of the investment, is a serious obstacle and threat in the development of the sector. On the other hand, the General Directorate of Forestry does not make the necessary researches nor provides necessary support to the activities carried out. When the residual material left in the forest is taken into account in calculating the result of producing wood raw material, it seems to be light in weight and value. Another obstacle in terms of sustainable supply of logging residues is the costs (Hacker 2005, Guo et al. 2007, Röser et al. 2008, Aguilar and Garrett 2009, Stidham and Simon-Brown 2011). Allen et al. (1998) and FAO (2009) states that supply costs have a high ratio in total costs of biomass use. One of the most important reasons for this lies in high logistic costs during supply. Eriksson and Björheden (1989) state that optimization of logging residues depends on minimization of transport costs. In developed countries, where logging residues are used as a source of bioenergy and industrial raw material, for the minimization of costs, especially of transport costs, the current production methods and technologies are changed and the mechanization level used in wood production system is improved (Karha et al. 2009). For example, logging residues are generally chipped in the area (Spinelli and Hartsough 2001). It is stated that the supply Croat. j. for. eng. 35(2014)2

Stakeholders’ Perspectives on Utilization of Logging Residues for Bioenergy in Turkey (153–165)

H. Alkan et al.

Table 8 Man-Whitney U test results Test Number

Groups

Mean ranks

Sum of ranks

Academic staff

51.88

3,994.50

Researchers

57.81

1,676.50

Total

106

R&D group

106

151.72

16,082.00

Execution group

181

139.48

25,246.00

Total

287

R&D group

106

161.36

17,104.00

Execution group

181

133.83

24,224.00

Total

287

R&D Group

106

155.92

16,528.00

Execution group

181

137.02

24,800.00

Total

287

R&D group

106

118.68

12,580.00

Execution group

181

158.83

28,748.00

Total

287

of 70% of woody biomass in Finland is obtained by this method (Junginger et al. 2005). The current production–transportation system in Turkey is an activity which is expected to be costly. Thus, alternative methods and techniques, like chipping in the forest after collecting and then transporting can be necessary. However, the lack of equipment of forest villagers and cooperatives that will be involved mandatorily in all related actions based on the low, and their inability to solve this lack of technology with their financial resources, etc. are considered to be the obstacles that can prevent the change. Whether extracting logging residues from the forest might have positive or negative effects on the forest ecosystem in terms of ecology is an issue commonly addressed in the current literature. Especially in places where growing environment is not fertile, if the organic material and food materials are removed from the forest in high amounts, the growing environment can grow poor (Sterba 2003, Hacker 2005). Stakeholders’ views as to whether extracting logging residues from the forest might have positive or negative effects Croat. j. for. eng. 35(2014)2

M-Whitney U

991.500

–1.187

0.027

8,775.000

–1.550

0.121

7,753.000

–3.036

0.002

8,329.000

–2.165

0.030

6,909.000

–4.246

0.000

on the forest ecosystem in terms of ecology was analyzed by Mann-Whitney U test (Table 8 – Test number 5) and statistically significant differences were found between the views of the groups (U=6,909.000, p=0.000). Members of R&D group are more negative concerning the effects on the forest ecosystem in terms of ecology. Respondents have emphasized the impact on site nutrient budgets. A similar situation is described in international literature. According to Hacker (2005), a majority of research on removal of logging residues is focused on this problem. In this reason, ecological researches should be focused on nutrient budget, before removal of logging residues from the stand. Although some job opportunities might emerge by the use of logging residues for the purpose of bioenergy production (Dirkswager et al. 2011), it can also cause problems concerning employment and work power (Cantor and Rizy 1991). In Turkey, another problem related to this issue is the difficulty of finding the right number of skilful workers. The reason the forest villagers think that they will be paid less for these activities discourages their involvement. If these

161

H. Alkan et al.

Stakeholders’ Perspectives on Utilization of Logging Residues for Bioenergy in Turkey (153–165)

Table 9 SWOT matrix Strengths

Weaknesses

Þ Adequacy of forest assets and resources organized for the purpose of wood raw material in Turkey.

Þ Authorized forest institutions not making necessary efforts.

Þ Promotes energy security: Legitimizing the use of wood biomass in energy production by the law No. 5346 concerning »Use of Renewable Energy Resources in Energy Production«. Þ Level of knowledge and willingness of stakeholders related to the necessity of producing energy from logging residues. Þ Logging residues have a potential of forming a market supported by energy, forestry and other applications.

Þ Unavailability of policy or strategy by Forestry General Directorate about using logging residues in energy production. Þ Lack of R&D activities related to using the energy obtained from logging residues. Þ Lack of current activities concerning the establishment and development of wood based bioenergy sector in our country. Þ Lack of development of methods and techniques necessary for supply of logging residues (difficulty of supply and chopping systems for forest operations). Þ Lack of finance and technology in forest village cooperatives. Þ Lack of acknowledgement about the issue in related groups. Þ Presence of topographic problems due to distribution of forest resources in mountainous regions of our country.

Opportunities

Threats

Þ Proliferation of using biomass in bioenergy production in the world.

Þ Competes with conventional forest products industry: Increase in raw material requirement of chip board sector and raw material competition of forest products (for example flake board sector) with industrial ones.

Þ Having potential of supporting local energy production. Þ Having potential of supporting rural development. Þ Help to decrease the deposition of fuel material in fire risk zones and number of fires. Þ Supplying new vocational opportunities in terms of energy industry. Þ Supplying new vocational opportunities to private forestry companies and their technical staff together with optimization of using forest resources. Þ Supplying forest managements with new sources of income. Þ Removing the doubts of forest villagers about production phase, and based on their contentment, positive views that logging residues can be collected.

Þ Forest villagers using a great amount of logging residues as source of heating wood. Þ Difficulty of finding adequate number of qualified workers caused by reluctance of villagers to collect these resources. Þ Possible damages to forest ecology: Negative views about the collection of logging residues that can have negative effects on forest ecosystems in terms of ecology. Beside, using logging residues can have potential negative effects on wild life habitats. Þ Difficulty of sustainable supply of logging residues. Þ Reduction in prices of fossil based energy resources. Þ Energy imports from other countries.

doubts are not clarified, forest villagers will object logging residues to be used for the purpose of producing energy and a potential source of workers will be lost. It is clear that logging residues to be used for the purpose of producing energy brings extra loads for forest management in terms of planning, applying and controlling. Forest managements that are not able to carry out their own activities due to the lack of personnel may have additional difficulties. Views of the respondents support this issue. Furthermore, if the necessary precautions are taken, it can be said that this issue will not constitute a problem in terms of forming and developing the bioenergy sector. Taking the above issues into consideration, SWOT matrix is shown in Table 9. As seen in SWOT matrix, Turkey has the strength to use logging residues for the

162

purpose of bioenergy production. Weaknesses, on the other hand, can be removed if the necessary attention is paid to them by related institutions and organizations, and comprehensive research related to the subject is carried out.

5. Conclusions Based on the current researches and views of the stakeholders, it can be said that forest assets of Turkey and required forest resources for the production in terms of raw material supply are adequate. Turkey should use biomass for the purpose of bioenergy production and purchase these resources in places where there is no disagreement in terms of financial, ecological and social requirements. Croat. j. for. eng. 35(2014)2

Stakeholders’ Perspectives on Utilization of Logging Residues for Bioenergy in Turkey (153–165)

A sector to be formed by widespread production – consumption of bioenergy and use of logging residues for bioenergy can: contribute and support local energy production, provide new employment opportunities, have positive effects on rural development, provide work and challenging opportunities for forest engineers, provide an extra source of income for forestry managements, and if the right techniques are applied, it can also prevent forest fires and have positive effects on forest ecosystem and reduce environmental pollution. To do so, the competent institutions led by the General Directorate of Forestry should focus their attention on this issue and start developing strategies and policies concerning the evaluation of logging residues as soon as possible. According to the stakeholders, the production of bioenergy and use of logging residues for this purpose is a subject that should be taken into consideration in Turkey. The importance of carrying out theoretical and practical researches about the issue is getting more and more importance and, however, it becomes obvious that the necessary level of knowledge is not even present at universities and research institutions and that there is no terminological understanding about the issue. The current forestry and wood production system does not let logging residues be procured economically and this prevents the development of an ideal supply chain system in Turkey. In order to solve this problem as done in developed countries, harvesting methods and production technologies can be changed and mechanization level might be improved. In this respect, forest villagers who are one of the most important links of the supply chain will be supported technically and financially.

Acknowledgement We thank to TÜBİTAK (Scientific and Technological Research Council of Turkey) for the financial support through Project No: 110O435. This study was derived from the part of socio-economic subjects of the Project. We hereby present our gratitude to everyone who has contributed to our project.

H. Alkan et al.

Aguilar, F., Garrett, H. E., 2009: Perspectives of woody biomass for energy: Survey of state foresters, state energy biomass contacts, and National Council of Forestry executives. Journal of Forestrty 107: 297–306. Akyüz, K., 2010: Avrupa Birliği’nde yenilenebilir enerji politikaları ve orman biyokütlesi. Ankara Üniversitesi Avrupa Toplulukları Araştırma ve Uygulama Merkezi, 46. Dönem AB Temel Eğitim Kursu, p.46. Allen, J., Browne, M., Hunter, A., Boyd, J., Palmer, H., 1998: Logistics management and costs of biomass fuel supply. International Journal of Physical Distribution and Logistics Management 28: 1–10. Aruga, K., Mrakami, A., Nakahata, C., Yamaguchi, R., Yoshioka, T., 2011: Discussion on economic and energy balances of forest biomass utilization for small-scale power generation in kanuma, togichi prefecture, Japan. Croatian Journal of Forest Engineering 32: 571–586. Ateş, S., Akyıldız, H. M., Vurdu, H., Akgül, M., 2007: Türkiye’de orman kesim artıkları ve değerlendirilmesi. KÜ Orman Fakültesi Dergisi 7(1): 94–103. Cantor, R. A., Rizy, C. G., 1991: Biomass energy: Exploring the risks of commercialization in the United States of America. Bioresource Technology 35: 1–13. Demirbaş, A., 2001: Energy balance, energy sources, energy policy future developments and energy investments in Turkey. Energy Conversion and Management 42: 1239–1258. Derinalp, S., 2007: Dünya’da ve Türkiye’de lojistik hizmetlerde dış kaynak kullanımı: Türkiye’de faaliyet gösteren endüstriyel işletmelerde bir uygulama. Yüksek lisans tezi, Çukurova Üniversitesi, Sosyal Bilimler Enstitüsü, Adana, p 85. Dirkswager, A. L., Kilgore, M. A., Becker, D. R., Blinn, C., Ek, A., 2011: Logging business practices and perspectives on harvesting forest residues for energy: a Minnesota case study. Northern Journal of Applied Forestry 28(1): 41–46. Dwivedy, P., Alavalapati, J. R. R., 2009: Stakeholders’ perceptions on forest biomass-based bioenergy development in the southern US. Energy Policy 37: 1999–2007. Dyson, R. G., 2004: Strategic development and SWOT analysis at the University of Warwick. European Journal of Operational Research 152: 631–640. Eker, M., 2011: Assessment of procurement systems for unutilized logging residues for Brutian pine forest of Turkey. African Journal of Biotechnology 10(13): 2455–2468.

6. References

Eker, M., Acar, H., Ozcelik, R., Alkan, H., Gurlevik, N., Coban, H. O., Korkmaz, M., Yılmaztürk, A., 2013: Ormancılıkta hasat artıklarının tedarik edilebilirliğinin araştırılması. TUBITAK, Proje (110O435) Raporu, Turkey 2013.

Acar, H. H., Kalaycıoğlu, H., Topalak, O., 2001: Harvesting of Rhododendron Wood by Chippers. Third Balkan Scientific Conference, October 2–6, 2001, Sofia, Bulgaria, Proceedings, Volume IV: 248–256.

Eker, M., Coban, H. O., Acar, H. H., 2009: The Procurement Problem of Logging Residues in Turkish Forestry. Proceedings of FORMEC ´09, Prague, Czech Republic 2009, p. 115– 124.

Croat. j. for. eng. 35(2014)2

163

H. Alkan et al.

Stakeholders’ Perspectives on Utilization of Logging Residues for Bioenergy in Turkey (153–165)

Eriksson, O., Björheden, R., 1989: Optimal Storing, Transporting and processing for a forest fuel supplier. European Journal of Operational Research 43: 26–33. Ertürk, F., Akkoyunlu, A., Varınca, B. K., 2006: Enerji üretimi ve çevresel etkileri, Türkasya Stratejik Araştırmalar Merkezi, Stratejik Rapor No:14, 88 p. FAO, 2009: State of the World’s Forests-2009. FAO, İtaly, 152 p. Gan, J., Smith, C. T., 2006: Availability of logging residues and potential for electricity production and carbon displacement in the USA. Biomass and Bioenergy 30: 1011–1020. GDF, 2009: Yenilenebilir enerjide orman biyokütlesinin durumu. Biyoenerji çalışma grubu raporu, Orman Genel Müdürlüğü, Turkey, 134 p. GDF, 2012: Türkiye orman varlığı, GDF Publication. Yayın No. 85, Turkey, 23 p. Gökcöl, C., Dursun, B., Alboyaci, B., Sunan, E., 2009: Importance of biomass energy as alternative to other sources in Turkey. Energy Policy 37(2): 424–431. Guo, Z., Sun, C., Grebner, D. L., 2007: Utilization of forest derived biomass for energy production in the USA: Status, challenges, and public policies. International Forestry Review 9(3):748–758. Hacker, J. J., 2005: Effects of logging residue removal on forest sites a literature review. West Central Wisconsin Regional Planning Commission, USA, p. 19–25. Hakkila, P., 2006: Factors driving the development of forest energy in Finland. Biomass and Bioenergy 30: 281–288. Halder, P., Prokop, P., Chang, C. Y., Usak, M., Pietarinen, J., Nuutinen, S. H., Pelkonen, P., Çakır, M., 2012: International survey on bioenergy knowledge, perceptions, and attitudes among young citizens. Bioenergy Research 5: 247–261. Junginger, M., Faaija, A., Bjorhaden, R., Turkenburg W. C., 2005: Technological learning and cost reductions in wood fuel supply chains in Sweden. Biomass and Bioenergy 29: 399–418. Karha, K., Jylhä, P., Laitila, J., 2009: The Fixteri – A Novel Machine Concept for Integrated Pulpwood and Energy Wood Harvesting in Early Thinnings. Proceedings of FORMEC´09, Kostelec/Prague, Czech Republic, 2009, p. 220–228. Kalaycı, S., 2010: SPSS uygulamalı çok değişkenli istatistik teknikleri. Asil Yayın Dağıtım, Turkey, 426 p. Karasar, N., 2005: Bilimsel araştırma yöntemi. Nobel Yayın Dağıtım, Turkey, 292 p. Kühmaier, M., Stampfer, K., 2012: Development of a multicriteria decision support tool for energy wood supply management. Croatian Journal of Forest Engineering 33(2): 181–198. Ladanai, S., Vinterbäck, J. 2009: Global potential of sustainable biomass for energy report 013. Swedish University of

164

Agricultural Sciences, Department of Energy and Technology, Sweden, 30 p. Laitila, J., Kilponen, M., Nuutinen Y., 2013: Productivity and cost-efficiency of bundling logging residues at roadside landing. Croatian Journal of Forest Engineering 34: 175–187. Malinen, J., Pesonen, M., Maatta, T., Kajanus, M., 2001: Potential harvest for wood fuels (energy wood) from logging residues and first thinnings in Southern Finland. Biomass and Bioenergy 20: 189–196. Nachar, N., 2008: The Mann – Whitney U: A test for assessing whether two independent samples come from the same distribution. Tutorials in Quantitative Methods for Psychology 4: 13–20. Özdamar, K., 1999: Paket programlar ile istatistik veri analizi. Kaan Kitabevi, 1. Cilt (2. Baskı), 335 p. Öztürk, A., Saglam, B., Barlı, Ö., 2010: Attitudes and perceptions of rural people towards forest protection within the scope of participatory, forest management: a case study from Artvin, Turkey. African Journal of Agricultural Research 5(12): 1399–1411. Resch, G., Held, A., Faber, T., Panzer, C., Toro, F., 2008: Potentials and prospects for renewable energies at global scale, Energy Policy 36(11): 4048–4056. Röser, D., Asikainen, A., Rasmussen, K. R., Stupak, I., 2008: Sustainable use of forest biomass for energy: A synthesis with focus on the Baltic and Nordic Countries. Managing Forest Ecosystems, 12 p. Saraçoğlu, N., 2006a: Enerji Ormancılığının Kırsal Kalkınmaya Katkısı. Ormancılıkta Sosyo-Ekonomik Sorunlar Kongresi, Çankırı, Turkey, p. 7–12. Saraçoğlu, N., 2006b: Modern enerji ormancılığının Türkiye ormancılığı, kırsal kalkınma ve ülke ekonomisine katkısı. Orman ve Av 83(1): 33–38. Saraçoğlu, N., 2008: Modern Enerji Ormanciliğı–Ormanlardan Biyokütle Enerjisi Üretimi ve Çözümlemeler. Orman Genel Müdürlüğü, Turkey, 26 p. Saraçoğlu, N., 2010: Küresel İklim Değişikliği, Biyoenerji ve Enerji Ormancılığı. Efil Yayınevi, Turkey, 300 p. Spinelli, R., Hartsough, B., 2001: A survey of Italian chipping operations. Biomass and Bioenergy 21:433–444. Sterba, H., 2003: Growth After Biomass Removal During Precommercial Thinning. In: Limbeck-Lilineau, B., Steinmüller, Th., Stampfer, K. (Hrsg.), Austro2003: High Tech Forest Operations for Mountainous Terrain, Institute of Forest Engineering, p. 1–9. Stidham, M., Simon-Brown, V., 2011: Stakeholder perspectives on converting forest biomass to energy in Oregon, USA. Biomass and Bioenergy 35: 203–213. Tekin, M., Zerenler, M., 2005: Konya otomotiv yan sanayinin rekabet gücü hakkında bir araştırma. V. Ulusal Üretim Araştırmaları Sempozyumu, İstanbul, Turkey, p. 75–81. Croat. j. for. eng. 35(2014)2

Stakeholders’ Perspectives on Utilization of Logging Residues for Bioenergy in Turkey (153–165) Yılmaz, I., Ilbas, M., Su, S., 2003: Türkiye Rüzgar Enerjisi Potansiyelinin Değerlendirilmesi. In: Yeni ve Yenilenebilir Enerji Kaynakları Sempozyumu, Kayseri, Turkey, p. 399– 401.

H. Alkan et al.

Yoshioka, T., Aruga, K., Sakai, H., Kobayashi, H., Nitami, T., 2002: Cost, energy and carbon dioxide effectiveness of a harvesting and transporting system for residual forest biomass, Journal of Forestry Research 7: 157–163.

Authors’ address:

Received: December 2, 2013 Accepted: February 28, 2014 Croat. j. for. eng. 35(2014)2

Assoc. Prof. Hasan Alkan, PhD.* e-mail: hasanalkan@sdu.edu.tr; hasanalkan07@gmail.com Assoc. Prof. Mehmet Korkmaz, PhD. e-mail: mehmetkorkmaz@sdu.edu.tr Assoc. Prof. Mehmet Eker, PhD. E-mail: mehmeteker@sdu.edu.tr Suleyman Demirel University Faculty of Forestry Department of Forest Engineering 32260 Isparta TURKEY * Corresponding author

165

Original scientific paper

Planning Best Management Practices to Reduce Sediment Delivery from Forest Roads Using WEPP:Road Erosion Modeling and Simulated Annealing Optimization James A. Efta, Woodam Chung Abstract Planning and implementation of road BMPs on a watershed scale can be a difficult task because of the need to prioritize locations while accounting for multiple constraints, such as the available budget, continuous maintenance, and equipment scheduling. Using simulated annealing (SA) as its heuristic optimizer, BMP-SA accounts for sediment being delivered to the stream network through incorporation of modeled road erosion predictions and alternative BMP options and scheduling for problematic road segments. BMP-SA was applied to the Glenbrook Creek watershed in the Lake Tahoe Basin in Nevada, US. WEPP:Road predictions were used to identify road segments posing an erosion risk and appropriate BMPs were identified for problematic segments. Using BMP-SA, modeled road-related sediment leaving the forest buffer, thus entering streams, was minimized over the course of the planning horizon while considering budget constraints and equipment scheduling concerns. BMP-SA can be applied to any watershed but relies heavily on the perceived accuracy of road erosion predictions. Keywords: forest roads, BMPs, simulated annealing, WEPP:Road, erosion modeling, road management, budget planning

1. Introduction To minimize sediment-related impacts of forest roads, Best Management Practices (BMPs) are frequently implemented on forest road networks. While BMPs may consist of a planning practice or mitigation strategy (e.g. maintaining a set buffer distance from a stream channel), the term is also broadly used in reference to specific structures or road network attributes that address sedimentation issues. Examples include, but are not limited to: sediment traps; drain dips; vegetated or rock-lined ditches, and/or road surfacing. Field research supports the effectiveness of specific BMPs and the physics behind them (e.g., Clinton and Vose 2003, Foltz and Truebe 2003, Luce and Black 2001, Megahan and Ketcheson 1996). In practice, physical BMPs are generally prescribed using expert judgment in the field, inevitably under limited budget conditions. Regardless of whether eroCroat. j. for. eng. 35(2014)2

sion risk is evaluated, not all potential BMP options may be explored at a given site, for reasons ranging from inexperience of the prescriber to budget limitations. While a BMP or set of BMPs may be ideal for a given site, selection of BMPs at the one- to few-segment scale may not effectively minimize erosion and sedimentation at the watershed scale. Optimization strategies have been employed since the 1970s to address multiple management goals and environmental constraints in forest planning (Rรถnnqvist 2003, Weintraub et al. 1995, Weintraub 2006). Application of heuristic optimization specifically to environmental concerns, including sedimentation associated with roads, has only occurred more recently due to the complexity of such spatially-explicit planning problems (Coulter et al. 2006, Contreras and Chung 2009, Weintraub et al. 2000). Multiple projects have incorporated BMPs and/or associated erosion

167

J. A. Efta and W. Chung Planning Best Management Practices to Reduce Sediment Delivery from Forest Roads ... (167–178)

potential into cost-benefit analyses using heuristics for road management planning (e.g., Aruga et al. 2005, Madej et al. 2006, Rackley and Chung 2008). Of these projects, however, none has directly addressed BMP implementation and maintenance on an existing road network. The research presented here fills this need with a decision support tool designed to assist managers in formulating watershed-scale road BMP installation and maintenance plans. The decision support toolcalled BMP-SA uses simulated annealing (SA) as its heuristic solver to minimize sediment contribution to downstream water bodies by prioritizing road BMP installations while accounting for budget constraints, maintenance requirements, and equipment scheduling concerns.

2. Study area The study area for this project is located in the Lake Tahoe Basin. Losing its famed clarity at a rate of approximately one-quarter meter per year for the past 25 years, the lake is currently designated as an impaired water body under Section 303(d) of the Clean Water Act (Roberts and Reuter 2007). Sediment from the basin road network has been identified as a negative contributor to the lake water clarity (Murphy and Knopp 2000).

The Glenbrook Creek watershed encompassed the majority of the study area (Fig. 1). Glenbrook Creek lies approximately 24 km west of Carson City, NV and 32 km north of South Lake Tahoe, CA on the east side of the Lake Tahoe Basin. Elevations in the Glenbrook Creek watershed range from approximately 1900 m to 2,700 m. Soils are volcanic and granitic in origin (Grismer and Hogan 2004). Average annual precipitation at the Marlette Lake SNOTEL weather station site, which lies at 2,400 m 5.6 km north of the Glenbrook watershed boundary, is approximately 84 cm. Most precipitation in the Glenbrook watershed falls as snow (Rowe et al. 2002). Monthly average maximum temperature at the Marlette Lake SNOTEL site between 1989 and 2008 was 11° C and monthly average minimum temperature was –1° C.

3. Materials and methods 3.1 Problem formulation Simulated annealing, developed by Kirkpatrick and others (1983), uses a modified Monte Carlo simulation that is analogous to a metal cooling, or annealing, after leaving a forge. Initial temperature and cooling rate are variables which control the number of iterations and range of acceptable solution values. This optimization technique is well suited to this problem type for multiple reasons: 1) It can readily be scaled to large and

Fig. 1 Map of Glenbrook Creek watershed vicinity, Nevada USA

168

Croat. j. for. eng. 35(2014)2

Planning Best Management Practices to Reduce Sediment Delivery from Forest Roads ... (167–178) J. A. Efta and W. Chung

Fig. 2 Changes in objective function value of the current solution during the simulated annealing optimization process small datasets; and 2) it is a relatively simple yet efficient heuristic solution technique for combinatorial optimization problems (Kirkpatric et al. 1983, Tarp and Helles 1997). During comparison of neighbourhood solutions, if the current solution is superior to the alternative, an acceptance probability is calculated in order to determine whether the alternative should be accepted despite its inferiority. This unique heuristic component is linked to the temperature variable:

sediment is discounted over time in order to promote early BMP installation and subsequent sediment reduction. A discount rate of four percent was used as it is standard practice in natural resource economic analysis involving U.S. Forest Service investments (Row et al. 1981). Minimize H

p(new) = e

temp

(1)

Where p(new) is the probability of accepting the new solution, current is the objective function value of the current solution, new is the objective function value of the new solution, and temp is the value of the temperature variable at the time of comparison. When compared against a randomly generated value, the solution may be accepted as the »new« current solution, from which a subsequent neighborhood solution will be formulated. In doing so, a near-optimal solution may be reached faster than if solutions were formulated randomly, since the »worse« solution may provide a bridge to a superior solution more quickly. As temperature decreases (more iterations are run), so too will the probability of acceptance of an inferior solution, thereby reducing solution variability (Fig. 2). To apply this heuristic framework to the issue at hand, the planning problem can be formulated as a minimization problem with the amount of sediment entering streams as the objective function (Equations 2 and 3, Fig. 3). The formulation below assumes that Croat. j. for. eng. 35(2014)2

Z = ∑∑

current − new

j =1 i =1

Subject to N

∑ cost Where Z

i =1

i,j

sedimenti,j 1.04( j−1)

≤ budget j

j ∈H

(2)

(3)

Total sediment leaving the buffer through the course of the planning horizon Planning period j Total number of planning periods H Segment number i Total number of segments on the road netN work sedimenti,j Sediment entering the nearest waterway from segment i during planning period j costi,j Cost of BMP treatment or maintenance scheduled for segment i in planning period j. budgetj Budget for planning period j 1.04(j – 1) Discount term

169

J. A. Efta and W. Chung Planning Best Management Practices to Reduce Sediment Delivery from Forest Roads ... (167â&#x20AC;&#x201C;178)

Fig. 3 Flowchart describing adapted simulated annealing optimization

170

Croat. j. for. eng. 35(2014)2

Planning Best Management Practices to Reduce Sediment Delivery from Forest Roads ... (167–178) J. A. Efta and W. Chung

Estimates of sediment entering the stream network from each road segment of interest must be established prior to execution of the heuristic. Currently, there are a suite of empirically-based and processbased erosion models being used for estimating road erosion, each with varying degrees of user accessibility (Fu et al. 2010). Among these, WEPP:Road has gained traction among US land management agencies including the Lake Tahoe Basin Management Unit (LTBMU) and continues to gain widespread use among academic and nonprofit organizations (e.g.

Table 1 WEPP:Road input parameters and possible values or parameter ranges WEPP:Road input parameter

Possible values/ allowable range

Climate

N/A Silt loam

Soil type

Sandy loam Clay loam Loam Insloped, bare ditch

Road design

Insloped, vegetated or rocked ditch Outsloped, unrutted Outsloped, rutted Native

Surface type

Graveled Paved None

Traffic level

Low High

Road width

1 ft – 300 ft

Road length

1 ft – 999 ft

Road gradient

3% – 99%

Fill slope length

1 ft – 999 ft

Fill slope gradient

3% – 99%

Buffer length

1 ft – 999 ft

Buffer gradient

3% – 99%

Coarse rock content

0% – 100%

Years of simulation time

Croat. j. for. eng. 35(2014)2

1 yr – 200 yrs

Briebart et al. 2007, Contreras et al. 2008, Inlander et al. 2007). WEPP:Road, a user-friendly interface to the Water Erosion Prediction Project (WEPP) Model (Flanagan and Nearing 1995), provides a web browserbased interface that requires few input parameters. In this study, we used WEPP:Road to assess potential erosion risk and threat of sedimentation from each road segment in a watershed as well as appropriate BMPs to address the erosion risk factor(s) for that segment. WEPP:Road estimates runoff and soil loss on three overland flow elements: the roadbed itself, a fill slope, and the buffer (hill slope area between the base of the fill slope and the nearest water source) (Elliot et al. 1999). Four soil types can be modeled by WEPP:Road, along with four road designs, three road surface types, and three traffic levels (Table 1). A variety of other parameters are also required by the model, some of which are best gathered in the field and some of which are best collected using a GIS or other data sources.

3.2 WEPP:Road data preparation A total of 173 road segments were identified on 12.5 km of road. Segments were delineated between two existing drainage structures, from a slope break or high point to a drainage structure, from a high point to a low point, or between a drainage structure and a low point. WEPP:Road input parameters were determined or measured through a combination of field data collection and geoprocessing using datasets acquired from the Lake Tahoe GIS Data Clearinghouse (http://tahoe.usgs.gov/). »From« nodes comprised the entrance or beginning segment locations for runoff and sediment entrainment. »To« nodes were delivery points, or the perceived segment outlet for runoff and sediment. Analysis of coarse rock content and soil texturing were performed on soil adjacent to the road grade itself. As WEPP:Road only accepts one of four soil textures (Table 1), soil textures evaluated in the field were matched as closely as possible to one of those four textures available in the standard WEPP:Road interface. The Tahoe CA SNOTEL site, the closest available long-term climate station was used as the climate input for WEPP:Road. While the model is running, WEPP:Road uses the CLIGEN weather generator to stochastically generate daily climate data for the desired simulation time (Elliot et al. 1999). Thirty years of daily climate data were generated for these simulations. Per WEPP:Road Documentation, thirty years of simulation is generally adequate for obtaining reasonable erosion estimates (Elliot et al. 1999). Road traffic level was held constant at »low« for all segments (Briebart et al. 2007).

171

J. A. Efta and W. Chung Planning Best Management Practices to Reduce Sediment Delivery from Forest Roads ... (167–178)

Table 2 Priority of BMP assignment for a given road segment Condition

Priority 1

Buffer slope > fill slope Road slope > 17%

Priority 2

Priority 3 †

Outslope*

Drain dips

Pave‡

Drain dips†

Outslope*

Notes: *Delivery point reassigned to center of road segment. Not applicable on paved or graveled segments. †Drain dips applicable on any segment greater than 46 meters (150 feet) in length. Segment length iteratively divided in half until segment length is less than 46 meters or sediment leaving buffer is zero. Not applicable on outsloped segments. ‡ If paving is already installed on a segment, no further BMPs can be installed

Table 3 Installation costs, maintenance costs, and maintenance frequencies associated with assigned BMPs Installation cost

Equipment move-in costs

Maintenance cost

Equipment move-in costs for maintenance

Maintenance frequency

yrs

Outsloping

1,865/km

1,000

622/km

500

Drain dips

100/each

500

100/each

500

Pavement

15,2269/km

1,500

9,323/km

500

BMP

Each road segment was modeled multiple times using WEPP:Road, first to simulate erosion under existing conditions then under the range of possible BMPs that were then incorporated into BMP-SA model inputs.

3.3 BMP-SA model input formulation During a field visit with a Lake Tahoe Basin road engineer, site-specific BMP options were prescribed for those segments predicted to yield the most sediment. From this visit, along with personal communication with other engineers, guidelines were established for installing BMPs on those sites not visited in the field (Table 2) (Catherine Schoen and Paul Potts, pers. comm., USFS Lake Tahoe Basin Management Unit, July 2008). BMP installation costs, maintenance costs, and maintenance frequencies associated with a given BMP were also obtained through personal communication (Paul Potts, pers. comm., USFS Lake Tahoe Basin Management Unit, July 2008) along with the Region Four Cost Estimating Guide for Road Construction (Table 3; USDA FS 2009). For each potential BMP scenario on a given segment, WEPP:Road was used to predict the effectiveness of each potential BMP installation. Up to four BMP options were assigned to each segment, including no treatment. To account for equipment scheduling costs (in doing so favoring solutions where BMPs are installed in close proximity in the same time period), a clustering subroutine was developed as a part of the model.

172

For testing purposes, cluster diameter was fixed at 305 meters. If the same type of treatments were scheduled on road segments that are located within the cluster diameter, equipment move-in cost was counted only once for those treatments. Planning horizon was assumed to be 20 years with a planning period of one year. Three budget expenditure scenarios for BMP implementation and maintenance were modeled: a given annual budget is used for 1) new BMP installation only, 2) new BMP installation and maintenance, and 3) existing BMP maintenance along with new BMP installation and maintenance. In all modeling scenarios, BMPs were assumed to be maintained in perpetuity at their assigned frequencies (Table 3). When included, existing BMPs were assumed to start their maintenance cycle in period one. Each scenario was modeled at multiple initial budgets to assess model behavior under different budget levels. At each level, it was assumed that the annual budget was constant throughout the planning horizon and unspent budget was not carried over into future years.

4. Results and Discussion 4.1 WEPP:Road results Of the 173 segments analyzed in the study area, 74 of them (accounting for 6.7 km, or 53 percent, of roads in the study area) were predicted to produce sediment leaving the buffer over the 30-year modeling period. Croat. j. for. eng. 35(2014)2

Planning Best Management Practices to Reduce Sediment Delivery from Forest Roads ... (167–178) J. A. Efta and W. Chung

Table 4 Predicted sediment leaving road and sediment leaving buffer in Mg yr–1 and Mg ha–1 yr–1 across the study area, Glenbrook Creek, NV. Average road width across each area of interest was used to calculate Mg ha–1 yr–1 values Study area

Sediment leaving buffer

Sediment leaving road –1

–1

Mg ha–1 yr–1

Mg yr

Mg ha yr

Mg yr

Entire study area

49.9

13.6

2.7

0.7

Glenbrook watershed

19.0

7.0

1.4

0.5

Per-segment sediment outputs ranged from 0 Mg yr–1 to 0.6 Mg yr–1 with a mean of less than 0.1 Mg yr–1. WEPP:Road predicted a total of 49.9 Mg yr–1 of sediment leaving the road and 2.7 Mg yr–1 sediment leaving the buffer from the study area (Table 4). The well-maintained existing BMP infrastructure on Glenbrook Creek forest roads, as well as the watershed dry climate, partially explains the minimal amount of sediment predicted to be leaving the buffer. Predicted average erosion rate from native surface roads in sandy loam soils - the predominant soil texture found within the Glenbrook Creek watershed was 8.1 Mg ha-1 yr-1. In comparison to regional empirical values, on the west slope of the Sierra Nevada range Coe (2006) observed an erosion rate of 8.1 Mg ha-1 yr-1 on native surface roads during one wet season (October through June) of data collection. Annual precipitation during this wet season was near the longterm average of 1,300 mm. Coe’s study segments were in primarily loam soils. Average road gradients, segment lengths, and parent materials were comparable for both studies.

4.2 Alternative BMP assignment Of the 74 road segments producing greater than zero sediment leaving the buffer per year, 38 were assigned treatments. Thirty-six segments could not be assigned BMPs because they were either paved (assumed to be an »end point« BMP) or had some combination of conditions which prevented assignment of BMP treatments. For example, outsloping was not considered an appropriate BMP for graveled segments and drain dips were not applied to segments already outsloped. The effectiveness of BMPs, defined in this study as predicted reduction in sediment delivery, varied depending on the characteristics of the road segment to which a BMP was assigned. Drain dips showed an exponential increase in effectiveness as segments were divided into two (one drain dip), four (three drain dips), and eight (seven drain dips), respectively (Table 5). Outsloping was modeled as being three times more Croat. j. for. eng. 35(2014)2

effective at reducing sediment than one drain dip, but an order of magnitude less effective than three drain dips and two orders of magnitude less effective than seven drain dips. There were no instances where paving was chosen as an applicable new BMP. In every instance where pavement was a potential BMP option, WEPP:Road model outputs indicated that pavement increased sediment leaving the buffer above existing levels potentially due to increased runoff exacerbating ditch and buffer erosion (Table 5). Other researchers have had similar results when applying WEPP:Road to paved road segments in the Lake Tahoe Basin (Briebart et al. 2007).

4.3 New BMP installation only In this modeling scenario, it was assumed that the annual budget can be used for new BMP installation only. The results show that sediment leaving the buffer produced a negative exponential trend with increasing annual budget (Fig. 4). The minimum sediment delivery solution was achieved when the annual budget reached $20,000. In this solution, all 38 segments had BMPs applied to them in period one (Fig. 5). Discounted sediment was reduced from 37.8 Mg when no treatment was applied over the 20 year planning horizon to 10.4 Mg. Increasing budget beyond this level yielded no reduction in sediment leaving the buffer. Proportion of segments with outsloping, chosen as an appropriate BMP, also increased with the budget. In several instances, solutions with two different budgets yielded decreases in sediment leaving the buffer while having the same number or fewer BMPs installed in period one. In all of these instances, the number of segments, where outsloping was installed as a BMP, was greater in the solution producing less sediment leaving the buffer. Outsloping is a highly effective BMP for reducing sediment leaving the road and buffer but also tends to be more expensive than single drain dips (Luce and Black 2001, Elliot et al. 2009, USDA FS 2009). In addition, there are limitations for where and when outsloping may be an applicable BMP. Fig. 6 shows the number of segments where BMPs were installed in each period for all three mod-

173

J. A. Efta and W. Chung Planning Best Management Practices to Reduce Sediment Delivery from Forest Roads ... (167â&#x20AC;&#x201C;178)

Fig. 4 Sediment leaving buffer through the 20 year planning horizon at various budgets under three modeling scenarios

Fig. 5 Number and type of BMP installed in period one at varying initial budget per period resulting from the new BMP installation only scenario eling scenarios. At $3,000, all BMPs were installed in the first seven periods under the new BMP installation only scenario.

4.4 New BMP installation and maintenance With maintenance costs incorporated into the model, a greater initial budget was required to achieve the same reduction in sediment as that found under the

174

new BMP installation only scenario. At a budget level of $6,000, the model failed to produce feasible solutions because no possible combination of BMPs existed below this budget threshold. Minimum sediment leaving the buffer through the course of the planning horizon was 10.4 Mg, the same solution as that found in the previously modeled scenario. At $6,000 annual budget, two segments had no Croat. j. for. eng. 35(2014)2

Planning Best Management Practices to Reduce Sediment Delivery from Forest Roads ... (167â&#x20AC;&#x201C;178) J. A. Efta and W. Chung

Fig. 6 With varying initial budget, number of periods required by BMP-SA to install all new BMPs. Black bars represent the new BMP installation only scenario, white bars represent the new BMP installation and maintenance scenario, and gray bars represent the existing BMP maintenance, new BMP installation, and new BMP maintenance scenario Croat. j. for. eng. 35(2014)2

175

J. A. Efta and W. Chung Planning Best Management Practices to Reduce Sediment Delivery from Forest Roads ... (167–178)

Table 5 Predicted effectiveness of BMPs (in terms of sediment savings) assigned to problematic road segments. Average road width across all segments with the same BMP assignment was used to calculate erosion rates BMP

Number assigned

Minimum effectiveness

Maximum effectiveness

Mean effectiveness

Mg ha–1 yr–1

One drain dip

0.00

0.11

0.02

Three drain dips

0.02

1.51

0.53

Seven drain dips

0.77

2.56

1.65

Outslope segment

0.00

0.59

0.06

Pave segment

-1.37

-0.01

-0.52

treatment chosen as the best possible option. This result indicates that the budget was so limited that neither BMP installation nor maintenance was feasible for these two segments. The number and types of BMPs installed in period one at varying budgets for the new BMP installation and maintenance scenario was very similar to that seen with the new BMP installation only scenario.

4.5 Existing BMP maintenance and new BMP installation and maintenance The lowest sediment delivery solution was achieved with a higher annual budget than the previous two modeled scenarios (Fig. 4). Maintenance of existing BMPs required approximately $35,000 minimum annual budget. Period 13 required the greatest annual budget as a result of numerous preexisting BMPs having maintenance frequencies of two, three, or four years. As a result, any new BMPs with a threeyear maintenance frequency (such as outsloping) could not be installed in period one until the budget was increased beyond this minimum level. A budget of $57,000 was necessary for all BMPs to be installed in period one. As a result of accounting for maintenance costs, the solution became more constrained, making this BMP installation scenario less variable than the previous scenario. In general, the number of periods required to install BMPs on all segments decreased with the increase in the budget (Fig. 6). 4.6 Discussion of BMP-SA modeling results Optimized solutions by BMP-SA for different budgets show that the model was able to produce costefficient BMP locations, types and implementation periods in reducing sediment delivery under limited budgets. High cost efficiency of BMPs was realized at low budget levels, but increases in the budget yielded diminishing returns in sediment reduction (Fig. 4). When solutions are constrained (as in these types of scenarios), BMP-SA has the potential to provide the

176

most benefit. A BMP may be chosen that is not necessarily the best for a given location due to the budget constraint but serves to maximize sediment savings across the area of interest. This highlights the importance of the clustering subroutine in BMP-SA. While a sensitivity analysis of cluster size is beyond the scope of this study, it provides one future avenue of research using this tool. Discount rate plays a role in dictating the budget at which BMPs can be most cost-effectively implemented (indicated by an inflection point in budgetsediment reduction plots for the three model scenarios in Fig. 4). If discount rate were increased, an inflection point would also be reached more quickly but would require a greater initial budget. Conversely, a lower discount rate would result in a less rapid reduction in sediment leaving the buffer. The maximum possible sediment reduction was achieved with all three modeling scenarios, albeit at higher budget levels when maintenance of new and existing BMPs was accounted for within the given budget. Under the best possible solution, sediment leaving the buffer was reduced by 72% if compared to buffer sediment outputs should no treatments be installed through the course of the planning horizon. With respect to the 38 treated segments, sediment was reduced by nearly 93% if compared to buffer sediment outputs with no treatment. Estimated sediment savings with new BMP installation was considerable, in part because little sediment was predicted to leave the buffer from the Glenbrook Creek watershed. If predicted sediment leaving the buffer was greater, percent decrease in sediment leaving the buffer as a result of BMP installation could be lower. In a watershed without a well-developed BMP infrastructure, however, there would be more potential for a tool like BMP-SA that can assist field managers in prescribing effective BMPs to minimize sediment leaving the buffer, especially when the budget is constrained. Croat. j. for. eng. 35(2014)2

Planning Best Management Practices to Reduce Sediment Delivery from Forest Roads ... (167–178) J. A. Efta and W. Chung

5. Conclusions

6. References

This research developed a decision support tool designed to increase the efficiency of BMP planning on a forest road network. Modeled road-related sediment leaving the forest buffer was minimized over the course of a planning horizon while accounting for budget constraints as well as equipment scheduling considerations. The solutions presented here used modeled WEPP:Road erosion estimates as well as guidelines for prioritizing appropriate BMPs for a given road segment. To minimize sediment leaving the buffer, these data were input into a model using an adapted simulated annealing optimization algorithm. Under limited budgets, the model was able to prioritize BMP placements and types through a tradeoff analysis between costs and effectiveness of BMPs. While the data used here is from the Lake Tahoe Basin, BMP-SA can be applied to any watershed. The model is also applicable at a scale greater than a single watershed and can be easily modified to accommodate non-linear spatial constraints, such as scheduling of equipment and BMP maintenance, though problem complexity may substantially increase if more BMP locations and options exist. There are two critical assumptions implicit in BMPSA. One is that BMPs must be maintained at appropriate intervals in perpetuity, otherwise money spent installing BMPs is not worthwhile. In addition, this modeling process relies on the accuracy of road sedimentation prediction for determining problematic road segments and the effects of BMP installation on sediment savings. BMP implementation is often sitespecific in nature. For that reason, some road segments may not be able to be treated using one of only a handful of generic BMPs; only professional judgment in the field may provide the ideal option in such situations.

Aruga, K., Sessions, J., Miyata, E. S., 2005: Forest road design with soil sediment evaluation using a high-resolution DEM. Journal of Forest Research 10: 471–479.

Acknowledgements We would like to thank Bill Elliot, Randy Foltz, and Jun Rhee at the Rocky Mountain Research Station in Moscow, ID, for their cooperation and insight through the course of this collaboration. Paul Potts, Catherine Schoen, and Jim Harris of the USFS Lake Tahoe Basin Management Unit also have our thanks for their time on the phone, in meetings, and in person through the duration of this project. This study was funded by the Tahoe Science Program and administered by the USDA Forest Service Pacific Southwest Research Station in cooperation with the Rocky Mountain Research Station (Project No. P010) along with significant cost match and in-kind contribution from The University of Montana. Croat. j. for. eng. 35(2014)2

Briebart, A., Harris, J., Norman, S., 2007: Forest Road BMP Upgrade Monitoring Report 2003-2005.USDA Forest Service, Lake Tahoe Basin Management Unit. Available online at http://www.fs.usda.gov/Internet/FSE_DOCUMENTS/ fsm9_045815.pdf; last accessed Aug. 8, 2012. Clinton, B. D., Vose, J. M., 2004: Differences in Surface Water Quality Draining Four Road Surface Types in the Southern Appalachians. Southern Journal of Applied Forestry 27: 100– 106. Coe, D., 2006: Sediment production and delivery from forest roads in the Sierra Nevada, California. M.Sc. thesis, Colorado State University, Fort Collins, CO., 110 p. Coulter, E. D., Sessions, J., Wing, M. G., 2006: Scheduling forest road maintenance using the analytic hierarchy process and heuristics. Silva Fennica 40(1): 143–160. Contreras, M. A., Chung, W., Jones, G., 2008: Applying ant colony optimization metaheuristic to solve forest transportation planning problems with side constraints. Canadian Journal of Forest Research 38: 2896–2910. Contreras, M. A., Chung, W., 2009: Designing skid-trail networks to minimize skidding cost and soil disturbances. In Proc. of the 32nd Annual Meeting of the Council on Forest Engineering, June 15–18, 2009, Kings Beach, CA. Elliot, W. J., Hall, D. E., Scheele, D. L., 1999: WEPP:Road: WEPP interface for predicting forest road runoff, erosion and sediment delivery. Available online at http://forest.moscowfsl. wsu.edu/fswepp/docs/wepproaddoc.html; last accessed Jun. 2, 2012. Elliot, W. E., Foltz, R. B., Robichaud, P. R., 2009: Recent findings related to measuring and modeling forest road erosion. In Proc. of the 18th World IMACS / MODSIM Congress on International Congress on Modelling and Simulation, Anderssen, R.S., R.D. Braddock, L.T.H. Newham (eds.). Cairns, Australia, July 13–17, 2009. Flanagan, D. C., Nearing, M. A. (eds.), 1995: USDA-water erosion prediction project: hillslope profile and watershed model documentation. NSERL Report #10, USDA-ARS National Soil Erosion Research Laboratory, West Lafayette, Indiana, 298 p. Foltz, R. B., Truebe, M. A., 2003: Locally Available Aggregate and Sediment Production. Transportation Research Record 1819, Paper No. LVR8-1050. Fu, B., Newham, L. T. H., Ramos-Scharron, C., 2010: A review of surface erosion and sediment delivery models for unsealed roads. Environmental Modelling and Software 25: 1–14. Grismer, M. E., Hogan, M. P., 2004: Simulated Rainfall Evaluation of Revegetation/Mulch Erosion Control in the Lake Tahoe Basin – 1: Method Assessment. Land Degradation and Development 15: 573–588. Inlander, E., Clingenpeel, A., Crump, M. A., Van Epps, M., Formica, S., 2007: Inventory and Sediment Modeling of Un-

177

J. A. Efta and W. Chung Planning Best Management Practices to Reduce Sediment Delivery from Forest Roads ... (167–178) paved Roads for Stream Conservation Planning. P. 156–165 in Proc. of the 2007 International Conference on Ecology and Transportation, Leroy Irwin, C., D. Nelson, and K. P. McDermott. (eds.). Center for Transportation and the Environment, North Carolina State University, Raleigh, NC. Kirkpatrick, S., Gelatt, C. D., Vecchi, M. P., 1983: Optimization by Simulated Annealing. Science 220: 671–680. Luce, C. H., Black, T. A., 2001: Spatial and Temporal Patterns in Erosion from Forest Roads. P. 165–178 in Influence of Urban and Forest Land Uses on the Hydrologic-Geomorphic Responses of Watersheds, Wigmosta, M. S. and S. J. Burges. (eds.). Water Resources Monographs, American Geophysical Union, Washington, D.C. Madej, M., Eschenbach, E. A., Diaz, C., Teasley, R., Baker, K., 2006: Optimization strategies for sediment reduction practices on roads in steep forested terrain. Earth Surface Processes and Landforms 31: 1643–1656. Megahan, W. F., Ketcheson, G. L., 1996: Predicting Downslope Travel of Granitic Sediments from Forest Roads in Idaho. Water Resources Bulletin 32: 371–382. Murphy, D. D., Knopp, C. M., 2000: Lake Tahoe Watershed Assessment: Volume I. USDA For. Serv. Gen. Tech. Rep. PSWGTR-175, 753 p. Rackley, J., Chung, W., 2008: Incorporating forest road erosion into forest resource transportation planning: a case study in the Mica Creek watershed in Northern Idaho. Transactions of the ASABE 51: 115–127. Roberts, D. M., Reuter, J. E., 2007: Draft Lake Tahoe Total Maximum Daily Load Technical Report California and Nevada. Available online at http://terc.ucdavis.edu/publications/

LakeTahoeTMDLTechnicalReport.pdf; last accessed Aug. 8, 2012. Rönnqvist, M., 2003: Optimization in Forestry. Mathematical Programming 97: 267–284. Row, C., Kaiser, H. F., Sessions, J., 1981: Discount Rate for Long-Term Forest Service Investments. Journal of Forestry 79: 367–369. Rowe, T. G., Saleh, D. K., Watkins, S. A., Kratzer, C. R., 2002: Streamflow and Water-Quality Data for Selected Watersheds in the Lake Tahoe Basin, California and Nevada, through September 1998. U.S. Geological Survey Water-Resources Investigations Report 02-4030, 118 p. Tarp, P., Helles, F., 1997: Spatial Optimization by Simulated Annealing and Linear Programming. Scandinavian Journal of Forest Research 12: 390–402. USDA Forest Service, 2009: Cost Estimating Guide for Road Construction. Available online at http://www.fs.usda.gov/ Internet/FSE_DOCUMENTS/fsbdev3_015406.pdf; last accessed July 5, 2012. Weintraub, A., 2006: Integer programming in forestry. Annals of Operations Research 149: 209–216. Weintraub, A. P., Church, R. L., Murray, A. T., Guignard, M., 2000: Forest management models and combinatorial algorithms: analysis of state of the art. Annals of Operations Research 96: 271–285. Weintraub, A., Jones, G., Meacham, M., Magendzo, A., Malchauk, D., 1995: Heuristic procedures for solving mixed-integer harvest scheduling-transportation planning models. Canadian Journal of Forest Research 25: 1618–1626.

Authors’ addresses: James A. Efta, Forest Hydrologist* e-mail: jefta@fs.fed.us USDA Forest Service- CusterNational Forest. 1310 Main Street Billings MT 59102 USA * Formerly Graduate Research Assistant The University of Montana 32 Campus Drive Missoula MT 59812 USA

Received: March 4, 2014 Accepted: July 11, 2014

178

Assoc. Prof. Woodam Chung, PhD. e-mail: woodam.chung@oregonstate.edu Department of Forest Engineering Resources and Management Oregon State University 267 Peavy Hall Corvallis, OR97331-5703 USA * Corresponding author Croat. j. for. eng. 35(2014)2

Original scientific paper

A Survey Analysis of Forest Harvesting and Transportation Operations in Michigan Dalia Abbas, Robert Handler, Bruce Hartsough, Dennis Dykstra, Pasi Lautala, Larry Hembroff Abstract This paper assesses the technology involved in commercial forest harvesting and delivery operations. It investigates existing forest-based production capacity and its potential to supply the startup of large scale forest-based industries. A survey of harvesting and transportation workforce and technology was mailed to 1,130 logging firms operating in Michigan and four Wisconsin counties that adjoin Michiganâ&#x20AC;&#x2122;s Upper Peninsula. The response rate received was 28%. The paper details and analyses the different operational matters, conditions, equipment and transportation use reported by logging firms. The study provides technical forest products operations information and methods for assessing the capacity of logging firms and markets looking to expand their businesses. Keywords: loggers, operations, products, equipment, transportation

1. Introduction Employment in the logging industry in the United States has been heavily impacted in recent years by the economic downturn, which has greatly reduced the demand for wood-frame housing (Drapala 2009), and through more systematic declines in the pulp and paper and furniture industries (Grushecky et al. 2006, JylhĂ¤ et al. 2010). One response has been to advocate for the growth of an industry to produce fuel and energy from forest woody biomass material (ESIA 2007, IRGC 2008), which would further increase efforts to promote the development of alternatives to fossil fuel as outlined in the 2007 US Energy Security and Independence Act (ESIA 2007). However, without understanding the supply logistics and operations that influence utilization of wood products from commercial harvesting operations, the economic potential for forestry-dependent industries cannot be accurately assessed. Commercial forest operations supply chains, which provide raw materials for forest industries, include harvesting, forwarding and transportation operations. The supply chain of forest pulpwood and small diameter trees is similar to that of larger trees but potentially includes specialized equipment and techniques to handle and produce value added prodCroat. j. for. eng. 35(2014)2

ucts such as woodchips. For example, supplying woodchips from harvesting residues and small diameter trees requires a chipper or grinder unit in addition to the standard forest harvesting equipment. As available information of logging firmsâ&#x20AC;&#x2122; technological capabilities was limited prior to this study, the objective was to investigate existing forest-based production capacity to supply the startup of large scale forest-based industries in Michigan. To examine this capacity, especially in relation to growing interest in cellulosic ethanol production, a survey instrument was developed that inquired about work force characteristics and conditions, logging equipment and productivity, production rates per harvest conditions and prescriptions, and transportation equipment use. The survey instrument was used in the study to identify potential for the growth of a cellulosic ethanol facility in the Upper Peninsula of Michigan. Results were provided to assist in understanding the overall logging capacity surrounding the proposed facility. The information related to the Upper Peninsula portion solely was reported to the facility in question, and were proprietary. Since the unique requirements of the proposed facility are expected to be requirements for other potential large scale wood products industries,

179

D. Abbas et al.

A Survey Analysis of Forest Harvesting and Transportation Operations in Michigan (179–192)

the survey instrument was expanded to cover a statewide database of loggers, for the potential growth of similar, or other wood products industries, in the state. Results in this study are based on survey responses and analysis from a subset of the logging industries. They do not intend to present a full description of logging firms in Michigan. However, they contribute to the knowledge of the operating conditions and equipment productivity of the state.

2. Background In 2009, researchers from Michigan State University and Michigan Technological University began studying the operations of the state’s forest feedstock supply of pulpwood and small diameter trees; focusing on harvesting, forwarding, on-site processing and transportation operations from natural forest stands to processing facilities. This paper examines the results of a questionnaire carried out to assess the workforce and equipment capacities available for supplying pulpwood, small diameter trees and woodchips from natural forests in the state. Research on logging equipment productivity has ranged from detailed studies using time and motion analysis (Hartsough et al. 2001, Spinelli and Hartsough 2001, Harris 2003, Hunsberger et al. 2003), to assessments of the cost of the total supply chain and shift level analysis (Han et al. 2004, Spinelli and Visser 2008, Abbas et al. 2011A), to broader-scale analyses involving survey focusing on logging production, business management, and ownership of logging firms (McNeel and Dudd 1996, Luppold et al. 1998, Egan 2001, Rickenbach et al. 2005, Drolet and LeBel 2010, Egan 2011, G.C. and Potter-Witter 2011). The study reported in this paper takes a different approach from previous studies in that it uses a survey instrument to integrate information about harvesting technology, the logging workforce, and the operating environment faced by logging firms in Michigan.

3. Materials and methods Dillman’s »Total Design Method« (Dillman 2000) was implemented. This methodology, as opposed to face-to-face interviews or direct observations, was identified as the most effective method to meet research objectives. The survey instrument allowed the research group to reach the largest number of logging firms possible in the state within a limited time frame, preserved anonymity, facilitated data analysis and captured the opinions of different logging firms interested in the survey questions regardless of their strat-

180

ification into different sized or targeted groups. The survey questionnaire was sent to 1,130 logging firms operating in Michigan and four bordering Wisconsin counties (Marinette, Florence, Vilas and Iron). The addressee list was obtained from a dataset from the Department of Forestry at Michigan State University that integrated a database from the Michigan Department of Natural Resources and state forest timber sale bidders (G.C. and Potter-Witter 2011). Development of the survey involved identifying the information needed, writing questions that would contribute to the project objectives and pilot-testing a series of drafts in consultation with logging machine operators, forestry and forest engineering experts. Pilot testing was carried out with several local logging firms. The final product was a 14-page survey questionnaire booklet that was mailed to logging firms in Michigan. The survey was conducted in two parts. The first, in 2009–2010, covered listings of loggers in the Eastern Upper and Northern Lower Peninsula of Michigan. The second, in 2010–2011, covered listings of loggers in the remaining parts of the state and the four neighboring counties in Wisconsin. The funding for the entire project was received over two years from two different sources. The first part was surveyed for a private industry that was interested in investigating supply potentials in the northeastern parts of Michigan. The process was then replicated to cover the entire state once funding from the Department of Energy was received. Each addressee was given a unique ID code. Most of the received questionnaire responses were mailed in, but there was an option provided for a web-based response instead. Respondents were permitted to access the online questionnaire by using their assigned ID code. The Michigan State University Office for Survey Research (an independent third party) mailed the questionnaire and received completed questionnaires. A $20.00 incentive check was mailed for each returned and completed survey. The first-phase mailing was sent to 612 logging firms and 112 responded. The second phase mailing was sent to 518 logging firms operating in counties that were not covered in the first phase and 110 responded. The survey was relatively complex, required detailed technical and comprehensive information, and was considerably long (14 pages), which required extensive input from logging firms. Using the American Association for Public Opinion Research response rate calculation methods (AAPOR 2010), the total response rate was calculated at approximately 28%. The response rate was found to be consistent with similar Croat. j. for. eng. 35(2014)2

A Survey Analysis of Forest Harvesting and Transportation Operations in Michigan (179–192)

loggers’ surveys publications that targeted data collection from logging firms (Luppold et al. 1998, Greene et al. 2001, Milauskas and Wang 2006). The response rate was calculated as the number of respondents who returned completed questionnaires divided by the number of eligible prospective respondents in the sample. However, some listings in the full listing of the initial sample were determined to be no longer in business and therefore were not eligible. Mailings to some other listings were returned as not deliverable or not valid. These were also considered ineligible. The number of ineligible listings was subtracted from the number of listings in the initial sample in computing the percent that responded. Additionally, the AAPOR Standard Definitions Response Rate 4 formula makes one other adjustment to the number of eligible prospective respondents in the sample. There were three broad categories of outcomes based on the results of the mailings. Listings that are clearly determined to be eligible listings, listings that are clearly determined to be ineligible listings, and listings where there is no response that indicates whether the listing is actually eligible or not, i.e., the status is ambiguous. Response Rate 4 presumes that a fraction of the listings that are ambiguous are probably ineligible. The proportion of logging firm listings that are eligible divided by the sum of those that are eligible and those that are not eligible is used to multiply the number of ambiguous listings. The sum of this adjusted number of ambiguous listings plus the number of eligible listings is used as the denominator to calculate the percentage of eligible listings that responded, as indicated in equation (1). Using this method allowed us to account for illegible, ineligible and ambiguous groups. It was not possible to compute nonresponse bias because of the lack in information of the non-respondents, and therefore, we opted to use eligibility standards instead. RR4 =

(Number of Completed Surveys) ((Eligible + ((Eligible/(Eligible + noneligible) ´ Ambiguous))

D. Abbas et al.

Þ Postcard reminder/thank you note, containing the URL to the online survey site; sent to the entire mailing list two weeks after initial mailing; Þ Reminder sent to non-respondents about two weeks after the previous reminder, with a replacement questionnaire and cover letter including the URL to the online survey site, and a postage-paid return envelope; Þ Mailing of incentive checks to all respondents. Responses from each participant were cataloged and analyzed using SPSS/PASW Statistics 18.0 and then transferred to Microsoft® Excel® 2010. Survey questions that involved units of measure permitted responses in English units (short tons, gallons, miles, acres) or common forest industry units (cords) in order to obtain accurate responses from loggers. Units were later converted into metric units for this publication; 1 short ton=0.9072 metric tonnes (t) and 1 cord=2.09 metric tonnes. Statistical analysis used in the publication permitted a description of real data. The database was inclusive of the logging firms of the entire state and the four Wisconsin neighboring counties. We could not control the response rate, but targeted the entire logging population. Statistical analysis of survey results described the count of respondents and used mean, mode, median, minimum, maximum, standard deviation and percentages of operations functions. Data analysis linked productivity data per equipment types to develop entire supply chain productivity estimates. The questionnaire requested information on the following: Þ Workforce characteristics and conditions; Þ Logging production capacity; Þ Equipment used for all supply chain activities; Þ Production rates for several different harvesting systems, configurations, conditions and prescriptions; Þ Mean hauling distance and preference for various modes of transportation.

(1)

Responses were received from loggers in 40 out of 83 counties. Nine respondents completed the online version of the survey. Most of the responses were received from the Upper Peninsula of Michigan, where the highest concentration of timber resources in the state exists. Contact with each survey participant involved up to five contact attempts, all by mail, as follows: Þ Preliminary notice by mail to notify respondents about the survey and its objectives; Þ The actual questionnaire booklet with a cover letter and a postage-paid return envelope; Croat. j. for. eng. 35(2014)2

4. Results and discussion 4.1 Operational matters Respondents were asked to relate their current actual production to their firm’s total operating capacity given the current crew and equipment. On average, respondents reported operating at 73% of their total capacity. For the most part, 87% of the respondents were owners and operators of logging firms. Only 8% of the respondents reported working as logging operators who do not own the logging firm, and 5% reported they were owners who do not operate their own equipment.

181

D. Abbas et al.

A Survey Analysis of Forest Harvesting and Transportation Operations in Michigan (179–192)

The survey asked questions about current and under normal conditions number of employees. It was difficult to compare the 2010 and 2011 surveys in that respect, because the year in question was different and the economic situation was dynamically changing due to the decline in the housing and the pulp and paper industries. To account for this bias, we aggregated value pertaining to number of employees per firm under normal conditions. On average, the respondents’ firms have been in business for 28 years and under normal conditions averaged 6.5 employees. This response is similar to a previous study carried out in 2008 within Michigan that identified firms having been in business for 29 years with an average of 7 employees per firm (G.C. and Potter-Witter 2011). An earlier survey study, however, carried out in 2003 in Wisconsin and the Upper Peninsula of Michigan reported there were only 4.8 employees per firm on average, including 0.7 part-time employees (Rickenbach et al. 2005). Stumpage price is a very significant part of the total harvesting and supply cost incurred. Stumpage price contributed about 11% of the harvest and delivery supply cost of pulpwood in Michigan (Abbas et al. 2013). Operations that required stumpage purchase averaged 70% of the operations of respondents, with a standard deviation of 39%, based on 180 responses.

4.2 Operational conditions 4.2.1 Landownership Michigan’s forests cover over 19 million acres. More than 12 million acres are privately owned. The State of Michigan owns 4 million acres (Pedersen 2005). Since land ownership and decisions can impact the amount of forest resources available for removal, the survey inquired about the ownership patterns. Almost 60% of harvest volumes came from private nonindustrial lands, based on the responses for that parTable 1 Percentage of wood volumes harvested from each property type Wood property type sources

Percent

Non-industrial private lands

59.2%

Forest industry or real estate timber

10.0%

State forest lands

22.8%

National forest lands

4.4%

Other public lands

2.5%

Tribal lands

0.1%

Unsure of ownership

1.1%

182

ticular question (Table 1). These results were consistent with an earlier survey that reported 64% of harvested volumes came from non-industrial private forests (G.C. and Potter-Witter 2011). 4.2.2 Terrain The type of terrain has a major influence on the cost and productivity of harvesting operations. More difficult terrain yields higher operating costs because of the site conditions, time spent in extraction and maneuverability of equipment. The survey results suggested that most operations in Michigan are performed on flat ground (34% of reported operations) and rolling hills (32% of reported operations), with smaller fractions of operations run in lowland terrain (wetter grounds – 24% of reported operations) and steep hilly areas (10% of reported operations). This is not unusual, since harvesting operations need to be performed under suitable soil conditions. For example, Minnesota timber harvesting guidelines recommend entering lowlands and wet soils only under frozen or dry site conditions to avoid displacing the soil (Abbas et al. 2011b). The fraction of firms that indicated a percentage of harvesting that takes place under wetter conditions, needs to be investigated further, since such operations could displace soil properties, as the cited guidelines suggest. Results, on the other hand, concerning where most of the operations were performed, could aid in determining locations for the startup of new facilities, since the concentration of logging operations would unlikely be in these lowland or steep terrains. 4.2.3 Winter and summer operations Mean shift hours were collected for both summer and winter operations. Results, at a 99% confidence level, did not differ significantly for the two seasons. This finding was unexpected since winter day times are shorter. Respondents indicated that their operations were limited to a single daily shift with mean summer hours of 37.6 per week with a standard deviation of 19.1 from 193 respondents, and mean winter hours of 37.4 per week with a standard deviation of 18.2 from 192 respondents. 4.2.4 Product types and the size of harvested stands The survey queried loggers about the types of products they processed. Results were structured into the percentage of their operations that involved sawlogs, pulpwood and woodchips products. While 97% of 206 respondents reported working with sawlogs, and 95% of these respondents reported working with pulpwood products, only 15% of 132 respondents indicated working with woodchips for at least part of their operations. Croat. j. for. eng. 35(2014)2

A Survey Analysis of Forest Harvesting and Transportation Operations in Michigan (179â&#x20AC;&#x201C;192)

Survey responses of the average harvest stand size were analyzed and the mean, median, and mode were reported. Also, responses for the two different questions reported the minimum and maximum stand size harvests were analyzed. The responses averaged 19 ha for the average stand harvested, and 7 and 64 ha for the minimum and maximum stands harvested, respectively (Table 2). The reported average harvested area was unsurprising considering: the cost of moving equipment to the site, potentially building roads and a landing/processing area that is accessible regularly in the harvest area, and because data were collected for different logging firms with different operational business sizes. Table 2 Maximum, minimum and mean size of harvested stands in 2009 and 2010, hectares Minimum stand size harvested, ha

Maximum stand size harvested, ha

Mean stand size harvested, ha

Mean

Standard deviation

Median

100

Mode

Minimum

Maximum

275

101

Number of respondents

177

179

185

A broader question inquired about the smallest operations in terms of volume and area on which operators would be willing to bid. The smallest volume 143 respondents were willing to bid on averaged 494 t (~237 cords), with a standard deviation of 435 t (~208 cords). On the other hand, the smallest area 192 respondents were willing to bid on averaged 9 ha. 4.2.5 Harvest types The largest percentage of harvesting prescriptions (45.7%) involved removing 30% of the harvestable volume from site. The next larger percentage (27.4%) involved removing 70% of the harvestable volume, followed closely by clearcutting (26.9%). This type of information is critical when considering the supply radius for a sustainable quantity of forest products for a potential new facility, since assuming clearcut sizes of removals would not be a practical option. To better understand the extent to which operations involved residue removal, survey questions inquired about percentages of clearcut and partial reCroat. j. for. eng. 35(2014)2

D. Abbas et al.

moval operations that removed residue. Most operations (>78%) did not involve residue removal. Responses indicated that over half of the logging operations involved partial cut harvesting with residues left on site (Table 3). Table 3 Residue management options, percent Residue management options

Percent

Clearcut and leave residue

27.8%

Clearcut and remove residue

9.9%

Partial removal and leave residue

50.9%

Partial removal and remove residue

9.7%

Other method

1.7%

4.2.6 Skidding and forwarding distance Since skidding and forwarding contribute significantly to the supply operations because of the fuel use and the labor involved, the survey enquired about the distance travelled using this equipment type. In the analysis, the mean of the skidding/forwarding distances responses were filtered to a maximum of 3.2 km (2 miles). Three eliminated responses from 186 responses reported an unrealistic average extraction distance of 5, 6 and 8 km. whereas, the maximum reported skidding/forwarding distances was 8.0 km (5 miles). Three eliminated responses from 179 responses reported an unrealistic maximum extraction distance of 145, 241 and 515 km. The survey question did not make a distinction between the extraction equipment used, be it a forwarder or a skidder. We could assume that the larger distances are more likely travelled by forwarders, the most popular skidding/forwarding system in the state and the unit that permits longer travel distance because of its truck-like features. The reported skidding/forwarding distance averaged 520 meters (0.25 miles) with a standard deviation of 870 meters (0.5 miles) based on 177 responses. Whereas the reported maximum forwarding/skidding distance averaged 1.16 km (0.72 miles) with a standard deviation of 1.3 km (0.8 miles) based on 176 responses. 4.2.7 Logging equipment 4.2.7.1 Equipment types Respondents were asked to provide information about the type and number of harvesting equipment they owned, along with key descriptive information of this equipment. Reporting these data required a detailed technical understanding of equipment used by operators. A large variety of equipment was re-

183

D. Abbas et al.

A Survey Analysis of Forest Harvesting and Transportation Operations in Michigan (179–192)

Table 4 Technical analysis of harvesting equipment Number of units

Model year

Total machine hours

Fuel use (l hr–1)

Cut-to-length

191

2003 ± 4.1* (132)

9,286 ± 6,543 (151)

18.55 ± 8.71 (142)

Feller buncher harvester

1996 ± 8.3 (35)

9,384 ± 6,696 (38)

23.85 ± 9.84 (37)

Feller delimber

1988 (1)

12,467 ± 6,788 (3)

10.22 ± 2.27 (3)

Forwarder

247

1997 ± 9.5 (153)

10,666 ± 6,138 (165)

12.11 ± 7.19 (159)

Harwarder

2001 ± 5 (5)

9,053 ± 7,586 (6)

8.33 ± 1.89 (5)

Chainsaw

569

2006 ± 4.9 (113)

668 ± 990 (36)

4.16 ± 2.27 (35)

Grapple skidder

1995 ± 8.1 (47)

11,583 ± 6,116 (31)

19.31 ± 8.71 (33)

Cable skidder

1976 ± 8.8 (17)

8,889 ± 3,772 (9)

9.08 ± 3.79 (11)

Loader

1996 ± 6.7 (30)

7,525 ± 7,429 (24)

14.38 ± 7.19 (26)

Grinder

2003 ± 1.0 (5)

2,459 ± 772 (6)

30.28 ± 3.41 (4)

Slasher

1995 ± 7.6 (14)

9,607 ± 7,140 (15)

14.76 ± 6.81 (18)

Delimber

1996 ± 8 (3)

6,220 ± 3,561 (5)

11.36 ± 0 (3)

Debarker

1997 ± 2 (2)

7,333 ± 1,155 (3)

50.35 ± 10.98 (3)

Chippers

1997 ± 9.1 (18)

8,798 ± 8,584 (17)

54.89 ± 33.69 (18)

Bulldozers

132

1992 ± 14.0 (72)

4,866 ± 3,226 (87)

14.38 ± 7.95 (79)

Equipment type

* For each item, the number following ± is the sample standard deviation and the value in parentheses is the number of survey responses from which the mean and standard deviation were calculated

ported in use among the different respondents (Table 4). Cut-to-length equipment was the predominant type of harvesting equipment throughout the state, with forwarders the most common off-road transport equipment. Cut-to-length equipment was reported to be the most popular style of mechanized harvesting equipment, outnumbering feller bunchers by nearly 3:1. Similarly, forwarders were the most common extraction equipment used, followed by skidders. Chainsaws represented the most ubiquitous type of equipment reported by respondents; used not only for felling trees but also for removing tops and branches, releasing vegetation and removing brush when it obstructs heavy machinery. 4.2.7.2 Equipment use In one of the longer questions, survey respondents were asked to provide mechanical details about the equipment used for their operations (Table 5). Respondents reported the total number of equipment, equipment model, type and year, total hours on equipment, hours on equipment and hours operated in the survey years (2009/2010), fuel use, equipment head type, and whether the equipment used tires or tracks, from a list of available harvesting equipment. Most reported equipment types were depreciated beyond the 5 years

184

life time expectancy of new equipment. The age of the full range of equipment averaged 15 years old. Cut-tolength equipment was among the youngest equipment age group and averaged 10 years old. Cable skidders were among the oldest equipment used, averaging 35 years old. On average the equipment with highest machine hours were the feller-delimbers and grapple skidders. The maximum of the reported hours per year per equipment use averaged 1,161 hrs. Based on a full time utilization rate of equipment of 2,000 hrs. yr-1, on average the highest utilization rate was reported by the cut-to-length equipment users, that runs at 58% and the harwarder (combined harvester/forwarder equipment) users’ utilization rate that runs at 45%. Reported utilization rates in this study could aid in the analysis of depreciated equipment and associated hours of operation to estimate the cost of harvesting operations. As production was reported on an hourly basis from a regular workday, these values were assumed to be scheduled machine hours used to determine harvest cost (Abbas et al. 2013). However, it is important to note that these utilization rates are reported within the limited number of responses, and under market conditions that have not fared well due to the economic downturn in the forest products industries (Table 5). Croat. j. for. eng. 35(2014)2

A Survey Analysis of Forest Harvesting and Transportation Operations in Michigan (179–192)

Table 5 Equipment mean age and utilization rates per year Equipment type

Mean age of equipment

Mean h yr–1

Utilization rates, 2000 h yr–1*

Cut-to-length

1,161

58%

Feller buncher harvester

626

31%

Feller-delimber

542

27%

Forwarder

762

38%

Harwarder

905

45%

Chainsaw

134

Grapple skidder

724

36%

Cable skidder

254

13%

Loader

502

25%

Grinder

307

15%

Slasher

600

30%

Delimber

415

21%

Debarker

524

26%

Chippers

628

31%

Bulldozers

256

13%

* This is based on 8 hours per day for 5 days. There are 52 weeks in a year with a two week vacation. Therefore: 8 h X 5 days X 50 weeks = 2,000 hours per year

4.2.7.3 Repair and maintenance An analysis was conducted to identify the type of equipment that required the most repairs and the repair time of equipment per day, with particular attention to equipment that requires the most repair time. The time spent repairing and maintaining equipment reduces operating, production and efficiency time for both the operators and the equipment. The questions for this section were phrased as follows: 1) in an average work day, approximately how many hours do you allocate for repairs and maintenance?; and 2) which of your equipment types requires the most repair time per day? On average, operators reported repairing and maintaining equipment on site for about 1.3 hours per day. Out of 171 responses, 32% reported that the cutto-length equipment required the most repair, 22% reported that the feller buncher equipment required the most repair, but only 14% reported that chainsaws required the most repair. These cutting systems that required the most repair time, were followed by the Croat. j. for. eng. 35(2014)2

D. Abbas et al.

skidding/forwarding systems that required the second most repair time; 11% of the respondents reported that the forwarder required the most repair, followed by the skidder (9%). Although the cut-to-length– forwarder configuration was reported by the larger number of respondents as the system that required the most repairs, it was also cited as the most used system. Therefore, having most repairs associated with this system does not necessarily reflect on the functionality of this equipment type, rather that it is more widely used and as a result the system more likely to be identified within the responses. 4.2.8 Production rates of equipment configurations Respondents were asked which harvesting equipment configuration they utilized for different harvesting scenarios and in different forest types. Responses were reported in cords or short tons of wood per hour and converted to metric units in this publication. Equipment configurations were grouped into three main categories: A – Cut-to-length and forwarder; B – Whole tree feller buncher with skidder and slasher; C – Chainsaws and skidder. To minimize error from including respondents that owned many pieces of equipment but did not use them equally or at all times, the analysis focused on respondents who owned only one or two pieces of cutting equipment (cut-to-length, and whole tree feller bunchers). Productivity was normalized to a single unit of harvesting equipment (i.e., production rates reported by respondents with two sets of harvesting equipment were divided by 2). Responses that indicated chainsaw use were also included. This analysis procedure, however, was not followed for chainsawbased harvesting, as multiple chainsaws are typically used by logging crews relying on this equipment configuration. Operators reporting the chainsaws-skidder configuration for >50% of their operations reported using 2.6±2.0 chainsaws. As a result, reported productivity for chainsaws is based on the assumed production of 2.6±2.0 chainsaws plus one skidder. For cut-tolength equipment and forwarders, the mean reported productivity increased as harvesting intensity increased from 30% selective cut up to clearcutting, as would be expected. In almost every case, productivity in each harvesting scenario was highest in softwood plantations, which are typically on even terrain, easier on equipment and stocked with optimal timber for harvesting. Tables 6, 7 and 8 summarize the reported roundwood harvesting productivity of different equipment configurations.

185

D. Abbas et al.

A Survey Analysis of Forest Harvesting and Transportation Operations in Michigan (179–192)

Table 6 Cut-to-length and forwarder productivity Treatment

30% cut (selective)

70% cut (shelterwood)

Clearcutting

Productivity per harvester, t system h-1

Forest type

Mean

Std. dev.

Natural hardwoods

6.97

2.88

Mixed hardwood/softwood

7.99

3.09

Natural softwoods

8.24

4.51

Softwood plantations

9.54

4.40

Natural hardwoods

8.53

3.76

Mixed hardwood/softwood

9.41

3.78

Natural softwoods

9.72

4.49

Softwood plantations

10.37

4.44

Natural hardwoods

11.50

5.72

Mixed hardwood/softwood

11.83

5.22

Natural softwoods

12.67

5.82

Softwood plantations

14.54

8.39

* N is the number of harvesting equipment units included in the analysis

Table 7 Feller buncher harvester, skidder and slasher productivity Treatment

30% cut (selective)

70% cut (shelterwood)

Clearcutting

Productivity per harvester, t system h-1

Forest type

Mean

Std. dev.

Natural hardwoods

7.76

3.17

Mixed hardwood/softwood

7.64

2.73

Natural softwoods

7.03

2.75

Softwood plantations

8.37

1.94

Natural hardwoods

9.89

2.98

Mixed hardwood/softwood

9.66

2.96

Natural softwoods

10.47

3.34

Softwood plantations

11.25

3.61

Natural hardwoods

14.23

5.59

Mixed hardwood/softwood

13.75

6.22

Natural softwoods

13.40

5.90

Softwood plantations

14.81

8.74

* N is the number of harvesting equipment units included in the analysis

Feller buncher-skidder-slasher system operators reported highest productivity per t hr–1 system under clearcut conditions. The maximum productivity of the whole tree system came close to 14 t hr–1. Cut-to-

186

length-forward system operations reported the maximum productivity of about 13 t hr–1. An analysis of the lowest production tonnes per hour was reported by chainsaw users (3 t hr–1, under clearcut conditions). Croat. j. for. eng. 35(2014)2

A Survey Analysis of Forest Harvesting and Transportation Operations in Michigan (179–192)

D. Abbas et al.

Table 8 Chainsaw and skidder productivity Treatment

30% cut (Selective)

70% cut (shelterwood)

Clearcutting

Productivity**, t system h-1

Forest type

Mean

Std. dev.

Natural hardwoods

4.21

2.78

Mixed hardwood/softwood

4.07

3.05

Natural softwoods

3.84

3.26

Softwood plantations

3.67

1.79

Natural hardwoods

4.59

3.36

Mixed hardwood/softwood

4.05

2.92

Natural softwoods

3.92

3.09

Softwood plantations

3.63

2.21

Natural hardwoods

4.17

2.34

Mixed hardwood/softwood

3.99

1.92

Natural softwoods

2.96

1.25

Softwood plantations

3.71

2.30

* N is the number of harvesting equipment units included in the analysis ** Operators reporting this configuration for >50% of operations reported using 2.6±2.0 chainsaws in this equipment configuration

The reported productivity within 30%, 70% and clearcut treatments using all three harvesting systems averaged 8 t hr–1. 4.2.9 Transportation 4.2.9.1 Trucks The survey inquired about trucks used by respondents to transport pulpwood, small diameter trees and woodchips. Truck data were analyzed for the larger 10–11 axle log trucks, smaller 2–9 axle log trucks, and chip vans. Large log trucks were found to be younger in age, with lower average fuel use and larger annual usage than other trucks represented in the survey, but the distribution of annual mileage data for trucks varied considerably in all truck classes. Table 9 summarizes the main characteristics of trucks owned by survey respondents. On average, 86% of roundwood was reported to be transported by self-loading trucks. Based on responses, most log trucks in the state of MI are equipped with self-loaders. Over 70% of respondents indicated that 100% of their roundwood production was transported with self-loading trucks. Forest products are delivered to a variety of endusers and intermediate supply chain points. Respondents were asked to report percentages of products that were delivered to different facilities. Pulpwood Croat. j. for. eng. 35(2014)2

Table 9 Trucking equipment descriptive summary Fuel use, km l-1

Annual use, km yr-1

All trucks reported: 2000±7* (156)

1.9±0.76 (148)

88,417±96,348 (150)

Large log trucks (10–11 axles): 2003±6 (76)

1.56±0.37 (71)

105,132±63,209 (66)

Small log trucks (2–9 axles): 1997±8 (84)

2.24±0.83 (74)

75,500±117,928 (76)

Chip Vans: 1998±7 (15)

1.78±0.42 (21)

68,880±45,636 (20)

Year

* Numbers following ± represent standard deviations based on indicated number of respondents inside parentheses

and hardwood sawmills received the largest percentage of products totaling 59% of the products generated from logging firms (Table 10). Results were consistent with results from 2008 survey work that reported 58% of production was delivered to pulp and paper mills and hardwood sawmills (G.C. and PotterWitter 2011). Survey respondents were asked about the percentage of their annual production delivered to different destinations (Table 10).

187

D. Abbas et al.

A Survey Analysis of Forest Harvesting and Transportation Operations in Michigan (179â&#x20AC;&#x201C;192)

Table 10 Percentage of production delivered to different facilities Recipient forest products industries

Percent

study that has shown that 90% of sawlogs were delivered to mills within approx. 145 km (90 miles) from logging sites (G.C. and Potter-Witter 2011). Wood chips were the product transported the longest distance, with over 27% of production traveled more than approx. 145 km by truck.

Hardwood sawmill

23.9%

Softwood sawmill

12.5%

Veneer mill

4.1%

4.2.9.2 Rail

Pulp mill

36.1%

Other panel mill

4.0%

Oriented strand board mill

6.0%

Wood pellet fuel mill

1.1%

Wood power generator

1.6%

Truck/rail landing

3.3%

Other location

7.5%

Rail transport was utilized unevenly throughout the state of MI for the transportation of forest products. In 2010 survey, respondents were asked about their most recent use of rail transportation, and over 60% of the respondents who answered the question indicated that they had never used rail transport (Table 11). Respondents were asked about the percentage of their annual production that was moved using rail transport. The 25 respondents that indicated some portion of their production had been moved by rail (12.7% of responses) moved approximately 22.1%Âą19.2% of their annual production by rail. It should be noted that all of the respondents indicating a use of rail transport were based in the Upper Peninsula of Michigan.

Pulp mills were reported to be the most popular destination for forest products in the state of Michigan. Transport distances for each of the three main forest products (sawlogs, pulpwood, chips) followed a similar pattern (Fig. 1). Respondents were asked to list what percentage of their annual production of sawlogs, pulpwood, and chips was transported by truck for several mileage categories. Over 55% of the respondents transported saw logs within 60 miles (approx. 97 km). Whereas only 45% of the responses for pulpwood and chips reported they transported less than 60 miles. Results coincide with results from an earlier

Table 11 Most recent use of rail transportation Responses

Percent of respondents

(1) In past 6 months

14.1%

(2) In past year

3.1%

(3) In past 3 years

4.7%

(4) In past 5 years

6.3%

(5) In past 10 years

7.8%

(6) In past 15 years

3.1%

(7) Not at all

60.9%

(8 ) No response

Time frame

* Data originated from 2010 respondents to logger survey, 2nd phase

Fig. 1 Truck transport distances for main forest products

188

When asked about the factors that limited their use of rail transport for moving forest products, survey respondents indicated that reliability of service and limited access in main work areas were the primary reasons that rail transport was not used more extensively (Table 12). Existing transport contracts and lack of knowledge about rail were the factors that limited use of rail transport the least among survey respondents (Table 12), indicating that loggers might have been familiar with the operations of the rail industry within the state, and they were not meeting the needs of the loggers. Croat. j. for. eng. 35(2014)2

A Survey Analysis of Forest Harvesting and Transportation Operations in Michigan (179–192)

Table 12 Reasons that limit use of rail transport by Michigan loggers Reason

Mean Response* 1= not limiting, 5 = extremely limiting

Lack of knowledge rail contractual arrangements

2.48±1.58

Reliability of service

3.53±1.47

Speed of delivery

3.39±1.45

Limited rail access in main work areas

3.49±1.64

Prices not competitive with other modes

3.03±1.51

Minimum shipment too large

2.49±1.69

Existing contract with other provider

2.12±1.57

* No. of responses ranged from 68 to 78 loggers for each reason listed

5. Conclusion The study helped provide detailed information of the state of forest products harvesting and transportation industries in the State of Michigan and four adjoining Wisconsin counties. A survey instrument was used to compile operational factors. Results, based on responses, helped identify interconnectedness between key operational matters such as work conditions, product types, and equipment and transportation logistics. The response rate was 28% from the respondents who received the survey from the database of loggers. Based on the extensive details required in the survey and other logging surveys sited in the study, this response rate is not uncommon. With this in mind, results do not provide a 100% analysis of the logging firms of the state. Conclusions were drawn from the analysis of responses to help explain operating conditions of a subset of loggers who responded to the survey. Even though the paper started with pointing to the decline in the harvesting industry because of recent economic downturn impacts, the results of the study intend to help attract further industries to replace this short coming. Based on survey results the problems for logging firms are larger than merely replacing, or introducing, forest products markets. Many issues emerged in this study that shed light on the importance of paying attention to the logging workforce and their operating conditions. For example, the prices of timber are widely accessible. The productivity of equipment from the survey results offer a chance to forest equipment operators to look into what they Croat. j. for. eng. 35(2014)2

D. Abbas et al.

might be able to produce, as well as, what their peers’ capacity looks like. Detailing operational logistics and the workforce capacity could help industries make a more informed decision when starting a forest products industry in the state and elsewhere. Different studies have explored the rate of employees per firm in Michigan. The rate is slightly lower in this study than previously reported. However, because of the recent shutdown of pulp and paper industries, the difficult and expensive work conditions, it is assumed that unless markets are in place, equipment operators would further be leaving the industry. Most logging firms identified in this study were found to be located in the upper peninsula of Michigan and were run by owners who operate their own equipment; at a reported operating capacity averaging 73%. This result is lower than the percentage reported in an earlier study that reported that logging firms worked at 82% of their full capacity (G.C. and Potter-Witter 2011). Interest in working within a fuller capacity needs to account for the full logistics conditions involved in harvest operations. Further, most of the harvest volume came from partial cut treatments, especially within 30% selected cut treatment types. Industries may not assume a natural area would be clearcut to fit their wood supply objectives. The equipment of highest use in Michigan was the cut-to-length equipment, but the largest productivities were reported by the whole tree feller buncher under clearcut conditions to be up to 14 t hr–1. A significant finding in the study was that most operations did not involve residue removal or chipping of material. In fact, grinders and chippers were used by only a few number respondents, since only 8 grinders and 14 chippers were reported. It is unclear from the responses why chipping and grinding equipment were very few in the responses. Hypothetically, this could be due to multiple factors, including: the large number of cut-to-length systems, or that most harvests fell within 30% cut treatments or that the residue material left behind from these operations did not justify investing in a chipping system. Further explanation could also be attributed to the lack of sufficient wood chips industries and markets that account for the cost of purchasing chipping equipment. Based on the survey responses, non-industrial private land owners were found to be the source for most of the products extracted within Michigan. Most products were transported within 97–145 km. Most of reported products were delivered to pulpwood mills, and most respondents who responded to the use of rail in transportation reported they never used rail, citing the reliability of service and accessibility as the

189

D. Abbas et al.

A Survey Analysis of Forest Harvesting and Transportation Operations in Michigan (179–192)

reasons most limiting to the use of rail transportation. The survey inquired about the terrain types linked to harvesting operations. Most operations were reported to be within flat and rolling ground operations, which is a key factor when determining where to deploy operations when identifying a location for a forest products facility, since starting a facility on the foothills may not be the ideal location, based on productivity details.

prove the harvest and delivery options in the state. Productivity from this study helped develop a cost of the supply chain logistics of forest material (Abbas et al. 2013). Attention to improving the efficiencies and conditions of these harvesting and transportation operations offer opportunities that can make the harvesting and delivery options more efficient for logging firms and more attractive to forest products industries.

The survey instrument used is transferable to other regions since even though responses might be different, they shed light on the factors that need to be considered when investigating forest harvesting and transportation operations. Findings are relevant to other regions and countries beyond the study because they pull together interpretations of key operational matters, conditions, work seasons, product types, equipment and transportation in an integrated manner. Results could be compared with other regions to draw a more realistic representation of logging firms and operations. For example, in southern states it would be rare to detect cut-to-length operations. The stark differences between harvesting equipment, and their productivity with existing equipment, are worthy of investigation to determine productivity and existing equipment potentials.

Acknowledgements

Data in the study applies to harvesting under both natural and plantation stand types. Results could be compared against similar variables from other regions, to determine unique productivity rates. Unique productivity rates could be used to promote wider involvement of stakeholders in the supply chain to build stronger forest products businesses. For example, one purpose of this study was to help inform the logging community about the productivity of their peers under their similar operating conditions. Promoting further knowledge transfer among the logging firms would help promote a more informed, integrated and rounded approach to logging operations analysis to build stronger forest products businesses. Results analyzed helped develop a fuller description of operating conditions to identify to what extent the logging community is responsive to the potential for the startup or expansion of new industries. The reported productivity of existing systems in different forest types, and the age of the technology running the operations, can help determine the extent of the size of facilities and the required attention needed to im-

190

The authors would like to extend a special thank you for the support provided by the Department of Energy award number DE-EE-0000280; forest machine operators; and logging firms consulted from the State of Michigan, forestry consultants, officials from Michigan Department of Natural Resources and to many others. The authors would like to acknowledge support from multiple experts including Dr. Donna La Courte (Michigan Centers of Excellence); Mr. Art Abramson (Frontier Renewable Resources); from Michigan State University: Dr. Karen Potter Witter, Mr. Mike Schira, and Ms. Shivan G.C.; Dr. Robert Rummer (U.S. Forest Service, Auburn, AL); and from Michigan Technological University: Mr. Stephen Chartier Jr., Mr. Hamed Pouryoudsef. We would also like to acknowledge constructive and helpful remarks from anonymous reviewers of the Croatian Journal of Forestry Engineering and the support provided by the journal’s editorial board.

Disclaimer »This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, or service by trade name, trademark, manufactured, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.«

Croat. j. for. eng. 35(2014)2

A Survey Analysis of Forest Harvesting and Transportation Operations in Michigan (179–192)

6. References AAPOR, 2010: Outcome Rate Calculator. American Association for Public Opinion Research. Version 3.1 November, 2010. Available online at: http://www.aapor.org; Last accessed 04 16 2012. Abbas, D., Current, D., Ryans, M., Taff, S., Hoganson, H., Brooks K., 2011a: Harvesting Forest Biomass for Energy: an Alternative to Conventional Fuel Treatments: Trials in the Superior National Forest, USA. Biomass and Bioenergy 35(11): 4557–4564. Abbas D., Current D., Phillips M., Rossman R., Brooks K., Hoganson H., 2011b: Guidelines for Harvesting Forest Biomass for Energy: A Synthesis of Environmental Considerations. Biomass and Bioenergy 35(11): 4538–4546. Abbas, D., Handler, R., Dykstra, D., Hartsough, D., Lautala, P., 2013: Cost Analysis of Forest Biomass Supply. Journal of Forestry 111(4): 271–281. Dillman, D. A., 2000: Mail and Internet Surveys: The Tailored Design Method. Second Edition. New York: John Wiley Co., USA. 464 p. Drapala, P., 2009: Decline in Housing Market Hits Forestry Industry Hard. Mississippi Agricultural News. Available online at: http://msucares.com/news/print/agnews/ an09/091217_forestry.html; last accessed March 10, 2012. Drolet, S., LeBel, L., 2010: Forest Harvesting Entrepreneurs, Perception of their Business Status and its Influence on Performance Evaluation. Forest Policy and Economics 12(4): 287–298. Egan, A., 2001: Clearcutting and Forest Regulation in the »New« Forestry: Views from Professional Foresters in the Northeastern U.S.. International Journal of Forest Engineering 12(2):19–25. Egan, A., 2011: Characteristics of and Challenges Faced by Logging Business Owners in Southern New England. Northern Journal of Applied Forestry, 28(4): 180–185. EISA 2007: Energy Independence and Security Act of 2007 (Enrolled as Agreed to or Passed by Both House and Senate), Library of Congress H.R.6. G.C., S., Potter-Witter, K., 2011: An examination of Michigan’s Logging Sector in the Emerging Bioenergy Market. Forest Products Journal 61(6): 459–465. Greene, W. D., Jackson, B. D., Culpepper, J. D., 2001: Georgia’s Logging Businesses, 1987 to 1997. Forest Products Journal 51(1): 25–28. Grushecky, S., Buehlmass, U., Schuler, A., Luppold, W., Cesa, E. 2006: Decline in the U.S. Furniture Industry: A Case Study of the Impacts to the Hardwood Lumber Supply Chain. Wood and Fiber Science 38(2): 365–376. Han, H. S., Lee, H. W., Johnson, L. R., 2004: Economic Feasibility of an Integrated Harvesting System for Small-Diameter Trees in Southwest Idaho. Forest Products Journal 54(2): 21–27. Croat. j. for. eng. 35(2014)2

D. Abbas et al.

Harris, R., Adams, T., Hiott, V., Van Lear, D., Wang, G., Tanner, T., 2005: Potential for Biomass Energy Development in South Carolina. Final report to the South Carolina Forestry Commission, 23 p. Hartsough, B. R., Zhang, X., Fight, R. D., 2001: Harvesting Cost Model for Small Trees in Natural Stands in the Interior Northwest. Forest Products Journal 51(4): 54–61. Hunsberger, R., Haase, S., Rooney, T., 2003: Evaluating Biomass Utilization Options for Colorado: Summit and Eagle Counties. Western Regional Biomass Energy Program, Final Report, McNiel Technologies, 73 p. IRGC 2008: Risk governance guidelines for bioenergy policies: Policy Brief. International Risk Governance Council 2008. Available online at: http://www.irgc.org/IMG/pdf/ IRGC_PB_Bioenergy_WEB-2.pdf; last accessed March 10, 2012. Jylhä, P., Dahl, O., Laitila, J., Kärhä, K., 2010: The Effect of the Supply System on the Wood Paying Capability of a Kraft Pulp Mill using Scots Pine Harvested from First Thinning. Silva Fennica 44(4): 695–714. Luppold, W., Hassler, C., Grushecky, S., 1998: An Examination of West Virginia’s logging industry. Forest Products Journal 48(2): 60–64. McNeel, J. F., Dodd, K., 1996: A Survey of Commercial Thinning Practices in the Coastal Region of Washington. Forest Products Journal 46(11/12): 33−39. MFRC 2005: Minnesota Forest Resources Council. Sustaining Minnesota Forest Resources: Voluntary Site-Level Forest Management Guidelines for Landowners, Loggers and Resource Managers (document on the Internet, cited May 3, 2011). 55 p., from: http://www.frc.state.mn.us/initiatives_ sitelevel_management.html. Milauskas, S. J., Wang, J., 2006: West Virginia Logger Characteristics. Forest Products Journal 56(2):19–24. Norusis, M. J., 2010: PASW Statistics 18 Guide to Data Analysis. Prentice Hall Press Upper Saddle River, NJ, USA, 672 p. Pedersen, L., 2005: Michigan State Forest Timber Harvest Trends. A Review of Recent Harvest Levels and Factors Influencing Future Levels. Report Submitted to Chief Lynne Boyd, FMFM, MI DNR 09/16/2005, 11 p. Spinelli, R., Hartsough, B., 2001: A Survey of Italian Chipping Operations. Biomass and Bioenergy 21(6): 433– 444. Rickenbach, M., Steele, T.W., Schira, M., 2005: Status of the Logging Sector in Wisconsin and Michigan’s Upper Peninsula – 2003. University of Wisconsin-Extension Cooperative Extension Service, Madison, WI, USA, 40 p. Rummer, R., Prestemon, J., May, D., Miles, P., Vissage, J., McRoberts, R., Liknes, G., Sheppard, W., Ferguson, D., Elliot, W., Miller, S., Reutebuch, S., Barbour, J., Fried, J., Stoke, B., Bilek, E., Skog, K., Hartsough, B., Murphy, G., 2005: A Strategic Assessment of Forest Biomass and Fuel Reduction

191

D. Abbas et al.

A Survey Analysis of Forest Harvesting and Transportation Operations in Michigan (179â&#x20AC;&#x201C;192)

Treatments in Western States. United States Department of Agriculture Forest Service Rocky Mountain Research Station General in partnership with the Western Forestry Leadership Coalition Technical Report RMRS-GTR-149 March 2005, 20 p.

Authorâ&#x20AC;&#x2122;s address: Asst. Prof. Dalia Abbas, PhD.* e-mail: sale0056@umn.edu. Tennessee State University 3500 John A. Merritt Blvd. Nashville, TN 37209 USA Asst. Prof. Robert Handler. PhD. e-mail: rhandler@mtu.edu Asst. Prof. Pasi Lautala, PhD. e-mail: ptlautal@mtu.edu Michigan Technological University Houghton, MI 49931 USA Prof. Bruce Hartsough, PhD. e-mail: brhartsough@ucdavis.edu. University of California Davis Davis, CA 95616 USA Dennis Dykstra, PhD. e-mail: dennisdykstra@blueoxforestry.com Blue Ox Forestry Portland, OR 97201 USA

Received: August 13, 2013. Accepted: May 30, 2014.

192

Larry Hembroff, PhD. e-mail: hembroff@msu.edu Michigan State University East Lansing, MI 48824 USA * Corresponding author Croat. j. for. eng. 35(2014)2

Original scientific paper

Impact of Slope on Productivity of a Self-levelling Processor Martin Strandgard, Muhammad Alam, Rick Mitchell Abstract Slope is a major factor affecting forest harvesting machine productivity. As ground-based harvesting methods are generally cheaper than the alternatives, forest managers need to know when ground-based harvesting equipment can be used on sloping sites. The study objective was to determine the effect of slope on the productivity, cycle time and elemental times of a Valmet 450 FXL self-levelling processor processing a 24 year-old, unthinned radiata pine plantation previously felled and stacked by a feller-buncher. The study site slope was estimated using a LiDAR (Light Detection and Ranging) derived digital terrain model and classified using the regional terrain classification system. Study trees were selected from areas predominantly in the hilly (12–19°) and steep (20–26°) slope classes, as these classes made up the majority of the study site area. In contrast to previous research, no significant differences were found between the processor productivity, cycle time and elemental times (moving/positioning, swinging and processing) between the slope classes. This was believed to result from the processor working well within its capabilities processing the relatively small trees on the study site. Other important factors may have included that the trees were pre-felled by a feller-buncher and placed in high density rows with their butt ends aligned, which minimised the processor boom and track movements, and that steep slope trees were selected from areas at the lower end of the steep slope class (20–23°). Further research is needed to determine whether the processor productivity would be significantly lower when processing larger trees on steeper slopes. Keywords: self-levelling, processor, slope, productivity, radiata pine, LiDAR, Australia

1. Introduction Single-grip harvester productivity is affected by many factors related to stand (tree size, form and spacing), terrain (slope, ground strength and roughness), machine (type, size, boom reach, etc) and operator (experience, technique and attitude) characteristics. Tree size has been shown in numerous studies to be the most important factor affecting harvester productivity, with productivity increasing with increasing tree size (Kellogg and Bettinger 1994, Acuna and Kellogg 2009, Visser and Spinelli 2012, Ghaffariyan et al. 2012). Operator performance is the other major factor in determining harvester productivity. Variability in productivity between skilled operators can be over 40% (Kärhä et al. 2004, Ovaskainen et al. 2004, Hogg et al. 2011). Croat. j. for. eng. 35(2014)2

Of the other factors affecting harvester productivity, slope has an important role in both the selection and productivity of harvesting equipment, as it is a major determinant of machine travel speed and stability (Davis and Reisinger 1990). Given good soil conditions, tracked harvesters with self-levelling cabins can operate on slopes up to 60% (31°), whereas specialised steep slope harvesters (such as the Komatsu 911.3 X3M) or cable-tethered harvesters can operate on slopes up to 70% (35°) (Stampfer and Steinmüller 2001) (wheeled harvesters are restricted to less steep slopes). However, in practice harvester slope constraints are generally set lower to maintain safety and reduce soil damage (MacDonald 1999, Sutherland 2012). On steeper slopes (>35°) or with poorer soil conditions, other, more expensive harvesting methods, such as cable-harvesting, must be used. As ground-

193

M. Strandgard et al.

Impact of Slope on Productivity of a Self-levelling Processor (193–200)

based harvesting systems can generally deliver logs more cheaply to roadside than the alternatives, forest owners aim to maximise the use of ground-based equipment in steep terrain (Fight et al. 2006). In order to determine when it is both technically feasible and cheaper to use ground-based harvesting equipment, more research into its performance on steep slopes is required. Harvesters in Pinus radiata (radiata pine) groundbased harvesting operations, either fell and process trees at the stump or process trees felled and stacked by a feller-buncher. The latter approach is typically used on sites where trees need to be extracted from steeper sections and stream reserves to minimise or eliminate machine movements in these areas (Spinelli et al. 2002). Differences in the type and relative duration of the activities performed by a harvester in these two roles may affect the impact of slope on its productivity. Evanson and McConchie (1996) reported that a harvester processing radiata pine at roadside was considerably more productive than when felling and processing trees in the stand because of the time saved not felling trees. Previous research on the impact of slope on harvester productivity has largely focused on machines harvesting rather than processing. A number of trials of both thinning and clearfell harvesting operations have reported that increasing slope decreases harvester productivity. However, Stampfer (1999) found that only the harvester movement was significantly affected by slope, whereas Bolding and Lanford (2002) found slope also affected tree swing time and Spinelli et al. (2010) found that it also affected felling and processing times. Acuna and Kellogg (2009) found increasing slope significantly decreased the productivity of a harvester processing trees felled by a fellerbuncher because the processor spent more time positioning the machine and ensuring the logs were piled correctly when operating on steep slopes. In contrast to these findings, Robert et al. (2013) reported that slope had no impact on the productivity of a Komatsu 911.3 X3M steep slope harvester operating on slopes from less than 20° to over 27°. Assessing the slope experienced by an operating harvester is difficult. Traditionally, the slope of study sites has been estimated using a clinometer, though this approach is limited to measuring the slope between a small number of points. Recently, LiDAR (Light Detection and Ranging) has become more commonly used to estimate slope for harvest planning as it can provide accurate »wall-to-wall« slope maps of a harvesting area (Sessions et al. 2006). However, Berkett and Visser (2013) suggest that the actual slope experienced by a harvesting machine can vary significantly

194

from that predicted from digital slope maps, though this may depend on the resolution of the digital map. The objective of this study was to compare the productivity, cycle time and elemental times of a self-levelling processor processing trees felled by a fellerbuncher when operating on 12–19° slopes and on 20–26° slopes. The hypothesis was that the processor productivity would be significantly lower and its cycle times and elemental times significantly longer when operating on 20–26° slopes compared with when it was operating on 12–19° slopes.

2. Materials and Methods 2.1 Study site The study was located approximately 6 km west of Port Arthur, Tasmania, Australia. The study site was an area of approximately 1 ha within a radiata pine plantation being clearfelled for pulp wood production (Table 1). Diameter at breast height (1.3 m) over bark (DBHOB) of all trees on the site was measured with a diameter tape to the nearest 0.1 cm. The heights of 100 trees spread across the site and covering the range of DBHOB values at the site were measured with a Vertex hypsometer and Impulse 200 laser to the nearest Table 1 Description of study site Attribute Species Plantation age at harvest, years Tree form Branchiness Merchantable stocking, trees/ha Thinning Undergrowth

Value Pinus radiata 24 Good Light branching 1,057 Unthinned None

Soil composition

Clay loam

Ground strength

Moderate

Ground roughness

Even with scattered small rocks

Mean slope range, degrees

21 (18–25)

Mean tree height range, m

26.1 (15.8–37.0)

Mean DBHOB range, cm

29.0 (10.3–61.0)

Mean merchantable tree volume range, m3

0.63 (0.04–3.47)

Croat. j. for. eng. 35(2014)2

Impact of Slope on Productivity of a Self-levelling Processor (193–200)

M. Strandgard et al.

0.1 m. An individual tree volume function supplied by the forest owner (Norske Skog Australasia) was used to estimate the merchantable volume of each tree. Merchantable tree volume is referred to as tree volume in the paper. A unique number was painted on each tree to identify it during the study. Tree measurements are summarised in Table 1.

2.2 Slope derivation The slope over the study site was derived from LiDAR data supplied by Forestry Tasmania, Australia with the following specifications (Table 2). Table 2 LiDAR parameters and scanning system settings LiDAR attribute Date of flight System

Value 25 May 2011 ALTM (Airborne Laser Terrain Mapping) Gemini

Beam divergence, milliradian

0.20

Footprint diameter, cm Laser mode

20 Single pulse st

Pulse return density range, m–2

>3 (1 , 2 , 3 and last) (2.3–3.2)

Horizontal accuracy, m

0.15

Vertical accuracy, m

0.15

Pulse rate frequency, kHz

LiDAR data were supplied in .LAS format and were classified into ground and non-ground points. LiDAR data accuracy was verified by the data provider. Slope was derived from a digital terrain model with a 2 m cell size constructed from the ground LiDAR points. The slope of the study site was classified into Flat-rolling = 0–11°, Hilly = 12–19°, Steep = 20–26°, Very Steep = >27° slope classes using the Tasmanian Forest Practices Code terrain slope classification (Forest Practices Board 2000), which is applicable to all Tasmanian timber production forests (Fig. 1).

2.3 Time and motion study The harvesting system consisted of a feller-buncher, a processor and a forwarder. Immediately prior to the processor study, the feller-buncher felled the trees on the study site and placed each tree across the slope with their butts aligned to form rows of felled trees running Croat. j. for. eng. 35(2014)2

Fig. 1 Processor study area showing slope classes and tree selection areas up and down the slope (Alam et al. 2013). The processor used in the study was a Valmet 450 FXL self-levelling processor with a 224 kW engine manufactured in 2010 with 2,408 engine hours. It was equipped with a Southstar 585 felling and processing head. The operator had four years experience in operating processors. The processor worked uphill processing felled trees on the right of the processor to logs (predominantly 5.4 m in length with a minimum small end diameter of 100 mm) piled to the left. At the completion of each strip of felled trees, the processor travelled down the slope to commence the next strip. Processing took place from the fourth to the sixth of April 2011 in overcast conditions and was filmed using a digital video recorder. Cycle time commenced when the processor or boom started to move towards a felled tree and ended when the processor had completed processing the tree and was about to move to the next felled tree. Cycles were divided into the following time elements: moving/positioning, swinging, processing, stacking/bunching, brushing/clearing and delays (Table 3). Elemental times were recorded from the video recordings using TimerPro Professional software (www.acsco.com). The time elements stacking/bunching, brushing/clearing, travel and delays were excluded from the analysis as they occurred infrequently and were unrelated to tree volume and slope. Trees used in the study were selected from sections of the site which were predominantly in the 12–19° or

195

M. Strandgard et al.

Impact of Slope on Productivity of a Self-levelling Processor (193–200)

Table 3 Description of processor time elements Time element

Definition Starts when the processor begins to move and/or swing its boom towards a felled tree and ends when the head clamps onto the tree Starts when head clamps onto a felled tree and ends when feed rollers are activated, or the first cut is made to reset the processor length measurement (whichever occurs first) Starts when feed rollers are activated, or the first cut is made to reset the processor length measurement (whichever occurs first) and ends when the last log is cut and dropped on the log pile

Moving/positioning Swinging Processing Brushing/Clearing

Any interruption to other elements to remove unmerchantable trees or clear processing debris Time taken to turn around to start new stack or move to and from break. Starts when wheels/tracks begin to rotate. Ends when boom begins its swing towards first tree on new stack Starts when the boom commences a swing to retrieve move or »stack« any processed logs. Ends when the boom moves to perform some other activity Any interruption to the previous time elements. The cause of the delay (e.g. operational, personal, mechanical, or study induced) is recorded

Travel Stacking/Bunching Delay

20–26° slope classes as these slope classes accounted for the majority of the study site area (Fig. 1). To improve the representativeness of the sample, trees were selected from several sections in each slope class and across all three days of the trial. Trees were excluded from the study when the tree number could not be identified during processing or the tree had multiple leaders and each leader was processed separately. Seventy trees were selected for analysis in the 12–19° slope class and sixty-nine in the 20–26° slope class. Trees in the >27° slope class had been moved by the fellerbuncher to adjacent, less steep areas.

against tree volume using Microsoft Excel 2007 and Minitab 16 Ltd. Various model forms and variable transformations were tested to identify models with the best goodness of fit (R2, root mean square error (RMSE), and mean absolute error (MAE) which also achieved homogeneity of variance of the residuals. To determine whether processor cycle time, elemental times or processor productivity differed between slope classes, the best-fit models for each slope class were compared using an F-test (p<0.05) (Motulsky and Christopoulos 2003) if the models were significant or with a Mann-Whitney test (p < 0.05) if they were not.

2.4 Data analysis

3. Results

Regression models were developed for each slope class for processor cycle time against tree volume, for moving/positioning, swinging and processing times against tree volume and for processor productivity

The processor work elements, cycle times and productivity are summarised in Table 4. With the exception of processing time, the relationships between

Table 4 Mean, Standard Deviation (SD) and range of processor time elements, cycle times, productivities and tree volumes for the 12–19° and 20–26° slope classes Slope class 12–19° Time element, minute

20–26°

Mean (SD)

Range

Mean (SD)

Range

Moving/positioning time

0.11 (0.04)

0.03–0.32

0.11 (0.05)

0.04–0.37

Swinging time

0.11 (0.04)

0.05–0.28

0.09 (0.03)

0.04–0.16

Processing time

0.3 (0.13)

0.10–0.74

0.31 (0.12)

0.09–0.64

0.51 (0.14)

0.24–0.98

0.51 (0.13)

0.27–0.86

Productivity, m PMH

69.4 (35.8)

14.0–167.1

59.5 (31.5)

15.7–154.0

Tree volume, m3

0.63 (0.42)

0.09–1.75

0.53 (0.35)

0.09–1.57

Cycle time, minute 3

196

–1 0

Croat. j. for. eng. 35(2014)2

Impact of Slope on Productivity of a Self-levelling Processor (193–200)

Fig. 2 Processor cycle time (minutes) against tree volume (m3) for the 12–19° and 20–26° slope classes

M. Strandgard et al.

Fig. 4 Productivity (m3 PMH0–1) against tree volume (m3) for the 12–19° and 20–26° slope classes slope classes was a linear regression of the dependent variable (Cycle time (minutes) or Processing time (minutes)) and Tree volume (m3) (Fig. 2 and Fig. 3, respectively): Cycle time = b0 + b1 × Tree Volume

(1)

Processing time = b0 + b1 × Tree Volume

(2)

Model coefficients and fit statistics are in Table 5. There was no significant difference between the models for each slope class. The model form which best fitted the data for both slope classes for productivity against tree volume was a natural logarithmic transformation of productivity (m3 PMH0–1) and of tree volume (m3) (Fig. 4): ln(Productivity) = b0 + b1 × ln(Tree Volume)

Fig. 3 Processing time (minutes) against tree volume (m3) for the 12–19° and 20–26° slope classes each time element and tree volume were not significant. No significant differences were found between mean processor moving/positioning time for each slope class and between mean swinging time for each slope class. The model form which best fitted the data for cycle time and processing time against tree volume for both Croat. j. for. eng. 35(2014)2

(3)

Model coefficients and fit statistics are in Table 5. There was no significant difference between the Productivity models for each slope class. As logarithmic transformation of the dependent variable introduces a negative bias, the predicted productivity values were corrected following back-transformation using the method of Snowdon (1991). The correction factors were 1.011 (12–19° slope class) and 1.018 (20–26° slope class).

4. Discussion Significant relationships were found between the cycle time and productivity of the processor and tree

197

M. Strandgard et al.

Impact of Slope on Productivity of a Self-levelling Processor (193–200)

Table 5 Processor cycle time, processing time and productivity model coefficients and goodness of fit statistics for each slope class Model Cycle time Processing time Productivity

Model coefficients Slope class

Mean bias

MAE

RMSE

12–19°

0.377

0.219

0.08

0.1

0.42

20–26°

0.377

0.259

0.08

0.1

0.43

12–19°

0.168

0.21

0.07

0.09

0.47

20–26°

0.177

0.244

0.07

0.08

0.50

12–19°

4.639

0.777

11.2

15.1

0.82

20–26°

4.571

0.736

9.5

13.4

0.81

volume (Table 5), with the productivity of the processor increasing with increasing tree size (Fig. 4), as found in numerous previous studies (Kellogg and Bettinger 1994, Acuna and Kellogg 2009, Visser and Spinelli 2012, Ghaffariyan et al. 2012). However, in this study the productivity of the processor was not significantly different when operating in the 12–19° slope class and in the 20–26° slope class. This is in contrast to the findings of previous research trials, which found (with the exception of a trial of a specialised steep slope harvester (Robert et al. 2013)) that productivity decreased as slope increased (trial slope ranges shown) (Stampfer 1999 (6–26°), Bolding and Lanford 2002 (0–25°), Acuna and Kellogg 2009 (0–20°), Spinelli et al. 2010 (0–27°)). The near linear relationship between productivity and tree volume for trees with a volume greater than 0.5 m3 in the current study (Fig. 4) suggests the volume and weight of the majority of the trees were well within the capabilities of the machine. This is the probable cause of the lack of significant difference in the processor productivity between the two slope classes. Spinelli et al. (2010) noted that engine power has a significant effect on the productivity of a harvester, but no interactions between slope and engine power were reported in that study. The divergence of the cycle time and productivity models for each slope class with increasing tree volume (Fig. 2 and Fig. 4) suggests that a site with a larger mean tree size may have resulted in a significant difference between the processor productivity in each slope class. The relatively small sample size and observation time in the study may also have been insufficient to detect differences in the performance of the processor between the slope classes. Any »observer« effect on the operator’s performance was believed to be insignificant as the observations were made over a period of three days whereas Makkonen (1954) reported that the observer effect did not last beyond the first day. The majority of previous trials have reported moving time to be significantly affected by changes in

198

Goodness of fit statistics

slope (e.g. Stampfer 1999, Bolding and Landford 2002, Spinelli et al. 2010). However, these trials were mostly of harvesters felling and processing trees whereas the current trial was of a processor processing trees felled and stacked by a feller-buncher. Typical operation of a harvester is to fell and process one or more trees from a stationary position and then move to a new position, with the number of trees felled and the distance moved depending on the density of trees and the proportion of trees being removed. In contrast, the processor in the current study performed most movements of its tracks while simultaneously swinging the boom to pick up the next tree for processing (the moving and positioning time element). The proportion of time spent moving and positioning was low (~21–22%) because the stand was unthinned with little mortality resulting in a high density of felled trees along the stacks created by the feller-buncher. Slope has also been reported in previous trials to have a significant effect on swinging (Bolding and Landford 2002), felling and processing time elements (Spinelli et al. 2010) and the time taken to position logs (Acuna and Kellogg 2009). In the study, felling was not performed by the processor and positioning logs was a rare event. Mean swinging and processing times were not significantly different between the slope classes. However, operating the machine on steeper slopes or with larger trees than in the current study may increase the swinging time because of the increased difficulty in swinging trees from the felled pile to be processed. In the study, slope was classified into broad classes defined by the Tasmanian Forest Practices Board (2000). However, the majority of the area in the steep slope class from which the study trees were selected was at the lower end of this class (20–23°), which may be another factor explaining the lack of significant impact of slope on the performance of the processor in the study. At the mean pooled tree volume for this study, the productivity of the processor was greater than that reCroat. j. for. eng. 35(2014)2

Impact of Slope on Productivity of a Self-levelling Processor (193–200)

ported by Strandgard et al. (2012) for three harvesters felling and processing radiata pine on relatively flat sites (48.4–55.9 m3 PMH0–1). This was expected as the processor in the current study did not have to fell trees and had a high density of trees in the stacks minimising the boom and track movements required to reach each tree. However, the processor in the study also had a greater productivity than a processor processing radiata pine infield on gentler slopes (41 m3 PMH0–1) at the mean pooled tree volume for the current study (Ghaffariyan et al. 2012). The lower productivity of the harvester in that operation may be due to it being an excavator-based machine with a less powerful engine (180 kW). The high density of trees along each stack and the arrangement of the felled trees in rows with their butt ends alongside the processor made processing in the current study more analogous to roadside processing than infield processing. FPInnovations (2007) reported the productivity of a processor at roadside to be 48.4 m3 PMH0–1 (logs <8 m) and 72.4 m3 PMH0–1 (logs >8 m) for trees at the mean pooled tree volume. Log length clearly had a significant impact on the productivity in these trials and may have been a factor in the high productivity of the processor in the current trial because most trees were processed into several logs of the longest allowable length (5.4 m) with only an occasional shorter log being cut.

5. Conclusion The lack of a significant impact of slope on the cycle time and productivity of the processor and on the individual time elements in the study suggests that the tree size at the site was well within the capabilities of the processor. Other important factors may have included that the trees were pre-felled by a fellerbuncher and placed in high density rows with their butt ends aligned, which minimised the processor boom and track movements, and that the steep slope trees were selected from areas at the lower end of the steep slope class. Further research is needed to determine whether the productivity of the processor would be significantly lower when processing trees with a larger mean volume on steeper slopes.

Acknowledgements The authors would like to thank the staff of Norske Skog, Australasia and their harvesting contractor BR & KF Muskett and Sons for their assistance in ensuring the success of this trial and the assistance of David Mannes (Forestry Tasmania) in providing the LiDAR data. Croat. j. for. eng. 35(2014)2

M. Strandgard et al.

6. References Acuna, M., Kellogg, L., 2009: An Evaluation of Alternative Cut-To-Length Harvesting Technology for Native Forest Thinning in Australia. International Journal of Forest Engineering 20(2): 17–25. Alam, M., Acuna, M., Brown, M., 2013: Self-Levelling FellerBuncher Productivity Based On Lidar-Derived Slope. Croatian Journal of Forest Engineering 34(2): 273–281. Berkett, H., Visser, R., 2013: Measuring Machine Slope When Harvesting on Steep Terrain. Proceedings of the International Conference on Forest Operations in Mountainous Conditions, Honne, Norway, June 2 – 5, 50–52. Bolding, M. C., Lanford, B. L., 2002: Productivity of a Ponsse Ergo Harvester Working on Steep Terrain. Proceedings of the Council of Forest Engineering 25th Annual Meeting »Forest Engineering Challenges: A Global Perspective«, Auburn, Alabama, June 16–20, 1–5. Davis, C. J., Reisinger, T. W., 1990: Evaluating Terrain for Harvesting Equipment Selection. Journal of Forest Engineering 2(1): 9–16. Evanson, T., McConchie, M., 1996: Productivity Measurements of Two Waratah 234 Hydraulic Tree Harvesters in Radiata Pine in New Zealand. Journal of Forest Engineering 7(3): 41–52. Fight, R. D., Hartsough, B. R., Noordijk, P., 2006: Users Guide for FRCS: Fuel Reduction Cost Simulator Software. General Technical Report. PNW-GTR- 668. Portland, Oregon, U.S. Department of Agriculture, Forest Service, Pacific Northwest Research Station. 23 p. Forest Practices Board, 2000: Forest Practices Code, Forest Practices Board, Hobart, Tasmania. Australia 7100. FPInnovations, 2007: Harvester Studies. Progress Report No. 12, Saint-Jean Pointe-Claire, QC, H9R 3J9, Canada. Ghaffariyan, M. R., Sessions, J., Brown, M., 2012: Machine Productivity and Residual Harvesting Residues Associated with a Cut-To-Length Harvest System in Southern Tasmania. Southern Forests: a Journal of Forest Science 74(4): 229–235. Hogg, G., Pulkki, R., Ackerman, P., 2011: Excavator-Based Processor Operator Productivity and Cost Analysis in Zululand, South Africa. Southern Forests: a Journal of Forest Science 73(2): 109–115. Kärhä, K., Rönkkö, E., Gumse, S., 2004: Productivity and Cutting Costs of Thinning Harvesters. International Journal of Forest Engineering 15(2): 43–56. Kellogg, L. D., Bettinger, P., 1994: Thinning Productivity and Cost for Mechanized Cut-To- Length System in the Northwest Pacific Coast Region of the USA. International Journal of Forest Engineering 5(2): 43–54. MacDonald, A. J., 1999: Harvesting Systems and Equipment in British Columbia. FERIC Handbook No. HB–12, p 1–197. Makkonen, O., 1954: Metsätöiden vertailevan aikatutkirnuks- periaate. (The Principle of Comparative Time Studies in Forest Work). Acta Forestalia Fennica 61, p. 1-18 (in Finnish with an English summary).

199

M. Strandgard et al.

Impact of Slope on Productivity of a Self-levelling Processor (193–200)

Motulsky, H. J., Christopoulos, A., 2003: Fitting Models to Biological Data Using Linear and Nonlinear Regression: A Practical Guide to Curve Fitting. GraphPad Software Inc., San Diego, California, USA, 269 p.

Spinelli, R., Owende, P. M. O., Ward, S. M., 2002: Productivity and Cost Of CTL Harvesting Of Eucalyptus globulus Stands Using Excavator-Based Harvesters. Forest Products Journal 52(1): 67–77.

Ovaskainen, H., Uusitalo, J., Väätäinen, K., 2004: Characteristics and Significance of a Harvester Operators’ Working Technique in Thinning. International Journal of Forest Engineering 15(2): 67–77.

Stampfer, K., 1999: Influence of Terrain Conditions and Thinning Regimes on Productivity of a Track-Based Steep Slope Harvester. Proceedings of the International Mountain Logging and 10th Pacific Northwest Skyline Symposium, Corvallis, Oregon, USA, March 28th – April 1st, 78–87.

Robert, R. C. G., Jaeger, D., Becker, G., 2013: Mechanization of Harvesting in Eucalyptus Spp. Plantation Forests Using a Harvester in Mountainous Areas in Brazil. Proceedings of the International Conference on Forest Operations in Mountainous Conditions, Honne, Norway, June 2–5, 4–43. Sessions, J., Akay, A., Murphy, G., Chung, C., Aruga, K., 2006: Road and Harvesting Planning and Operations in Computer Applications in Sustainable Forest Management: Including Perspectives on Collaboration and Integration. Series: Managing Forest Ecosystems, Vol. 11. G. Shao and K. Reynolds (eds.). Springer. Chapter 5: 83–99. Snowdon, P., 1991: A Ratio Estimator for Bias Correction in Logarithmic Regressions. Canadian Journal of Forest Research 21(5): 720–724. Spinelli, R., Hartsough, B., Magagnotti, N., 2010: Productivity Standards for Harvesters and Processors in Italy. Forest Products Journal 60(3): 226–235.

Stampfer, K., Steinmüller, T., 2001: A New Approach to Derive a Productivity Model for the Harvester Valmet 911 Snake. Proceedings of the International Mountain Logging and 11th Pacific Northwest Skyline Symposium. Seattle, Washington, USA, December 10–12: 254–262. Strandgard, M., Walsh, D., Acuna, M., 2012: Estimating Harvester Productivity in Pinus radiata Plantations Using Stanford Stem Files. Scandinavian Journal of Forest Research 28(1): 73–80. Sutherland, B., 2012: Review of Tethered Equipment for Steep-Slope Operations. Steep slope workshop, FPInnovations, Vancouver, 30th October 2012. Visser, R., Spinelli, R., 2012: Determining the Shape of the Productivity Function for Mechanized Felling And FellingProcessing. Journal of Forest Research 17(5): 397–402.

Authors’ address: Martin Strandgard* e-mail: mstrandg@usc.edu.au Australia Forest Operations Research Alliance (AFORA) University of the Sunshine Coast 500 Yarra Boulevard 3121 Richmond AUSTRALIA Muhammad Alam, PhD. e-mail: mmalam@student.unimelb.edu.au University of Melbourne 500 Yarra Boulevard 3121 Richmond AUSTRALIA

Received: December 24, 2013 Accepted: March 14, 2014

200

Rick Mitchell e-mail: rmitchel@usc.edu.au Australia Forest Operations Research Alliance (AFORA) University of the Sunshine Coast 35 Shorts Place 6330 Albany AUSTRALIA * Corresponding author Croat. j. for. eng. 35(2014)2

Original scientific paper

Productivity Analysis of an Un-Guyed Integrated Yarder-Processor with Running Skyline Bruce Talbot, Giovanna Ottaviani Aalmo, Karl Stampfer Abstract An excavator-based integrated yarder-processor was evaluated in a clearfelling in central Norway. The machine is unique because, as it uses a running skyline setup, yarding and processing cannot take place simultaneously as is the case with many European integrated tower yarders. Felling productivity was 10.6 m3 E15h-1, yarding 9.2 m3 E15h-1 and processing 10.9 m3 E15h-1. Given that yarding and processing take place alternately accounting for 54% and 46% of a system hour, the overall system productivity was 4.9 m3 E15h-1 (processed and stacked). The processing rate was approximately 30% of what is achieved by single grip harvesters, indicating the effect of space limitations, a possible over-dimensioned processing head, and the need to simplify the assortment list under such conditions. An increase in processing productivity would require a second feller-chokersetter in the crew, although neither would then be used to full capacity. Un-choking alone accounted for 19% of the yarding cycle time and might be reduced by applying self-releasing chokers. System productivity needs to be increased by 30â&#x20AC;&#x201C;50% to make it competitive. Much of this could be achieved simply by deploying the machine in stands with larger mean tree volumes than those observed (0.27 m3). Keywords: steep terrain, timber harvesting, cable yarding, un-guyed yarder

1. Introduction Excavator-based forest machines are an alternative to purpose built machines and, where terrain allows, can be used in applications ranging from drainage maintenance through site preparation and planting, as tracked harvesters, roadside processors, stump harvesters and cable yarders (Johansson 1997). Their popularity can likely be explained by their global availability, low cost, robustness, ease of operation and large interface with other sectors, such as earth moving, construction, and road building. Cable yarding is a specific application of excavators in forestry, but is widely applied in Japan (Yoshimura and Noba 2013) and gaining ground in countries like the UK (Tuer et al. 2013), Ireland (Devlin and Klvac 2013), South Africa (McEwan et al. 2013) and Canada (Gingras 2013). The mass of the base machine and support of the boom arm as an outrigger allow for excavator based yarders to operate un-guyed. Croat. j. for. eng. 35(2014)2

Un-guyed yarders are considered to offer a number of advantages under conditions of (i) space restrictions: mobility on the landing during operation to allow for trucks to pass, to prevent congestion by moving continuously away from tree piles or log stacks or obstructions experienced in the corridor, un-guyed yarders allow for the forest road to effectively be used as a continuous landing, (ii) Short corridors or low volume densities: only the tail-spar needs to be rigged, un-guyed yarders have a lower setup time and therewith a competitive advantage on shorter corridors where higher rigging times are not justified by the limited volume extracted, and (iii) local and seasonal availability: excavators are relatively low cost and readily available base machines that have a range of applications and can be used seasonally for forest work by e.g. farmers (Johansson 1997). The configuration and functionality of un-guyed excavator based yarders varies considerably. Each concept offers benefits and restrictions pertaining to com-

201

B. Talbot et al.

Productivity Analysis of an Un-Guyed Integrated Yarder-Processor with Running Skyline (201–210)

plexity, versatility, stability, productivity and economy. Base machines range from c. 15 to 40 tonnes. Winches range from single drum to 3-drum systems, both mechanically and hydraulically driven. Some use a block mounted on the boom for lift, while others are fitted with towers of varying height, mounted on the machine, the boom, or the boom tip. Some of the configurations retain the bucket for stability, others have replaced this with a timber grapple, while yet others have a felling/processing head attached to the boom tip. Excavator-based yarders can be distinguished from other yarders built on similar undercarriages, in that part or all of the boom is retained and not replaced with a gantry setup as are the Madill type yarders. Torgersen and Lisland (2002) provided an overview of the perceived advantages and disadvantages of these configurations in considering their potential application in Norway. Excavator-based yarders are considered to be well suited to the inland conditions in Norway with small crews (2–3 people), small trees in (0.2–0.4 m3 per stem), generally small work objects (c. 1–3 ha.) and with low harvesting volumes (150–220 m3 ha–1). There is a need to develop more versatile harvesting systems in Norway where some 150 million m3 of timber is mature or maturing on slopes with an inclination steeper than 33% (Larsson and Hylen 2007) equating to the volume of 15 national annual cuts. Recent cable yarding productivity studies of relevance include Spinelli et al. (2010) who studied two small-scale units in hardwood stands the Apennine’s,

Ghaffariyan et al. (2009) who developed production equations for two tower yarders in predominantly fir stands in Alpine conditions, and Zimbalatti and Proto (2009) who report on productivity rates for three different tower yarders extracting timber for firewood production in Calabria. However, apart from Torgersen and Lisland (2002), only limited work has been published on the productivity rates achieved by this machine genre. Largo et al. (2004) studied a Timbco fellerbuncher based yarder and a Caterpillar excavator based yarder in thinning operations in Idaho. Both were fitted with two-drum winches and used in a live, gravity system, and operated with 2-man crews. The work of Stampfer et al. (2006) is relevant in that they studied installation times for tower yarders, an important potential area for time saving on un-guyed yarders. The lack of literature addressing this specific topic indicates that no previous productivity studies have been published for this type of fully integrated machine. As the use of excavator-based yarders appears to be on the increase globally, results from the present work might be useful in identifying areas for improvement or application.

1.1 Aim The aim of the present paper was to analyze the productivity levels achieved by a new fully integrated yarder-processor combination operating in a clear cut in the inland forest region of Norway.

Fig. 1 Distribution of trees to volume intervals

202

Croat. j. for. eng. 35(2014)2

Productivity Analysis of an Un-Guyed Integrated Yarder-Processor with Running Skyline (201–210)

2. Materials and methods 2.1 Site information and work conditions Studies were carried out over 4 days in a mixed Norway spruce (Picea abies) and Scots pine (Pinus sylvestris) stand located in the upper Gudbrandsdalen valley in central Norway (UTM N 6,835,676 m, E 531,062 m). Working conditions on the site could be classified as good, an even and moderate north facing slope of ~43% with no notable surface obstacles. The diameters of 98 trees were measured, and diameter-height relationships of an additional 20 trees were measured in calculating single tree volumes. The mean tree volume was 0.27 m3 (s.d. 0.21), while the smallest was 0.04 m3 and the largest 1.12 m3 (Fig. 1). For the time study, trees were classified into 3 size classes with the following mid volumes; (1) small 0.17 m3, (2) medium 0.31 m3 and (3) large 0.56 m3. Stumps in three randomly located circular plots (r=10 m) were counted after harvesting, indicating a stand density of 610 stems ha–1 and a stand volume of roughly 140 m3 ha–1, a poor stand equating to a site index40 of 11 m (Tveite 1977). The operation studied was a clearcut. Weather conditions were warm and dry.

B. Talbot et al.

Table 1 Technical information on the winch Manufacturer

Zöggeler Forsttechnikk (http://www.zoeggeler.at/)

Drums

3, hydrostatically driven with auto-tensioning

Haulback line

500 m, 11 mm

Mainline

250 m, 11 mm

Slackpulling line

500 m, 6 mm (also used as rigging winch) Max 4 m s–1

Line speed Carriage

Zöggeler carriage with slackpulling capacity

Carriage mass

150 kg

2.2 Technical machine data The machine studied was an excavator-based yarder developed by Zöggeler Forsttechnik in Austria (Fig. 2), which is unique in that it is fully integrated with both yarding and processing capability, but unlike similar tower yarders, these operations cannot take place simultaneously. The hydraulic winch (Table 1) has 3 in-line drums mounted on the boom, each fitted with auto-tensioning, which allows for slack to be spent or taken up continuously while slewing during processing or stacking without pulling up the tail spar or applying undue tension on the boom (Fig. 3). The lightweight carriage uses the slackpulling line in feeding the mainline out to be used as a skid line. The winch was mounted on a 21 t Doosan DX210W wheeled excavator, stabilized with a dozer blade in the

Fig. 2 The Zöggeler yarder at work on the study site

Table 2 Technical information on the base machine and processing head Base machine

Processing head

Model

Doosan DX210W

Model

Zöggeler ZBH58

Mass

20,500 kg

Mass

1,480 kg

Motor

Doosan 6 cyl. 6 liter

Maximum cut diameter

70 cm

Rated power (gross)

127 kW at 2000 rpm

Optimal oil supply

300 l min–1

2x232 l min–1

Loading grapple

150 cm/0.7 m3

Hydraulic pumps

Croat. j. for. eng. 35(2014)2

203

B. Talbot et al.

Productivity Analysis of an Un-Guyed Integrated Yarder-Processor with Running Skyline (201–210)

After chokersetting and yarding, felling (motormanual) and processing (mechanized) took place simultaneously and were studied individually. Felling alternated with chokersetting approximately every 20–25 minutes and so provided the feller-chokersetter with a varying workload over the day. Felling cycles started and ended when the tree hit the ground, and included elements such as moving, clearing underbrush, and brushing low branches (Table 3). Felling times for 217 trees were included in the final analysis.

Fig. 3 Illustration of the 3-drum inline winch and hydraulically lifted tower with butterfly pulleys mounted on the boom, as well as the (A) slackpulling line, (B) mainline and (C) haulback lines (Copyright Zöggeler Forsttechnik) front and outrigger at the rear. The excavator is fitted with a two-piece boom, and a telescopic replacement for the boom arm. A specially designed processing head with loading arms (Zöggeler ZBH58) was fitted to the boom arm for processing the trees and stacking logs (Table 2).

2.3 Time and productivity studies The machine yarded uphill in a running skyline setup. Corridor length was short, varying between 80 and 120 m. The operation involved a 2-man crew, one machine operator and one feller-chokersetter, with multiple years of experience on tower yarders, but only around 4 months of operating experience on the Zöggeler machine. The standard work method adopted by the crew was studied. This alternated between yarding (involving both crew members) and processing (machine operator) with simultaneous felling (feller-chokersetter). The switch between work functions took place for every 5–7 loads (11–16 trees). Time studies were carried out using Haglöf SDI® software running in a Windows CE® environment on an Allegro MX datalogger from Juniper Systems™, which allows for continuous recording at the centiminute (min × 10–2) level. Work elements and variables measured for each operation are provided in Table 3 in the results section to avoid duplication. The number of trees in each size class was recorded for each load. Estimates of haul-out distance (distance carriage travels into the stand) were calibrated intermittently using a laser range finder sighting back to the base machine. Lateral distance was estimated visually as the time keeper worked in close proximity to the feller-chokersetter.

204

Processing was recorded at tree level but time for the processing of individual logs within each tree was also recorded. Processing commenced when the processing head took hold of a new tree from the landing, and included other functionality such as the handling of residues, sorting, stacking and clearing the landing. Processing times for 254 trees and 745 logs were included in the final analysis. Down-rigging, moving and rigging of new corridors was measured for 3 moves using wristwatch time. To minimize waiting time on the yarder, only the centerline of the new corridor was felled for a new rigging, the remaining trees were felled during normal operation. Time study data was cleaned of outliers, the distributions of individual time elements checked, and the regression models were developed and adapted using R statistical software.

3. Results Results are presented separately for each of the 3 discrete operations: felling, yarding, and processing. Mean E15 times were 91.5 s tree–1 for felling, 240 s cycle–1 for yarding, and 88.3 s tree–1 for processing (Table 3). For felling, cutting out the felling notch and performing the felling cut was the single most time consuming element, at c. 36 s tree–1. Values are here averaged out over all effective observations and can therefore appear shorter than their actual duration when occurring – e.g. the felling wedge was used 99 times out of 217 observations with a mean of 24.7 s per time used, but 11.3 s per observation mean. Felling productivity was 10.6 m3 E15h–1. For yarding, mean cycle time was 240 E15s and mean extraction distance 75.4 m, requiring 27 s for the outhaul and twice that for the inhaul under load, as can be seen in the simple regression on time for hauling-out empty and hauling-in under load (carriage speed 1.67 ms–1) in Fig. 4. At 42 s per load, un-choking was the second highest single time element after hauling in. Overall yarding productivity was 9.2 m3 E15h–1. Croat. j. for. eng. 35(2014)2

Productivity Analysis of an Un-Guyed Integrated Yarder-Processor with Running Skyline (201–210)

B. Talbot et al.

Table 3 Means and standard deviations (in parentheses) for all work elements and numerical variables measured in the field study Felling (n=217) Move to tree, s

18.3 (16.2)

Clear brush, s

Prepare, s

Yarding (n=149)

4.4 (6.6)

11.4 (15.9)

Processing (n=254)

Haul-out, s

27.2 (8.7)

Prepare, s

17.1 (14.8)

Haul distance, m

75.4 (28.7)

Lateral-out, s

23.2 (9.6)

Process logs, s

45.0 (25.4)

Lateral distance, m

6.5 (3.9)

Logs per tree, n

2.9 (1.4)

Choke, s

24.8 (14.7)

Residue handling, s

2.8 (2.7)

Trees per load, n*

2.27 (0.99) Stacking logs, s

17.6 (45.1)

Cut, s

35.9 (22.8)

Lateral-in**, s

35.5 (18.1)

Wedge, s

11.3 (8.6)

Haul-in, s

58.7 (24.9)

Un-choke, s

42.2 (11.2)

–1

Time tree , E0s

81.3 (70.2)

Time load , E0s

212 (59.8)

Time tree–1, E0s

82.5 (109.5)

Delay time, s

10.2 (56.3)

Delay time, s

27.8 (113)

Delay time, s

5.8 (5.3)

–1

Time tree , E15 s

91.5 (118.5)

–1

Trees, E0 hr 3

–1

44.3

–1

Time load , E15 s –1

Trees, E0 h 3

–1

240 (131.6) 38.6

–1

Time tree , E15s –1

Trees, E0 h 3

–1

88.3 (120.9) 43.6

Prod. m , E0 h

11.9

Prod. m , E0 h

10.4

Prod. m , E0 h

11.7

Trees, E15h–1

39.3

Trees, E15h–1

40.8

Prod. m3, E15h–1

10.6

Prod. m3, E15h–1

9.2

Prod. m3, E15h–1

10.9

* Movement between multiple trees during choking was accrued to lateral-out time ** Lateral-in is not a discrete element when winching with a running skyline as the load is hauled tangentially toward the yarder, and not to the corridor centerline first. In this study, lateral-in was used to record the break-out process, i.e. the time taken to get the load into motion toward the tower, thereby maintaining integrity of the distance based haul-in component

For processing, it took around 17 s to take hold of the tree, get it into position and dress the butt-end when necessary, and a further 45 s on average to process the logs. Sorting logs into the correct stacks was time consuming, adding another 18 s per tree. Processing productivity was relatively low at 10.9 m3 E15h–1, but highly dependent on tree and log size. Fig. 5 shows how the processing time per log is relatively constant, while the productivity in m3 E0h–1 decreases exponentially. Here the common preparation time per tree is distributed to the logs by their volume proportion. Time elements associated with processing, such as stacking and handling biomass, are not included in Fig. 5. Time consumption models were developed against effective time (ET) per unit. Various models were tested and those reported here were selected on their goodness of fit and F-value. Fig. 4 Carriage travel time as a function of distance, where haul-in is travelling under load toward the yarder, and haul-out is travelling empty out into the stand Croat. j. for. eng. 35(2014)2

3.1 Felling Only two independent variables could be included in the effective time consumption model for felling:

205

B. Talbot et al.

Productivity Analysis of an Un-Guyed Integrated Yarder-Processor with Running Skyline (201–210)

Fig. 5 Processing time (E0s) per log, by tree size category and log sequence in the stem. The lines represent logarithmic approximations of processing productivity rates achieved by tree size and log sequence, as read against the right hand vertical axis in m3 E0h–1is travelling under load toward the yarder, and haul-out is travelling empty out into the stand

TS, a categorical variable explaining the tree size classes and WDG, a binary variable indicating whether a wedge was used for directional felling or not. The general model to predict effective time consumption for felling per tree (ETfell) is given by equation 1, where b0 is the estimate of the intercept and b1–2 are the coefficients to be estimated. The model assumptions were checked using a full residual analysis: ETfell ~ b0 + b1TS + b2WDG + e

Regression results for the effective time consumption model for felling are reported in Table 4. This regression model produced R2 = 0.35, F (3,162)= =29.8, p<0.001. All independent variables were significant and positive confirming that the effective time to implement the felling operation increases with increasing tree sizes and with the use of the wedge. The low R2 is likely due to the fact that the moving distance between the trees was not recorded, but accounted for a relatively large part of the effective time.

(1)

Table 4 Regression model parameters for felling Coefficients

Standard error

t stat

P value

Intercept b0

43.09

4.89

8.80

<0.001***

Treesize 2 b1

16.17

7.28

2.22

<0.05**

Treesize 3 b1

42.34

12.01

3.52

<0.001***

Wedge (1) b2

41.59

7.28

5.71

<0.001***

R-squared

0.35

Adjusted R-squared

0.34

F-statistic No. observations

29.8 (on 6 and 162 DF)

<0.001

217

Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

206

Croat. j. for. eng. 35(2014)2

Productivity Analysis of an Un-Guyed Integrated Yarder-Processor with Running Skyline (201–210)

B. Talbot et al.

Table 5 Regression model parameters for yarding Coefficients

Standard error

t stat

P value

Intercept b0

73.6

12.78

5.76

<0.001***

Hauling Distance b1

1.17

0.11

10.35

<0.001***

Lateral Distance b2

1.61

0.82

1.97

<0.01.

Trees/Cycle b3

19.19

3.28

5.85

<0.001***

R-squared

0.56

Adjusted R-squared

0.55

F-statistic

51.59 (on 3 and 121 DF)

No. observations

149

<0.001

Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

3.2 Yarding The general model for predicting the effective time for yarding was given by equation 2 as: ETyard ~ b0 + b1HD + b2LatDist + b3TC + e

(2)

Where ETyard is the effective time for yarding, b0 estimates the intercept and b1–3 are the coefficients to be estimated, HD is the haul distance, LD is the lateral distance and TC is the number of trees per cycle. The results for this multiple linear regression model are given in Table 5. The regression model yields an adjusted R2 of 0.55, F (3,121)=51.59, p<0.001. Results show that the variable lateral distance was significant at a 10% level, while the other variables were all statistically significant at 1%. All coefficients were positive, confirming the positive correlation between time needed to perform the yarding task and the distances of the trees yarded from the tower and the line in addition to the number of tree parts of each load.

3.3 Processing The time consumption prediction model for processing that best fitted the data consisted of a categorical variable representing the tree size TS and the number of logs obtained for each tree, LOGN (equation 3). b0 is the estimate of the intercept and b1–2 are the coefficients to be estimated. ETproc ~ b0 + b1TS + b2LOGN + e

(3)

Analysis of the residual plots indicated no systematic pattern and the underlying assumptions for regression were supported. The coefficients, all significant, are also all positive following the expected result of an effective time increase with increasing tree sizes Croat. j. for. eng. 35(2014)2

and number of logs obtained per tree. Note that time for sorting and stacking logs, and handling biomass are not included in this model.

Fig. 6 Distribution of the system hour to yarding, and the slowest of processing/felling

3.4 System performance Table 3 showed the time consumption for each work phase individually. As the machine cannot yard and process simultaneously, system productivity is limited by the least productive work phase. Each system hour was made up of yarding (54%) and the slower of felling or processing (46%), which in this case are almost identical at E15 time (Fig. 6). The resultant system productivity was 4.9 m3 E15h-1. Relocation, rigging of the tail spar and corridor changes (it took roughly 2.5 h with 1 person) are not included in the E15h.

4. Discussion A fully integrated machine configuration such as this that cannot yard and process simultaneously is

207

B. Talbot et al.

Productivity Analysis of an Un-Guyed Integrated Yarder-Processor with Running Skyline (201–210)

Table 6 Regression model parameters for processing Intercept b0

Coefficients

Standard error

t stat

P value

19.07

3.31

5.94

<0.001***

Treesize 2 b1

6.39

3.407

1.88

<0.1.

Treesize 3 b1

30.26

6.07

4.98

<0.001***

Number of logs/tree b2

12.80

1.21

10.16

<0.001***

R-squared

0.52

Adjusted R-squared

0.51

F-statistic No. observations

91.5 (on 3 and 250 DF)

<0.001

254

Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

restricted by the weakest performing work phase. Felling productivity should largely coincide with processing to avoid operational delays. If unavoidable, it is preferable that delays befall the feller-chokersetter as that represents only c. 20% of the total system cost, and because such a delay implies rest time for this worker. As a yarder, the machine showed good performance rates. Cycle times were short, some 240 E15s on average (15 turns per hour), partially due to the short corridor lengths, limited lateral yarding, and the fact that the running skyline configuration makes the skidline immediately available to the chokersetter. The winch is powerful enough to yard larger loads than the 0.61 m3 observed, and this would significantly improve productivity, but the chokers were well utilized at 2.3 trees on average, meaning that larger loads should come from larger trees. Roughly 19% of the yarding cycle time was used for unchoking, and this required the operator to climb up and down from the cab frequently, not without some risk. Research suggests that the use of self-releasing chokers could be useful in a setting such as this (Stampfer et al. 2010). While felling and yarding are relatively effective, processing at 10.9 m3 E15 h-1 is considerably slower than that for single grip harvesters. Gerasimov et al. (2012) found mean productivities of 31.3 m3 E15h-1 during processing of trees of 0.16–0.30 m3 and 46.1 m3 per processing machine hour for stem volumes ranging between 0.31 and 0.5 m3, in a study including over 4 million trees. This is roughly 3 times higher than the processing productivity observed in this study. While processing is somewhat constrained by the working position and limited space available to the yarder, this considerable gap can likely only be explained by the operator and processing head, which might be better suited to larger trees found in central Europe. The operator was highly skilled on the Konrad Woody 60™ processing head, but the controls for the Zöggeler

208

head are configured differently and the operator might have required a longer period of adaption. An increase in processing speed would require another worker in the system as the feller appears to be working near maximum speed. However, a second fellerchokersetter would only be partly employed. An increase in mean tree size would likely provide the easiest path to increasing system productivity, especially with regard to processing. The system hour consisted of 54% yarding and 46% processing/felling. With their similar machine, Torgersen and Lisland (2002) found the opposite distribution of 41–59%, probably as yarding was carried out over longer distances and the processor was more rudimentary (i.e. stroke delimber). However, their results at 6.2 m3 E0 h-1 sorted at roadside, were similar with those presented in this study. Fig. 7 shows the influence of an increase or decrease in yarding or processing/felling productivity on overall system productivity. It illustrates how considerable increases in either or both dimensions are required in making marginal increases in system productivity. Rigging was generally handled by the feller alone. The machine operator used the time to clean up on the landing, and mark timber piles for different customers. Corridors were short (80–120 m), no intermediate supports were used, all corridors were for uphill yarding, and the low cable tensions during operations allowed for light equipment and limited efforts on tail spar rigging. Most of the 2.5 h rigging time involved felling the centreline, and so was productive. By comparison, Stampfer et al. (2006) show how a small tower yarder working under similar conditions would require roughly 5 h installation time with a crew of 2. In their study of two non-guyed yarders in Idaho, Largo et al. (2004) report corridor changes of as low as 30 minutes. Detailed system costs were not calculated, but estimates indicate required hourly prices of roughly Croat. j. for. eng. 35(2014)2

Productivity Analysis of an Un-Guyed Integrated Yarder-Processor with Running Skyline (201–210)

B. Talbot et al.

Fig. 7 Sensitivity analysis showing potential system productivity in relation to relative increases or decreases in yarding or processing/felling productivity US$ 200 for the machine and operator, US$ 45 for the feller-chokersetter, including the chainsaw and all social on-costs (1 US$=6.12 NOK or 0.74 EUR). At a cost of roughly US$ 245 and a productivity of 4.9 m3 E15h–1, the system is some way from being profitable in the present application, but is currently applied in areas with subsidies for special harvesting conditions. System productivity would need to be raised by 30–50% or the capital outlay reduced, to make the machine competitive in the free market. Opportunities for achieving this might include using a cheaper, reconditioned base machine, deploying the machine in stands with larger mean tree sizes, increasing operator productivity in processing through training and simplifying the somewhat complex number of assortments made.

5. Conclusions The single machine system works well in terms of balance with a 2-man crew, but system productivity remains too low. The simplest method of increasing productivity while maintaining balance would be to deploy the machine in stands with slightly larger trees. Processing rate was approximately 30% of that of a single grip harvester in similar tree sizes, and is the main bottleneck to increased system performance. A higher processing rate would result in the need for a second worker in the field, as the feller already works at or near the maximum rate. With two workers inCroat. j. for. eng. 35(2014)2

field, neither would be fully employed. While this may still be economically beneficial, even given the high cost of workers in Norway, an important motivation for purchasing this system was the fact that it could be operated by a 2-man team. To fully understand the potential of this interesting machine concept, more studies under varying conditions would be required. A full system analysis would also be required considering the costs, workload and productivity of a second man in the field and the separation of the yarding and processing functionality to two base machines.

Acknowledgements The authors wish to acknowledge the financial support received from the Research Council of Norway (projects ES500223 and 225329), and project owner Mjøsen Skog SA. Thanks also to contractor T. Frivik Taubanedrift AS for permitting the study of their machine, the manufacturer, Zöggeler Forsttechnik for an open process of cooperation, and forestry students Julian della Pietra (BOKU, Vienna), and Even Hoffart (NMBU, Ås) for assistance with the fieldwork.

6. References Devlin, G., Klvac, R., 2013: Opportunities for developing excavator based cable logging operations in Ireland – A pro-

209

B. Talbot et al.

Productivity Analysis of an Un-Guyed Integrated Yarder-Processor with Running Skyline (201–210)

ductivity analysis. In: Talbot, B. & Berkett, H., editors. Proceedings of the IUFRO unit 3.06 Conference on Forest Operations in Mountainous Conditions; June 2–5; Honne, Norway; 58–61. Gerasimov, Y., Senkin, V., Väätäinen, K., 2012: Productivity of single-grip harvesters in clear-cutting operations in the northern European part of Russia. Eur J Forest Res 131(3): 647–654. Ghaffariyan, M. R., Stampfer, K., Sessions, J., 2009: Production Equations for Tower Yarders in Austria. Int J For Eng. 20(1): 17–21. Gingras, J-F., 2013: Update on steep slopes operations research at FP-Innovations in Canada. In: Talbot, B. & Berkett, H., editors. Proceedings of the IUFRO unit 3.06 Conference on Forest Operations in Mountainous Conditions; June 2–5; Honne, Norway; 15–17. Johansson, J., 1997: Earth-Moving Equipment as Base Machines in Forest Work: Final Report of an NSR Project (NSR 37/93). SLU / Dept. of operational efficiency Research Note No. 294, 1–75. Largo, S., Han H-S., Johnson, L., 2004: Productivity and Cost Evaluation for Non-guyline Yarders in Northern Idaho. In: Proceedings of the Council on Forest Engineering (COFE) conference – Machines and People, The Interface; April 27–30; Hot Springs, Arkansas; 1–6. Larsson, J. Y., Hylen, G., 2007: Skogen i Norge. Statistikk over skogforhold og skogressurser i Norge registrert i perioden 2000-2004 [Statistics of Forest Conditions and Forest Resources in Norway]. Viten fra Skog og landskap 1/07. Norwegian Forest and Landscape Institute. ISBN 978-82-311-0006-5. In Norwegian with English summary. 1–91. McEwan, A., Brink, M., van Zyl, R., 2013: Guidelines for difficult terrain ground-based harvesting operations in South

Received: March 15, 2014 Accepted: June 13, 2014

210

Africa. In: Talbot, B. & Berkett, H., editors. Proceedings of the IUFRO unit 3.06 Conference on Forest Operations in Mountainous Conditions; June 2–5; Honne, Norway; 27–28. Spinelli, R., Magagnotti, N., Lombardini, C., 2010: Performance, Capability and Costs of Small-Scale Cable Yarding Technology. Small-Scale Forestry 9(1): 123–135. Stampfer, K., Leitner, T., Visser, R., 2010: Efficiency and Ergonomic Benefits of Using Radio Controlled Chokers in Cable Yarding. Croat.j.for. eng. 31(1): 1–9. Stampfer, K., Visser, R., Kanzian, C., 2006: Cable Corridor Installation Times For European Yarders. Int J For Eng. 17(2): 71–77. Torgersen, H., Lisland, T., 2002: Excavator-Based Cable Logging and Processing System: A Norwegian Case Study. Int J For Eng. 13(1): 11–16. Tuer, K., Saunders, C., MacIntosh, G., 2013: Steep ground harvesting project – Forestry Commission Scotland: Evaluation, Innovation and Development in Scottish Skyline Operations. In: Talbot, B. & Berkett, H., editors. Proceedings of the IUFRO unit 3.06 Conference on Forest Operations in Mountainous Conditions; June 2–5; Honne, Norway; 21–23. Tveite B., 1977: Bonitetskurver for gran [Site index curves for spruce]. Meddelelser fra Norsk institutt for skogforsking 33.1: 1–84. In Norwegian. Yoshimura, T., Noba, T., 2013: Productivity analysis of thinning operation using a swing yarder on steep slopes in western Japan. In: Talbot, B. & Berkett, H., editors. Proceedings of the IUFRO unit 3.06 Conference on Forest Operations in Mountainous Conditions; June 2–5; Honne, Norway; 35–36. Zimbalatti, G., Proto, A. R., 2009: Cable logging opportunities for firewood in Calabrian forests. Biosyst Eng 102(1): 63–68.

Authors’ address: Bruce Talbot, PhD.* e-mail: bta@skogoglandskap.no Giovanna Ottaviani Aalmo, PhD. e-mail: gio@skogoglandskap.no Norwegian Institute for Forest and Landscape Section for Forest Technology and Economics P.O. Box 115 1431 Ås NORWAY Prof. Karl Stampfer, PhD. e-mail: karl.stampfer@boku.ac.at University of Natural Resources and Applied Life Sciences Vienna Department of Forest and Soil Sciences Institute of Forest Engineering Peter Jordan Straße 82/3 1190 Vienna AUSTRIA * Corresponding author Croat. j. for. eng. 35(2014)2

Original scientific paper

Theoretical Potentials of Forwarder Trailers with and without Axle Load Restrictions Ola Lindroos, Iwan WĂ¤sterlund Abstract In mechanized ground-based forestry, machines operate on rough soils that, ideally, should remain unaffected by the operation. This implies small (that is, light) loads and careful driving are required. However, economical rationality implies large loads and high speeds. Recently, the concept of adding a trailer to a conventional forwarder has been revived, with the objective of addressing both concerns, and fitting into the current, mechanized, cut-to-length system. Here we present the theoretical benefits of the forwarder-trailer concept compared to conventional forwarding for final-felling operations. The analysis addresses the trailer potential in terms of break-even extraction distances under different scenarios, and estimates the abundance of favorable conditions (as a percentage of final-felling volume) in Swedish final fellings. The results show that the forwarder-trailer concept has potential to reduce costs, and especially if there are restrictions on axle loads. However, the viability of the trailer concept is highly sensitive to changes in the increased purchase costs and the increased work-element time-consumption. That is, small changes in these variables result in large changes in viability. In the scenarios presented here, the increase in time consumption was more influential than the purchase cost. It can be concluded that there are potential economic and possibly also environmental benefits that warrant further investigation of the forwarder-trailer concept, which is currently being evaluated in practice in Sweden. Keywords: forwarder, ground pressure, productivity, cost-efficiency, fuel consumption, theoretical potentials, comparative study

1. Introduction In mechanized forestry, machines operate on rough soils that, ideally, should remain unaffected by the operation. However, large masses such as trees and logs are handled and, thus, the machines are often heavy. Machine masses are especially high in the work of transporting trees or logs from the terrain to roadside landing points, as it is generally time- and cost-effective to maximize payloads as well as transport speed. Thus, there is generally a conflict between minimizing soil disturbance and maximizing operational efficiency. A cause of soil damage is the year-round harvesting employed to supply industry with timber. The increased frequency of rainy periods and the reduction in frozen ground in northern Europe expected as a result of ongoing climate change will effect forest operations (Goltsev and Lopatin 2013) by increasing soil moisture content and reducing its bearing capacity Croat. j. for. eng. 35(2014)2

and tensile strength. Good planning before harvesting should steer the operation towards better areas, but heavy rains can alter conditions very fast. Thus, even with good planning, the axle loads may become too heavy for the machinery used and there may be too little traction for the soil characteristics, which would create deep rutting and soil compaction (WĂ¤sterlund 1992, Nadezhdina et al. 2006, SirĂŠn et al. 2013). Lately, there have been increased concerns about soil damage from harvesting operations, resulting in restrictions on where machines can travel and on their ground pressure (Horn et al. 2004, 2007). As ground pressure is the product of the force applied and the size of the supporting area, a decrease in ground pressure can be achieved by decreasing the force (i.e. axle loads) and increasing the supporting area, either separately or in conjunction. This implies that the total mass of the machines should be decreased, which can be achieved with a decreased payload for a given machine, or with

211

O. Lindroos and I. Wästerlund

Theoretical Potentials of Forwarder Trailers with and without Axle ... (211–219)

a constant payload but with a lighter machine. The supporting area can also be increased, by increasing the wheel diameter and widths, using bogie tracks, or even by adding extra axles/wheels. An alternative to the modification of currently used machines is to introduce new machine concepts. For instance, it has previously been suggested that forwarders could be equipped with a semi-trailer to increase the total payload transported and tests have indicated that the use of semi-trailers is cost-effective if transport distances are long (Eriksson 1998). However, the cost-effectiveness is very sensitive to the level of the increased cost implied by using the semitrailer compared to only using a conventional forwarder. From here on, the term trailer will be used for all vehicles with increased load-space that can be attached to a forwarder, irrespective of type and of the fact whether they are powered or not (for example, including semi-trailers). The use of a trailer can be reconsidered despite the fact that this has previously been found to be not economically viable compared to a conventional forwarder (Eriksson 1998). First, with potential restrictions on maximum ground pressure during forwarding, conventional forwarders might not be able to fully use their load capacity, which would increase the cost per transported unit for such a conventional system. Second, trailers admit larger payloads that can be distributed on additional axles and larger supporting areas than with conventional forwarders. Hence, trailers might admit larger loads with decreased ground pressure. Moreover, one of the previously found limitations of the use of trailers was insufficient crane capacity, resulting in decreased efficiency when operating at the required, full crane reach during trailer loading and unloading (Eriksson 1998); whereas technical developments have resulted in more powerful cranes (Nordfjell et al. 2010). Some trailer solutions are already available on the market in both the Northern and Southern hemispheres (for example, Timbear Lightlogg C (Timbear 2011) and Bell’s long range forwarders (Bell 2010), respectively). New inventions also circumvent the need to work at long crane reaches, by having a trailer reversing parallel with the forwarder during loading, powered by the forwarder engine (Volungholen 2008). Thus, there are both environmental and technical reasons to re-evaluate the forwardertrailer concept. The objective of the study was to analyze the potential benefits of forwarder trailers in terms of time consumption, cost-efficiency, and fuel consumption compared to conventional forwarders, with and without axle load restrictions. The restrictions were moti-

212

vated by the assumption that increased axle loads may increase soil damage (Håkansson 1994, Jansson and Johansson 1998), and that environmental concerns might eventually result in such restrictions. Thus, our evaluation addresses whether or not it would be more efficient to just reduce payloads on normal forwarders, or to use forwarder trailers. However, the restrictions are complementing and motivating the analysis, but to fully evaluate the possible machine-soil interactions when using a forwarder trailer is not within the scope of the study. The analyses were conducted by use of theoretical modelling to identify stand conditions in which the use of trailers may be viable compared with conventional forwarders in final felling. Moreover, the abundance of Swedish final fellings with favorable conditions for forwarder trailers was assessed.

2. Materials and methods To fully evaluate the impact of the examined forwarding concepts, two general assumptions were made concerning the similarity of concepts. First, it was assumed that the outcomes of work were identical in terms of effect on the roundwood transported and unloaded at roadside landings. However, the impact on stand environment (for example, rutting and soil compaction) might vary, but is only taken into account here in terms of analyzing various measures to decrease axle loads. Second, it was assumed that it generally takes the same amount of time for the same type of work, but when differences are expected the employed methodology allows the parameters to be adjusted appropriately. Aggregated machine time-consumption functions are first presented for productive machine (PM) time in minutes per produced solid m3 of roundwood under bark (PMmin m–3), which is defined as the delayfree machine time that directly contributes to the completion of the intended work task (cf. Björheden 1991). Then, the level of technical utilization is included, giving the time consumption per scheduled machine (SM) time in minutes per produced m3 (SMmin m–3). Finally, costs per m3 are calculated based on scheduled-machine time-consumption. Costs were calculated in Swedish crowns (SEK), and converted to euros (€) using an average exchange rate of 10 SEK=1 € during 2010 (Sweden’s Central Bank 2011). 1 m3 of wood was assumed to have a mass of 900 kg.

2.1 Machine combinations and scenarios All forwarders included in the study were assumed to be eight-wheelers, with tracks on all four bogies. The trailer was four-wheeled, with tracks on both bogies. In Croat. j. for. eng. 35(2014)2

Theoretical Potentials of Forwarder Trailers with and without Axle ... (211–219)

the comparisons, we assumed that the forwarder trailer was combined with a medium-sized forwarder with a reduced payload. The performance of that combination was compared with the performance of medium-sized and large forwarders with full payload and reduced payload, respectively, resulting in five different machine combinations in our study. Details of the combinations are presented below. To include uncertainties in the cost and performance of a forwarder trailer, four scenarios were evaluated to cover the expected speed and price ranges. For time consumption, the fast scenario assumed that the use of a trailer increased the time required for all of the medium forwarder work elements by 5%. In the slow scenario, the use of a trailer was assumed to require 10% more time, plus an additional extra PM minute per load to account for eventual arrangements required for the loading and unloading of the trailer (for example, turning the trailer (Volungholen 2008) or adjusting the distance between the trailer and forwarder). In the cheap trailer price scenario, it was assumed that the trailer price was 30,000€ (10% of the cost of a medium forwarder), whereas a trailer in the expensive scenario was assumed to cost 70,000 € (23.3% of a medium-sized forwarder). Altogether, the scenarios were: fast-cheap, fast-expensive, slow-cheap, and slow-expensive. Thus, in total the four forwarder trailer scenarios were compared with the four normal forwarder combinations.

2.2 Estimation of time consumption Total forwarding time-consumption for a given machine was computed as: TTotal = TDriving,Empty + TDriving,Full + TLoading + , PMmin m–3 (1) TDriving,Loading + TUnloading where: TDriving, Empty TDriving, Full TLoading TDriving, Loading TUnloading

time consumption of pure driving when empty (that is, from roadside landing and until loading starts), pure driving with full payload, loading time, pure driving when loading, and unloading time.

Time consumption for the work elements was based on equations provided by Nurminen et al. (2006) for loads with several assortments:

TDirving,Empty

 V ´ l  Max  0 , dm − F r  2VR   = , PMmin m–3 vE ´ VF

Croat. j. for. eng. 35(2014)2

(2)

TDirving,Full

O. Lindroos and I. Wästerlund

 V ´ l  Max  0 , dm − F r  2VR   = , PMmin m–3 vF ´ VF

TDirving,Loading =

TLoading = 1 +

lr vL ´ VR

(3)

, PMmin m–3(4)

0.155   100VR   Exp  −0.447 + 0.3 ´ Ln    lr   F  PMmin m–3

(5)

where: VR abundance of loaded assortment(s), m3 ha–1, lr total length of strip road network, m ha–1, VF (full) forwarder load volume (payload), vE average speed when driving empty, m min–1, vF average speed when driving full, m min–1, vL average speed while loading, m min–1, dm mean extraction distance one way, m. The total strip-road length (lr) was set to 769 m, based on the assumption that there would be 13 m between roads in final felling (cf. Nurminen et al. 2006). It was assumed that all assortments were loaded together, and thus, VR was equal to the stand density. Moreover, it was assumed that the distance driven loaded was equal to the distance driven unloaded. Given the assumptions, the distances driven full and unloaded could be estimated to be negative for small values of VR and dm, and for large values of VF; hence, the use of the max function in equations 2 and 3. In practice, TUnloading varies depending on the mixture of assortments in loads. Although loads of different mixtures can be created, load mixtures were here assumed to be identical for all machine combinations. TUnloading was set to 0.657 PMmin m–3; the highest mean value in the range, 0.547–0.657 PMmin m–3, suggested by Nurminen et al. (2006). For the medium-sized forwarder, vE, vF, vL, and TUnloading were set to 56 m min–1, 44 m min–1, 27 m min–1, and 0.657 PMmin m–3, respectively (cf. Nurminen et al. 2006). To accommodate for larger engine and grapple, the large forwarder was assumed to be slightly faster, with the corresponding driving speed values being set to 58 m min–1, 46 m min–1, and 27 m min–1, respectively, while TLoading and TUnloading were taken to be 5% less than for the medium-sized forwarder.

213

O. Lindroos and I. Wästerlund

Theoretical Potentials of Forwarder Trailers with and without Axle ... (211–219)

2.4 Fuel consumption

Technical utilization was set to 90% for all machine combinations, that is, PM-time was transformed to SM-time by dividing it by 0.9. It was assumed that the reduction of payload did not affect the time consumption per m3.

It was assumed that, owing to the hydrostatic-mechanical transmission of forwarders, each engine would work at a given number of revolutions per minute, with the forwarder speed reduced for heavier loads. Thus, the fuel consumption was estimated as:

2.3 Costs

where C is the fuel consumption and W is the engine output power in kW (Klvac and Skoupy 2009). The reduction of loads was assumed not to change the fuel consumption, whereas the use of a trailer was assumed to increase the fuel consumption by 5%. The estimated fuel consumption is shown in Table 2.

2.5 Axle load and ground pressure For each axle of a machine combination, the axle load was calculated based on the machine mass and the number of axles to distribute the load on. It was assumed that 40% of the mass was on the front axle when a forwarder was loaded. Valmet 860 and 890 (Komatsu Forest, Umeå, Sweden) were used as model machines for the calculations, and they were assumed to be equipped with 0.81 m wide ECO-Track bogie tracks (Olofsfors AB, Olofsfors, Sweden) each weighing 895 kg. The trailer was assumed to have a mass of 7 t including bogie tracks, and being loaded with 9.5 t. The calculated axle load pressures are presented in Table 2. To give a rough estimate of the ground pressure, the axle load can be divided by area covered by the axle bogie bands (ca. 1.2 m2 for each of the two bogie bands on an axle). The restriction on the medium sized forwarder axle loads were set to approximately reflect ground pressures of maximum 70 MPa (Wästerlund 1992). However, if the large forwarder should meet the same restriction, it should have a payload of only 8 t (44% of full payload). This

Table 1 Costs for the machine sizes and the trailer combination scenarios Purchase cost

Hourly cost

10 €

€ SMh–1

Large

500

95.84

Medium

400

85.55

Cheap

430

88.46

Expensive

470

91.39

Forwarder combination

(6)

C = 0.046W + 7.222 , l PMh–1

Fixed costs for the machines were calculated according to Miyata (1980), applying the straight line method of depreciation and an approximate annuity method for interest. For all machine combinations, the interest rate was set to 6.5%, the expected service life was set to 6 years with 2600 scheduled hours per year, and the salvage value was taken to be 10% of the purchase cost. The labor cost was set to 37.8 € SMh–1. Operating costs excluding fuel were set to 13.0 and 14.3 € SMh–1 for the medium-sized and large forwarder, respectively. The fuel cost was set to 1.1 € per liter and the hourly cost for fuel depended on fuel consumption (and hence on the engine size, see section 2.4). The total hourly costs for the machine combinations are given in Table 1.

Medium+trailer

Table 2 Machine parameters

Forwarder combination

Acronym

Engine output power

Fuel consumption

Ratio of payload to unloaded

l PMh–1

Total

Payload

mass

Front

Back

Trailer

Mass, t

Axle load, t

Large

190

16.0

38.0

18.0

0.90

15.2

22.8

–

Large – reduced

190

16.0

33.0

13.0

0.65

13.2

19.8

–

Medium

150

14.2

31.0

14.0

0.82

12.4

18.6

–

Medium reduced

150

14.2

28.9

11.9

0.70

11.6

17.3

–

Medium reduced+trailer

MRT

150

14.9

45.4

21.4

0.89

11.6

17.3

16.5

214

Croat. j. for. eng. 35(2014)2

Theoretical Potentials of Forwarder Trailers with and without Axle ... (211–219)

was considered unrealistic, so its restriction was set to reflect ground pressures of maximum 80 MPa.

2.6 Stand data Follow-up data for finally felled stands harvested by conventional systems were gathered from forestry companies for three regions of Sweden: Northern (Norrbotten, ca. 66° N, 22° E), Central (Medelpad, ca. 62° N, 16° E), and Southern (Östergötland-Sörmland, ca. 58° N, 16° E). For each stand, these data included information on the stand volume (m3), stand density (m3 ha–1), mean harvested stem size (m3), and mean extraction distance one way (m) (Table 3). The time-consumption functions used here were not adapted to stands with densities less than 100 and more than 1,000 m3 ha–1, and such stands were therefore excluded. This resulted in the exclusion of 7.5% of the harvested volume from the pooled, original data. Stands with more than 1,000 m3 ha–1 corresponded to 0.6% of the pooled data and only occurred in the Southern dataset. The dataset used contained ca. 1.6 million m3.

2.7 Data analysis A deterministic, spreadsheet based model was constructed based on the abovementioned equations and assumptions for time and fuel consumptions as well as costs. In the analysis of favorable conditions, the model was used to systematically investigate the effects of various levels of extraction distances, stand

O. Lindroos and I. Wästerlund

Table 3 Characteristics of 1 129 stands (containing 1 624 004 m3) included in the follow-up dataset of Swedish final fellings Volumeweighted mean

Range

Mean stem size, m3

0.41

0.05–2.78

Mean extraction distance, m

389

20–1500

Mean stand density, m3 ha–1

250

100–952

Variable

volumes and stand densities. Subsequently, the model was applied to the stand data set, in order to investigate the abundance of favorable conditions. Thus, the former step aimed at finding the conditions where the forwarder trailer should be competitive. The latter step indicated how common such trailer favorable conditions were, based on a large sample of conditions occurring in Sweden.

3. Results 3.1 Favorable conditions Compared to the extraction distance, the stand density had only minor effects on the time and fuel consumption of the machine combinations. The effects were largest at small stand densities; the time consumption per m3 was ca. 5–6% higher with a density of 50 m3 ha–1 than with a density of 100 m3 ha–1, and

Fig. 1 Time (left panel) and fuel consumption (right panel) in final fellings (at 250 m3 ha–1) as a function of extraction distance for two payload scenarios, for large and medium-sized conventional forwarders, and for the forwarder-trailer combination (MRT) in the two speed scenarios Croat. j. for. eng. 35(2014)2

215

O. Lindroos and I. Wästerlund

Theoretical Potentials of Forwarder Trailers with and without Axle ... (211–219)

was 5–6% higher at 100 m3 ha–1 than at 500 m3 ha–1 (data not shown). Looking at costs, the effects were even less distinguishable. For instance, the distance at which it was equally expensive to use the trailer combination as it was to use the fully loaded large forwarder only marginally varied over densities from 50 m3 ha–1 to 500 m3 ha–1 (the break-even distance was between 750 and 760 m; data not shown). Thus, further analyses of favorable conditions focused on the influence of extraction distance at a given stand density (namely, 250 m3 ha–1, that is, the mean of the stand data). As could be expected, the longer the extraction distance, the more time and fuel were consumed and the higher the costs were (Figs. 1 and 2). The fully loaded medium-sized forwarder was cheaper than both the large forwarder and all trailer scenarios at distances shorter than ca. 200 m but was more expensive at distances longer than ca. 650 m (Fig. 2). The large forwarder was cheaper than the fast-cheap trailer scenario only at short distances (less than ca. 150 m) whereas it was cheaper than the slow-expensive trailer scenario for all tested distances. The costs of the payload-reduced forwarders were very close to each other, and were less than those of all trailer scenarios for distances less than ca. 80–150 m but the payloadreduced forwarders were more expensive for distances longer than 300–400 m.

Fig. 2 Forwarding cost for final fellings (250 m3 ha-1) as a function of one way distance for the four conventional forwarder-payload scenarios and the four forwarder-trailer (MRT) speed-cost scenarios (exp.=expensive)

216

3.2 Abundance of favourable conditions The proportion of the final-felling volume for which the trailer combination was cheaper to use than the conventional forwarders varied considerably across the speed-price scenarios; from 10 to 79% and from 0 to 98% for the medium-sized and large fully loaded forwarders, respectively (Fig. 3). The variation across speed-price scenarios persisted for the payload-reduced forwarders but was somewhat less dramatic; the trailer combination was cheaper than the medium-sized forwarder for at least 52% of the volume, and for at least 78% of the volume when compared to the large forwarder (Fig. 3). On the total volume of final fellings, the fast-trailer combination was always cheaper to use than a conventional forwarder with either a full or reduced payload, irrespective of trailer cost (Fig. 4). In the slow scenario, the trailer was at least 6.4% more expensive to use than the conventional, fully loaded forwarders. Compared to a fully loaded conventional forwarder, the trailer combination generally consumed less time and fuel (up to 8% less), except for in the slow scenario, in which fuel consumption was 3% higher. Conversely, the trailer combination generally consumed more time and fuel (up to 11.6% more) than a fully loaded large forwarder, except in the fast scenario, in which fuel consumption was decreased by 2.5%.

Fig. 3 Proportion of total final-felling volume in four speed-price scenarios for which the forwarder-trailer combination (MRT) was cheaper to use than each of the four conventional forwarder-payload scenarios (M/MR=medium-sized forwarder with full/reduced payload; L/LR=large forwarder with full/reduced payload) Croat. j. for. eng. 35(2014)2

Theoretical Potentials of Forwarder Trailers with and without Axle ... (211–219)

O. Lindroos and I. Wästerlund

Fig. 4 Relative differences in costs (across speed-price scenarios, left panel) and consumption of time and fuel (across speed scenarios, right panel) between the forwarder-trailer combination (MRT) and conventional forwarder combinations when applied to the total final-felling volume. Negative values indicate that MRT was cheaper or less time- or fuel-consuming. Horizontal lines indicate the increased levels caused by payload reduction on conventional forwarders. M/MR=medium-sized forwarder with full/reduced payload; L/LR=large forwarder with full/reduced payload In comparison with the payload-reduced conventional forwarders, all the trailer speed-price scenarios were viable in terms of lowering total costs (1.5–13.6% cheaper), time consumption (0.2–13.5% faster), and fuel consumption (a 3.2–12.7% reduction). Since costs and fuel consumption were directly dependent on time consumption for a conventional forwarder, the reduction in payload resulted in 6.4% higher costs, time, and fuel consumption than when using full payloads with the medium-sized forwarder. The corresponding increase was 11.8% for the large forwarder.

4. Discussion 4.1 Results All together, there seems to be a substantial theoretical potential for the forwarder-trailer concept, especially so when restrictions on conventional forwarders axle load (that is, reduced payloads) apply. The cost of reducing axle load was considerably cheaper with the trailer combination (between 4.9% lower and 6.4% higher than the conventional cost) than when reducing the conventional forwarders payload (Fig. 4). In half of the speed-cost scenarios, the use of a trailer combination resulted in reduction of both axle load and costs, even when compared to fully loaded conventional forwarders (Fig. 4). However, the analysis indicates that the level of increased purchase costs and work-element time-consumption are crucial for the Croat. j. for. eng. 35(2014)2

trailer concept viability when competing with conventional full-payload forwarders. Apparently, even rather small alterations in these levels result in large changes in the viability (for example, in the abundance of suitable conditions, Fig. 3). In the scenarios considered here, the time-consumption increase (fast vs. slow) was more influential than the purchase cost (cheap vs. expensive). The current results agree with previous field studies (Eriksson 1998) in which a trailer combination was viable for extraction distances longer than 300 m under the assumption that the trailer only resulted in a higher hourly cost but not in increased time-consumption. When the observed increased time requirements were included in the calculation, the distance had to be at least 850 m for the trailer to be viable. The studies differ in terms of hourly costs for the trailer, since Eriksson (1998) assumed that the trailer would be used for only 50% of the work time (implying a higher fixed cost), whereas here it was assumed that the trailer would be used throughout the work time. However, assuming that the trailer is easy to attach and detach, it would be possible to use the trailer only on the most suitable harvesting sites. Indeed, the forwarder trailer combination would have the same hourly fixed cost irrespective of whether or not the trailer was used, but would have a lower fuel cost and allow faster work. An intelligent selection of where to use the trailer, and where not, could slightly improve the trailer combina-

217

O. Lindroos and I. WĂ¤sterlund

Theoretical Potentials of Forwarder Trailers with and without Axle ... (211â&#x20AC;&#x201C;219)

tion viability. The estimation of such breakpoints was beyond the scope of this study, but would be of interest in studies based on empirical data.

4.2 Strengths and weaknesses of the study In the methodology applied, the principal differences between forwarder combinations were addressed theoretically. Hence, the risk of confounding the effects of differences with noise intrinsically present in field studies was avoided. For instance, the influence of variations in the work environment, technical maturity, and operator influences did not affect the analysis. Moreover, this kind of theoretical approach enables analysis of machine concepts even when they are merely ideas (e.g. JundĂŠn et al. 2013). Thus, this kind of analysis is beneficial for technological development because it can be used for early evaluations of, and subsequent concentration of resources on, systems with the highest theoretical potentials (Lindroos 2012). However, all theoretical analyses are intrinsically dependent on the constructed models and input data used. Clearly, it is important to rigorously construct logical, realistic theoretical models and carefully evaluate the influence of variations in input levels and assumptions. In this study, these requirements were met by basing the model on generic forwarding models, with appropriate adjustments, and by addressing various scenarios to cover the uncertainties in time consumption and prices. Moreover, time-consumption differences were mainly expressed in relation to each other. Hence, changes in variables that are likely to affect all systems were also changed accordingly. This should minimize the risks of confounding differences between machine combinations with those related to low quality of available input data for the combinations (for example, unrelated and, thus, unharmonized data). In the analysis, it was assumed that all combinations load similarly (for example, the same number and proportion of assortments are used for each combination), although there is the possibility of using different load mixtures for the different load spaces. This would alter the work time for loading, driving while loading, and unloading, with a general trade-off in time saved between work related to loading and to unloading (Manner et al. 2013). Hence, estimating the effects of this would not be straightforward and, moreover, the number of load mixtures grows rapidly as the number of assortments increases. Thus, load mixtures were excluded here for the sake of simplicity, but should be of interest to address in future studies. However, it is most likely that the time-consumption scenarios used here do not favor the trailer. One might argue that the increased payload volume should enable the trailer to take additional assort-

218

ments and thereby decrease the loading time. However, despite the theoretical potential it is likely that there would be practical limitations in terms of crane capacity, as pointed out previously (Eriksson 1998). In practice, it might turn out that only small-sized wood (for example, pulpwood) could be loaded onto the trailer. Under the assumption that the trailer has to be loaded first, since loading the bunk first would severely impede on the visibility in trailer loading, the practical loading possibilities might be considerably reduced and this was therefore not addressed here. Additional possible practical limitations that were not considered here are whether or not the conventional forwarders have to be modified in order to be capable of pulling the trailer. Although this might require only minor modifications, it is likely to increase the purchase cost. Although not specifically addressed here, the assumed increased price for the trailer combination includes the cost of the trailer and of the forwarder modifications, up to the price levels specified in the scenarios. Even though the viability of the trailer combination was analyzed mainly for Nordic conditions, it is very likely that the relationships found between factors also apply under other conditions. However, the specific outcome in terms of favorable conditions and their abundance is intrinsically site-specific and would have to be assessed for any given location under consideration. Thus, future studies should focus on analysis at an enhanced level of detail and/or applications and under other environmental settings. The trailer combination might, for instance, be of interest in the recovery of logging residues. It would also be of interest to go forward with field studies to gather contemporary empirical data on costs and time consumption. Some trailer prototypes are already in use in Sweden, and will be subject to field studies.

5. Conclusions Based on this theoretical analysis, it can be concluded that there are potential economic and possibly also environmental benefits that warrant further investigation of the forwarder-trailer concept. Prototypes are already being tested in practice, which will contribute to such investigations by providing empirical data on practical limitations and actual costs and time-consumption.

Acknowledgements Rolf Volungholen and Lennart Hult are kindly thanked for valuable and creative discussions on the topic. Sees-Editing Ltd is also thanked for revising the English. Croat. j. for. eng. 35(2014)2

Theoretical Potentials of Forwarder Trailers with and without Axle ... (211–219)

6. References Bell, 2010: Bell Timber haulers and long range forwarders; T403 & T302. Bell Equipment. 4 p. Available at: http://www. bellequipment.com, under Products-Forestry-Bell Timbertrucks. Björheden, R., 1991: Basic time concepts for international comparisons of time study reports. Journal of Forest Engineering 2(2): 33–39. Eriksson, P., 1998. Påhängsvagn för skotare (Semi-trailers for forwarders: potential exists for long extraction distances). Resultat 25. Skogforsk. Uppsala. 4 p. (In Swedish with English summary). Goltsev, V., Lopatin, E., 2013: The impact of climate change on the technical accessibility of forests in the Tikhvin District of the Leningrad Region of Russia. International Journal of Forest Engineering 24(2): 148–160. Håkansson, I., 1994: Subsoil compaction caused by heavy vehicles—a long-term threat to soil productivity. Soil and Tillage Research 29(2–3): 105–110. Horn, R., Vossbrink, J., Becker, S., 2004: Modern forestry vehicles and their impacts on soil physical properties. Soil and Tillage Research 79(2): 207–219. Horn, R., Vossbrink, J., Peth, S., Becker, S., 2007: Impact of modern forest vehicles on soil physical properties. Forest Ecology and Management 248(1–2): 56–63. Jansson, K.-J., Johansson, J., 1998: Soil changes after traffic with a tracked and a wheeled forest machine: a case study on a silt loam in Sweden. Forestry 71(1): 57–66. Jundén, L., Bergström, D., Servin, M., Bergsten, U., 2013: Simulation of boom-corridor thinning using a double-crane system and different levels of automation. International Journal of Forest Engineering 24(1): 16–23. Klvac, R., Skoupy, A., 2009: Characteristic fuel consumption and exhaust emissions in fully mechanized logging operations. Journal of Forest Research 14(6): 328–334.

O. Lindroos and I. Wästerlund

Lindroos, O., 2012: Evaluation of technical and organizational approaches for directly loading logs in mechanized CTL harvesting. Forest Science 58(4):326–341. Manner, J., Nordfjell, T., Lindroos, O., 2013: Effects of the number of assortments and log concentration on time consumption for forwarding. Silva Fennica 47(4): article id 1030. Miyata, E. S., 1980: Determining fixed and operating costs of logging equipment. General Technical Report GTR-NC-55. USDA Forest Service, North Central Forest Experiment Station. St. Paul, MN. 20 p. Nadezhdina, N., Čermák, J., Neruda, J., Prax, A., Ulrich, R., Nadezhdin, V., Gašpárek, J., Pokorný, E., 2006: Roots under the load of heavy machinery in spruce trees. European Journal of Forest Research 125(2): 111–128. Nordfjell, T., Björheden, R., Thor, M., Wästerlund, I., 2010: Changes in technical performance, mechanical availability and prices of machines used in forest operations in Sweden from 1985 to 2010. Scandinavian Journal of Forest Research 25(4): 382–389. Nurminen, T., Korpunen, H., Uusitalo, J., 2006: Time consumption analysis of the mechanized cut-to-length harvesting system. Silva Fennica 40(2):335–363. Sirén, M., Ala-Ilomäki, J., Mäkinen, H., Lamminen, S., Mikkola, T., 2013: Harvesting damage caused by thinning of Norway spruce in unfrozen soil. International Journal of Forest Engineering 24(1): 60–75. Sweden’s Central Bank 2011: Exchange rates / Annual aggregate. www.riksbank.com. [accessed on 6 September 2011]. Timbear 2011: Products: Lightlogg C. Available at: www.timbear.se. [accessed 26 September 2011]. Volungholen, R., 2008: Vehicle for felling trees and/or transport of pieces of timber. Patent PCT/SE2007/050068, WO 2008/097146 A1, World Intellectual Property Organization, Wästerlund, I., 1992: Extent and causes of site damage due to forestry traffic. Scandinavian Journal of Forest Research 7(1): 135–142.

Authors’ address:

Received: December 19, 2012 Accepted: February 17, 2014 Croat. j. for. eng. 35(2014)2

Assoc. Prof. Ola Lindroos, PhD.* e-mail: ola.lindroos@slu.se Department of Forest Biomaterials and Technology Swedish University of Agricultural Sciences 90183 Umeå SWEDEN Prof. Iwän Wasterlund, PhD. e-mail: iwanolasgarden@telia.com Olasgarden Forest and Roads Solvägen 9 91832 Sävar SWEDEN * Corresponding author

219

Original scientific paper

Effect of Transmission Type on Wheel Slip under Overload – Presented on the Example of the AGT 835 T Tractors Jurij Marenče Abstract During their occasional work on small forest holdings, forest owners often use tractors that are, as a rule, not intended for professional forest purposes. Due to their small size, these tractors are appropriate for cultivating smaller agricultural areas and, with additional forestry equipment, also for forestry operations. This paper analyses their performance at the capacity limits, since this type of use is possible but very limited due to their low technical characteristics. Here, a comparison is made between two AGT 835 T tractors produced in Slovenia, with the same basic characteristics but different types of the engine power transmission to the forest ground (comparison between a machine with a standard mechanical transmission system and a machine with a newer version of a hydro-mechanical transmission system). The analysis focuses on the wheel slip – this time only in the last meters of skidding when the slip reaches its peak and the tractors stop because of excessively demanding working conditions. Both tractors were used for skidding timber in the same working conditions – the same skid trail and the same load size. On the steepest section with a 27% longitudinal incline and under the load of 1 m3, both tractors stopped due to excessively demanding working conditions. However, there was a fundamental difference between the two machines in the final section of skidding. The mechanical transmission system enabled rotation of tractor wheels, which led to a multi-fold increase in slip values (remarkable 80% in the last metre of movement). Contrary to that, the system with hydro-mechanical power transmission resulted in a substantially lower wheel slip (no more than 31%). In the latter case the tractor stopped due to excessively demanding working conditions but the hydro-mechanical steering system reduced the wheel slip. It is important to know that the selected transmission system can significantly influence the efficiency of transmitting power to the ground surface – with a smaller slip, which is also important for the forest ground and the environment. Keywords: small forest holding, tractor harvesting, wheel slip, transmission type, overload

1. Introduction In Slovenian forests, the majority of timber is skidded by tractors. They vary in size, technical capacities and equipment. During their development, these machines have undergone numerous changes. It all started with small adapted agricultural tractors that were later replaced by bigger skidders – mostly for operations in difficult working conditions – (Krivec 1967). They did not only differ in size, mass and engine power but also in ergonomic characteristics, forestry upCroat. j. for. eng. 35(2014)2

grade and, recently, in the transmission of engine power to the wheels and ground (Košir 1997, Tomašić 2006, Rebula and Košir 1988). Skidders are mostly used for operations in difficult conditions (Košir 1997, Košir 2000, Košir and Krč 2000, Marenče 2005, Obranovič 2013), as on the majority of worksites – where this is feasible considering the terrain and the load size – smaller adapted agricultural tractors are used (Zupančič 2008). This is possible due to the technical characteristics of the latter enabling

221

J. Marenče

Effect of Transmission Type on Wheel Slip under Overload ... (221–231)

efficient and safe work on less demanding worksites. Such tractors are more affordable mainly due to their lower purchase prices. This is a particularly important factor in case of small and fragmented forest holdings, of course, provided that working conditions permit the use of smaller tractors. Most agricultural tractors that can also be used for forestry operations with additional equipment tend to have engines with less power. The data (Poje 2012) shows the prevalence of tractors with 37 kW of engine power (Fig. 1). Almost two thirds of all tractors do not exceed this engine power. The majority of these tractors are mainly used

Fig. 1 Number of farm tractors according to their engine power on the agricultural land, while some owners use them to conduct various forest operations. It is therefore understandable that the usability of such machines is limited. They are used occasionally, mostly on smaller private forest holdings, but nevertheless, they might be interesting despite their limitations. However, to be used in forests, they need to be appropriately upgraded. This paper was drafted on the basis of the above stated fact that these smaller tractors are relatively frequently used but at the same time limited in the scope of operations, and therefore it presents some characteristic principles of uphill skidding. By doing so, the engine power transmission to the forest ground is analysed with an exclusive focus on the section when working conditions become excessive-

222

ly difficult for such tractors. At this point the longitudinal incline of skid trail and the load size become too big an obstacle to continue skidding. In demanding working conditions, the tractor can no longer operate, and therefore it stops. This is a limit point that cannot be surmounted.

2. Problem definition Most studies have so far examined the equipment used for measuring different technical parameters (Jejčič et al. 2003), and their interrelations and interdependencies (Horvat 1996, Košir et al. 2005, Marenče and Košir 2006, Wong 2001). By doing so, those parameters were analysed in various working conditions, mostly by using different sizes of loads skidded and by changing longitudinal inclines of the skid trail. The studies dealt with different machines used for transporting timber. In most cases the entire timber transportation route was covered, from the stump to the truck road, focusing on the environmental impacts (Wasterlund 1992, Šušnjar 2005, Najafi 2009). Few are studies that only cover a single segment of harvesting (e.g. the moment when the machine stops due to excessively demanding working conditions). These studies focus on changes in technical parameters (slip, weight distribution, torque, winch pulling forces) only occurring in the last meters of timber skidding before the machine stops (Marenče and Košir 2008). The quoted study examined a skidder with a hydro-mechanical transmission gear. The focus was on its technical parameters in the period immediately before the machine stopped, i.e. when the tractor was still moving along the skid trail but operated at its maximum capacity. The present paper does not examine the machines used for professional purposes but rather focuses on tractors operated by private forest owners during their occasional forestry operations. The aim is to upgrade the present tractors skidding at their capacity limits by comparing the operation and performance of two tractors that have identical basic characteristics but different types of transmission gear or different ways of power transmission to the ground. For this purpose, the results were used of field measurements of two tractors – one with a standard mechanical transmission system and the other with a newer version of a hydro-mechanical transmission system. At present, the following three types of transmission systems installed in tractors prevail: mechanical, hydrostatic and hydro-mechanical. The selected type depends on the designated use of a tractor; however, Croat. j. for. eng. 35(2014)2

Effect of Transmission Type on Wheel Slip under Overload ... (221–231)

it also influences its price and operation. The mechanical version comprises a standard cogwheel transmission with the highest efficiency rates and a finite number of gear ratios. The hydro-mechanical version, however, combines the advantages of mechanical transmission (high efficiency rate) and of hydrostatic transmission (lower efficiency rate, but an infinite number of gear ratios). This study does not deal with the hydrostatic versions. The data analysis mainly focuses on the wheel slip, which is one of the most important factors of the environmental impact assessment (forest ground). However, the energy consumption is also considered here as an important factor. During timber skidding, both tractors operated in the same working conditions, i.e. loads of the same size were skidded on the same skid trail. As expected, the biggest slip occurred in the most difficult section of skidding. The above stated different types of transmission have substantially influenced the value of the slip. Hence, the slip does not only depend on working conditions (mostly longitudinal incline and load size) but can also be influenced by the type of tractor or transmission. Today, tractors with mechanical transmission still prevail. There are no substantial differences between private owners and professional providers. The important fact is that the selected transmission system notably influences the efficiency of the power transmission to the ground. At low tractor speeds, minimum slips are desired. Slips do not only cause energy losses, but also damage the upper layers of the ground. Therefore, when working in forests, it is not only important to select the appropriate tractor, but also the right transmission type. The study aims at presenting: Þ the work involving tractors with less power and tractive force that are appropriate for forestry operations under certain conditions, Þ the efficiency of engine power transmission via the wheel to the ground, Þ the analysis of their operation at the capacity limit when skidding becomes impossible due to the excessive longitudinal incline. It is important to answer the following questions: to what extent are such tractors appropriate for forestry operations considering their technical characteristics and how acceptable are they for the environment in the most difficult working conditions considering different wheel slips? Many authors have recommended the ways for selecting the appropriate mechanisation for forest operations, especially on the soft forest ground. In this regard, they emphasise that all forest Croat. j. for. eng. 35(2014)2

J. Marenče

operations need to be conducted with the most suitable machinery and technology at the most convenient time (Owende et al. 2002). In the studies dealing with the slip (Horvat 1993, Wasterlund 2003), the authors established values between 10 and 30% for wheel tractors. Horvat (1193) indicates that tractive forces are comparatively small when the slip values are small. The largest tractive forces occur with higher slip values, which however depend on the characteristics of the ground. Sever (1980) and Saarilahti (2002) mention the slip threshold value of 40% – higher values cause a dramatic drop of effective power of wheels and subsequently cause excessive damages to the ground, mainly due to the ground shifting and deep, long ruts. Horvat (2003) states that with slip values exceeding 33%, the rut depth also increases significantly (by 60%). The approach to this kind of issues is always multifaceted: in terms of energy, the tractive forces necessary for timber skidding are observed, while in terms of environmental protection, the wheel slip value should be kept as low as possible. Wheel slip also represents a loss of energy that decreases the machine speed. Therefore, the aim is to find a sort of optimum value that would be acceptable in terms of energy and environment.

3. Methods The study examined two AGT tractors produced in Slovenia. All tractors of this brand are equipped with less powerful engines (from 13.2 to 26.4 kW); they are classified as adapted agricultural tractors, and have four equal-size wheels and a two-axle drive system. Being small, they can be used on smaller agricultural areas and, with additional forestry equipment, also for forestry operations in worksites with smaller longitudinal inclines and load sizes appropriate for skidding. Such working conditions were also selected for this study. The AGT tractors are available with the mechanical and hydro-mechanical types of transmission (Jejčič 2001), both of which are examined in the study (Fig. 1). Both tractors have identical dimensions, mass and 16-inch pneumatic tyres. They have a watercooled, three-cylinder diesel engine of 26.4 kW produced by Lombardini. The tractors fall into the category of adapted agricultural tractors with front-wheel steering. The mechanical transmission system has 6 gears for driving forward and 3 gears for driving backward, with a maximum speed of approximately 20 km/h. The hydro-mechanical transmission system is based on the consecutive operation of a diesel traction engine, an oil pump and a hydro engine, connected by

223

J. Marenče

Effect of Transmission Type on Wheel Slip under Overload ... (221–231)

Fig. 2 Adapted agricultural tractor AGT 835 T with measurement equipment mechanical transmission. The mechanical part of transmission covers three ranges, selected with a gearshift lever, with infinitely variable transmission ratio (Jejčič 2001). Both tractors were upgraded with the following equipment: a safety cab, a front and rear blade, a sin-

gle-drum winch Krpan with pulling force of 30 kN and wheel chains. A tractor equipped like this can also be used for forestry operations. The skid trail used in the test was selected with regard to the capacities of the tested tractor and was divided into several sections depending on its longitudinal incline. It was concave in shape, 191 m long, with an increasing incline that reached its maximum value just before the trail end (Fig. 3), thus meeting the trial conditions that, when skidding the selected load, the tractor would stop due to excessively demanding working conditions. The analysis of developments on the entire skid trail has already been presented in other papers (Marenče and Košir 2006b, Marenče and Košir 2007); this time the focus will be strictly on the developments from profile No. 7 onwards. This section of the skid trail (20.8 m) is the steepest (27%) and, with the selected load (1 m3), it was too big of an obstacle for the tractor, so it stopped. As already stressed, the tractor is intended for operations in less demanding working conditions, the level of difficulty being determined by the longitudinal incline and the load size. Based on the experiences gathered by operating such tractors, it was assumed that, on the selected skid trail, the load size of 1 m3 would be an insurmountable obstacle for the tractor. This assumption was confirmed during the test. The load was a single fir 8 m log, with bark. In both cases (hydro-mechanical and mechanical version) the load was skidded with butt-end forward in the driving direction. This paper does not deal with different load orientations because this issue has already been presented in other publications (Marenče 2005, Marenče and Košir 2007). Before the measurements took place, the test load was weighed (Fig. 4). Its weight was 770 kg (1 m3).

Fig. 3 Data on the test skid trail and a photo of the skid trail upper part where it is the steepest

224

Croat. j. for. eng. 35(2014)2

Effect of Transmission Type on Wheel Slip under Overload ... (221–231)

J. Marenče

Fig. 5 The fifth wheel used on the front part of the tractor

The wheel slip value derived from the research can be defined in different ways: Košir (Košir 1997) defines it as follows:

( st − ss ) st

(1)

Which can also be expressed as: Fig. 4 Weighing the load before the test Recording was carried out with the measuring instruments connected to a measuring chain. The required equipment was installed on the tractor. The detailed description of data collection and of the measurement equipment has already been presented (Jejčič et al. 2003). For both tractors, the distance driven was measured by using the fifth wheel (Fig. 5). In general, the application of this method for measuring uneven and longer sections could be questionable. However, in the present case, the skid trail on the measured section was straight and short (8.9 or 11.2 m). In this way the measured section could be subdivided into 1 m sections, within which the analytical tasks set for the test were carried out. Particular attention was devoted to the last few metres, where the tractor gradually stopped. The slip occurs when the force applied to the wheel exceeds the cohesion force of the ground surface (Sever 1980). Its size also critically depends on the type and characteristics of the surface. In this research the measurements with both tractors were carried out in the same conditions: on the same skid trail and with the same soil moisture. Croat. j. for. eng. 35(2014)2

d = 1−

sr st

(2)

Where: d wheel slip coefficient, ss, sr actual distance travelled, st theoretical distance travelled. The wheel slip can be expressed as a relative number or as a percentage. In this research it is expressed as a percentage. The distance travelled by the tractor was measured by using the above mentioned fifth wheel, whereas the path travelled by a tractor wheel was measured with special sensors individually mounted on the axle of each wheel (Marenče 2000, Marenče 2005). By doing so, the distance travelled by a wheel, including the slip, was measured (Fig. 6). Pressure in the pneumatic tyres was equalized before the test started. Both tractors were equipped with chains on all four wheels. In this research, the Tempo chains produced by CMC System, d.d., Lesce, were used – Fig. 7.

4. Results In the analysis of this kind, the data on the tractor mass distribution is also important, i.e. it is important

225

J. Marenče

Effect of Transmission Type on Wheel Slip under Overload ... (221–231)

Fig. 6 Sensor (rotary optical generator) mounted on the wheel axle

to determine the front axle load and the rear axle load. This is of crucial importance for the transmission of engine power to the forest ground. For this purpose, both tractors were weighed on a weighbridge together with the entire measurement equipment. The established mass ratio between the front and rear axles was 64:36 (Fig. 8). This means that when the tractor was in a horizontal position and unloaded, the front axle carried almost two thirds of the total mass. The data was identical for both tractors examined in this paper. Table 1 presents the values obtained on the weighbridge and during the operation on the steepest section of the skid trail. Just for comparison and as a point of interest: if unloaded, the front part of the tractor carries more mass that the rear one, even in the steepest section of the trail. However, when skidding the biggest load (1 m3) in the steepest section (27%), with its front part leaning against the tractor, the ratio substantially changes. The front axle carries approximately one third of the entire load. It must be stressed here that, when skidding with these tractors, there is no disburdening and consequently no lifting of the tractor’s front. Therefore, the transmission of engine power to the ground is ensured even in the most difficult working conditions. Thus, the tractor longitudinal stability and the tractor driver’s safety are never jeopardized in any of the skidding sections. Table 1 Tractor mass distribution when operating on the steepest section of the skid trail Front axle : Rear axle

Fig. 7 Chains on a tractor drive wheel

Fig. 8 Weighing the front part of the AGT 835 T tractor

226

Weighed in a horizontal position

64:36

Operating unloaded (at an incline of 27%)

52:48

Operating loaded (at an incline of 27%)

34:66

This paper deals with the agricultural tractors with similar weight distribution as articulated tractors. This feature enables them to be more successful in uphill skidding operations. Due to this unique characteristic, these two tractors are more suitable for forest operations that other tractors from this group. All obtained data are related to the last segment of skidding, which is the most difficult section of the skid trail due to its longitudinal incline. During the test, the distance travelled by tractors was different. The one with the hydro-mechanical transmission stopped at the 9th meter, whereas the one with the mechanical transmission made it further (to the 12th meter). It must be stressed that the measurements of the tractor with mechanical Croat. j. for. eng. 35(2014)2

Effect of Transmission Type on Wheel Slip under Overload ... (221â&#x20AC;&#x201C;231)

transmission ended at the moment when the wheel reached the maximum slip value (100%), which means that the wheel was still rotating, but the tractor was no longer moving. As mentioned above, in both cases the skidding conditions were the same: measurements were carried out on the same skid trail, with the same ground conditions and load mass, the only difference was in the tractor transmission systems. The analysed values are presented for each individual meter of skidding, with particular attention on the last two meters of the skid trail, where the biggest changes were expected due to the tractor limitations. The majority of published papers dealing with a selected tractor, examined the tractor operation on the entire skid trail, and they never particularly focused on the final part of skidding, when the tractor is overloaded. The speed analysis of both tractors shows that their average speed in the measured section ranged from 3 to 4 km/h (Fig. 9). There were no substantial changes of these values in the first part of the examined section - they remained on the similar level. The tractor with mechanical transmission was somewhat faster (on average by 0.5 km/h). In the last meters of movement, the machines gradually stopped. The tractor with hydro-mechanical transmission came to a stop relatively quickly in the last meter. Namely, the hydro-mechanical transmission system no longer enabled the excessive rotation of wheels at the bottom limit of the machine move-

Fig. 9 Driving speed of both tractors in the test skid trail section Croat. j. for. eng. 35(2014)2

J. MarenÄ?e

ment. In case of the tractor with the mechanical transmission, the situation was quite different: here the speed started reducing already 2â&#x20AC;&#x201C;3 meters before stopping and the tractor was still able to proceed on the steep slope while operating at its capacity limit. By doing so, the tractor was able to travel a longer distance, and therefore the stopping time was longer, too. This analysis focused exactly on the section where the tractor was coming to a stop, because it was expected that the substantial changes of some technical parameters of the machine would occur in this section. In the forefront, the size of the wheel slip demonstrated the efficiency of the engine power transmission to the ground. Many authors have examined the occurrence of wheel slip and its impact on the forest ground. However, very limited data is available on what is going on between the wheel and the ground surface in extreme working conditions, i.e. in a few meters where a machine is operating at its capacity limit. Do the slip values change substantially in this case? Does the transmission type make any difference and what is its role? Can the machine selection by foresters influence the scope of harmful environmental impacts in such cases? The answers to the set questions were sought by a thorough analysis of the wheel slip of both tractors. In the first part, some principles of the slip occurrence were analysed for each tractor separately, whereas in the second part, the comparison was made between the two of them. According to the assumptions, and supported by experiences gathered while operating such tractors, substantial increases in the wheel slip values were anticipated just before the tractor comes to a halt. The tractor transmission type was expected to play a crucial role in this increase. The tractor with the hydro-mechanical transmission travelled a distance of 9 meters in the last section of the skid trail. The wheel slip values were established separately for each driven meter and ranged from 10 to 25% (Fig. 10A). It is assumed that, with the same longitudinal incline and ground surface, it was the microrelief that caused the differences established during the test. This assumption is also supported by the fact that similar value oscillations were recorded for both tractors, i.e. in the same sections. The highest slip value was reached in the last metre of movement (30 to 32%, Fig. 10A). It was also established that the slip value was always somewhat higher on the front axle. This means that the higher the load on the rear axle of the tractor, the more effective the transmission moment to the ground surface (the mass ratio in this skidding section was 34:66). The last two metres of skidding were the crucial segment of the analyses. The

227

J. MarenÄ?e

Effect of Transmission Type on Wheel Slip under Overload ... (221â&#x20AC;&#x201C;231)

Fig. 10 AGT tractor (hydro-mechanical transmission) biggest change in case of the hydro-mechanical transmission occurred in the last meter. To illustrate the development and to compare values, the average value in the analysed section from the first to the seventh metre was taken into account (to simplify, this value

was marked as level 100). This value was compared with the values in the last two meters of load skidding. An approximately twofold increase in the slip value was recorded in the last meter of movement (Fig. 10B). At this point the tractor stopped.

Fig. 11 AGT tractor (mechanical transmission)

228

Croat. j. for. eng. 35(2014)2

Effect of Transmission Type on Wheel Slip under Overload ... (221–231)

Fig. 12 Comparison of wheel slip between tractors It must be stressed that this was not a laboratory test. The section where measurements took place was a part of a skid trail on a selected worksite, otherwise used for regular forest operations. On the same skid trail section, the tractor with the mechanical transmission travelled a distance of 9 meters, i.e. a longer distance than in the first case. In this case, the wheel slip values measured in the predominant part of the analysed section ranged from 10 to 30% (Fig. 11A), and were again higher on the front axle, which was less loaded. Substantial changes were explicitly recorded in the last few meters of skidding. This time, the development was very much different than in the first case. In this part, the speed of movement was gradually decreasing, while the slip values were increasing. Just before stopping, the slip value was close to 80%. The measurement was interrupted when the tractor stopped. An additional comparison presented in Fig. 11B shows that the mechanical transmission substantially increased the slip. Compared to the increase in the last ten meters, a fourfold slip increase was recorded immediately before stopping.

5. Conclusion When harvesting timber, foresters often wonder which tractor is the most appropriate for use. An equally important question is when and under what working conditions their use is most appropriate. Studies like this can help provide at least partial answers to these questions. The analyses made so far, Croat. j. for. eng. 35(2014)2

J. Marenče

mostly for uphill timber skidding on longer sections (Marenče and Košir 2006, Šušnjar 2005), have revealed that tractors with the hydrostatic transmission are more appropriate because they cause less environmental damage in forests. The main goal of this paper was to present two machines that only differ in the type of engine power transmission, and to compare them when operating under the same working conditions. The study does not particularly cover changes in the skid trail and removing of the ground due to the wheel slip. It is only focused on the wheel slip itself. By comparing the two tractors (Fig. 12), some similarities and some crucial differences between them can be highlighted. There were no substantial differences in the wheel slip between the two tractors at the beginning of the analysed section. In most cases, the slip value for the mechanical transmission system was somewhat higher. The values reached by this tractor mostly ranged between 10 and 30%. However, with gradual speed reduction of the tractor in the last 2–3 m, the slip considerably increased. The crucial differences between the two tractors were observed in this final part of skidding. The mechanical transmission system enabled a continued rotation of tractor wheels and slow movement of the machine, but this was accompanied by a multi-fold increase in the wheel slip values. The measurement ended when the tractor stopped (at the 100% slip), the average slip value in the last meter being close to 80% (Fig. 12). Consequently, the tractor with mechanical transmission was able to travel 3 m longer under the same working conditions. Quite the opposite, the hydrostatic transmission system, and other settings on the respective tractor, prevented its wheels from rotating and therefore the machine stopped. This also resulted in proportionally smaller shear forces transmitted from wheels to the forest ground. In this case, contrary to the tractor with mechanical transmission, no substantially increased wheel slip was observed. Its value in the last metre, just before stopping, was 31%. The tractor stopped due to overload and the hydrostatic steering system prevented an increased slip. The measurement results showed that the hydro-mechanical transmission system was more successful in preventing overloading of the tractor and thereby reduced the impact on the forest ground and environment. This type of data clearly demonstrates the differences between individual machines that can be used for forestry operations, as well as their appropriateness. It is very important and necessary to know these differences when selecting tractors. Besides, it is also inevitable to question the appropriateness of using

229

J. Marenče

Effect of Transmission Type on Wheel Slip under Overload ... (221–231)

machines at their capacity limit in general. Is this really appropriate and, above all, necessary? The case examined in this paper has clearly demonstrated that, from a strictly technical point of view, the mechanical type performed better. However, this is surely not enough. From the perspective of energy, the increase of slip up to a certain point also means a higher skidding efficiency. However, from the environmental point of view, it is not preferable as it causes the ground shifting and track prints. In order to assess the suitability of forest operations, the energy and environmental issues need to be considered, i.e. the optimum wheel slip interval of the forest operation must be determined. To conclude, considering their technical capacities, the presented tractors can also be used for forestry operations. They are mostly appropriate for work on small forest holdings, i.e. for occasional forestry operations in less difficult working conditions. However, they should not be operated at their capacity limit, because of the excessive wheel slip and thereby intolerable environmental impact. This is particularly true for the mechanical transmission system.

6. References Horvat, D., 1993: Prilog proučavanju prohodnosti vozila na šumskom tlu. Disertacija, Fakultet Strojarstva i brodogradnje Sveučilišta u Zagrebu, 1–234. Horvat, D., 1996: Traction parameters of four skidders used for wood transportation in mountain forest thinning. Proceedings of the Seminar on environmentally sound forest roads and wood transport, June 17–22, 1996, Sinnaia, Romania, FAO Rome, 377–381. Jejčič, V., 2001: Hidrostatični traktor AGT 835T. Tehnika in narava 5(3): 4–6. Jejčič, V., Poje, T., Marenče, J., Košir, B., 2003: Development of measuring equipment for forest tractor AGT 835 with mechanical and hydromechanical transmission. Proceedings of the 31. International Symposium on Agricultural Engineering, February 24–28, 2003, Opatija, Croatia, Zavod za mehanizaciju poljoprivrede, Agronomski Fakultet Sveučilišta u Zagrebu, 65–74. Košir, B., 1997: Pridobivanje lesa (študijsko gradivo), Biotehniška fakulteta – Oddelek za gozdarstvo in obnovljive gozdne vire, Ljubljana, 330 p.

Košir, B., Marenče, J., Jejčič, V., Poje, T., 2005: Determining technical parameters in tractor skidding – basis for the choice of tractor. FORMEC 2005: Innovationen in der Forsttechnik durch Wissenschaftliche Kooperation, Slovenia – September 26–28, Austria – September 29, Department of Forestry and Renewable Forest Resources, Biotechnical Faculty of Ljubljana University, 43–55. Krivec, A., 1967: Preučevanje mehanizacije transporta lesa. IGLG, Ljubljana, p. 203. Marenče, J., 2000: Ugotavljanje tehničnih parametrov traktorja Woody 110 (metodologija in merilni inštrumenti). Zbornik referatov, Kranjska gora, maj 2000. Marenče, J., 2005: Changes in technical parameters of tractors in timber skidding – a criterion for selecting work equipment. PhD thesis, Biotechnical Faculty of Ljubljana University, 271 p. Marenče, J., Košir, B., 2006a: Small tractors and small-scale forest property. Proceedings of 39th International Symposium on Forestry Mechanisation »FORMEC 2006« Sofia, Bulgaria, September 24–28, 2006, 221–228. Marenče, J., Košir, B., 2006b: Influence of forestry tractors technical parameters on tractor choice. Gozdarski vestnik 64(4): 213–226. Marenče, J., Košir, B., 2007: Wheelslip in skidding with the AGT 835 T adapted farm tractor. Zb. gozd. lesar. 82: 25–31. Marenče, J., Košir, B., 2008: Technical parameters dynamics of WOODY 110 cable skidder within the range of stopping due to overload in uphill wood skidding. Zb. gozd. lesar. 85: 39–48. Najafi, A.,2009: Soil disturbance following four wheel rubber skidder logging on the steep trail in the north mountainous forest of Iran. Soil and Tillage Research 103(1): 165–169. Obranovič, A., 2013: Obremenitev delavca z dejavniki delovnega okolja pri spravilu lesa z zgibnim traktorjem ECOTRAC 120V. Magistrsko delo, Biotehniška fakulteta – Oddelek za gozdarstvo in obnovljive gozdne vire, Ljubljana, 1–120. Owende, P. M. O., Lyons, J., Haarlaa, R., Peltola, A., Spinelli, R., Molano, J., Ward, S. M., 2002: Operations protocol for Eco-efficient Wood Harvesting on Sensitive Sites. Project ECOWOOD, Funded under the EU 5th Framework Project (Quality of Life and Management of Living Resources) Contract No. QLK5-1999-00991 (1999–2002), 1–74. Poje, T., 2012: The development trends of the fleet of tractors in Slovenia. 40. Symposium »Actual Tasks on Agricultural Engineering«, Opatija, Croatia, p. 23–31.

Košir, B., 2000: Lastnosti prenosa sil na podlago pri traktorju Woody 110. Gozdarski vestnik 58(3): 139–145.

Rebula, E., Košir, B., 1988: Gospodarnost različnih načinov spravila lesa. UL, IGLG, Strok. In znan. dela 96, Ljubljana, p. 123.

Košir, B., Krč, J., 2000: Študij časa pri spravilu lesa z WOODY 110. XX Gozd. štud. dnevi, Zbornik referatov, Kranjska Gora, Univerza v Ljubljani, Biotehniška fakulteta, Oddelek za gozdarstvo in obnovljive gozdne vire, Ljubljana, 151–168.

Saarilahti, M., 2002A: Soil interaction model. Project deliverable D2 (Work package No. 1) of the Development of a Protocol for Ecoefficient Wood Harvesting on Sensitive Sites (ECOWOOD). EU 5th Framework Project (Quality of Life

230

Croat. j. for. eng. 35(2014)2

Effect of Transmission Type on Wheel Slip under Overload ... (221–231)

J. Marenče

and Management of Living Resources) Contract No. QLK51999-00991 (1999–2002), 1–87.

– Oddelek za gozdarstvo in obnovljive gozdne vire, Ljubljana, 1–67.

Sever, S., 1980: Istraživanje nekih eksploatacijskih parametara traktora kod privlačenja drva. Disertacija, Šumarski fakultet Sveučilišta u Zagrebu, Zagreb, 1–301.

Wasterlund, I., 1992: Extent and causes of site damage due to forestry traffic. Scandinavian Journal of Forest Research 7(1–4): 135–142.

Šušnjar, M., 2005: Istraživanje međusobne ovisnosti značajki tla traktorske vlake i vučne značajke skidera. Disertacija, Šumarski fakultet Sveučilišta u Zagrebu, 1–146.

Wästerlund, I., 2003: Soil disturbance problems in forestry. Proceedings of the 2nd International Scientific Conference »Forest and Wood-Processing Technology vs. Environment – Fortechenvi Brno 2003«, May 26–30, 2003, Brno, Chech Republic, Mendel University of Agriculture and Forestry Brno & IUFRO WG 3.11.00, 491–495.

Tomašić, Ž., 2007: Istraživanje tehničko-eksploatacijskih značajki skidera za prorede. Disertacija, Šumarski fakultet Sveučilišta u Zagrebu, Zagreb, 1–316. Zupančič, M., 2008: Časovna študija spravila lesa s traktorjem John Deere 6220. Diplomsko delo, Biotehniška fakulteta

Wong, J. Y., 2001: Theory of Ground Vehicles. Ottawa, Carleton University, Department of Mechanical and Aerospace Engineering, 528 p.

Author’s address:

Received: December 20, 2013 Accepted: July 3, 2014 Croat. j. for. eng. 35(2014)2

Asst. Prof. Jurij Marenče, PhD. e-mail: jurij.marence@bf.uni-lj.si University of Ljubljana, Biotechnical Faculty Department of Forestry and Renewable Forest Resources Večna pot 83 1000 Ljubljana SLOVENIA

231

Original scientific paper

Analysis of Helicopter Activities in Forest Fire-Fighting Enrico Marchi, Francesco Neri, Enrico Tesi, Fabio Fabiano, Niccolò Brachetti Montorselli Abstract In Southern European countries wildfires are the most natural threat to forests and wooded areas. Over the last decade, public and scientific debates on forest fire management have increased. Helicopters and airtankers are extremely effective fire suppression means, but they are also very expensive. Studies on the improved performance of suppression for the enhancement of firefighting organization are still needed. Consequently to make a plan for the distribution of financial resources to be divided between fire suppression and fire prevention actions in terms of fuel management is not possible. The aim of this study is to compare the helicopter’s forest fire-fighting activity in Tuscany (central Italy) over two periods: between 1998–2000 and 2001–2005 when five and ten helicopters were respectively assigned. For both periods (1998–2000 and 2001–2005) the following were analyzed: the number of forest fires and the burned area with or without helicopter intervention and the position of the helicopter bases in relation to the fire. The results showed that a fleet of 10 helicopters is oversized, in relation to the fire regime of Tuscany, suggesting the need to evaluate a reduction in the fleet. Financial resources may be thus made available for more profitable fire prevention activities, such as, active fuel management. The results also showed where there is the need to improve the helicopter efficiency via the re–management as regards the positioning of their bases. Keywords: fire-fighting, helicopter, spatial distribution, maps, circular statistics

1. Introduction Wildfires are the most important natural threats to forests and wooded areas in Southern European countries (Spain, Portugal, Italy, Greece and France). Over the last decade (2002–2011), the average annual number of forest fires throughout Southern Europe exceeded 53,000, although 11.8% less than the previous decade (1992–2001) (European Commission 2010). The average annual burnt area in the period 2002–2011 was around 381,000 hectares, which resulted in 8.7% less than the previous decade (1992–2001). Portugal and France experienced an increase of burnt area respectively: 47.5% and 48.8% while Greece (– 33.6%), Spain (– 23.1%) and Italy (– 29.2%) saw a reduction. Over the last decade public and scientific debate about the fire management has intensified. Forest experts and managers of many European countries acknowledge the improvement of firefighting organizaCroat. j. for. eng. 35(2014)2

tion as the most important measure in forest fire prevention and suppression (Raftoyannis et al. 2014). Since ‘70s, airtankers and helicopters have been utilized in firefighting in several countries of Southern Europe (Vélez Munoz 2002). Helicopters and airtankers are extremely effective fire suppression means, especially when used during the early stages of fire growth (Vélez Munoz 2002) but they are also costly (Greulich and O’Regan 1982, Greulich 2003). In fact, several researchers argue that firefighting organizations should have had a decisive positive effect by now, at least in reducing the burnt area, if the problem was only a matter of fire suppression, especially when considering the corresponding global increase in firefighting budgets (Xanthopoulus 2007). Furthermore, even though fire suppression organization has improved, the frequency of occurrence of large fires increased in the last decades in many Mediterranean countries and a new approach to fire management has

233

E. Marchi et al.

Analysis of Helicopter Activities in Forest Fire-Fighting (233–243)

been recommended (Xanthopoulus 2007, Fernandes et al. 2011). Despite this ongoing debate, detailed studies on the performance improvement of suppression, due to an enhancement of firefighting organization, are still needed, and it is very difficult to develop a cost-benefit analysis related to firefighting organization improvement. This does not allow for the planning of a redistribution of financial resources between fire suppression and prevention, in terms of fuel management. The purpose of this study was to analyze the impact of helicopter fleet enhancement in fire suppression efficiency and effectiveness in Tuscany Region (central Italy), in order to contribute to public and scientific debate on forest fire management approach. The activity of the helicopters deployed in forest fire-fighting over two periods, with 5 helicopters between 1998–2000 and 10 between 2001–2005, was monitored. This analysis may also be useful in fire prevention management and may represent an effective dataset for planning helicopter’s use in forest fire fighting. The study combines the analysis of the following: helicopter’s pilot forms, fire databases, circular statistics and GIS.

2. Materials and methods 2.1 Study area The Region of Tuscany in central Italy has a surface of 22,998.24 km2 of which 10,861.60 km2 are covered by forests. Helicopters for forest fire-fighting have been used in Tuscany since the 1970s for slowing fire growth and helping ground suppression crews to improve their control action (Favilli and Barberis 1976, Boncompagni 1978, Marchi et al. 2013). In Italy both Regional helicopter fleets and national aerial means are used in forest firefighting activities. Both helicopters (such as Ericson S64, Agusta Bell AB 412, Boeing CH47) and airplanes (Canadair and Airtractor) are used in the national fleet. The deployment of the aerial means at nationwide level is planned by the National Civil Protection administration. The deployment of regional helicopters and the rules of helicopter management are decided by each Regional Administration. In Tuscany Region the evaluation of the needs and the request of the helicopter intervention is the responsibility of the incident commander who manages the firefighting resources when forest fires arise. All the aerial means are used in firefighting to drop water on the fire line, with or without fire retardant. In mountainous areas helicopters are sometimes used to transport firefighters close to the fire line.

234

When the regional helicopter system is working at full capacity or when airtankers or helitankers with a greater water capacity are needed, the incident commander may ask for the cooperation and support of the national aerial means.

2.2 Helicopter characteristics In the period 1998–2000 and 2001–2005, the Tuscany Region Administration was provided with 5 and 10 helicopters, respectively, during the period of high fire risk, i.e. the summer, according to the data of the fire risk model developed by the Tuscany Region. The helicopters were equipped with helibucket (capacity 800 litres) and were chartered from private companies, which is a usual practice in forest firefighting activities in Italy (Marchi et al. 2013). The helicopter fleet was made of Eurocopter SA315 B »Lama«, AS 350 B3 »Ecureuil« and Bell 407. Each helicopter was deployed in a different base. In July–August 1998–2000 three helicopters were deployed in the northern part of Tuscany (Mondeggi – close to the city of Florence; Villa Cognola – close to the city of Arezzo; Calci – close to the city of Pisa) one in the Island of Elba (Elba) and the last one in the south of Tuscany (Alberese – in Grosseto province) (Fig. 1). In July – August 2001–2005, one helicopter was deployed in each Tuscan province, except »Macchia Antonini« that covered two provinces (Prato and Pistoia), and »Elba«, which was in charge for the islands of the Tuscan Archipelago. In addition to these bases, other temporary helibases for refueling were available, allowing to reduce the refueling time during extinction operations and were located so that a maximum of 10 minutes was needed to reach a temporary base from any operation point. The Tuscany region chartered the firefighting helicopters by means of a call of bidders based on a set of specific parameters (number of helicopters availability in each season, maximum flight hours per each charter period of five years, etc.). In the first period (1998–2000) the contract included the availability of at least 1 helicopter for the whole year and up to five helicopters during the summer period (i.e. maximum of 5 helicopters during summer for 90 days). The average flight hours included in the contract were 900 per year (i.e. 4,500 hours in five years). The average hourly cost in this period was 1,161 Euros. In the second period (2001–2005) the contract included the availability of 2 helicopters for the whole year and up to ten helicopters during summer (i.e. maximum of 10 helicopters during summer for 90 days). The average flight hours included in the conCroat. j. for. eng. 35(2014)2

Analysis of Helicopter Activities in Forest Fire-Fighting (233–243)

tract were 1,600 per years (8,000 hours in five years). The average hourly cost in this period was 1,239 Euros. In general, the helicopter cost of a five–year contract increased by 90% between the first and the second period, while the number of the available flight hours increased by 77%.

2.3 Data collection and analysis Data about single operative helicopter flight concerning July – August of each year between 1998 and 2005, which are the most critical months in Tuscany with the highest number of forest fires and burned area (Marchi 2009) were collected from the forms filled in by the pilots. The analyzed data included: date, fire location, helicopter type, helicopter take off base, name of the pilot, time of take–off, operational time (time on fire), total time, and number of water drops. When more than one flight was needed for a single fire both in one as in more consecutive days, the data about operative times were added together and expressed as per single fire event. The burned forest and total area per fire event were obtained from a regional database, including date and fire location. The data collected were analyzed in order to highlight the differences between the two periods (1998– 2000 with 5 helicopters and 2001–2005 with 10 helicopters). The number of fires and burned areas of the two periods were compared by separating fires that required firefighting helicopter activities and fires that did not. A detailed description of the helicopter activity in the different periods was made, taking into account the number of fires per helicopter and year and the number of fires that require one or more helicopters. The total number of days that require helicopters activity was determined in relation to the number of helicopters used. Using a GIS software (ArcView Gis 3.2), fires and helicopter bases were georeferenced and the direction of each flight was calculated. The distance between the helicopter base and the fire location of each flight was also determined by GIS, and then analyzed. Finally the operational times, i.e. the time spent in firefighting, were analyzed in relation to both the number of helicopters on each fire and the time periods.

E. Marchi et al.

parametric U test was used to compare the significance of differences in the number of fires per year, annual number of days with flights and number of fires per helicopter between the two periods. A oneway ANOVA was applied in order to test the effect of the increased number of helicopters on the average distance between helicopter base and fire location, for either all helicopters or the only five helicopters allocated in both two periods. One–way ANOVA was also applied to distances, to verify differences among helicopters in the same period. The post hoc HSD test per unequal N was used to compare the significance of differences among means. The Kruskal–Wallis non– parametric multiple-comparison test was used to test the effect of the number of helicopters per fire on operational time, per fire or per helicopter, in regards to the forest burnt area. A circular statistics was applied to determine the mean vector of each helicopter activity. Two descriptive measures – namely the circular mean and circular variance – were then computed. GIS was finally applied to map helicopter base, fire distribution and mean vector, which may be useful in helicopter firefighting planning. This analysis was carried out by means of a spreadsheet specifically built in Excel. By circular distribution statistics (Batschelet 1981), azimuth angles were calculated in order to determine the mean direction of fire locations for each helicopter base. The circle origin was a single helicopter base. Angular data were grouped on the basis of arcs of equal length and the sampled points (fire location in which each helicopter was used) in each arc were counted. The arcs were established on a geographic basis: 0° corresponded to North and angle width increased in clockwise direction. The circle was subdivided into 12 arcs of equal length (30°). The mean values and the number of fires, which occurred in each sector, were used to determine the mean vector direction for each base. As a measure of dispersion, rc, corrected value of the mean vector length, was used (Batschelet 1981). The rc value spans from 0 to 1. When rc approaches 0, the distribution is dispersed; when rc approaches 1, the distribution is concentrated.

3. Results

2.4 Statistics

3.1 Distribution, number and extent of fires

Data were analyzed using Statistica 7.1 (2007) Software. All the data were checked for normality (Kolmogorov–Smirnov test) and homogeneity of variance (Levene test). Cubic root transformation was applied to normalize distances between helicopter base and fire location of each flight. The Mann–Whitney non-

In the periods 1998–2000 and 2001–2005, 688 and 1,361 forest fires occurred in Tuscany, respectively,

Croat. j. for. eng. 35(2014)2

In the summers 1998-2005, forest fires occurred in most of the Tuscan territory (Fig. 1). Most of the fire events, however, were located in Northern and Eastern Tuscany.

235

E. Marchi et al.

Analysis of Helicopter Activities in Forest Fire-Fighting (233–243)

Fig. 1 Distribution of forest fires with or without use of helicopters in the summers 1998–2005 in Tuscany

over a total burned forest area of 7,288 ha (Table 1). The average annual number of fires put out with or without helicopters, did not show significant differences between the two periods. Helicopters were used in 34% (years 1998–2000) and 36% (years 2001–2005) of fires, where 6,265 ha of forests were burned, i.e. 86% of the total burned forest. Helicopters, in fact, were used in the largest fires, as shown by the mean area per forest fire. Nevertheless, as suggested by the median value, helicopters were used in fires that were ≤ 1.0 ha for half of the fires. The average forest and total burnt areas, during fires that required the helicopter support, were not significantly different between the periods, while significant differences were recorded in fires without helicopter support. In fact, the average burnt area was higher in the first period (1998–2000) than in the second one (2001–2005). Taking into consideration the total number of fires, both total and forest average burnt areas were significantly higher in the first period. Taking into consideration the whole period, both total and forest average burnt areas were significantly higher when helicopters were used. In the period 1998–2000 the highest number of heliattacks was made by the Calci base, followed by Villa Cognola, and Mondeggi (Table 2). All these bases are located in the North–East of Tuscany. In the period 2001–2005 the highest number of heliattacks was

Table 1 Number of forest fires and burnt area (±SE) in July – August 1998–2000 (5 helicopters) and 2001–2005 (10 helicopters) in Tuscany* Burned area

Average

Total

Median

Max

Mean

N/year

ha/fire

1998–2000

232

77.3 (±42)

2,312

1.0

700

9.97 (±3.19)

13.29 (±4.05)

2001–2005

485

97.0 (±32)

3,953

1.0

471

8.15 (±2.21)

12.81 (±2.80)

>0.05

p level

>0.05

Total

717

89.6 (±19)

6,265

1.0

700

8.74 (±1.82)

1998–2000

456

152.0 (±67)

575

0.1

1.26 (±0.20)a

Fires without 2001–2005 helicopters p level

876

175.2(±52)

448

0.1

0.51 (±0.15)b

Total

>0.05

Total

1,332

1998–2000

688

2001–2005

1,361

166.5 (±46) 229.3 (±107.5) 272.2 (±83.3)

2,049

256.2 (±65)

p level 1998–2005

0.002

12.97 (±1.37) 2.02 (±0.24) a 0.75 (±1.08)b 0.000

1,023

0.1

0.77 (±0.12)

2,887

0.3

700

4.20 (±1.42)a

5.82 (±1.46)a

4,401

0.2

471

3.23 (±0.66)b

5.05 (±0.99)b

0.049

0.000

7,288

0.2

700

3.56 (±0.65)

5.31 (±0.82)

>0.05

Mean

1.18 (±1.01)

p level

Total

0.000

Fires with helicopters

Total

Forest p level

Period

a 0.000

Fires

*Different letters show significant differences among values

236

Croat. j. for. eng. 35(2014)2

Analysis of Helicopter Activities in Forest Fire-Fighting (233–243)

E. Marchi et al.

made by the Tassignano base, followed by Calci, and Macchia Antonini. Also these bases are located in the North–East of Tuscany. The average number of heliattacks did not significantly differ between the periods. In the two periods, 1998–2000 and 2001–2005, only one helicopter was used to support ground suppression crews in 82% and 66% of fires in Tuscany, respectively, (Table 3). In 15 and 23% of fires, two helicopters were used. Three or more helicopters on the same fire were rarely used and more than 5 helicopters were never used. During the summer period, however, helicopter activity was not continuous. In some days, all helicopters were inactive. In other days, several helicopters were active. Only in 63% (117 days) and 60% (186 days) of the July – August period, at least one helicopter was used in the first and second period, respectively (Table 4).

3.2 Heliattack distribution and distance Tuscan helicopters were generally used close to their bases, even though they can sometimes cover large areas (Table 5, Fig. 2 and 3). Therefore, the mean distance flight between the base and the fire was generally low, i.e. <52 km. The helicopters of Pentolina (Siena) and Riparbella (Livorno) showed the longest flight distances, while the helicopter of Elba showed

Fig. 2 Helicopter activity in Tuscany, July – August 1998–2000

Table 2 Number of fire per helicopter in July – August 1998–2005 in Tuscany* Helicopter base

Year

Sub–total

Total

136

171

149

n.a.

138

Cinquale

Macchia Antonini

Riparbella

n.a.

Tassignano

114

157

373

720

1,003

1998

1999

2000

Alberese

Calci

Elba

Mondeggi Villa Cognola

2001

2002

2003

2004

2005

n.a.

Pentolina

Total

Sub–total

Not present in these years

125

Average

67 94

283

144

p level >0,05

* Every base has only one helicopter. The sum of number of fire per helicopter is higher than the number of »fires with helicopters« showed in Table 1 because two or more helicopters were used together in many fires n.a. – not available

Croat. j. for. eng. 35(2014)2

237

E. Marchi et al.

Analysis of Helicopter Activities in Forest Fire-Fighting (233–243)

Table 3 Number of forest fires where 1–5 helicopters were used in July – August 1998–2005 in Tuscany* Helicopters per fire

Fires 1998

1999

2000

2001

2002

2003

2004

2005

Total 1998–2000

Total 2001–2005

155

191

82.3

321

66.2

14.7

112

23.1

2.2

7.6

0.4

2.3

0.4

0.8

Total

101

106

242

232

100

485

100

*No more than 5 helicopters were used on the same fire in the whole period

Table 4 Number of days with helicopter flights in July – August 1998–2005 in Tuscany Helicopters

Days with flight 1998–2000

2001–2005

37.61

29.57

33.33

20.43

22.22

13.44

5.13

11.29

1.71

10.22

8.06

4.30

1.61

0.54

117

100

186

100

Total Average, N/year Total days available, 31*2month*years

39.0 (±8.9)

37.2 (±6.9) p level >0.05

186

310

62.9%

60%

the shortest distances. Comparing the distances of the helicopters used in both periods, only Mondeggi showed a significant reduction of the average value in the second period. Black lines (Fig. 2 and 3) show the mean vector direction and length (x100 km) of each helicopter base.

238

Fig. 3 Helicopter activity in Tuscany, July – August 2001–2005 Starting from the base, each circle radius shows the distance corresponding to the 25th, 50th, and 75th percentile, respectively. Fire distribution around the bases was random (Fig. 2 and 3, Table 6). The highest dispersion was found in the Calci, Mondeggi and Villa Cognola bases. Croat. j. for. eng. 35(2014)2

Analysis of Helicopter Activities in Forest Fire-Fighting (233–243)

E. Marchi et al.

Table 5 Distance between base and fire location in July – August 1998–2000 (5 helicopters) and 2001–2005 (10 helicopters) in Tuscany* Base

Period 1998–2000 Mean, km

Period 2001–2005

Max., km

Alberese

47.5 (±3.1)

Calci

32.8 (±2.7)b b

Max., km

p level

167

41.4 (±2.7)

151

>0.05

157

30.8(±2.5)cd

142

>0.05

18.8 (±6.9)

146

>0.05

33.9 (±2.6)bc

143

36.9 (±2.8)abc

109

Elba

25.0 (±10.6)

117

Mondeggi

47.3 (±3.1)a

132

Villa Cognola

39.3 (±3.2)ab

136

Mean, km

Cinquale

23.4 (±2.9)

118

MacchiaAntonini

30.4 (±2.7)cd

139

50.6 (±2.7)

106

51.6 (±4.0)ab

106

160

Not available

Pentolina Riparbella Tassignano p level

23.4 (±2.2) 0.000

Period mean

0.000 >0.05

0.000 a

34.0 (±1.0)b

40.4 (±1.5)

0.000

*Different letters show significant differences among values

Table 6 Mean vector direction (a) and length (rc) for each helicopter base in the summers 1998–2000 and 2001–2005* Period 1998–2000 a

Base

Alberese

rad

(°)

0.56

31.8

0.43

1.13

S rad

(°)

1.06

61.0

Calci

5.95

341.1

0.46

1.07

1.03

59.3

Elba

1.09

62.7

0.32

1.37

1.17

67.0

Mondeggi

4.65

266.7

0.30

1.39

1.18

67.5

Villa Cognola

2.98

170.5

0.16

1.68

1.30

74.4

Alberese

1.16

66.5

0.34

1.15

65.9

Period 2001–2005 1.32

Calci

0.57

32.7

0.11

1.78

1.33

76.2

Cinquale

0.56

32.1

0.49

1.03

1.02

58.4

0.78

44.7

0.23

1.54

1.24

71.0

2.62

150.1

0.41

1.17

1.08

61.9

Elba Macchia Antonini Mondeggi

5.96

341.5

0.15

1.70

1.30

74.5

Pentolina

2.21

126.6

0.51

0.97

0.99

56.7

Riparbella

0.54

30.9

0.54

0.92

0.96

55.0

Tassignano

5.45

312.3

0.32

1.35

1.16

66.5

Villa Cognola

3.46

198.2

0.17

1.65

1.29

73.9

*S – angular variance, 2 * (1 – rc); S – mean angular deviation, RADQ (2 * (1 – rc))

Croat. j. for. eng. 35(2014)2

239

E. Marchi et al.

Analysis of Helicopter Activities in Forest Fire-Fighting (233–243)

Cinquale, Riparbella and Pentolina showed the most concentrated fire distribution.

3.3 Operational time The operational time is the time the helicopter works on a fire. If the operational time was greater than the operating range, time to go back and forth for refueling was also included. In both periods, the operational time per fire was significantly higher if two or more helicopters were used (Table 7). No statistical relationships were recorded between operational time and burned area in every fire. This discrepancy may be explained by the variability in fire spreading, eventual support of national aircrafts, and, most importantly, by the activity of ground suppression crews.

4. Discussion The results did not show significant statistical differences in the number of fires and burnt areas between the two time periods.

The performance of wildfire suppression is often monitored using statistics related to area burned and time to contain a fire (Plucinski 2012). A high number of helicopters should contribute to reduce the burnt area. However, in our study, the average forest and total burned area of fires with helicopter support did not show statistically significant differences between the two time periods, which suggests that an increasing number of helicopters did not have significant effect in terms of suppression effectiveness. However, in both periods, helicopters were used in ≤1.0 ha burnt area for half of the fires, as suggested by the median value, i.e. when used during the first phase of fire growth, a fast initial attack by the helicopter may prevent small fires becoming larger (Plucinski 2012). The average burnt area for fires without helicopter support was significantly higher in the first period (1998–2000) than in the second (2001–2005). This is most likely because of the increasing use of helicopters in small fires when ten helicopters were available. In fact, about 66% and 64% of fires did not require helicopter support in firefighting in the first and second period, respectively. However, median and average values of burnt areas, both forested and total, in fires

Table 7 Helicopter operational time (±SE) on forest fires in July – August 1998–2000 (5 helicopters) and 2001–2005 (10 helicopters) in Tuscany* Flight time Period

Helicopter per fire

ha/fire

min./fire

min./helic.

2.2 (±4.4)a

94.2 (±4.7)a

191

94.2 (±10)a

9.9 (±19.8)

339.5 (±45.6)

N 191 b

169.8 (±18.3)

42.9 (±71.3)

729.2 (±240.5)

243.1 (±53.6)

637.0 (-)

2760 (-)

690 (±160.9)b

p level

700.0 (-)

3173 (-)

0.000 1

p level

Per helicopter

2001–2005

Average operational time Per fire

2 1998–2000

Average forest burnt area

0.000 a

96.6 (±4.1) b

634.6 (±170.3)

0.001 a

1.4 (±3.5)

4 b

321 b

96.6 (±4.1)a

321 b

14.6 (±57.6)

331.8 (±30.3)

112

165.9 (±11.9)

224

33.4 (±92.1)c

684.4 (±76.0)b

228.1 (±20.3)c

111

43.1(±55.8)

1053.5 (±223.9)

263.4 (±38.1)

41.2 (±25.9)c

2018.5 (±681.3)b

403.7 (±80.5)c

0.000

*No more than 5 helicopters were used on the same fire in the whole period. Different letters show significant differences among values

240

Croat. j. for. eng. 35(2014)2

Analysis of Helicopter Activities in Forest Fire-Fighting (233–243)

without helicopters were lower than in fires with helicopters, suggesting that ground suppression provided the necessary effectiveness. The lack of significant statistical difference in the number of heliattacks between the two periods may be explained by the lack of difference in the number of fires. Usually, a large fire needs many helicopters in order to be suppressed. Nevertheless, even though the number of helicopters was double in the second period, the maximum number of helicopters per fire did not exceed five, i.e. the maximum number of helicopters per fire recorded in the previous period. This fact can be explained by difficulties in managing a large number of aircrafts on the same fire. Moreover, in very large fires the incident commander requires the support of more effective aerial means from the national fleet. However, the higher the number of available helicopters, the lower the number of fires where only one helicopter was used to support ground suppression crews. In the second period more than seven helicopters were used in a few days. Overall, the helicopters were used only in 186 days out of the 310 available. Therefore, 9 and 10 helicopters were rarely needed, suggesting the need to evaluate a reduction in the fleet. Financial resources may be thus available for more profitable fire prevention activities, such as, active fuel management (Agee and Skinner 2005). In fact, among the elements affecting the fire behavior, the fuel is the only factor that can be regulated in terms of quantity, spatial distribution and composition. Several studies have shown the role played by fuel reduction activity in terms of minor severity of the flame length (Agee and Skinner 2005, Agee et al. 2000, Omi and Martinson 2002, Pollet and Omi 2002, Martinson and Omi 2003), which also means a reduction of aerial and ground suppression needs. The forest fuel reduction treatment is one of the most valuable tools to effectively address the problem of forest fires (Xanthopoulus et al. 2006). Fuel management should be planned using principles of fire-safe forests: reduction of surface fuels, increasing the height to live crown, decreasing crown density, and retaining large trees of fire-resistant species. Thinning and prescribed fire can be useful tools to achieve these objectives. Low thinning will be more effective than crown or selection thinning, and management of surface fuels will increase the likelihood that the stand will survive a wildfire (Agee and Skinner 2005). However, its application does not exclude the start and spread of the fire, and it is therefore necessary to include it in the context of the planning of fire prevention activity (Corona 2004). Croat. j. for. eng. 35(2014)2

E. Marchi et al.

In this study, increasing the number of helicopters in the period 2001–2005, did not translate into a significant reduction of flight distances travelled by the same helicopters present in the first period (1998– 2000), with the exception of the helicopter at the base of Mondeggi. This suggests that the new helicopters were not well distributed among the bases. Considering all the helicopters in both periods, the average distance traveled to reach the fire was reduced significantly from 40 to 34 km. Often circular statistics is applied to very different biological (Batschelet 1981) or geographical problems (Mardia 1975) such as neuronal discharge patterns during locomotion (Drew and Doucet 1991) and journey to work (Corcoran et al. 2009). In forest science, circular statistics has been mostly used to study the effects of wind (Rentch 2010), the distribution of branching (Faravani et al. 2009) and plant species (Tremblay and Castro 2009). Its use in the analysis of helicopter activity in forest fires suppression is innovative and will help to assess the best location of the bases. Our results showed a high dispersion value of fire around the bases, which means a quite good positioning of the helicopters, as helicopters can operate efficiently in almost all directions. Higher values of rc were recorded at three bases (Cinquale, Riparbella and Pentolina), suggesting that the helicopters preferentially flew close to the mean vector direction. In particular, the bases of Cinquale and Riparbella were located close to the seaside and thus had a wide area on the West side where fire cannot occur. Therefore, they should be moved toward North-East, thus reducing the average distance to reach the fires. Moreover, the rc and mean vector direction values suggest that the Pentolina base should be moved toward South–East in order to improve its efficiency.

5. Conclusions This study analyzed the activities of the helicopters used in forest fire-fighting in Tuscany Region (Italy) during two periods, characterized by a fleet of 5 (1998– 2000) and 10 helicopters (2001–2005). The result showed that a fleet of 10 helicopters is oversized in relation to the current fire regime in the area, suggesting the need to evaluate a reduction in the fleet. However, it is important to highlight that our results do not consider the potential future fire regime, as a result of climate change. A planning of the size and spatial distribution of helicopter fleet should include: environmental, social and economic aspects; »vulnerable« areas, i.e. areas of special importance from an environmental or land-

241

E. Marchi et al.

Analysis of Helicopter Activities in Forest Fire-Fighting (233–243)

scape point of view; ground crew organization, with special reference to its travel time for reaching potential fire; availability of a suitable forest road network and finally available budget. Our results highlight the needs of a fire prevention and suppression planning revision for improving the financial resources distribution between suppression and prevention activities, with particular attention to fuel management. Information from the pilot forms is recommended as a useful tool for describing and evaluating the helicopter activity in forest firefighting. Processing such data, together with data about landscape and other components of firefighting organization, helps to assess and address the planning of this activity. The results for this case study in Tuscany also show that: Þ Helicopters were used in about 35% of fires and 86% of the total forest burned area, Þ The median burnt forest area shows that, in half of the fires, helicopters were used in fires that were ≤1 ha. This suggests that, when used during the first phase of fire growth, a fast initial attack by helicopter may prevent small fires becoming larger, Þ Only in 60% (186 days) of the July – August period, at least one helicopter was used, Þ The higher the number of helicopters deployed, the higher the number of helicopters used on the same fire to support ground suppression crews. The location of helicopter bases in Tuscany was usually well planned, according to both the distribution of fires and the circular statistics results. However, a reallocation of three bases is recommended, in order to minimize the flight distances and raise the helicopter effectiveness.

Acknowledgement The authors would like to thank the Tuscany Region administration for providing all the data (regional forest fire database, helicopter pilot forms, etc.) used in this study.

6. References Agee, J. K., Bahro, B., Finney, M. A., Omi, P. N., Sapsis, D. B., Skinner, C. N., van Wagtendonk, J. W., Weatherspoon, C. P., 2000: The use of shaded fuelbreaks in landscape fire management. Forest Ecology and Management 127: 55–66. Agee, J. K., Skinner, C. N., 2005: Basic principles of forest fuel reduction treatments. Forest Ecology and Management 211: 83–96.

242

Batschelet, E., 1981: Circular Statistics in Biology. Academic Press, London, 371 p. Boncompagni, T., 1978: Nuove tecniche e modalità d’impiego dei mezzi aerei nella lotta contro gli incendi boschivi nella regione Lombardia, Annali dell’Accademia Italiana di Scienze Forestali XXVII: 275–281. Corcoran, J., Chetri, P., Stimson, R., 2009: Using circular statistics to explore the geography of the journey to work. Papers in Regional Science 88: 119–132. Corona, P., 2004: Interventi selvicolturali. In: Incendi e complessità ecosistemica. Dalla pianificazione forestale al recupero ambientale. (Blasi, C., Bovio, G., Corona, P., Marchetti, M., Maturani, A.) Editors, Palombi, Rome. Drew, T., Doucet, S., 1991: Application of circular statistics to the study of neuronal discharge during locomotion. Journal of Neuroscience Methods 38(2–3): 171–181. European Commission, 2010: Forest Fire in Europe 2009. In: JRC Scientific and technical Reports Report No10. (European Communities, Joint Research Centre Institute for Environment and Sustainability, Italy). Faravani, M., Baki, B. B., Kato, S., Shimizu, K., Sim, C. H., 2009: Effects of intra–plant competition on the ensuing spatial branching patterns of straits rhododendron. Research Journal of Environmental Sciences 3(4): 427–438. Favilli, P., Barberis, G., 1976: L’uso dell’elicottero nella campagna antincendi 1975 in provincia di Nuoro, Monti e Boschi Anno XXVII (2): 3–22. Fernandes, P. M., Rego, F. C., Rigolot, E., 2011: The FIRE PARADOX project: Towards science–based fire management in Europe. Forest ecology and management 261: 2177– 2178. Greulich, F. E., O’Regan, W. G., 1982: Optimum use of airtankers in initial attack: Selection, basing, and transfer rules. Research paper PSW-163. Pacific Southwest Forest and Range Experiment Station. Berkeley, California. 8 p. Greulich, F., 2003: Airtanker Initial attack: a spreadsheetbased modeling procedure. Canadian Journal of Forest Research 33: 232–242. Mardia, K. V., 1975: Statistics of directional data. Journal of the Royal Statistical Society Series B Methodological 37(3): 349–393. Marchi, E., Boni, I., Brachetti, N. M., Tesi, E., 2003: Gli elicotteri nell’antincendio boschivo. Proposta metodologica per l’analisi dell’attività. Il caso della Regione Toscana. Sherwood 94: 19–26. Marchi, E., 2009: Protezione dagli incendi boschivi. In: Rapporto sullo stato delle foreste in Toscana 2008. Compagnia delle Foreste, Arezzo, p. 107–116. Martinson, E. J., Omi, P. N., 2003: Performance of fuel treatments subjected to wildfires. In: Proceedings RMRS. USDA Forest Service, Rocky Mountain Research Station. Fort Collins, Colorado. 29 p. Croat. j. for. eng. 35(2014)2

Analysis of Helicopter Activities in Forest Fire-Fighting (233–243) Omi, P. N., Martinson, E. J., 2002: Effect of fuels treatment on wildfire severity. In: Final report to the Joint Fire Science Program Governing Board. Western Forest Fire Research Center, Colorado State University. Fort Collins, Colorado. 40 p. Plucinski, M. P., 2012: Factors Affecting Containment Area and Time of Australian Forest Fires Featuring Aerial Suppression. Forest Science 58(4): 390–398. Pollet, J., Omi, P. N., 2002: Effect of thinning and prescribed burning on crownfire severity in ponderosa pine forests. International Journal of Wildland Fire 11: 1–10. Raftoyannis, Y., Nocentini, S., Marchi, E., Calama S. R., Garcia, G., Pilas, I., Peric, S., Amaral, P. J., Moreira-Marcelino, A. C., Costa–Ferreira, M., Kakouris, E., Lindner, M., 2014: Perceptions of forest experts on climate change and fire management in the European Mediterranean forests. Iforest 7: 33–41. Rentch, J. S., 2010: Relationship between tree fall direction, slope-aspect, and wind in eight old–growth oak stands in the Central Hardwood Forest, USA. Journal of the Torrey Botanical Society 137(4): 391–400.

E. Marchi et al.

Tremblay, R. L., Castro, J. V., 2009: Circular distribution of an epiphytic herb on trees in a subtropical rain forest. Tropical Ecology 50(2): 211–217. Vélez Munoz, R. D., 2002: Los medios aéreos en la ejecucion de lanes de ataque en la Cuenca Mediterranea. In: Proceeding of I Simposioum Internacional “La gestion de los medios aéreos en la defensa contra los incendios forestales, Cordoba. Xanthopoulos, G., Caballero, D., Galante, M., Alexandrian, D., Rigolot, E., Marzano, R., 2006: Forest Fuels Management in Europe. In: Proceedings of the 1st Fire Behavior and Fuels Conference, »Fuels Management – How to Measure Success«, Portland, Oregon, USA, March 27–30. Andrews, P. L. and B. W. Butler, compilers. USDA Forest Service. Rocky Mountain Research Station, Fort Collins, Colorado. RMRSP-41. 809 p. Xanthopoulos, G., 2007: Forest fire policy scenarios as a key element affecting the occurrence and characteristics of fire disasters. In: Proceedings of 4th International Wildland fire Conference, May13–17, Siviglia, Spain. 11 p.

Authors’ address: Prof. Enrico Marchi, PhD. e-mail: enrico.marchi@unifi.it Francesco Neri, PhD.* e-mail: francesco.neri@unifi.it Fabio Fabiano, PhD. e-mail: fabio.fabiano@unifi.it Niccolò Brachetti Montorselli, PhD. e-mail: n.motorselli@gmail.com Department of Agriculture Food and Forestry System University of Florence Via S. Bonaventura 13, 50145 Florence ITALY

Received: August 8, 2013 Accepted: February 2, 2014 Croat. j. for. eng. 35(2014)2

Enrico Tesi, PhD. e-mail: enrico.tesi@regione.toscana.it Tuscany Region, Sector »Forestry, Promotion of Innovation and EU Community Measures in Agri environment« Via di Novoli 26, 50127 Florence ITALY * Corresponding author

243

Original scientific paper

The Effect of Mounting Height on GNSS Receiver Positioning Accuracy in Forest Conditions Michał Brach, Michał Zasada Abstract In spite of the high prices of GNSS receivers, many users decide to invest in this equipment because of the high accuracy of X, Y and Z data capture. Measurements in forested environments are affected by the increased positional error because of the signal multipath effect caused by trees. The main idea of this paper is to raise the antenna of a GNSS receiver during measurements, in order to reduce the multipath effect in the highest part of forests. A 15 meter pole was used in order to capture the GNSS signal at a height of 5, 10 and 15 m above ground level, in various forest conditions. The main factor, which determines the precision and accuracy, is the operational mode of the receiver. When in the FIXED mode, the results obtained are more reliable than those obtained when in the FLOAT mode. Due to difficult conditions in the forest stand, FIXED mode occurrence is not always possible, but much more likely at higher elevations. The FLOAT mode, however, is more likely to occur in the forest conditions and the obtained accuracy of the X and Y coordinates was ±0.81 m and 1.11 m for the elevation (Z coordinate). The best results were achieved for X and Y coordinates at an altitude of 10 m in a leafless state with an average error of ±0.54 m for the FLOAT mode. We cannot assume, therefore, that raising the GNSS antenna will improve the precision and accuracy in every case. Keywords: GNSS, accuracy, precision forestry, survey, ANOVA

1. Introduction The Global Navigation Satellite System (GNSS) has become one of the most popular techniques for fast and accurate positioning in open spaces. This method has been used in many areas of mapping because of the low cost and simplicity of its use, compared to the standard way of surveying (Mauro et al. 2010). Taking into account the diversity of the forest structure (Puettmann et al. 2009) and accessibility of digital maps as a main source of land use information (Bach et al. 2006), there is a real need to be able to gather up-todate and accurate positioning data in forests (Suarez et al. 2005). The GNSS technology works well in unobstructed open spaces and all GNSS manufacturers provide the accuracy of their receivers assuming that they work without any obstacles. The fact that forest may supCroat. j. for. eng. 35(2014)2

press or even completely block the satellite signal is not taken into account (Næsset and Jonmeister 2002). The low accuracy of GNSS receivers in forestry conditions has been widely discussed in many research papers. There are a lot of factors caused by forest conditions which can influence positioning accuracy. Forests are a barrier for signal propagation so the final radio wave is weak and the reflection causes an elevated signal-to-noise ratio, which is caused by the so called multipath effect (Hasegawa and Yoshimura 2007, Pirti et al. 2010, Valbuena 2012). The base idea of multipath is strictly connected to signal reflections from objects located near the receiver, which ultimately causes an error in distance measurements. There are many software and hardware solutions to weaken this effect, however it still does not solve the strong forest influence (Valbuena 2014). Additionally, the multipath effect is multiplied by high moisture (Sigrist et al. 1999)

245

M. Brach and M. Zasada

The Effect of Mounting Height on GNSS Receiver Positioning Accuracy in Forest ... (245–253)

and the presence of leaves (Valbuena et al. 2012). One of the first conclusions concerning navigation accuracy in the forest was formulated by Næsset (1999), who concluded that high density sites, tree species, satellite constellation and observation times, are important factors for accurate positioning. When compared to open sky conditions, leaves decreased a number of visible satellites, and hence geometry (PDOP value) was deteriorating. The PDOP factor corresponds to the general uncertainty of coordinate calculations. The desirable low value of PDOP depends on angled satellite constellation (Valbuena 2014). Forest conditions have little impact on PDOP, so this is why the observation session should be carefully planned and the final data processed afterwards (Wing et al. 2009). Planning is also very important in mountainous terrains, which can block satellite signals (Deckert and Bolstad 1996). GNSS data capture may be obstructed by tall trees and large basal areas, especially in mixed coniferous stands (Georges et al. 2004). Analyzing the differences between deciduous and coniferous trees, it can be concluded that needles and trunks have significant influence on positioning accuracy (Sawaguchi et al. 2003). When considering the impact of forest characteristics on satellite navigation, Ordóñez Galán (2011) carried out complex research on dasymetric and GPS parameters. It turned out that the slenderness coefficient, Hart-Becking spacing index, wood volume and dominant height are the most important variables that can be used for the description and prediction of the horizontal and vertical measurement accuracy. Basal area can better describe forest conditions than tree height and density (Næsset 2001) and this value can be better for assessing the absolute error (Næsset 2000). The complex research by Valbuena et al. (2012) confirms the relative spacing index (RSI) and wood volume as values that can explain GNSS positional variability. It was also mentioned that the leaf area index (LAI) can be used as an accuracy prediction factor. Taking into account the close connection between signal-to-noise ratio (SNR) and positioning precision, it can be assumed that wood resistance, quantity and satellite elevation angles should also be considered (Sawaguchi et al. 2005). Finally, the best way to conduct data capture in forest conditions is to use dual frequency receivers operated in the fixed mode (Hasegawa 2007). The DGNSS technique is one of the best and economically justified methods to improve positioning accuracy (Næsset 2001) and it can reduce errors caused by atmospheric delay (Næsset and Jonmeister 2002). Nowadays there are a lot of reference stations with coverage varying from local to global, which can transmit corrections directly to the GNSS receiver by the Internet. This method is far more effi-

246

cient than the post-processing procedure (Andersen et al. 2009, Valbuena 2014) and it is better than setting up your own reference station (Valbuena et al. 2010). In most known cases, researches used standard GNSS receivers mounted on poles up to 5 m high (Næsset 1999, Næsset and Jonmeister 2002, Cole 2004, Rodriguez-Perez et al. 2006, Wing 2009, Valbuena et al. 2010). Such antenna locations cannot guarantee access to open sky in the forest. Næsset and Jonmeister (2002) suggest that receiver location is changed to close open areas, which is not easy in dense forests. The alternative approach was presented by Sigrist et al. (1999), who recommended using a mast for the antenna as a solution to open sky access. One of the most well-known and comprehensive studies considering the influence of antenna height on positioning accuracy was presented by Yoshimura (2005). In his research he concluded that using a sophisticated GNSS receiver mounted at high heights and DGNSS correction can result in a significant increase in the accuracy of the position measurements in forest conditions. Our goal was to determine whether the measurement height has an effect on the positioning accuracy in the forest stands that are typical for Eastern Europe. The key to improve accuracy is to increase the access to open sky. This can be realized by using an exceptionally high mast. In this way the significant influence of forest characteristics on GNSS positioning accuracy can be reduced. The second goal was to assess the possibility of using the high accuracy FIXED GNSS receiver mode and its comparison with the less accurate FLOAT mode.

2. Materials and methods The experimental site was located in the Głuchów forest district belonging to the Warsaw University of Life Sciences-SGGW. The site location was 51°45’13.01” N and 20° 6’33.72” E. The network of reference points was created and stabilized along main roads crossing the Głuchów forest district. The traverse technique was used in order to assess error propagation. All points were measured using classical geodetic surveying techniques and adjusted to reference points located outside of the forest (Valbuena 2014). They formed a reference base in order to set 36 sample plots in different forestry sites. The sample plots were measured using trigonometric (for Z value - elevation) and polar (for X, Y values) methods. All coordinates were expressed and calculated in the 2000 zone 7 Polish coordinate system (EPSG code 2178). The maximum horizontal error for the sample point was 0.09 m. In order to correct any elevation errors, additional levelCroat. j. for. eng. 35(2014)2

The Effect of Mounting Height on GNSS Receiver Positioning Accuracy in Forest ... (245–253)

M. Brach and M. Zasada

ing measurements were carried out. The final results of the vertical calculations gave a maximum elevation error of 0.05 m, expressed and calculated in Kroonstad 1986 reference frame. At each of the 36 sample points grouped by main species, age and stand height, an aluminum mast with a GNSS receiver was set. The detailed characteristics of the research stands are given in Table 1. Table 1 Categorization of sample points based on stand characteristics Main species

Age, years

Average height, m

Stocking

Birch

19–42

14–20

Full

Spruce

Full

So 72–110

Pine

72–110

25–26

Moderate

So 17

Pine

Moderate

So 61–88

Pine

61–88

23–25

Full

Db 80–84

Oak

80–84

23–24

Full

Stand category Brz 19–42 Św 50

The dual-frequency, geodetic class surveying GNSS receiver Topcon HiperPro was used for obtaining the positioning coordinates. Valbuena et al. (2010) research based on this receiver, reports it as one of the best in horizontal and vertical absolute error and practically independent on data capture time in forest conditions. The receiver was mounted on an aluminum mast at three different heights: 5, 10 and 15 m (Fig. 1). All measurements were differentially corrected by the Polish network of reference stations (ASG-EUPOS) by the NAWGEO service. This service provides real-time correction data by using a virtual reference station (VRS) technique (Landau et al. 2002). During measurements in the forest, two basic measurement modes were used: FLOAT and FIXED, which provide significantly different positioning accuracy (Teunissen et al. 2008). FLOAT mode is based on a real value of sequences for the carrier phase between the receiver and satellite. This is only an approximate value which does not correspond to the reality, so the expected positioning accuracy is about 0.5 m (Cellmer et al. 2010). FIXED mode is a measurement mode in which the search for the carrier phase ambiguity has been solved as an integer value. This solution is the most reliable and most accurate. Depending on the data correction service, it is possible to achieve a 0.03 m horizontal position accuracy and 0.05 m vertical position accuracy (by the accuracy of Croat. j. for. eng. 35(2014)2

Fig. 1 An aluminum telescopic mast with the GNSS receiver in the Głuchów forest district (photo: Dariusz Górscy) the system claimed by ASG-EUPOS) (Oruba et al. 2009). The default measurement mode is FIXED, but if the phase ambiguity has not been solved because of forest conditions, the mode is automatically switched to FLOAT. The construction was stabilized by an aluminum tripod. Considering the mast weight (18 kg), height (15 m) and weight of GNSS antenna (1.74 kg), it was necessary to check whether the mast was vertically oriented at 5, 10 and 15 m. In order to receive the final results, the measurements by total station were carried out in an open space from two directions. It was found that, at a height of 10 m and 15 m, the average GNSS receiver deflection was 0.04 m and 0.08 m, respectively. There was no deflection when the mast was extended to 5 m. The complete error analysis consists of a calculation for accuracy and precision (Yoshimura 2003, Valbuena 2014). In order to analyze horizontal positioning accuracy the following equation was used:

247

M. Brach and M. Zasada

The Effect of Mounting Height on GNSS Receiver Positioning Accuracy in Forest ... (245–253)

s H_accuracy

_  _  =  x − xtrue  +  y − ytrue     

(1)

Where: xtrue, ytrue reference coordinates, _ _ x, y mean coordinates captured at different heights. Similarly for the elevation: s v_accuracy = z − ztrue

Where:

(2)

ztrue reference elevation, z

mean elevation value at different heights.

In order to assess the relation of the measurements to the true value, the root mean square (RMS) estimator can be used (Sigrist et al. 1999, Rodrigez-Perez et al. 2006). The precision calculation by the RMS error was used for horizontal coordinates: s H_precision = s x2 + s y2

(3)

The values of sx and sy were calculated using the following equations: n

_   ∑  xk − x s x2 = k =1 n−1 n

_  ∑  yk − y s y2 = k =1 n−1

(4) 2

(5)

Where: xk, yk horizontal coordinates captured for a sample point, _ _ x , y mean horizontal coordinates captured for a sample point, n a number of observations made at every mast height. The precision for elevation was calculated using the following formula:

∑ (z n

2 v_precision

k =1

−z

n−1

)

(6)

Where: zk elevation captured for a sample point, z N

248

mean elevation captured for a sample point, number of observations made at every mast height.

Analyses of accuracy and precision were made separately for GNSS receiver modes FLOAT and FIXED. Because of a large number of deciduous stands in the Głuchów district, the observations were made in two vegetation seasons, so the final result could be analyzed in leaf on and leaf off conditions. Measurements at all sample points were repeated 10 times, so all in all, considering 36 sample points and 3 antenna heights (5, 10 and 15 m), a total of 1,080 observations were recorded. One of the main hypothesis for this research is that the amount of light, which can reach the ground in the forest, may be correlated with positioning accuracy (Valbuena et al. 2012). This factor was examined by hemispherical photographs, which allowed us to determine a percentage of light (canopy openness) on the ground at each sample point in leaf on and leaf off conditions (Jonckheere et al. 2005). The idea of hemispherical photography was successfully used by Sigrist et al. (1999), who highlighted canopy closure as a good estimator of satellite signal blocking. The complete set of photographs was used for leaf-on and leafoff seasons in order to compare final positioning errors with canopy openness. During the GNSS measurements, the dilution of precision (DOP) value was also recorded. This value describes the satellite geometry. The high value of DOP means that the satellites are located close to the straight line and a potential error propagation by the triangulation calculation can lead to higher positioning uncertainty (Valbuena et al. 2014). Under the forest canopy, the DOP value can increase, so it can be an important factor in final positioning results (Lewis et al. 2007). The influence of the antenna height and the presence of leaves (leaf-on/leaf-off season) on the horizontal and vertical accuracy was investigated by the twoway analysis of variance (ANOVA) with power transformation of dependent variables to comply with ANOVA assumptions.

3. Results The mean canopy openness for sample points depending on antenna height was 19.2–41.7% for a leafon season and 56.7–68.7% for a leaf-off season. These values varied depending on the stand category (Fig. 2, 3). The mean value of DOP was 3.31, so it was close to optimal (Duncan et al. 2013, Puente et al. 2013). Peyret (2000) states that DOP value can give information about the repeatability of measurements, so a slight advantage of leafless and bigger antenna height can be observed (Fig. 4). Croat. j. for. eng. 35(2014)2

The Effect of Mounting Height on GNSS Receiver Positioning Accuracy in Forest ... (245–253)

M. Brach and M. Zasada

Fig. 2 Canopy openness for different stand category and hemispherical camera height in a leaf-off season

Fig. 4 PDOP value depending on antenna height and vegetation season

Fig. 3 Canopy openness for different stand category and hemispherical camera height in a leaf-on season

Fig. 5 Number of observations depending on measurement mode in leaf season

The main factor which affects GNSS accuracy in the forest is the measurement mode. The total number of FIXED mode measurements was 344 (32%) and 736 (68%) for FLOAT mode, respectively. These values change, however, depending on antenna height and vegetation season. The number of FIXED positions

increase very slowly with the height, but in practice FLOAT mode is more probable (Fig. 5, 6).

Croat. j. for. eng. 35(2014)2

Depending on the antenna height, the mean horizontal accuracy for a FIXED mode varied from ±0.09 to ±0.25 m and 0.13 to 0.15 m for elevation, respectively. Considering the FLOAT mode, the mean error

249

M. Brach and M. Zasada

The Effect of Mounting Height on GNSS Receiver Positioning Accuracy in Forest ... (245–253)

Fig. 6 Number of observations depending on measurement mode in leafless season

Fig. 8 RMS value depending on measurement mode and antenna height

Fig. 7 Mean horizontal and vertical accuracy depending on measurement mode and antenna height

Fig. 9 Relationship between the mean error of the XY coordinate determination and antenna height for various seasons in the FLOAT mode

was a few times higher and decreased from ±0.96 m for low antenna locations to ±0.70 m for the antenna located on the mast extended to the maximum. For an elevation, the error in the FLOAT mode varied from 0.74 to 1.35 m (Fig. 7). Analysis of the horizontal precision gave the RMS error for the FIXED mode ranging

from ±0.05 meter for a leaf-off season to ±0.09 m for a leaf-on season and for the FLOAT mode from ±0.36 to ±0.48 m, respectively. The RMS error for elevation in the FIXED mode varied from 0.07 meter in a leaf-off season to 0.10 m for a leaf-on season and from 0.46 to 0.63 m for the FLOAT mode, respectively (Fig. 8).

250

Croat. j. for. eng. 35(2014)2

The Effect of Mounting Height on GNSS Receiver Positioning Accuracy in Forest ... (245–253)

Fig. 10 Relationship between the mean error of the Z coordinate determination and antenna height for various seasons in the FLOAT mode No significant relationships were found between the accuracy and DOP or canopy openness value. Measurements taken in the FIXED mode had very low variability. Taking into account these factors and the small number of FIXED observations (32%), it was decided to make statistical analysis for the FLOAT mode only. The analysis showed that, for the FLOAT mode, both factors: antenna height and season (presence of leaves) significantly influence the measurement accuracy of XY coordinates (p=0.0000, Fig. 9). There is an interaction between the analyzed factors: the influence of the antenna height on the accuracy differs significantly for leaf-on and leaf-off seasons (p=0.0000). The ANOVA analysis for the height coordinate (Z) revealed that, in the FLOAT mode, the influence of both factors was also significant (p=0.0000, Fig. 10). The influence of the antenna height on the accuracy of Zcoordinate determination had the same characteristics in both analyzed seasons (p=0.126), i.e. there was no interaction between the two observed factors.

4. Discussion and conclusions Increasing the height of the GNSS receiver antenna enables achieving significantly better horizontal and vertical positioning accuracy in the stands in both seasons (leaf-on and leaf-off). These conclusions correspond to other published research: by elevating the Croat. j. for. eng. 35(2014)2

M. Brach and M. Zasada

antenna one can expect the increased number of visible satellites (Næsset 2001, Arslan and Demirel 2008), decreased influence of foliage (Valbuena 2014) and faster activation of the receiver (Sigrist et al. 1999), hence, improved positioning accuracy (Yoshimura and Nakanishi 2005). One of the main factors determining the accuracy and precision of coordinates is the measurement mode. The best results are expected in the FIXED mode supported by the DGNSS technique. In forest conditions, however, this mode may be uncertain in some situations (Næsset 2001), due to the relatively small number of FIXED observations. The conditions for collecting data in the stands are much worse than in open spaces, so the measurement mode of the receiver is random and does not depend on the observer. The leafless state of the forests has a good influence on the total accuracy because the number of FIXED measurements is higher, which was confirmed by Sigrist et al. (1999) amongst others. Taking into account the statement by Deckert and Bolstad (1996), that the number of fixes are very important for accurate measurements, we can assume the leaf-off season as the best time for taking measurements. It was also found that the number of observations in the FIXED mode is associated with canopy openness and the height of the measurement. This is because access to satellite signals becomes easier, hence the probability of making measurements in the FIXED mode increases and thus replaces the FLOAT mode. It is also worth mentioning that in some cases elevating the antenna does not necessarily lead to accuracy improvement. At 15 meters over ground the accuracy of the FIXED measurement is slightly decreased compared to lower antenna positions. This can be explained by the presence of the antenna in the zone of tree crowns, in the close proximity of leaves. This may cause an increased sensitivity of the receiver readings to the multipath especially in wet conditions, which was also noticed by Sigrist et al. (1999) and Valbuena et al. (2012). Thus, when taking measurements in forest conditions with an elevated antenna, it is necessary to pay attention to its location in relation to tree crowns, i.e. to take into account not only measures of stocking (wood volume, tree density, basal area), but also stand and tree characteristics, such as the dominant height, crown base height, crown length, density of foliage, etc. Taking into account all the sample points in all the stands, the horizontal coordinates may be affected by an error of ±0.17 m in the FIXED mode and ±0.81 m in the FLOAT mode. In the case of elevation, these values were 0.14 m in the FIXED mode and 1.11 m in the FLOAT mode, respectively. Results similar to Næsset

251

M. Brach and M. Zasada

The Effect of Mounting Height on GNSS Receiver Positioning Accuracy in Forest ... (245–253)

and Jonmeister (2002) and Sigrist (1999) were observed for DOP value, which has no effect on the accuracy. Very poor results in terms of measurements were obtained in the spruce stands. Data gathered in these conditions should be subject to further analysis, especially in mountainous conditions, where terrain shape is an important factor (Sigrist et al. 1999). The overall conclusion is that the use of an aluminum mast gave significant benefits in improving the accuracy of measurement in the FLOAT mode, which is the dominant part in the total number of measurements. This mode will actually be the main one used for taking measurements in forest conditions. In the FLOAT mode, the influence of multipath on the accuracy is almost nonexistent. Additionally, by extending the FLOAT observation time, it is possible to achieve accuracy similar to the FIXED mode in the forest conditions (Valbuena et al. 2010). It is worth mentioning, however, that increasing the height of the antenna does not affect the accuracy of measurement proportionally. In some stands, the density of the crowns and the proximity of leaves can influence measurement errors because of a low signal visibility. The final results allow us to conclude that the use of masts with GNSS receivers is justified in all types of forests and can significantly increase the accuracy at altitudes of 10 m in the leaf-off season. It is important, however, to take into account the characteristics of trees and stands, and to avoid placing the antenna in the close proximity of tree crowns and leaves causing the decrease of the measurement accuracy. Using higher masts brings an increase in operating costs as well as technical difficulties, e.g. the weight of the mast, problems with electrical power, the mast tilt, etc. and shows insignificant improvement in terms of accuracy, especially in the leaf-off season. That is why, in the forest conditions, the use of light and shorter telescopic poles is recommended. One possible option to help eliminate positional error is to use the HDGNSS technology (Carter 2013) or RTK-Net positioning (Bakula et al. 2012). These methods are expensive, however, as well as time-consuming and therefore need further analysis. Thus, using aluminum masts or poles currently seems to be the most efficient and accurate method of data capturing in the forests. The present technology cannot predict the GNSS accuracy with full confidence in the forest conditions (Valbuena et al. 2012) and the use of standard surveying methods as a source for real coordinates will still be needed for reliable accuracy assessment. The increasing access to new satellite navigation technologies, however, can limit terrestrial surveying (Valbuena et al. 2014).

252

Acknowledgements This research was funded by the Polish Ministry of Science and Higher Education in the frame of the N N309 114137 project titled »Accuracy analysis of GNSS receiver in forestry environment«. Many thanks to PhD Janusz Walo for his invaluble support in the realization of this project. Special appreciation and thanks to Natalia and Dariusz Górscy for their help in field work and data analysis.

5. References Arslan N., Demirel H., 2008: The impact of temporal ionospheric gradients in Northern Europe on relative GPS positioning. J Atmos Sol-Terr Phy 70: 1382–1400. Andersen H.E., Clarkin T., Witerberger K., Strunk J., 2009: An accuracy assessment of positions obtained using surveygrade global positioning system receivers across a range of forest conditions within the Tanana Valley of Interior Alaska. West J Appl For 24: 128–133. Bach, M., Breuer, L., Frede, H. G., Huisman, J. A., Otte, A., Waldhardt, R., 2006: Accuracy and congruency of three different digital land-use maps. Landscape Urban Plan 78: 289–299. Bakula, M., Pelc-Mieczkowska, R., Walawski, M., 2012: Reliable and redundant RTK positioning for applications in hard observational conditions. Artif satellites 47(1): 23–33. Carter, R. A., 2013: Exploring the dimensions of digital solutions in mine mapping. Eng min j 214(1): 40–41. Cellmer, S., Wielgosz, P., Rzepecka, Z., 2010: Modified ambiguity function approach for GPS carrier phase positioning. J of Geodesy 84(4): 267–275. Cole, J. A., 2004: Global Positioning System Accuracy Under Varying Forest Canopy Conditions. College of Environmental Science and Forestry State University of New York. Deckert, C., Bolstad, P.V., 1996: Forest canopy, terrain and distance effects on global positioning system point accuracy. Photogramm Eng Rem S 62: 317–321. Duncan, S., Stewart, T.I., Oliver, M., Mavoa, S., MacRae, D., Badland, H.M., Duncan, M.J., 2013: Portable Global Positioning System Receivers. Am J Prev Med 442 :19–29. Hasegawa, H., Yoshimura, T., 2007: Estimation of GPS positional accuracy under different forest conditions using signal interruption probability. J For Res 12: 1–7. Jonckheere, I., Nackaerts, K., Muys, B., Coppin, P., 2005: Assessment of automatic gap fraction estimation of forests from digital hemispherical photography. Agr Forest Meteorol 132: 96–114. Landau, H., Vollath, U., Chen, X., 2002: Virtual reference station systems. J Global Position Syst 1(2): 137–143. Lewis, J. S., Rachlow, J. L., Garton, E. O., Vierling, L. A., 2007: Effects of habitat on GPS collar performance: using data screening to reduce location error. J of App Eco 44: 63–671. Mauro, F., Valbuena, R., Manzanera, J. A., Garcı´a-Abril, A., 2010: Influence of Global Navigation Satellite System errors Croat. j. for. eng. 35(2014)2

The Effect of Mounting Height on GNSS Receiver Positioning Accuracy in Forest ... (245–253) in positioning inventory plots for tree height distribution studies. Can J of For Res 41: 11–23. Næsset, E., 1999: Point accuracy of combined pseudorange and carrier phase differential GPS under forest canopy. Can J of For Res 29: 547–553. Næsset, E., Bjerke, T., Øvstedal, O., Ryan, L.H., 2000: Contributions of differential GPS and GLONASS observations to point accuracy under forest canopies. Photogramm Eng Rem S 66: 403– 407. Næsset, E., 2001: Effects of differential single- and dualfrequency GPS and GLONASS observations on point accuracy under forest canopies. Photogramm Eng Rem S 67: 1021– 1026. Næsset, E., Jonmeister, T., 2002: Assessing point accuracy of DGPS under forest canopy before data acquisition, in the field and after postprocessing. Scand. J. For. Res. 17: 351–358. Ordóñez Galán, C., Rodríguez-Pérez, J. R., Martínez Torres, J., García Nieto, P. J., 2011: Analysis of the influence of forest environments on the accuracy of GPS measurements by using genetic algorithms. Math Comput Model 54: 1829–1834. Oruba, A., Leończyk, M., Ryczywolski, M., Wajda, S., 2009: ASG-EUPOS after year. Geodeta 4(167): 10–14. Pirti, A., Gümüs, K., Erkaya, H., Ramazan, G. H., 2010: Evaluating Repeatability of RTK GPS/GLONASS Near/Under Forest Environment. Croatian Journal of Forest Engineering 31(1): 23–33. Puente, I., González-Jorge, H., Martínez-Sánchez, J., Arias, P., 2013: Review of mobile mapping and surveying technologies. Measurement 46: 2127–2145. Puettmann, K. J., Coates, K. D., Messier, C., 2009: A Critique of Silviculture: Managing for Complexity. Island Press Washington-Covelo-London. Rodriguez-Perez, J. R., Alvarez, M. F., Sanz, E., Gavela, A., 2006: Comparison of GPS Receiver Accuracy and Precision in Forest Environments. Practical Recommendations Regarding Methods and Receiver Selection. Shaping the Changes, XXIII FIG Congress, Munich, Germany. Sawaguchi, I., Nishida, K., Shishiuchi, M., Tatsukawa, S., 2003: Positioning precision and sampling number of DGPS under forest canopies. J For Res 8: 133–137.

M. Brach and M. Zasada

Sawaguchi, I., Saitoh, Y., Tatsukawa, S., 2005: A study of the effects of stems and canopies on the signal to noise ratio of GPS signals. J For Res 10: 395–401. Sigrist P., Coppin P., Hermy M., 1999: Impact of forest canopy on quality and accuracy of GPS measurements. Int J Remote Sens 20, 3595–3610. Teunissen, P. J. G., Verhagen, S., 2008: GNSS Ambiguity Resolution: When and How to Fix or not to Fix? VI HotineMarussi Symposium on Theoretical and Computational Geodesy International Association of Geodesy Symposia 132: 143–148. Valbuena, R., Mauro, F., Rodríguez-Solano, R., Manzanera. J.A., 2010: Accuracy and precision of GPS receivers under forest canopies in a mountainous environment. S. J. Agr. Res. 8(4): 1047–1057. Valbuena, R., Mauro, F., Rodríguez-Solano, R., Manzanera. J.A., 2012: Partial Least Squares for Discriminating Variance Components in Global Navigation Satellite Systems Accuracy Obtained Under Scots Pine Canopies. Forest Science 582: 139–153. Valbuena, R., 2014: Integrating ALS with other data sources: field GNSS and optical imagery. In: Maltamo M., Naesset E. & Vauhkonen J. (Eds.) Forestry applications of LIDAR remote sensing. Managing Forest Ecosystems Vol 27. Springer. ISBN 978-94-017-8662-1 Wing, M. G., 2009: Consumer-Grade Global Positioning System Performance in an Urban Forest Setting. J of For 209(107): 307–312. Wing, M. G., Frank, J., 2011: Vertical measurement accuracy and reliability of mapping-grade GPS receivers. Comput Electro Agr 78:188–194. Yoshimura, T., Hasegawa, H., 2003: Comparing the precision and accuracy of GPS positioning in forested areas. J For Res 8(3): 147–152. Yoshimura, T., Nakanishi, A., 2005: Accuracy enhancement of GPS positioning antenna height. Asian Conference on Remote Sensing proceedings. Kyoto University.

Authors’ address: Michał Brach, PhD.* e-mail: michal.brach@wl.sggw.pl Department of Forest Management Geomatics and Forest Economics POLAND

Received: October 01, 2013 Accepted: April 10, 2014 Croat. j. for. eng. 35(2014)2

Michał Zasada PhD. e-mail: michal.zasada@wl.sggw.pl, Laboratory of Dendrometry and Forest Productivity Faculty of Forestry Warsaw University of Life Sciences – SGGW 159 Nowoursynowska str. 02-776 Warszawa POLAND * Corresponding author

253

Subject review

Trends in Woody Biomass Utilization in Turkish Forestry Mehmet Eker Abstract This study aims to provide information to all stakeholders and present an analysis of the trends in the biomass utilization for bioenergy generation to the forestry sector. The analysis focuses on forest resources, production and consumption of wood products, actual situation and trends in the bioenergy sector and forest services. One of the major challenges faced by the Turkish forestry sector is to meet the increasing demand for wood raw material in the wood products industry taking into consideration the trends in the bioenergy sector to promote the renewable energy sources. Therefore, another objective of the study is to determine the available biomass and to reveal its estimated theoretical potential as energy wood. Two projections were performed by using a scenario-based analysis (pessimistic and optimistic projections for bioenergy) of woody biomass supply based on the existing databases, outlook studies, financial balance sheets and progress in renewable energy generation. Special attention was paid to the impact of the forest industry factors that determine the woody biomass potential and to the gaps and uncertainties in the current situation. Consequently, it was found that the bioenergy production based on woody biomass has not been developed yet, although there was 1,494.5 million m3 of growing stock in nearly 21.7 million ha of forestland, in Turkey. However, the total amount of industrial roundwood production increased by approximately 2.12 fold while the fiber chip board production increased by 29 times in the Turkish forestry sector in the last three decades. Surprisingly, the traditional fuelwood production decreased by 69%. The findings reveal that fiber chip board industry is a competitor to the bioenergy sector and it seems to become an obstacle to the modern utilization of woody biomass for energy in near future. As the wood products favored by the forest industry sector, it can be assumed that logging residues will become a primary source of bioenergy without compromising the supply of the industrial roundwood and fuelwood. The estimated theoretical biomass potential that was only obtained from the logging residues and did not include secondary and tertiary wood residues and waste was estimated to be equivalent to 3.5â&#x20AC;&#x201C;5.5 million tons according to the short-scale scenarios. Keywords: woody biomass, biomass utilization, logging residues, wood supply and timber procurement

1. Introduction There is an increasing tendency all around the world to use the renewable energy sources instead of fossil fuels with a view to mitigate climate change, supply renewable energy, adapt to climate change and for other reasons. However, certain criteria such as protection of the forest, enhancement of biodiversity, competitiveness for forest products, sustainability, development of appropriate policies, etc. should be met to use the forest based woody biomass for bioenergy (WBfB) as a renewable energy source. The woody bioCroat. j. for. eng. 35(2014)2

mass is a main component of the forest biomass, a significant potential as a primary energy source in the world, that has been used in various forms ranging from industrial raw material to energy wood through modern and/or traditional ways. Woody biomass from forestry is defined as all of the aboveground and underground biomass of trees, including all by-products and residues (RĂśser et al. 2008). Woody biomass can be generated directly from harvest operations related to the commercial and precommercial forest management, forest restoration and

255

M. Eker

Trends in Woody Biomass Utilization in Turkish Forestry (255–270)

fuel reduction activities. The natural gross potential of biomass energy (including agricultural, forestry and other products) was calculated as 135–150 Mtoe (million tons of oil equivalent)/year while it was assumed that the net potential was 90 Mtoe/year, the technical potential was 40 Mtoe/year, and the economical potential was 25 Mtoe/year in Turkey. According to the data obtained from the Ministry of Energy and Natural Resources (MENR), the total available biomass potential was roughly 8.6 Mtoe per year (Karayılmazlar et al. 2011). Furthermore, the total woody biomass was 1,633 million tons only in the productive forest area and 160.5 tons per hectare according to the statistical data of FAO (Eker et al. 2009). The total recoverable bioenergy potential from agricultural residues, forestry wastes and wood processing residues was estimated to be 16.9 Mton in 2000. The total biomass production was anticipated to be 12.6 Mtoe in 2020 (Kaygusuz and Keleş 2009). For effective utilization of woody biomass, products with the highest added value are generated. Woody biomass utilization options can be grouped into four categories ranging from high value products (sawlog/lumber, veneer, poles, etc.), value added products (mining pole, engineered wood products, etc.), low value products (paper pulp and chips for board, etc.) to minimal value products (logging residues, etc.) (USDA 2007, Pincus and Moseley 2009). Besides several factors influencing the selection of biomass utilization, the current market demands and public necessities are the dominant factors that guide the sharing of the wood products generated from a forest tree. There is a common trend in favor of high value biomass products due to the economic balance, income and expenses. However, a part of woody biomass with low value or minimal value is often called fuel wood (fire wood or wood fuel) used as a traditional or classic source of energy. Moreover, woody biomass has been used as a primary source of renewable energy and heat, while it is expected to become the secondary source of energy to be generated through modern methods in the future. Therefore, it plays a considerable role in the integrated systems of energy and industrial wood supply (Ladanai and Vinterbäck 2009). Despite the increased value of woody biomass for various forms of utilization, the use of some parts such as logging residues has been recently growing in parallel to the development of modern processing and utilization technologies for energy generation worldwide (Aruga et al. 2011), whereas it has also slowly progressed in Turkey (GDF 2009a, Saraçoğlu 2010a, Eker 2011) for modern biomass utilization. The slow progress in the mod-

256

ern utilization of biomass for energy generation is caused by several challenges and barriers such as available biomass potential, subsidies, demand, operational cost efficiency, technological infrastructure, etc. The sustainability of woody biomass utilization depends on technical, economic, ecological and socioinstitutional factors. It is vital and key to achieve the sectoral balance and meet the supply and demand for all sectors in order to identify the raw material requirements of the forest product based sectors so that the woody biomass can also be used for bioenergy production. Therefore, it is important to take into account the available potential, actual production and consumption quantities as well as the trends in woody biomass utilization in order to predict the long term supply possibility and distribution of woody biomass between forest product industry and energy production sector. The wood industry and markets are under significant pressure due to the shortage of raw materials all around the world, especially in some developed countries where woody biomass is commonly used for energy production. The market demand fluctuations that depend on the increasing biomass costs, decreasing forest product prices, electricity prices, etc. have a potential to change the woody biomass utilization options. Furthermore, increased demand for energy wood could be the main driver behind the increase in the wood prices (NEP 2009). The periodical variation depends on time varying forestry paradigms and approaches can lead to the horizontal and/or vertical transfer of wood product range. This shows that the increased demand for woody biomass for bioenergy can also result in the increase of wood prices. Many developed countries promote the use of biomass and other renewable resources for energy generation. Some of the European Union Directives aim to ensure that 20% of the EU’s final consumption of energy should come from renewable sources by 2020 (Wunder 2012). To maintain the stability of wood supply in the forestry industry, many subsidies are provided in a balanced supply demand system. For example, in Germany, it was stated that the use of biomass as a raw material in manufacturing industry should be given priority because of greater value creation and benefit compared to the biomass energy production. The use of biomass as a raw material is guaranteed by the legislation and subsidies are also provided to prevent market distortion (BMU 2009). The literature review of the interactions between conventional use of wood raw material and modern biomass supply (Galik et al. 2009, Guo 2011, Smeets and Faaij 2007) revealed that the biomass availability Croat. j. for. eng. 35(2014)2

Trends in Woody Biomass Utilization in Turkish Forestry (255–270)

for energy production was based on industrial roundwood production. Sedjo (1997) found the interactions between the fuelwood prices and the prices of the traditional wood products, and highlighted that the traditional wood utilization was the main competitor of the biomass supply in energy production. Furthermore, it was stated that some wood products with low prices could also be used in bioenergy sector if the other sources of energy were not appropriate (Perlack et al. 2005). On the other hand, it was predicted that the sawlog and other high value and value added products would be too expensive to be used for bioenergy production due to the market conditions (Hazel 2006, Jonsson et al. 2011). Lundmark (2006) also suggested that the minimal value products such as logging residues could be used in energy sector without any adverse impact on the forest industry. There is a raising awareness in Turkey regarding the modern ways of energy generation from woody biomass. However, it can be claimed that it has not been given as much attention as other primary energy sources (hydro, wind and solar). This is caused by various macro and micro level external and internal factors as follows: Þ The available woody biomass potential is not clear, Þ there is an uncertainty regarding the supply and utilization costs, Þ there is no sectoral structure, Þ the institutional and social structure in forestry is not convenient, etc. The purpose of this study was to explore the following research questions: which type of woody biomass can be used for energy production, what is the theoretical biomass potential quantity, and which sectors might compete for raw material. The aim of the study was also to analyze using a bottom up analysis of some key factors such as the relationship between the industrial and fuelwood utilization. It was assumed that it was possible to project the type and proportion of woody biomass available for bioenergy production by using the wood production quantities at the scale of time series with historical data. To support the scenario-based assumption, the supply driven approach dependent on trend analysis was used in the study. Special attention was paid to the data and projections from existing database and outlook studies. Especially national trends and important drivers of change in woody biomass utilization were evaluated in addition to the past experience and projections of future use and supply of wood resources as well as the developments in the wood product market and bioenCroat. j. for. eng. 35(2014)2

M. Eker

ergy sector in Turkey for the last three decades. Thus, an attempt was made to fill the gaps and lack of knowledge on woody biomass utilization for the supply of biomass energy.

2. Material and Method The research material used in this study was based on the Turkish forest inventory and forestry sector data (GDF 2011a, FAO 2011, EFSOS 2012, FAO 2012, FS 2012). The national statistical data and information contained the forest resources, forest biomass potential, wood procurement system, production and consumption of forest products, supply and demand profile of the forestry sector, procurement cost and selling price of forest products, report on wood stream in the forest industry sector, overall status of the bioenergy sector, etc. Future outlook and technological prospects of woody biomass utilization were predicted on the basis of the data and reports. The time horizon for the study of past trends was based on data availability. In most cases, historical statistics dated back to the year 1973. However, the analysis of trends only covered the past 30–32 years. The MS Excel was the main tool used to store data, while the statistical data was processed and analyzed through SPSS statistical software. To assess the developments and changes in the forestry and bioenergy sectors, the following information was used:

2.1 Forest resources in Turkey In order to describe the trends and to argue the sustainability of the future wood supply, the data about forest resources was used in the study. Table 1 shows the data about the forest areas. The forestland has an area of approximately 21.67 Mha (million hectares) and covers 27.8% of the surface area of the entire country, 53% of which is productive forest land. (FS 2012). The productive and high forest area, where woody biomass could be produced, accounts for 10.3 Mha within the total forestland. Moreover, Table 2 shows the trends in growing stock for the period 1973 and 2012. The total growing stock accounts for 1.49 billion m3 among the forest resources. The growing stock has increased during the last four decades. It is also correlated with forest areas. In order to give an indication of the sustainability of forest management for a comparative assessment, Table 3 presents the historical data on annual increment within the time horizon. The annual increment in the productive forestland is 40.02 Mm3 (million cubic meter), 89 of which is from high and productive forest and the

257

M. Eker

Trends in Woody Biomass Utilization in Turkish Forestry (255â&#x20AC;&#x201C;270)

Table 1 Distribution of forest area between 1973 and 2012 (FS 2012) Total forest area

Forest form

Productive forest area

Degraded forest area

years

1973

20,199,296

100

8,856,457

11,342,839

1999

20,763,248

100

10,027,568

10,735,680

2005

21,188,747

100

10,621,221

10,567,526

2009

21,389,783

100

10,972,509

10,417,274

2010

21,537,091

100

11,202,837

10,334,254

2012

21,678,134

100

11,558,668

10,119,466

Table 2 Distribution of growing stock (FS 2012) Forest form years

Total

Productive forest

Degraded forest

1973

935,512,150

100

847,033,015

88,479,135

1999

1,200,791,637

100

1,113,612,229

87,179,408

2005

1,288,124,772

100

1,199,034,187

89,090,585

2009

1,374,240,926

100

1,290,450,115

83,790,811

2010

1,428,504,717

100

1,347,453,572

81,051,145

2012

1,494,454,538

100

1,417,482,684

76,971,854

Table 3 Distribution of annual increment (FS 2012) Forest form years

Total 3

Productive forest 3

1973

28,063,205

100

25,604,869

2,458,336

1999

34,269,650

100

31,306,039

2,963,611

2005

36,282,291

100

33,834,897

2,447,394

2009

38,454,916

100

36,156,989

2,297,927

2010

40,061,594

100

37,800,646

2,260,948

2012

42,179,115

100

40,020,179

2,158,936

rest is from coppice. Only less than 50% of the net annual increment in forests available for woody biomass supply is harvested in Turkey. Therefore, the annual increment is a limitation criterion for the wood production potential.

258

Degraded forest 3

2.2 Woody Biomass Demand and Supply The data used in this study regarding the wood resources was divided into two groups: the supply side containing the state forest resources but not private and illegal resources; and the demand side conCroat. j. for. eng. 35(2014)2

Trends in Woody Biomass Utilization in Turkish Forestry (255–270)

M. Eker

Table 4 Breakdown of wood production from 1980 to 2012 (GDF 2012a, FS 2012) Standing tree volume

Year

Log

Telephone pole

Mining pole

Other industrial wood

Pulp and paper wood

Total roundwood

Fiber chip wood

Thin pole

Total industrial wood

Fuel wood x1,000 stere

x1,000 m3 1980

11,225

5,343

111

621

398

144

6,617

164

–

6,781

21,949

1985

8,932

3,892

264

530

265

1,572

6,523

884

–

7,407

14,289

1990

7,560

3,310

513

639

923

5,445

1,113

6,581

12,145

1995

9,192

3,578

134

498

936

1,558

6,704

1,320

8,046

9,539

2000

8,880

3,007

155

413

830

1,533

5,938

1,371

7,329

7,861

2005

10,009

2,936

405

726

1,528

5,672

2,409

8,100

7,667

2010

16,424

4,375

577

788

2,146

7,940

4,608

12,569

7,194

2011

17,648

4,839

686

874

2,383

8,853

4,663

13,533

6,778

2012

16,700

5,028

693

875

2,334

8,990

5,425

14,424

6,432

m3 – cubic meter Stere – Stacked cubic meter =0.7 m3 =500 kg

Table 5 The supply and demand balance in last decade (MFWA 2013) Industrial wood, x1,000 m3

Fuel wood, x1,000 stere

Years

Supply

Demand

Difference

Supply

Demand

Difference

2002

11,305

12,359

–1,054

16,137

16,650

–513

2003

10,620

11,780

–1,160

15,981

16,359

–378

2004

11,553

13,189

–1,636

15,900

16,223

–323

2005

11,400

13,547

–2,147

15,067

15,519

–452

2006

12,599

14,440

–1,841

14,123

14,411

–288

2007

13,353

15,832

–2,479

13,717

14,093

–376

2008

14,841

15,297

–456

14,007

14,080

–73

2009

14,763

15,943

–1,180

14,101

14,081

2010

15,869

17,455

–1,586

13,897

14,357

–460

2011

16,832

17,705

–873

13,451

13,768

–317

* The supply and demand quantity includes state, private, off the record, import, export and illegal production and consumption

taining the industrial wood and fuel wood. The demand for woody biomass was defined as all wood raw materials supplied from forests. Table 4 shows a summarized part of statistical data regarding the production of industrial round and fuelwood obtained from forests via legal and planned allowable cuts in the last Croat. j. for. eng. 35(2014)2

three decades. The tabular data was used in econometric modeling in order to produce projections of the supply of wood products and the material use of woody biomass. Nearly all forests in Turkey are owned and managed by the state. 82% of the domestic industrial

259

M. Eker

Trends in Woody Biomass Utilization in Turkish Forestry (255–270)

Table 6 Average selling price (fixed price) of wood products in last decade (TL/m3 – stere) Wood product types and years

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

Log/sawnwood, Pine woods, TL/m

184

199

229

225

211

204

189

171

172

212

Mining pole, TL/m3

141

160

188

183

180

164

157

146

150

178

Other industrial wood, TL/m

131

151

170

169

161

150

142

127

133

154

Fiber chip wood, TL/stere

Pulp and paper wood, TL/m

104

125

130

124

117

105

119

134

Fuel wood, TL/stere

*1 € = 2.4 TL in 2011

able way. In this framework, the conventional wood procurement is realized through commercial harvesting (27,263 ha), thinning (466,427 ha), rehabilitation (320,525 ha) and conversion operations (81,416 ha), and fuel reduction in firebreaks and forest roadsides (GDF 2012a). The low and minimal value biomass may be left unexploited in the forest area if the selling price does not meet the harvesting and transportation cost due to economic balances. The high value and value added wood products, such as sawlog lumber and poles, etc. are primarily preferred and harvested in all operations. Logging residues are often left on the forest floor. The low-value branch woods are sold to the suppliers for fuel wood and chipboard production and majority of them are subsidized for rural residents with stumpage prices. Wood procurement operations have been mainly performed by forest villagers as per the legislation and rarely by forest contractors. The workforce potential and employment capacity of forest villagers is around 300,000 people for annual wood harvesting that account for 13–15 million men/day. Conventionally, cutto-length harvesting method has been preferred in

roundwood is supplied from the state forests. The total industrial and fuel wood supply and demand in the period from 2002 to 2011 are presented in Table 5. The state owned forests are the largest sources of wood supply. Therefore, the production from these forests account for 70–75% of the total domestic consumption. Domestic industrial wood production increased by 49% from 2002 to 2011, while the total consumption increased by 43%. However, the consumption of the fuel wood had a downward trend. In order to explain the reason why the high value wood products are preferred, a breakdown of the average selling price and procurement cost are illustrated in Table 6 (GDF 2012b). Significant fluctuations have been observed in the prices of the wood product. The price of roundwood decreased by around 70% in the period before 2001, while a horizontal trend was observed after 2001.

2.3 Wood Procurement System The forestry operation goals of the GDF dictate harvesting high quality wood products in a sustain-

Table 7 The annual private placement wood product demands in 2011 (GDF 2011b, GDF 2012b) Source of wood supply

Installed production capacity

Actually processed raw material

Capacity utilization ratio

GDF

Private sectors

Imports

x1,000 m3

Chip board

4,691

2,210

1,480

654

Fiber board

2,428

100

730

1,062

636

Fiber chipboard

7,119

4,683

2,210

1,716

712

Pulp and paper

1,754

1,098

485

232

381

Total

8,873

5,736

2,695

1,948

1,093

Shared wood product types

260

Croat. j. for. eng. 35(2014)2

Trends in Woody Biomass Utilization in Turkish Forestry (255–270)

M. Eker

Table 8 Wood industry sectors and use of raw material (Sakarya and Canli 2011, GDF 2011b) Wood industry sector demands Shared products

Institution number

Production quantity in 2010

Production capacity in 2010

Production quantity in 2011

Estimated capacity in 2011

Expected capacity

x1,000 m3/year

Issue

Chip board

3,100

5,040

4,978

6,500

8,100

Fiber board

3,300

6,600

3,983

7,170

7,966

Plywood, etc

110

180

250

360

450

Veneer

165

107

165

190

Lumber

10,000

6,243

9,200

9,000

9,200

15,000

Total

10,111

12,849

21,185

18,318

23,395

31,706

Turkish forestry; the harvesting residues are left in stand. The operations have been carried out through moderate to intermediate harvesting technology (Eker et al. 2009).

2.4 Forestry sector The forest industry demands for wood raw material could be supplied by GDF with both private placement method (20–22 %) and public action method (25–30%), while the rest of the demand is supplied by the private sector and through import (Table 7) (GDF 2012b). The production quantities and capacities are shown in Table 8 to describe the current market conditions of wood industry in Turkey.

2.5 Energy production and consumption in Turkey Turkey’s annual primary energy production was 29.3 Mtoe, while the consumption was 106.3 Mtoe in 2008. The primary energy consumption is expected to be 222.4 Mtoe in the year 2020 (Karataş and Gül 2012). Turkey’s demand for energy has rapidly increased and is expected to grow by 40% in 2035 (Yıldız 2011). Also, it was argued that the energy potential of the total woody biomass (including residues of wood product industry) was 7–8.5 Mtoe in Turkey. Table 9 depicts the annual biomass energy potential of forest residues and wood industry residues with other biomass.

Table 9 Annual biomass energy potential (Demirbaş 2008) Annual potential, Mtons

Energy potential, Mtoe

Forest residues (including logging residues)

5.5

Wood industry residues

1.8

Other

24.7

Total

117

Type of biomass

According to the statistics of the MENR (2012), the material quantity for fuel wood was 3.4 Toe as the primary source of energy, accounting for 10.4% of the total national sources for residential heating. However, the final energy consumption of fuel wood is expected to reach 3.075 Mtoe in 2020 (Gökçöl et al. 2009). The estimate of the traditional and modern biomass requirements is shown in Table 10.

2.6 Method Forest resources were taken into account as the only source of supply for woody biomass, while the demand side was divided into the industrial wood and fuel wood categories. Fuel wood (also referred to as energy wood or wood fuel) was classified into two groups: traditional fuel wood (fire wood) and modern

Table 10 Traditional and planned modern biomass requirements for energy production (Saraçogˇlu 2010a) Years

1999

2000

2005

2010

2015

2020

2025

2030

Traditional biomass, Mtoe

7.012

6.965

6.494

5.754

4.790

4.000

3.345

3.310

Modern biomass, Mtoe

0.005

0.017

0.776

1.660

2.530

3.520

4.465

4.895

Total biomass, Mtoe

7.017

6.982

7.260

7.414

7.320

7.520

7.810

8.210

Croat. j. for. eng. 35(2014)2

261

M. Eker

Trends in Woody Biomass Utilization in Turkish Forestry (255–270)

fuel wood. Multiple correlation analysis was performed in order to determine the sharing of woody biomass types and relation between the product components, and to interpret the trends. The fact was taken into account that demands for the wood raw material was limited to the supply, i.e. to total allowable cut/production quantity according to the management and silvicultural principles and to the annual increment quantity. Therefore, the historical course of production quantities was taken as the basis to analyze the trend in the utilization of woody biomass. For data reliability and acquisition, a table was prepared to show the wood raw material harvests in the period from 1973 and/or 1980 to 2012. The annual and periodical (with five year intervals) changes were calculated according to the sharing of the products by using this tabular data on MS Excel. In order to determine the trend in the annual supply of woody biomass, time series analysis was performed (Güngör et al. 2004). For goodness of fit statistics, regression analysis method (least square method) was used. In order to select the best mathematical trend model, among quadratic, linear, exponential, polynomial and S-curve trend models, the R2 and RMSE accuracy criteria were used having the highest coefficient of determination/R2 and the lowest root mean square error/RMSE. The trend analyses showed that Polynomial Trend Model usually had better accuracy values compared to the other models. In order to determine the periodical demand depending on the supply of woody biomass, the average value of »comparative annual variation (growth or shrinkage) quantity and ratio« was used to estimate the annual fixed changes in the periodical wood supply. As a different form of moving averages in a time series model, this method was based on the value of the differences in the supply between two consecutive periods to demonstrate the variation in the mix of wood products. It was claimed that the increase or decrease values found with this method could be a fixed coefficient to project the wood supply in the future. Such annual and periodical values were used to estimate the production amount of woody biomass by 2023 when the strategic forestry plan will expire. The period until 2023 also accounts for a decade required for the amortization of the bioenergy investment costs. Furthermore, two scenarios were developed to interpret the trend analysis of woody biomass utilization. The first scenario was based on the fixed structure of woody biomass supply and demand that was regarded as a supply driven projection scenario. The estimations in the scenario were made using the statistical trends and qualitative (by means of simple/

262

graphical regression) analysis of the annual production data regarding the state forestry sector. This scenario symbolized a pessimistic outlook (low potential for WBfB) in favor of the use of woody biomass for bioenergy production in Turkey. This scenario represented a condition in which the industrial wood utilization would increase and the ecological/environmental sensitivity would not surpass the socio economic objectives of the forestry industry. The second scenario was optimistic (high potential for WBfB) for the bioenergy sector. The percentage derived from the demand profile of the forest industry was assumed to reflect the impact of the increased demand on WBfB. Woody biomass demand for total industrial wood was predicted by means of sector reports (OAİB 2011) and planned goals of state forestry sector. The energy wood requirements and the woody biomass market demand for bioenergy could then be calculated by reviewing the literature on sectoral basis. In this scenario, an unsteady coefficient obtained from the actual structure of the energy market was used to determine the variable wood demand. In this scenario, it was assumed that the modern utilization of biomass would increase depending on the production of the industrial wood and bioenergy policies and that the traditional fuel wood would be completely or partially used as energy wood in addition to the logging residues. In both scenarios, a recovery factor (7–20% green tons of total wood production volume (Eker 2011, Eker et al. 2013, Çoban and Eker 2014) was used to estimate the theoretical potential of logging residues to substitute the energy wood supply.

3. Results and Discussion Turkish forestry has a growth trend with respect to the increasing forestlands and forest resources. Forest areas were enlarged by 7.3% in 39 years from 1973 to 2012 (Table 1) due to the aggregated forestry approach and best management practices. Furthermore, the area of the productive forests increased by 30.5% in the same time frame. According to the polynomial trend equation y = 1,194.9x 2 – 5E + 06x + 5E + 09; R2 = 0.997

(1)

Where: y forest area, x year. the area of the forests is expected to reach 22.5 Mha by 2023. This result guarantees the spatial sustainability of the forest resources for woody biomass. Furthermore, in line with the forest land, the total growing stock in all forestland increased by 59.7% Croat. j. for. eng. 35(2014)2

Trends in Woody Biomass Utilization in Turkish Forestry (255–270)

M. Eker

during 39 years (Table 2). The growing stock refers to the inventory of standing trees in a forest. Therefore, wood harvesting, harvest scheduling, wood production intensity, and all forestry operations are considered according to the growing stock. This is also an excellent indicator for capital investment into wood products industry in the future. According to polynomial trend model y = 332,521x 2 – 1E + 09x + 1E + 12; R2 = 0.992 (2)

the maximum growing stock is expected to reach 1.85 billion m3 by 2023. Moreover, the annual increment increased by 50% in the last 39 years and, and from 1.38 m3 to 1.86 m3 per hectare. According to the polynomial trend model y = 9,839.2x 2 – 4E + 07x + 4E + 10; R2 = 0.984

(3)

the annual increment is predicted to reach maximum 46–49 Mm3 by 2023. Depending on the incremental structure of forest resources, the annual allowable cut is also expected to increase as a function of standing tree volume. Eventually, the amount of standing tree volume has consistently soared since 2005, and peaked in recent years (Fig. 1). In Turkish forestry, the average annual allowable cut is approximately 16 Mm3 according to last five year data, and it increased by 75% in the last decade. The trend has a polynomial incremental momentum and it shows that the standing tree volume should be between 25–30 Mm3 by 2023. This symbolizes theoreti-

Fig. 2 Relationship between log volume and years

Fig. 3 Relationship between telephone pole volume and years

Fig. 1 Relationship between standing tree volume and years Croat. j. for. eng. 35(2014)2

cal wood raw material to be supplied from state forests as both industrial and fuelwood, and threshold value of the annual supply quantity of woody biomass. The available potential of woody biomass for bioenergy depends on distribution of wood products to be obtained by total wood supply, as shown through Fig. 2 to Fig. 10. Based on the production ratio and polynomial trend model of wood raw material, it was concluded that the industrial wood production

263

M. Eker

Trends in Woody Biomass Utilization in Turkish Forestry (255–270)

Fig. 4 Relationship between mining pole volume and years

Fig. 6 Relationship between pulp and paper wood volume and years

Fig. 5 Relationship between other industrial wood volume and years

Fig. 7 Relationship between total roundwood volume and years

y = 12.976x 2 – 51,645x + 5E + 07; R2 = 0.82

(4)

and especially sawlog production 2

y = 7.8263x – 31,258x + 3E + 07; R = 0.84

(5)

would continue to increase significantly. Furthermore, it was also estimated that there would be a tremendous increase in the fiber-chipboard production

264

y = 5.8288x 2 – 23,150x + 2E + 07; R2 = 0.91

(6)

as shown in the figures above. This is an overwhelming growth rate, higher than the growth in other wood products in the last three decades. This is also consistent with the data given in Table 4 - the fiber chipboard wood production increased by 33.1 times from 164 thousand cubic meters to 5,425 million cubic meCroat. j. for. eng. 35(2014)2

Trends in Woody Biomass Utilization in Turkish Forestry (255–270)

Fig. 8 Relationship between fiber-chipboard wood volume and years

M. Eker

Fig. 10 Relationship between fuelwood in stere and years increasing imports of manufactured wood products. Over 70% of the demand of the fiber chipboard industry could be supplied by GDF, and the rest by private plantations and imports (GDF 2011b). However, the traditional fuel wood production remarkably decreased 3.4 times, from 21,949 to 6,432 stacked cubic meters: y = 21.829x 2 – 87,542x + 9E + 07 ; R2 = 0.95

Fig. 9 Relationship between total industrial wood volume and years ters in the last 32 years. At the same time, fiber chipboard wood products had the highest proportion (24%) in total woody biomass (standing tree volume) in the period between 2001 and 2012. This is a result of the increased demand from the fiber chip board industry. This sector is the fourth biggest producer in the world after China, USA and Germany (OAİB 2011). Therefore, the demand of the fiber chipboard industry for wood raw material has been steadily increasing due to the Croat. j. for. eng. 35(2014)2

(7)

The decline has been supported by GDF through National Forestry Programs (ÇOB 2004) and Strategic Planning in Turkish Forestry (GDF 2009b) for fuel wood production. The aim was to reduce the fuel wood production to 3.5 Mm3 (5 million stere) in 2014. Depopulation within and near the forest areas, use of substitution or alternative energy sources instead of fuel wood for heating and cooking, seasonal emigration of local people living in the forest areas, etc. led to a decrease in the consumption of the fuel wood by around 70% in last three decades. This findings are supported by the fact that the logging residues as fuel wood (Eker et al. 2011) have not been collected any longer by rural people for traditional heating and cooking. When compared with the increase in the fiberchipboard production and decrease in the fuel wood production by means of correlation analysis, a negative/inverse significant (p<0.01; r=–0.674) relationship was found between the chipboard and fuel wood supply in Turkey. The severe decline in the fuel wood production demonstrates the horizontal shift in product

265

M. Eker

Trends in Woody Biomass Utilization in Turkish Forestry (255–270)

range from fuel wood to fiber chipboard. Fiber chipboard and fuel wood sector can use the same type of woody materials such as low and/or minimal value added products. According to the findings, it can be claimed that the fiber chipboard industry is a competitor to the fuel wood and other non-industrial wood products and has an extensive manipulation capacity for all woody biomass at low procurement cost. The results of the analysis, graphical inferences and market reports are also consistent with this assumption (Sakarya and Canlı 2011, GDF 2011b). On the other hand, the biomass has the potential to become a more reliable source of energy; therefore, there is a need for improvement in collaboration with all relevant sectors in order to increase the utilization of WBfB. Price competitiveness, raw material distribution, sustainable supply of woody biomass and market conditions in the forest industry have important roles in the growth of biomass in energy sector (Gökçöl et al. 2009, Heinimo 2011). Thus, it seems that the use of annual modern biomass demand will increase significantly up to almost 5 Mtoe by 2030 (Saraçoğlu 2010a, Saraçoğlu 2010b), although biomass energy only accounts for 4–5% of primary energy consumption in Turkey (Karataş and Gül 2012). However, to satisfy the demand, Turkish Assembly ratified a renewed legislation »Renewable Energy Law« at the end of 2010, which proposed that the highest purchase guarantee for the selling price of the energy produced from biomass biogas power would be 13.3 cent € /kWh for 10 years (Türkyılmaz 2011). This subsidy encouraged the energy producers, thus the number of power plants licensed for electricity generation from biomass resources increased to 11 and the capacity to roughly 50 MW (Gözen 2011). To promote the progress, the GDF, which was the only state owned wood producer, declared that maquies, rhdododendron woods and residues of red pine (Pinus brutia) forest could be used in bioenergy sector as raw material. As in most developed countries, biomass energy sector primarily uses the minimal and/or low value added products and by-products, including logging residues, small sized trees, mill residues and other coarse debris (Hakkila 2004, Mendell et al. 2011). Despite this trend in favor of WBfB, selling price is a key factor and deterministic component for utilization and sharing of woody biomass, such as fuel wood and logging residues, among the competing sectors. As stated by Spinelli (2011), if the selling price offered by a traditional supplier, vendor or user cannot cover the cost of recovering the main forest products, then the products will end up as by-products for energy or other utilization. The higher price proposed by the energy

266

supplier not only may contain the cost of recovering non-merchantable trees and tree parts, but it can also exceed the price of low value forest products. In Turkey, although the bioenergy sector has not been developed yet, the particleboard supplier has offered a higher purchase price for tree branches, small diameter trees and wood chips and this price is higher than the price offered by fuel wood supplier. The fiber chipboard industry sector is dominant in terms of the price competition and raw material demand. Therefore, there has always been a tendency towards the high value types of woody biomass. However, this is a driver of the change of wood product market and negative factor for WBfB potential. It can be claimed that the fast growing demand of particleboard industry has also influenced the fuel wood demands because of their diameters, which are greater than those of the logging residues. Therefore, logging residues and small diameter branches have a potential to be used as energy wood source, so that more forest biomass can be actually utilized to reduce competition between the industry and energy sectors. This trend confirmed that the demand of fiber chip board industry was a barrier to bioenergy sector in providing WBfB. Depending on the results on forest resources, woody biomass (theoretical) potential and development of bioenergy, the scenario analysis gave the following results: In the pessimistic scenario, a demand profile based on the increasing industrial wood production and consumption was taken into consideration for energy woody biomass. It was assumed that the demand for both industrial roundwood and fiber chipboard wood would increase, while the fuel wood demand would decrease. This scenario limited to the energy wood procurement from commercial harvesting operations. The growth of the fiber chipboard industry has strengthened the possibility that a part of logging residues can be used in the particleboard production. It makes it possible to use only the unutilized part of logging residues as bioenergy resource. The selling method, such as standing tree sale (time bargain) or public auction, behavior of the logging operators and the request of the manager or customer have an impact on the decision whether to leave the logging residues in the forest site or not. During the conventional harvesting operations, small diameter branches, with the diameter smaller than 4 cm, are not collected as fiber chipboard wood or fuel wood and they are left in the stand as residue because of their high procurement cost and non-merchantable value. If forest villagers do not collect the woody biomass and logging residues for fuel wood based on their legal rights, they can be used in the bioenergy sector. Croat. j. for. eng. 35(2014)2

Trends in Woody Biomass Utilization in Turkish Forestry (255–270)

M. Eker

Table 11 The estimated projection of the supply of woody biomass (x1000) Year

2023

Scenario

Standing tree volume

Sawlog

Leaf/fiber chipboard

Fuelwood

Logging residues

Mstere

Mton

Pessimistic

27.55–30.87

5.81–12.05

9.35–29.77

2.40–5.04

1.60–2.80

Optimistic

19.17–27.55

4.69–5.81

6.72–10.23

5.04–5.15

3.83–5.51

The theoretical potential of logging residues to be used in bioenergy sector depends on harvesting volume. Therefore, the estimated supply trend of woody biomass is given in Table 11. It was calculated that the theoretical potential of WBfB, including logging residues plus a portion of fuel wood, should be minimally 1.6 Mtonsgreen as energy wood. The available potential is not substantially sufficient to supply the demand of the large or medium scaled bioenergy sector, with the pessimistic scenario. With the optimistic scenario, the trend of the modern utilization of woody biomass for power and heat was taken into account. It was assumed that the international protocols (Kyoto and others) that Turkey signed promoted the use of woody biomass in bioenergy to reduce greenhouse gas emissions, provide cleaner air, reduce fuel materials on forest floor, etc. In the light of the development around the world, Turkey has a tendency to use the logging residues as a source of bioenergy (EFSOS 2012). According to the trends, it can be claimed that there will be a market opportunity for energy wood. The theoretical potential of WBfB dependent only on logging residues was estimated to be 3.8–5.51 Mtonsgreen according to the optimistic scenario (Table 11). It is possible to expand the annual WBfB potential range to 5.5–7 Mtonsgreen when adding thin branch wood, which is inutilizable for chipping, and other forest biomass resources such as shrubs, stumps, very small diameter tree residuals, fast growing species, energy forestry crops, etc. The use of branch woods and other parts of trees for fuelwood has been decreasing because there is a rural depopulation and coal and other fossil fuels (natural gas, fuel oil) are preferred for heating and cooking. However, the woody biomass that is not used for fuelwood can be used in the fiber chipboard industry or pulp and paper wood industry. In this sense, the horizontal variations in the sharing of the products depend on the market demands, production costs and selling prices. In this study, it was found that the roundwood production has increased and thus the production of the fuelwood decreased. This indicates that there has been a change in the sharCroat. j. for. eng. 35(2014)2

ing of the products although the production amount has increased. The quantities of utilizable WBfB wood are variable, and depend on some stand characteristics, types of silvicultural interventions and harvesting methods. 7–15% can be utilized as green tons of residual biomass from clear cutting (Eker et al. 2013) and 20–30% of the harvested small diameter trees from thinning operations. However, the policies encouraging the use of renewable natural resources for energy production can result in significantly higher demand for bioenergy and other renewables in the future. The rapid energy expansion in biomass energy production seems to continue until the electricity prices decrease, the wood supply exceeds the demand, and new fossil fuel based cogeneration plants have emission rights at no cost.

4. Conclusion The relationship between demand and supply in the wood industry has an impact on the development of WBfB market. The national woody biomass supply is a function of the current growing stock, standing tree volume, annual growth rate and demand profile with current wood prices. Turkey has a tremendous natural woody biomass potential, whereas technical, economic and ecological potential for energy wood is limited by various factors. The industrial wood supply is insufficient to meet the demand of the domestic wood industry. Although the annual wood procurement increased within the last three decades, the growing demand has not been satisfied, yet, because the wood industry has a growing trend. Particularly the fiber chipboard industry has increased the production capacity and demand. However, fuelwood production from state owned resources has a severe decline. There is a significant relationship between the fiber chipboard and fuelwood sectors. The GDF has partially guaranteed the supply to the fiber chipboard sector via private placement method. Furthermore, the GDF have to subsidize the fuelwood demands of the forest villagers. The grow-

267

M. Eker

Trends in Woody Biomass Utilization in Turkish Forestry (255–270)

ing demands lead to horizontal product variation between low and minimal value woody biomass. However, a number of tradeoffs between different needs of Turkish forest sector have also changed the woody biomass supply. On the other hand, Turkey is willing to mitigate the greenhouse gas emissions and to meet other environmental requirements by using the renewable energy resources such as biomass. Therefore, woody biomass should be utilized through modern ways. However, the utilizable WBfB is also insufficient due to the lack of supply. Logging residues as a type of biomass is a crucial alternative for small scale energy production, but it is insufficient to satisfy long term biomass energy production requirements. If the wood markets as well as the production and consumption trend remain unchanged, only the logging residues can be used for bioenergy production and thus it could be reasonable and economic to establish only local and small scale heat and power plants. If the demands for woody biomass for energy exceed the supply of logging residues, the entire wood products industry can be influenced by the increased raw material prices. There is uncertainty as to how the demand growth and price increases will influence the production of wood products, market conditions and supply agreements between GDF and forest products industry. As regards the demand, the global bioenergy production from woody biomass will have an impact on the fiber chipboard sector. The shrinkage of the fiber chipboard market and increased necessity to use the renewable energy sources have enabled the use of fuelwoods obtained from the branches for bioenergy and created an opportunity for sustainable raw material supply. The methodology used in this study was based on a static approach to the supply of forest growth, forest areas, wood production, etc. No attempt was made in the study to include technological and socio-economical system to calculate the implementation potential of WBfB supplied only from state owned forests. The study results present a rough estimate of the trend of woody biomass utilization for bioenergy in 2023. Further research should include all variables related to technical, economical and ecological limitations and other woody biomass resources to make a comprehensive projection of the future demand and supply.

Acknowledgment I would like to give a special thanks to the Scientific and Technological Research Council of Turkey (TUBITAK) for providing a financial support via the Project No. 110O435. Any data and result obtained from this project, except those referenced, were not used in the study.

268

5. References Aruga, K., Murakami, A., Nakahata, C., Yamagachi, R., Yoshioka, T., 2011: Discussion on Economic and Energy Balances of Forest Biomass Utilization for Small Scale Power Generation in Kanuma, Togichi Prefecture, Japan. Croatian Journal of Forest Engineering 32(2): 571–586. BMU, 2009: National Biomass Action Plan for Germany Bundesministerium für Umwelt, Naturschutz und Reaktorsicherheit (BMU) 11055 Berlin, Germany www.bmu.de (Accessed April 2012). Cilan, S., 2011: Yılı Üretim Faaliyetleri, Üretim Şube Müdürlüğü, İşletme ve Pazarlama Daire Başkanlığı, General Directorate of Forestry, Ankara. ÇOB, 2004: Türkiye ulusal ormancılık programı (National forestry program of Turkey) (2004–2023). Ministry of Environment and Forestry, Ankara, 95 p. Çoban, H.O., Eker, M., 2014: A hierarchical approach to estimate spatially available potential of primary forest residues for bioenergy, BioRes. 9(3): 4076–4093. Demirbaş, A., 2008: Importance of biomass energy sources for Turkey. Energy Policy 36(2): 834–842. EFSOS, 2012: The European forest sector outlook study-II (2010–2030, UN/ECE/TIM/SP/28, 111 p. Eker, M., Çoban, H.O., Acar, H.H., 2009: The procurement problem of logging residues in Turkish forestry. Proceedings in FORMEC, June 21st–24th, Prague-Czech Republic, 115–124. Eker, M., 2011. Assessment of procurement systems for unutilized logging residues for Brutian pine forest of Turkey. African Journal of Biotechnology 10(13): 2455–2468. Eker, M., Acar, H.H., Özçelik, R., Alkan, H., Gürlevik, N., Çoban, H.O., Korkmaz, M., Yılmaztürk, A., 2013: Ormancılıkta hasat artıklarının tedarik edilebilirliğinin araştırılması – Sonuç Raporu (Investigation on supply possibilities of logging residues in forestry – Final Report), TUBITAK, Project No:110O435, Ankara. FAO, 2011: State of the world’s forest. FAO, Rome, 179 p. FAO, 2012: Yearbook of Forest Products 2008–2012, FAO Forestry and Statistics Series No:47/203. Rome 358 p. FS, 2012: Forestry Statistics, Republic of Turkey Ministry of Forestry and Water Affairs. A Publication of Official Statistics Programme, Publication Number – 01, Ankara 84 p. Galik, C.S., Abt, R.C., Wu, Y., 2009: Forest Biomass Supply in the Southeastern United States – Implications for Industrial Roundwood and Bioenergy Production. Journal of Forestry 107(2): 69–77. GDF, 2009a: Türkiye’ de orman biyokütlesinden temiz enerji üretimi (Clean energy production from woody biomass in Turkey), Bioenergy Working Group, General Directorate of Forestry, www.ogm.gov.tr (Accessed 20 April 2012). GDF, 2009b: Stratejik Plan 2010–2014 (Strategical Plan 2010– 2014), Department of Strategy Development, General Directorate of Forestry, Ankara, 123 p. Croat. j. for. eng. 35(2014)2

Trends in Woody Biomass Utilization in Turkish Forestry (255–270) GDF, 2011a: Üretim ve pazarlamada mevzuat ve uygulamaları (Legislation and application in forest harvesting and marketing), Department of Business and Marketing, General Directorate of Forestry, Ankara. GDF, 2011b: 2010 Yılı Orman Ürünleri Piyasasindaki Gelişmeler ve 2011 Yili Beklentileri (Developments of forest products markets in 2010 and Expectations of 2011). General Directorate of Forestry, Ankara, 7 p. web.ogm.gov.tr/ birimler/merkez/isletmepazarlama/ (Accessed February 2012). GDF, 2012a: 1980–2012 Yılları Asli Orman Ürünleri Üretim Programı ve Gerçekleşmeleri (Timber Harvesting Program and Realizations Between 1800 and 2012). General Directorate of Forestry http://web.ogm.gov.tr/birimler/merkez/isletmepazarlama/ (Accessed March 2012). GDF, 2012b: 2011 Yılı İdare Faaliyet Raporu (Administration Annual Report for 2011). General Directorate of Forestry, Ankara, 90 p. Gokçöl, C., Dursun, B., Alboyacı, B., Sunan, E., 2009: Importance of biomass energy as alternative to other sources in Turkey. Energy Policy 37(2): 424–431. Gözen, M., 2011: Düzenleme Perspektifinden Elektrik Piyasasında Gelinen Aşama ve Hedefler (Stage Reached,and Goals in the Electricty Market Regulatory Perspective). The Energy market Regulatory Authority, ICCI, İstanbul. Guo, Z., 2011: Forest Biomass Utilization in the Southern United States: Resource Sustainability and Policy Impacts. PhD thesis, University of Tennessee. Güngör, İ., Kayacan, C., Korkmaz, M., 2004: Endüstriyel Odun Hammaddesi Talebinin Tahmininde Yapay Sinir Ağlarının Kullanımı ve Bazı Tahmin Yöntemleri İle Karşılaştırılması, Yöneylem Araştırması Endüstri Mühendisliği XXIV. Ulusal Kongresi, Adana, Türkiye, 166– 168. Hakkila, P., 2004: Developing technology for large-scale production of forest chips. Wood Energy Technology Programme 1999–2003, Final Report, Technology Report 5/2004, Tekes. 44 p. Hazel, D., 2006: How will our forests be impacted by a woody biomass energy market? Presentation given at Energy from Wood: Exploring the Issues and Impacts for North Carolina, Raleigh, NC, March 13–14. Heinimo, J., 2011: Developing Markets of Energy Biomass – Local And Global Perspectives PhD Thesis, Lappeenrante University of Tecnology, Finland, 78 p. Jonsson, R., Egnell, H., Baudin A., 2011: Swidish Forest Sector Outlook Study, Geneva Timber and Forest Discussion Paper 58, Forestry and Timber Section, United Nations, Geneva, Switzerland, 100 p. Karataş, H., Gül, S., 2012: Country Update Turkey. TUBİTAK MAM, IEA Bioenergy, Task 33 Meeting, April, İstanbul, 19 p. Karayılmazlar, S., Saraçoğlu, N., Çabuk, Y., Kurt, R., 2011: Utilizations of biomass as an energy source in Turkey. Journal of Bartın Faculty of Forestry 13(19): 63–75. Croat. j. for. eng. 35(2014)2

M. Eker

Kaygusuz, K., Turker, M.F., 2002: Biomass energy potential in Turkey. Renewable Energy 26(4): 661–678. Kaygusuz, K., Keleş, S., 2008: Use of biomass as a transitional strategy to a sustainable and clean energy system. Energy Sources, 31(1): 86–97. Ladanai, S., Vinterbäck, J., 2009: Global potential of sustainable biomass for energy report 013, Swedish University of Agricultural Sciences, Department of Energy and Technology ISSN 1654-9406, Uppsala, Sweeden. Lundmark, R. 2006: Cost structure of and competition for forest-based biomass. Scandinavian Journal of Forest Research 21(3): 272–280. Mendell, B., Hamsley, A., Sydor, T., 2011: Wood biomass as a forest product, wood supply and market implications. Forisk Consulting, 19 p. MENR, 2012: General Energy Balance of Turkey in 2010. Energy Statistics, Ministry of Energy and Natural Resources http://www.enerji.gov.tr/ (Accessed April 2012). MFWA, 2012: Council Working Report on Forest Utilization. Working Group 10, Council of Forestry and Water 2013. Ministry of Forest and Water Affairs, Ankara, 43 p. NEP (Nordic Energy Perspectives), 2009: Wood markets and the situation of the Forest industry in the Nordic countries. NEP Research Group www:nordicenergyperspectives.org (Accessed December 2011). OAİB, 2011: Tomruk Raporu (Reports of Sawlog). Central Anatolian Exporter Unions. www.turkishwood.org (Accessed February 2012). Perlack, R.D., Wright, L.L., Turhollow, A.F., Graham, R.L., Stokes, B.J., Erbach, D.C., 2005: Biomass as feedstock for a bioenergy and bioproducts industry: the technical feasibility of a billion ton annual supply. DOE/GO-102995-2135. ORNL/ TM-2005/66, 78 p. Pincus, M.N., Moseley, C., 2009: Social issues of woody biomass utilization: A review of the literature. EWP Working Paper Number: 20, Ecosystem Workforce Program, University of Oregon, 15 p. Röser, D., Asikainen, A., Rasmussen, K.R., Stupak, I., 2008: Sustainable Use of Forest Biomass for Energy: A Synthesis with Focus on the Baltic And Nordic Countries. Managing For. Ecosyst. Vol. 12, 232 p. Sakarya, S., Canlı, Ş., 2011: Levha Sanayi Raporu (Report of Fibre Board Industry). Wood and Forestry Products Exporters’ Assocation of Central Anatolia, Ankara, 22 p. www.turkishwood.org (Accessed January 2012). Saraçoğlu, N, 2010a.: Global Climate Change, Bioenergy and Energy Forestry. Efil Publishing, ISBN:978-605-4334-40-7, Ankara, 298 p. Saraçoğlu, N., 2010b: The Biomass Potential of Turkey for Energy Production: Part II. Energy Sources, Part B: Economics, Planning and Policy 5(4): 384–389. Sedjo, R.A., 1997: The economics of forest-based biomass supply. Energy Policy 25(6): 559–566.

269

M. Eker

Trends in Woody Biomass Utilization in Turkish Forestry (255–270)

Smeets, E., Faaij A., 2007: Bioenergy potentials from forestry in 2050 – An assessment of the drivers that determine the potentials. Climatic Change 81(3–4): 353–390. Sürmen, Y., 2003: The necessity of biomass energy for the Turkish economy. Energy Sources 25(2): 83–92. Spinelli, R , 2011: Supply of Wood Biomass for Energy Purpose: Global Trends and Perspectives. Proceedings of the 22nd Annual Meeting of the Club of Bologna, November 13-14, Hannover, Germany, 16 p. Türkyılmaz, O., 2011: Türkiye’ nin Enerji Görünümü (Turkey’s Energy Perspetive). Union of Chamber of Turkish En-

gineers and Architecs. The Chamber of Mechanical Engineers, Branch Office of Adana, 104 p. USDA, 2007: Woody Biomass Utilization Desk Guide. Forest Management Forest and Rangelands National Forest System. United States Departmen of Agriculture, Washington, D.C., 91 p. Wunder, S., 2012: Impact of EU bioenergy policy on developing countries – Focus: woody biomass, Presentation of the Briefing Paper DEVE/7/08607, January 25, Brussels, 16 p. Yıldız, T., 2011: Türkiye Enerji Politikalarımız (The Energy Policy of Turkey). Ministry of Energy and Natural Resources. http://www.enerji.gov.tr/index.php (Accessed May 2012).

Author’s address:

Received: June 22, 2012 Accepted: July 05, 2014 Subject review

270

Assoc. Prof. Mehmet Eker, PhD. e-mail: mehmeteker@sdu.edu.tr Süleyman Demirel University Faculty of Forestry Department of Forest Engineering East Campus, Cunur, Isparta – 32260 TURKEY Croat. j. for. eng. 35(2014)2

ISSN 1845-5719

9 771845 571000