Environmental & Engineering Geoscience by Association of Environmental & Engineering Geologists (AEG)

Environmental & Engineering Geoscience MAY 2016

VOLUME XXII, NUMBER 2

THE JOINT PUBLICATION OF THE ASSOCIATION OF ENVIRONMENTAL AND ENGINEERING GEOLOGISTS AND THE GEOLOGICAL SOCIETY OF AMERICA SERVING PROFESSIONALS IN ENGINEERING GEOLOGY, ENVIRONMENTAL GEOLOGY, AND HYDROGEOLOGY

Environmental & Engineering Geoscience (ISSN 1078-7275) is published quarterly by the Association of Environmental & Engineering Geologists (AEG) and the Geological Society of America (GSA). Periodicals postage paid at AEG, 1100 Brandywine Blvd, Suite H, Zanesville, OH 43701-7303 and additional mailing offices. EDITORIAL OFFICE: Environmental & Engineering Geoscience journal, Department of Geology, Kent State University, Kent, OH 44242, U.S.A. phone: 330-672-2968, fax: 330-672-7949, ashakoor@kent.edu. CLAIMS: Claims for damaged or not received issues will be honored for 6 months from date of publication. AEG members should contact AEG, 1100 Brandywine Blvd, Suite H, Zanesville, OH 43701-7303. Phone: 844-331-7867. GSA members who are not members of AEG should contact the GSA Member Service center. All claims must be submitted in writing. POSTMASTER: Send address changes to AEG, 1100 Brandywine Blvd, Suite H, Zanesville, OH 43701-7303. Phone: 844331-7867. Include both old and new addresses, with ZIP code. Canada agreement number PM40063731. Return undeliverable Canadian addresses to Station A P.O. Box 54, Windsor, ON N9A 6J5 Email: returnsil@imexpb.com. DISCLAIMER NOTICE: Authors alone are responsible for views expressed in articles. Advertisers and their agencies are solely responsible for the content of all advertisements printed and also assume responsibility for any claims arising therefrom against the publisher. AEG and Environmental & Engineering Geoscience reserve the right to reject any advertising copy. SUBSCRIPTIONS: Member subscriptions: AEG members automatically receive digital access to the journal as part of their AEG membership dues. Members may order print subscriptions for $60 per year. GSA members who are not members of AEG may order for $60 per year on their annual GSA dues statement or by contacting GSA. Nonmember subscriptions are $295 and may be ordered from the subscription department of either organization. A postage differential of $10 may apply to nonmember subscribers outside the United States, Canada, and Pan America. Contact AEG at 844-331-7867; contact GSA Subscription Services, Geological Society of America, P.O. Box 9140, Boulder, CO 80301. Single copies are $75.00 each. Requests for single copies should be sent to AEG, 1100 Brandywine Blvd, Suite H, Zanesville, OH 43701-7303. © 2016 by the Association of Environmental and Engineering Geologists All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or by any information storage and retrieval system, without permission in writing from AEG.

EDITORIAL BOARD ROBERT H. SYDNOR JEROME V. DEGRAFF USDA Forest Service Consulant THOMAS J. BURBEY CHESTER F. WATTS (SKIP) Virginia Polytechnic Institute Radford University SYED E. HASAN University of Missouri, Kansas City ASSOCIATE EDITORS JOHN W. BELL PAUL M. SANTI Nevada Bureau of Mines and Colorado School of Mines Geology ROBERT L. SCHUSTER U.S. Geological Survey RICHARD E. JACKSON (Book Reviews Editor) ROY J. SHLEMON R. J. Shlemon Geofirma Engineering, Ltd. & Associates, Inc. JEFFREY R. KEATON AMEC Americas GREG M. STOCK National Park Service PAUL G. MARINOS National Technical University RESAT ULUSAY Hacettepe University, Turkey of Athens, Greece CHESTER F. “SKIP” WATTS JUNE E. MIRECKI U.S. Army Corps of Radford University Engineers TERRY R. WEST Purdue University PETER PEHME Waterloo Geophysics, Inc NICHOLAS PINTER Southern Illinois University SUBMISSION OF MANUSCRIPTS Environmental & Engineering Geoscience (E&EG), is a quarterly journal devoted to the publication of original papers that are of potential interest to hydrogeologists, environmental and engineering geologists, and geological engineers working in site selection, feasibility studies, investigations, design or construction of civil engineering projects or in waste management, groundwater, and related environmental fields. All papers are peer reviewed. The editors invite contributions concerning all aspects of environmental and engineering geology and related disciplines. Recent abstracts can be viewed under “Archive” at the web site, “http://eeg.geoscienceworld.org”. Articles that report on research, case histories and new methods, and book reviews are welcome. Discussion papers, which are critiques of printed articles and are technical in nature, may be published with replies from the original author(s). Discussion papers and replies should be concise. To submit a manuscript go to http://eeg.allentrack.net. If you have not used the system before, follow the link at the bottom of the page that says New users should register for an account. Choose your own login and password. Further instructions will be available upon logging into the system. Please carefully read the “Instructions for Authors”. Authors do not pay any charge for color figures that are essential to the manuscript. Manuscripts of fewer than 10 pages may be published as Technical Notes. For further information, you may contact Dr. Abdul Shakoor at the editorial office.

THIS PUBLICATION IS PRINTED ON ACID-FREE PAPER EDITORS ABDUL SHAKOOR Department of Geology Kent State University Kent, OH 44242 330-672-2968 ashakoor@kent.edu

BRIAN G. KATZ Florida Department of Environmental Protection 2600 Blair Stone Rd. Tallahassee, FL 32399 850-245-8233 eegeditorbkatz@gmail.com

Cover photo A view of one of the alleys in Rock City, Mountain Lake, Virginia, formed by extension along discontinuities during lateral spreading of the rock mass. Photo courtesy of Abdul Shakoor. See article on page 93.

Environmental & Engineering Geoscience Volume 22, Number 2, May 2016 Table of Contents

Using Discontinuity Mapping to Investigate the Origins of Rock City and Mountain Lake, Giles County, Virginia Nidal W. Atallah, Abdul Shakoor, and Chester F. Watts

113

Discovering and Characterizing Abandoned Waste Disposal Sites Using LIDAR and Aerial Photography Andrew de Wet

131

Identification of Wall Tension Fractures Caused by Earthquakes, Blasting, and Pile Driving Jeffrey A. Johnson and Alan “Bob” Mutchnick

141

Geologic and Geotechnical Factors Controlling Incipient Slope Instability at a Gravel Quarry, Livermore Basin, California Philip L. Johnson, Patrick O. Shires, and Timothy P. Sneddon

157

Factors Affecting Failure by Internal Erosion of Geosynthetic Clay Liners Used in Freshwater Reservoirs Hakki O.Ozhan and Erol Guler

171

Book Review Geomodels In Engineering Geology—An Introduction By Peter Fookes, Geoff Pettifer, and Tony Waltham Review by: Richard Jackson

Using Discontinuity Mapping to Investigate the Origins of Rock City and Mountain Lake, Giles County, Virginia NIDAL W. ATALLAH Department of Geology, Kent State University, Kent, OH 44242, nidalatallah@gmail.com

ABDUL SHAKOOR1 Department of Geology, Kent State University, Kent, OH 44242, ashakoor@kent.edu

CHESTER F. WATTS Department of Geology, Radford University, Radford, VA 24142, cwatts@radford.edu

Key Terms: Mountain Lake, Rock City, Landslide Dam, Colluvial Deposits Lateral Spread, Discontinuity Mapping

extensive analysis must be done to rule out other interpretations.

ABSTRACT

INTRODUCTION

Mountain Lake’s unusual location, near the summit of Salt Pond Mountain, VA, in the non-glaciated portion of the Appalachian Mountains, has prompted geologists to study its origin for decades. The northeastern end of the lake abuts an area of heterogeneous colluvial deposits that contain large rectangular blocks of hard Tuscarora Sandstone. This area is known as “Rock City” because of the resemblance of the gaps between the rock blocks to streets and alleys in a city. The purpose of this study was to investigate the origin of Rock City and whether the colluvial deposits within its boundaries are part of a landslide that is possibly responsible for the formation of Mountain Lake. Mapping of Rock City included taking global positioning satellite readings at the corners of rock blocks and along the boundaries of other outcrops, then using ArcMap software to generate maps. Using stereonet analysis, the mode of rock-block displacement was investigated by comparing the measured orientations of principal discontinuity sets forming the rock-block boundaries with discontinuity orientations of undisturbed outcrops. Discontinuity data analysis indicates that Rock City is most likely a landslide that dammed the valley of Pond Drain, forming the lake. The primary mode of slope movement involves lateral spreading associated with extension occurring along discontinuities. The Tuscarora Sand-stone blocks comprising Rock City were detached from a scarp face along a northwest-southeast–trending joint set and were displaced laterally toward the west. A seis-mic event may have triggered slope movement; however, more

Mountain Lake is located in Giles County, VA, about 18 km east of Pearisburg (Figure 1). It is situated approximately 1,200 m above mean sea level (AMSL) AMSL, between the crest and the northwest limb of a gently plunging anticline (Figure 2). The anticline is on the Narrows thrust sheet in the Valley and Ridge physiographic province. The axial plane of the anticline trends N60u E (Mills, 1989), and its axis plunges toward the northeast at 7.5u (Parker et al., 1975). Three geologic units underlie the lake: the Silurian Tuscarora Sandstone (Stu) at the northern end, the Ordovician Juniata Sandstone (Oj) under the middle portion, and the Ordovician ReedsvilleTrenton Formation (Ort) at the southern end (Figure 2 and Table 1). Mountain Lake is an elongated body of water oriented south to northwest. The lake, fed by precipitation, surface runoff, and a line of lake-bottom springs, covers a total surface area of about 1.9 6 105 m2 within a small watershed (1.3 km2) (Cawley et al., 2001). The headwaters of four major streams lie close to the lake: Sartain Branch to the east, Doe Creek to the southwest, Johns Creek to the south, and Pond Drain to the northwest. Mountain Lake has a history of unusual self-draining behavior. Recent episodes in 2008, 2011, and 2012 almost completely drained the lake. Draining of the lake revealed the presence of four sinkhole-like depressions in the lake sediment that has accumulated over colluvium, with piping holes at their bottoms and sides, near the northeastern and northwestern margins of the lake (Figure 3). Aside from water loss through evapotranspiration and surface runoff through Pond

Corresponding author, email: ashakoor@kent.edu

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 93–111

Atallah, Shakoor, and Watts

Figure 1. Location of the study area.

Drain (only when the lake is full), many studies (Marland, 1967; Parker et al., 1975; Cawley, 1999; Cawley et al., 2001; Jansons et al., 2004; Roningen, 2011; and Joyce, 2012) have suggested that lake-level fluctuations are caused by water seeping out of the lake basin through subterranean pathways at the northern end of the lake. Rock City is the area at the northern end of the lake (Figure 4a) featuring large rectangular rock blocks, measuring up to approximately 140 m2, and other colluvial deposits of rock fragments of varying sizes. The objective of this study was to investigate whether Rock City is part of a landslide that

contributed material for damming Pond Drain and to determine the mode of displacement of large blocks within Rock City. PREVIOUS HYPOTHESES ON THE ORIGIN OF MOUNTAIN LAKE Karst-Related Hypothesis Holden (1938) hypothesized that Mountain Lake formed as a result of dissolution of a calcareous layer in the upper Reedsville-Trenton Formation. Ferguson

Figure 2. Geologic map of the study area; modiďŹ ed from geologic map of Giles County, VA, by Schultz et al. (1986).

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 93â&#x20AC;&#x201C;111

Origins of Rock City and Mountain Lake Table 1. Simplified stratigraphic column in the study area. Formation Rose Hill (Srh) Tuscarora (Stu) Juniata (Oj) Reedsville-Trenton (Ort)

Age

Rock Type

Thickness (m) (Mills, 1990)

Silurian Silurian Ordovician Ordovician

Sandstone Sandstone/orthoquartzite Sandstone Shale (mostly) and limestone

45–60 15–45 60–110 425–490

et al. (1939) supported this hypothesis, referring to the origin as a “natural solution collapse basin.” The existence and depth of calcareous strata in the ReedsvilleTrenton Formation has not been confirmed. Holden (1938), Butts (1940), Marland (1967), and Roningen (2011) suggest its presence within the uppermost portion of the formation, while others, such as Eckroade (1962) and Parker et al. (1975), propose its presence at a considerable depth. Sharp (1933) rejected the possibility of a natural collapse basin origin, asserting that dissolution could not have reached the surface because the depth of the soluble limestone is approximately 305 m. He also argued that Mountain Lake did not resemble other limestone sinks in the area. Parker et al. (1975) used calculations of formation thicknesses to dismiss the collapse basin hypothesis. He argued that “a true sinkhole in the upper Martinsburg [Reedsville-Trenton] would have to extend upward through at least 15 m of Juniata to reach the bottom of Mountain Lake, and then further dissolve through 457 m of Martinsburg [Reedsville- Trenton] to make a sinkhole.” However, Williams (2003), in his Encyclopedia of Caves and Karst Science, provides an account of subjacent karst collapse sinkholes in caverns deeper than 1,000 m in Canada and Russia. Finally, Roningen (2011) identified the presence of “significant carbonate content” in the uppermost section of the Reedsville-Trenton Formation in Narrows, VA. Based on this observation, and considering the alkalinity of

Mountain Lake, she argued that karst dissolution should not be ruled out as a possible hypothesis for Mountain Lake’s formation. Landslide-Dam–Related Hypothesis The major studies promoting the landslide dam hypothesis are those of Rogers (1884), Hutchinson and Pickford (1932), Sharp (1933), Eckroade (1962), Marland (1967), and Parker et al. (1975). Most of these studies suggest that the headwaters of the northwesterly flowing Pond Drain cut through the resistant sandstone ridges, breaching the northwestern end of the anticline and carving out a narrow valley. The narrow valley was then dammed by the colluvial blocks of the Tuscarora Sandstone that make up Rock City. The mode of displacement by which the damming occurred has been attributed to different types of mass movement and is the primary focus of this study. Hutchinson and Pickford (1932) and Hutchinson (1957) suggested that the damming occurred as a result of the “caving in of overhanging ledges of hard rock that were undermined by the stream.” Sharp (1933) suggested that Tuscarora blocks crept downward from the ridge in the form of talus, but he mentioned the possibility of a rockslide as well. Eckroade (1962) agreed with Sharp’s explanation but added that frost heaving generated additional Tuscarora blocks that were displaced, probably by solifluction, further

Figure 3. Drastic water level drop in December 2012 showing the lake-bottom depressions (left) compared to lake level in April 2012 (right).

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 93–111

Atallah, Shakoor, and Watts

Figure 4. (a) Aerial view and bathymetric map of Mountain Lake, Rock City, and Pond Drain. North is at top of image. Yellow line on map marks location of the cross section shown in b; (b) southwest (left) to northeast (right) cross section of the north end of Mountain Lake at Pond Drain, obtained by seismic refraction (C. F. Watts, Radford University, 2013).

damming the stream valley. In this case, solifluction corresponds to the definition provided by Easterbrook (1999): the “downslope movement induced by alternate freezing and thawing of debris slopes” or “gelifluction.” Marland (1967) used 14C-dating of core bottoms to confirm that the colluvial material was produced by solifluction during climate with repeated freeze-thaw cycles about 9,180 ¡ 330 years before present (YBP) YBP. He also added that the leaky nature

of the damming material prevented the formation of a “permanent lake” until about 2,000 YBP. Parker et al. (1975) synthesized a number of modes of displacement suggested by earlier works in explaining the damming process and consequent lake formation. These authors agreed that the damming occurred via talus or slide rock (Sharp, 1933; Eckroade, 1962) through mass movement by solifluction (Eckroade, 1962; Marland, 1967) and vertical collapse of ledges

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 93–111

Origins of Rock City and Mountain Lake

due to undercutting (Rogers, 1884; Hutchinson and Pickford, 1932). However, they suggested that vertical collapse, promoted by undercutting of the resistant Tuscarora and Rose Hill sandstone beds by erosion of the less resistant Juniata Sandstone, was the primary mode of displacement. Investigations of colluvial deposits in the Mountain Lake area, not far from the study site, led Mills (1981, 1988, 1989, 1990) to suggest their lateral transport and the retreat of Tuscarora escarpments through the process of “topographic inversion of hollows and noses.” Recent seismic refraction studies, performed by Watts (2013), confirm the presence of a colluviumfilled valley at the northwestern end of the lake (C. F. Watts, oral communication, 2013). The study revealed a narrow gorge in the bedrock, at least 30 m deep and filled with colluvium, at the north end of the lake (Figure 4b).When the lake is full, the water overflows into Pond Drain (Freeman et al., 2012). This supports the various landslide hypotheses. Fracture-Lineation–Related Hypothesis Cawley (1999) conducted fracture trace analysis that documented a lineament trending from SE to NW. He interpreted the lineament as a “fracture and probable fault associated with the regional Appalachian fold and thrust tectonics.” Cawley (1999) studied this feature using direct current resistivity techniques and concluded that pronounced resistivity lows along the feature were indicative of a “water-filled fracture zone.” Cawley (1999) proposed that the regional fracture had a role in carving out the valley of Pond Drain at the northwestern end of the lake as well as in its damming and consequent lake formation. In a later publication, Cawley et al. (2001) described the damming process as “incremental settling and breakup of an overlying resistant rib of Clinch [Tuscarora Sandstone] bedrock in physical contact with the fault lineation.” Additionally, Cawley (1999) and Cawley et al. (2001) suggested that the basin of Mountain Lake was formed when fine sediments were eroded away by water seeping through the fracture. This also explains periodic drops in lake water levels. However, in a study by Roningen (2011), the presence of the fracture identified by Cawley (1999) could not be confirmed using electrical resistivity tomography, joint sampling, or lineament analysis. RESEARCH METHODS Investigations of Rock City included mapping the locations and orientations of the large rectangular rock blocks, the “alleys” and “streets” separating the

blocks along major discontinuities, and other colluvial deposits. The study also included an investigation of a complex of scarps consisting of irregular, discontinuous cliff outcrops of Tuscarora Sandstone just upslope of Rock City (Figure 5). Discontinuity data, including both bedding planes and joint sets in the rock blocks and outcrops upslope, were analyzed in order to evaluate the mode and extent of displacement exhibited by the rock blocks. Rock City Mapping A GIS map (Figure 6) was generated using global positioning satellite (GPS) readings taken at the corners of rock blocks, at the outcrops upslope of the rock blocks, and at the boundaries of boulder fields containing a substantial number of small to mediumsized colluvial boulders. The GPS unit used for mapping included a Trimble Pro XRT Backpack and a NOMAD Data Logger with Terrasync V6.x. Field Software. The data points collected were differentially corrected using GPS PathfinderH Office Software and entered into ESRI ArcMap 10 Office Software to generate the Rock City map. By connecting the coordinate points at the rock-block corners, the boundaries of large rock blocks were drawn on the map as idealized rectangular-shaped polygons (Figure 6), matching the orientations of two principal and nearly vertical joint sets representing the sides of each block (Figure 7). Tape measurements of block dimensions were taken to supplement the GPS readings and to correct for block shapes that were irregular as a result of weathering and/or disintegration of the original joint surfaces. The rock blocks were categorized into the following size-based groups: very large (with footprints .50 m2), large (20–50 m2), medium (2.5–20 m2), and small (,2.5 m2). In this study, all blocks with a base area of $2.5 m2 were considered “rock blocks,” and their dimensions were measured in detail. Rock blocks with footprint areas of ,2.5 m2 were considered “boulders” and were mapped only where concentrated into sizeable boulder fields. Discontinuity Measurements The orientations of discontinuities within the scarp outcrops upslope were measured using the Window Mapping Method (Wyllie and Mah, 2004). All joints encountered within the selected area were measured. These measurements were used to assess whether these outcrops represented the in situ bedrock and to establish a baseline for comparing the orientations of discontinuities within the detached and transported

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 93–111

Atallah, Shakoor, and Watts

Figure 5. Examples of scarp outcrops: (a, b) small separate scarp outcrops and (câ&#x20AC;&#x201C;f) continuous cliff-like walls of scarp of varying heights. The scale in (a) through (f) is approximately 1 cm 5 0.7 m.

blocks in Rock City. Similarly, the orientations of all joints within the rock blocks were measured, especially those clearly making up the sides of the blocks (Figure 7). Discontinuity measurements were made using both a Brunton compass and an Ipad (application GeoID 1.61). DIPS 6.0 software (Rocscience, 2012)

was used to generate stereonet plots of discontinuity orientations that were used to identify bedding and principal joint sets for both the scarp outcrops and for individual rock blocks. The dip and dip directions for both bedding planes and joint sets were compared between the rock blocks and the scarp outcrops using

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 93â&#x20AC;&#x201C;111

Origins of Rock City and Mountain Lake

Figure 6. Rock City boulder ﬁeld and large block locations.

DIPS and RockPack III (Watts et al., 2012) software programs, respectively, in order to determine the following: 1) The direction of tilt exhibited by each rock block; 2) The direction, clockwise or counterclockwise, and angle of lateral rotation experienced by rock blocks during displacement away from the scarp outcrops. The two principal orthogonal joint sets, identified in both the scarp outcrops and in each of the blocks, were used to determine rock-block rotation; and

3) The type of mass movement that resulted in the formation of Rock City and its connection to Mountain Lake’s formation.

ROCK CITY MAPPING RESULTS Rock City can be divided into two main components: (1) a belt of multiple scarps of Tuscarora Sandstone upslope, referred to as the scarp complex in the following discussion, and (2) the overall debris field containing the large rectangular rock blocks as well as other coarse colluvial deposits. Scarp Complex

Figure 7. An example of the two principal joint sets marking the boundaries of a rock block in Rock City.

The scarp complex bounds the debris field to the north and consists of partly exposed, cliff outcrops that are discontinuous and non-linear in nature but that have an overall trend of N40u–70u W (Figure 6). The scarps vary in height, width, elevation, and aspect. They range from small separate outcrops to sizable cliff walls (Figure 5). The scarps crop out at different elevations with variable setbacks along any particular stretch, in places giving the appearance of steps or benches. To help visualize the exposure of outcrops at several levels, Figure 8 shows the GPS readings taken at outcrops along the scarp complex during various trips. Bedding attitudes in the scarp complex are nearly horizontal (Figure 9). Bedding dips vary from 0u to

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 93–111

Atallah, Shakoor, and Watts

Figure 8. Locations of GPS readings along scarp outcrops where discontinuity measurements were taken.

24u, with an average dip of 7u and a standard deviation of 4.5u. Dip directions in the scarp complex average 331u but vary somewhat, with a circular standard deviation of 72u. At least four principal joint sets are present in the scarp complex (Figure 10). The near-vertical joint

sets strike: (1) N35uW, (2) N15uE, (3) N57uE, and (4) N83uE. Sets 1 and 3 are distinctly orthogonal. The N35uW set may be stress relief joints that developed when Pond Drain carved a sandstone-walled valley prior to the collapse of a portion of the valley wall.

Figure 9. Bedding plane attitudes for scarp outcrops plotted as poles. Plot generated by DIPS software (Rocscience, 2012).

100

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 93â&#x20AC;&#x201C;111

Origins of Rock City and Mountain Lake

Figure 10. Density concentrations of poles for joint set orientations in the scarp outcrops, with the corresponding great circles of the principal joint sets labeled in blue. Plot generated using DIPS software (Rocscience, 2012).

Variations in bedding-plane and joint orientations within the scarp complex most likely exist because some blocks that have started to detach from in situ strata within Salt Pond Mountain have tilted or rotated by a small amount. The irregular shape of the scarp line (Figure 6) implies that blocks detached along pre-existing fractures and joints. The scarp is generally a continuous line of outcrops that extends the full width of the area of the large rock blocks and boulder fields comprising Rock City, so it can be considered the head scarp for the entire debris field. Also, notwithstanding some irregularities, the standard deviation of 4.5u for bedding dip in the scarp outcrops means that the average dip of 7u can be used for comparison with bedding-plane orientations within the detached blocks in Rock City. Bartholomew et al. (2000), authors of the geologic map for the Radford quadrangle, indicated local dip to the NE near the northeastern corner of Mountain Lake. This study found dip directions primarily to the NE and to the NW and SW. The NW directions represent the NW limb of the anticline. The SW dips are contrary to the NE plunge of the anticline, and they possibly indicate cambering out of the local hillside, away from the scarp. The NE dips may be the result of some backrotation from slope movement.

Debris Field The Rock City portion of the debris field extends about 170 m E-W along the lake shore and about 100 m N-S from the lake shore to the scarp complex (Figure 6). Its main components are the rock blocks, “streets” and “alleys” separating the blocks, and boulder fields or “jumbles” containing boulders of all sizes. Mapping of Rock City revealed the presence of at least 57 rock blocks, of which 10 are very large, 21 are large, and 26 are medium-sized (Figure 11). Most rock blocks are rectangular. The passageways between rock blocks, referred to as “streets” and “alleys,” vary in width from a few to several meters (Figure 12). They vary from rectilinear to nearly random. Rocks with footprint areas of less than 2.5 m2, designated as small boulders, are scattered over the entire area of Rock City, but the map documents only concentrated boulder fields (Figure 13). These boulder fields originated from scarp outcrops, collapse of overhangs between adjacent blocks, and breakdown of larger rock blocks (Figure 13). The bedding planes of rock blocks dip in all directions; however, the majority of the dip directions exceed 180u, indicating that dips toward the NW, W, and SW are more common, with a slight predominance to the NW (Table 2). Exceptions to the overall

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 93–111

101

Atallah, Shakoor, and Watts

Figure 11. Examples of rock-block sizes: rock block #16 (top) is of medium size, and rock block #28 (bottom) is of very large size.

Figure 12. Rock City streets and alleys of varying widths: (top) a small alley (3 m wide) and (bottom) a large alley “Main Street” (14 m wide).

westward tilting include two main clusters of rock blocks present at the southeastern edge of Rock City and another near the scarp complex, which dip mostly south and north, respectively. Measurements of joint orientations in rock blocks indicated that each of the blocks has two principal joint sets and in some cases other minor/secondary sets (Table 2). The two principal joint sets were, in the majority of cases, nearly orthogonal, similar to the orthogonal sets 1 and 3 identified in the scarp complex.

planes (Figure 14) indicate primarily outward tilting, away from the head scarp (Table 2). Tilt appears to increase with distance from the scarp and to be more randomized with smaller blocks. However, the cluster of blocks at the southeastern edge of Rock City (Figure 6) experienced greater tilting toward the SE and SW, most likely because of steeper slopes in that area (Figure 15). The small number of northward-dipping blocks, located very close to the scarp face, suggests that after initial detachment, blocks back-rotated. Blocks farther away from the scarp moved by lateral spreading and by outward tilting from the scarp.

DISCONTINUITY MAPPING AND ANALYSIS RESULTS Analysis of Bedding Orientation Data Bedding planes within the rock blocks in Rock City are more scattered than those in the scarp outcrops (Figure 14), suggesting that block orientations became randomized during transport. The rock-block bedding planes also have steeper dips, ranging from 5u to 82u, compared to the scarp area in which bedding dips range from 0u to 24u and average 7u. The predominantly westward orientations of rock-block bedding

102

Analysis of Joint Orientation Data The orthogonal joint sets 1 and 3 in the scarp face were used as references to compare to joint orientations in rock blocks. The purpose was to determine the degree of lateral rotation experienced by the blocks during displacement away from the scarp. As lateral rotation exceeds 40u, distinguishing between the joint faces and between clockwise and counterclockwise rotation becomes very difficult. The same problem may occur for rock blocks that do not form a perfect

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 93–111

Origins of Rock City and Mountain Lake

Figure 13. Examples of boulder ﬁelds (a, b) and the processes supplying material for boulder ﬁelds, such as collapsing roofs between rock blocks (c, d), collapsing overhangs (e), and sliding rock blocks from the scarp (f). Approximate scale: 1 cm 5 0.7 m.

90u orthogonal system of joints. For example, rock block #32 shows that it might have experienced either a clockwise or counterclockwise rotation of 3u (Table 3). This study focused in part on major rock blocks relatively close to the head scarp, namely those in Main Street, because of their limited rotation and their low degree of tilting. Orientations of joints in

the rock blocks clustered around Main Street (Figure 16) demonstrate that these blocks have rotated from 10u to 30u in both clockwise and counterclockwise directions. Additionally, the rock blocks that are adjacent to one another seem to exhibit similar rotational movement. For example, rock blocks 24A, 24B, and 25 (Figure 16) rotated counterclockwise about 15u–

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 93–111

103

Atallah, Shakoor, and Watts Table 2. Rock-block sizes and discontinuity orientations. Discontinuity Orientations* Joint Sets (JS) Rock Bock (RB) No. RB#1 RB#2A RB#2B RB#3 RB#4A RB#4B RB#5A RB#5B RB#6 RB#7 RB#8 RB#9 RB#10 RB#11 RB#12A RB#12B RB#12C RB#12D RB#12E RB#12F RB#12G RB#13 RB#14 RB#15 RB#16 RB#17 RB#18 RB#19 RB#20 RB#21A RB#21B RB#22 RB#23 RB#24A RB#24B RB#25 RB#26 RB#27 RB#28 RB#29A RB#29B RB#30A RB#30B RB#31 RB#32 RB#33 RB#34A RB#34B RB#35A RB#35B RB#35C RB#36 RB#37 RB#38 RB#39 RB#40 RB#41

Block Size (m2) 21.2 35.0 28.1 9.8 34 9.3 18.0 9.0 19.0 26.2 10.5 14.2 31.8 31.9 14.4 9.7 6.3 13.6 8.8 5.5 10.6 12.6 48.1 19.3 7.8 25.3 38.5 36.6 34.6 52.4 2.7 54.6 9.6 66 47.3 93.2 42.9 31.9 102.4 64.9 57.4 104.0 19.3 11.2 142.5 57.1 35.1 45.3 6.6 4.3 6.5 18.5 18.6 33.1 26.7 41.3 26.5

*Strike directions (degrees); dips **Dips/dip directions (degrees).

104

Size Category Large Large Large Medium Large Medium Medium Medium Medium Large Medium Medium Large Large Medium Medium Medium Medium Medium Medium Medium Medium Large Medium Medium Large Large Large Large Very large Medium Very large Medium Very large Very large Very large Large Large Very large Very large Very large Very large Medium Medium Very large Very large Large Large Medium Medium Medium Medium Medium Large Large Large Large

Bedding Plane**

JS#1

JS#2

JS#3

JS#4

13/213 21/299 13/322 17/339 11/221 21/259 13/300 10/006 06/314 06/267 29/354 28/276 9/344 11/310 82/201 23/21 30/196 16/255 36/314 75/197 50/200 15/280 32/192 18/232 16/260 63/238 26/203 20/169 11/150 6/158 50/084 71/318 32/358 11/315 10/013 08/318 8/096 14/325 07/305 09/317 09/025 12/305 20/311 05/293 05/167 24/324 25/318 27/315 13/281 12/277 12/285 19/068 07/342 09/359 18/026 20/012 33/310

N70E N40E N40E N70E N45E N60E N48E N35E N45E N30E N85E N23E N77E N55E N25E N25E N5E N85E N18E N12E N76E N35E N40E N10E N84E N35E N70E N73E N30E N54E N5E N46E N35E N40E N40E N35E N70E N70E N65E N86E N78E N20W N5E N82W N60E N60E N42E N48E N45E N70E N65E N70E N80E N70E N50E N20E N45E

N20W N50W N53W N13W N52W N20W N42W N55W N45W N60W N5E N67W N14W N35W N65W N65W N80W N5W N72W N85E N10W N65W N50W N80W N6W N80W N3E N17W N60W N36W N85E N44W N40W N50W N50W N55W N20W N15W N25W N5W N15W N85W N70W N8W N38W N30W N48W N42W N42W N20W N25W N15W N30W N25W N30W N70E N35W

N15E N4W North-South N85E N70E N5E — N70W N12W — — — — — — — N42E — — N15W — — N75E — — — N35W N82W

N60W — N34E — N6W — —

N15E N25W — — N63W N70W N34W — — — N30E N8W — — — — — — — N7W — — — N55E N55W N10W — —

are near vertical.

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 93–111

— — — — — — — — — — — — — — — — — — N40E

N20E — — — — N66W — — — — — — — — — — — — — — — — — — — — —

Origins of Rock City and Mountain Lake

Figure 14. Comparison between poles of bedding plane attitudes for the scarp outcrops (red) and the rock blocks mapped in Rock City (black). Plot generated using DIPS software (Rocscience, 2012).

20u, while rock blocks 26 and 27 rotated clockwise about 13u (Table 3). Rock blocks 29A and 29B, at the western end of Main Street, show a clockwise rotation of 29u and 20u, respectively (Table 3). Joint set orientations for a number of selected rock blocks from other locations in Rock City were also compared to the joint orientations in the scarp outcrops (Figure 17 and Table 3). The direction and degree of rotation for some of these rock blocks can be described as follows: rock blocks that exhibit a counterclockwise rotation are #7 (25u), #9 (32u), #14 (15u), #20 (25u), #21A (1u), #22 (9u), #34A (13u), and #34B (7u), whereas those that exhibit a clockwise rotation are #10 (20u), #33 (3u), and #19 (16u) (Table 3). The comparison does not support a

relationship between the locations of the rock blocks and their degrees of rotation. However, rock blocks immediately adjacent to Main Street, both to the east and west, exhibit minimal rotation of just 1u to 15u. The rock blocks in the cluster at the southeastern edge of Rock City had high rotation angles, ranging from 13u to 45u. However, the usefulness of comparing discontinuity orientations in the scarp to those in the rock blocks is limited. ORIGINS OF ROCK CITY AND MOUNTAIN LAKE Head scarp and rock-block discontinuity orientation data support the interpretation that Rock City

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 93â&#x20AC;&#x201C;111

105

Atallah, Shakoor, and Watts

Figure 15. An east-facing photo of the lake and the southeastern edge of Rock City showing a block from the cluster of blocks at the southeastern edge of Rock City experiencing greater tilting toward the SE and SW caused by steeper slopes into Mountain Lake.

comprises the remnants of a landslide that dammed Pond Drain, forming Mountain Lake. This interpretation agrees with several previous studies (Rogers, 1884; Hutchinson and Pickford, 1932; Sharp, 1933; Eckroade, 1962; Marland, 1967; and Parker et al., 1975). In addition, seismic refraction data indicate the presence of a colluvium-filled, buried gorge near the northern end of the lake (Figure 4b). The predominance of blocks tilted outward from the scarp and down the hillside, with principal lateral rotation around an axis perpendicular to slope, in both clockwise and counterclockwise directions, also supports the landslide hypothesis. The presence of a scarp at the northern edge of Rock City suggests that displacement of the Tuscarora rock blocks occurred from slope movement rather than from karst-related subsidence. The absence of a scarp feature along the entire perimeter of Mountain Lake suggests that it is unlikely for such a collapse to have occurred. However, without further subsurface investigations, the possibility of karst collapse within the Reedsville-Trenton Formation cannot be discarded. Thus, a catastrophic slope failure appears to be the most likely mechanism by which the rock blocks of Rock City were displaced. The blocks of Tuscarora Sandstone, which make up Rock City, broke loose from the scarp face, generally along a northwest-southeast–trending joint set, one of the two orthogonal joint sets. After the initial detachment, the rock mass broke into rectangular blocks along the orthogonal joint faces. Although many Tuscarora boulders and blocks may have

106

detached from the scarp as rock falls and collapsing overhangs, this was not the primary mode of origin of Rock City. Because adjacent blocks can generally be fitted back together across the “alleys,” especially the bigger blocks near the scarp complex, most likely the rock mass moved to the west as either a translational slide or a lateral spread (Cruden and Varnes, 1996). Kinematically, translational sliding seems unlikely for the following reasons: (1) bedding dips within and below the Tuscarora Sandstone are relatively gentle (7u); (2) beds dip into the hillside; (3) the Tuscarora Sandstone is underlain by a relatively resistant sandstone of the Juniata Formation; (4) the blocks are separated in Rock City, negating the possibility of sliding as a single sheet; and (5) the blocks are not concentrated in the toe area. Therefore, lateral spreading (Cruden and Varnes, 1996) may have been the primary mode of movement. Lateral spreads are common on gentle slopes and typically involve extensional movement similar to the formation of the “streets” and “alleys” in Rock City (Figure 12). Variations in the orientations of alleys and streets document little to considerable independent rotation of rock blocks. Soeters and Van Westen (1996) describe the morphology of lateral spreads as “irregular arrangement of large blocks tilting in various directions; block size decreases with distance and morphology becomes chaotic; large cracks and linear depressions separating blocks ….” This description is quite similar to the morphological features observed in Rock City. Most lateral spreads involve liquefaction or plastic flow of a subjacent layer (Cruden and Varnes, 1996).

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 93–111

Origins of Rock City and Mountain Lake Table 3. Direction and angle of lateral rotation for rock blocks compared to the outcrop scarp complex. Rock-Block Joint Sets and Angle of Rotation with Respect to Scarp Orthogonal Joints (u)

Rock-Block Rotation Summary (u)

Rock Block JS#1 (NE Strike)

Scarp Outcrop JS#1 (N57E)

Rotation of RB Relative to JS#1*

Rock Block JS#2 (NW Strike)

Scarp Outcrop JS#3 (N35W)

Rotation of RB Relative to JS#2*

Rotation Relative to Smaller Angle*

Average Rotation (JS1 and JS2)*

RB#1 RB#10 RB#18 RB#19 RB#26 RB#27 RB#28 RB#29A RB#29B RB#30A RB#33 RB#36 RB#37 RB#38 RB#39 RB#2A RB#2B RB#4A RB#5A RB#6 RB#7 RB#9 RB#12A RB#14 RB#15 RB#17 RB#20 RB#21A RB#22 RB#24A RB#24B RB#25 RB#30B RB#34A RB#34B RB#40 RB#11 RB#32

N70E N77E N70E N73E N70E N70E N65E N86E N78E N85E N60E N70E N80E N70E N50E N40E N40E N45E N48E N45E N30E N23E N25E N40E N10E N35E N30E N54E N46E N40E N40E N35E N5E N42E N48E N20E N55E N60E

N57E N57E N57E N57E N57E N57E N57E N57E N57E N57E N57E N57E N57E N57E N57E N57E N57E N57E N57E N57E N57E N57E N57E N57E N57E N57E N57E N57E N57E N57E N57E N57E N57E N57E N57E N57E N57E N57E

13 20 13 16 13 13 8 29 21 38 3 13 23 13 −7 −17 −17 −12 −9 −12 −27 −34 −32 −17 −47 −22 −27 −3 −11 −17 −17 −22 −52 −15 −9 −37 −2 3

N20W N14W N3E N17W N20W N15W N25W N5W N15W N20W N30W N15W N30W N25W N30W N50W N53W N52W N42W N45W N60W N67W N65W N50W N80W N80W N60W N36W N44W N50W N50W N55W N70W N48W N42W N70W N35W N38W

N35W N35W N35W N35W N35W N35W N35W N35W N35W N35W N35W N35W N35W N35W N35W N35W N35W N35W N35W N35W N35W N35W N35W N35W N35W N35W N35W N35W N35W N35W N35W N35W N35W N35W N35W N35W N35W N35W

15 21 38 18 15 20 10 30 20 15 5 20 5 10 5 −15 −18 −17 −7 −10 −25 −32 −30 −15 −45 −45 −25 −1 −9 −15 −15 −20 −35 −13 −7 −35 0 −3

13 20 13 16 13 13 8 29 20 15 3 13 5 10 5 −15 −17 −12 −7 −10 −25 −32 −30 −15 −45 −22 −25 −1 −9 −15 −15 −20 −35 −13 −7 −35 0 −3

14 20.5 25.5 17 14 16.5 9 29.5 20.5 26.5 4 16.5 14 11.5 −1 −16 −17.5 −14.5 −8 −11 −26 −33 −31 −16 −46 −33.5 −26 −2 −10 −16 −16 −21 −43.5 −14 −8 −36 −1 0

RB#41

N45E

N57E

−12

N35W

Rock Block No.

*Positive

−6

Direction of Rotation Clockwise Clockwise Clockwise Clockwise Clockwise Clockwise Clockwise Clockwise Clockwise Clockwise Clockwise Clockwise Clockwise Clockwise Clockwise Counterclockwise Counterclockwise Counterclockwise Counterclockwise Counterclockwise Counterclockwise Counterclockwise Counterclockwise Counterclockwise Counterclockwise Counterclockwise Counterclockwise Counterclockwise Counterclockwise Counterclockwise Counterclockwise Counterclockwise Counterclockwise Counterclockwise Counterclockwise Counterclockwise No rotation shown Either clockwise or counterclockwise No rotation shown

values indicate a clockwise rotation, while negative values indicate a counterclockwise rotation.

At Rock City, Juniata Sandstone underlies Tuscarora Sandstone, and shale and limestone of the ReedsvilleTrenton Formation underlie the Juniata Sandstone (Table 1). The Reedsville-Trenton Formation is at considerable depth and is unlikely to have served as a basal zone of plastic flow. Therefore, the Tuscarora sandstone blocks most probably moved over Juniata Sandstone. Another possibility is that the blocks moved over pre-existing colluvial debris covering the slope. The Mountain Lake area has significant

thicknesses of colluvial deposits containing materials of all sizes (Mills, 1981, 1988, 1990). We have no subsurface data below Rock City to document the nature of the failure zone. However, Varnes (1978) described spreads of rock that experienced lateral extension, similar to Rock City, without a well-defined basal shear surface or a zone of plastic flow. Such lateral spreads usually occur near ridge crests (Varnes, 1978). It is not clear what triggered the slope movement that resulted in the formation of Rock City.

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 93–111

107

Atallah, Shakoor, and Watts

Figure 16. Comparisons between the principal joint sets within the rock blocks in Main Street (black) and the principal joint sets in the head scarp (red). Plots generated using RockPack III software (Watts et al., 2012).

Considering that weaker formations, such as the Juniata Sandstone, lie directly beneath the Tuscarora Sandstone, that the Reedsville-Trenton Shale [Martinsburg] lies at a significant depth below the Tuscarora Sandstone, and that the bedding along which a translational movement could occur dips very gently, a seismic event probably triggered the movement. According to Andrus and Youd (1987), “spreads are the most common ground failure during earthquakes.” Lateral spreads require considerably strong ground shaking (Keefer, 1984). Giles County, VA, has a

108

documented history of seismicity, including a massive earthquake of 1897. This event was the second-largest recorded earthquake in the southeastern United States (Bollinger and Wheeler, 1988) and is now the third largest, following the August 23, 2011, earthquake in Mineral, VA. The U.S. Geological Survey (USGS) estimates that the 1897 earthquake had a Richter magnitude of 5.7 and that the 2011 earthquake had a Richter magnitude of 5.8 (USGS, 2011). Mountain Lake is located approximately 17 km from Pearisburg, VA, the presumed epicenter of the 1897 earthquake.

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 93–111

Origins of Rock City and Mountain Lake

Figure 17. Comparison between the principal joint sets of selected rock blocks from all portions of Rock City (black) and the principal joint sets in the head scarp (red). Plots generated using RockPack III software (Watts et al., 2012).

However, the Giles County seismic zone, as defined by Bollinger and Wheeler (1988), has generated earthquakes even closer to Mountain Lake in historic time. Based on data from 40 historical world-wide earthquakes, supplemented with intensity data from several hundred U.S. earthquakes, Keefer (1984) estimated that a magnitude 5.5 earthquake can trigger lateral spreads at distances of 12–15 km. As the lake clearly formed long before earthquake records were kept, it is impossible to judge what magnitudes and distances to earthquakes may have existed in the Giles

County Seismic Zone, including Mountain Lake, at the time. Previous studies (Parker et al., 1975; Cawley, 1999; and Cawley et al., 2001) have suggested the possibility of earthquakes playing a role in Mountain Lake’s formation. However, this role was within the context of lake-level fluctuations. Earthquakes may have adjusted boulders within the landslide dam causing changes in the amount of seepage leaving the lake. Previous studies have suggested rock-block displacement from freeze-thaw cycles (Eckroade, 1962; Marland, 1967; and Parker et al., 1975) and runoff from

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 93–111

109

Atallah, Shakoor, and Watts

storm events (Mills, 1981). Other types of mass movement mechanisms may have been creep (Sharp, 1933), solifluction (Eckroade, 1962; Marland, 1967; and Parker et al., 1975), and vertical collapse due to undercutting (Rogers, 1884; Hutchinson and Pickford, 1932; and Parker et al., 1975). However, none of these mechanisms explain the extensional features observed in Rock City. Mills (1981, 1988, 1989, 1990) favored lateral displacement of colluvial deposits in the Mountain Lake area; however, he did not suggest lateral spreading. LIMITATIONS AND FUTURE RESEARCH The research presented herein has the following limitations: 1. No subsurface data are available to document the nature of material that may have served as the basal zone of shear or plastic flow for a lateral spread to occur. Our hypothesis that Rock City comprises a lateral spread is based on the extensional features observed within the Rock City area and based on comparisons of discontinuity orientations within rock blocks to discontinuity orientations within head scarp. Future research should look into the possibility of drilling within the Rock City area to confirm the nature of the material below the displaced mass. 2. Although most lateral spreads are caused by earthquakes, and although Giles County has a documented history of seismic activity, our hypothesis of a seismic event being the triggering mechanism for lateral spreading, resulting in the formation of Rock City, requires additional documentation. Future research should apply landslide shaking models to document whether a seismic event can trigger lateral spreading in the stratigraphic and structural settings of the Rock City area and to determine the magnitude of such an event. CONCLUSIONS The conclusions of this study can be summarized as follows: 1. Rock City may be the remnant of a catastrophic ancient slope movement. The large rectangular blocks of Tuscarora Sandstone and other colluvial deposits comprising Rock City dammed the valley of Pond Drain, forming Mountain Lake. 2. The primary mode of slope movement in Rock City is lateral extension from lateral spreading. The landslide material detached from a bedrock scarp along northwest-southeast–trending joints

110

as well as other pre-existing fractures and joints, with individual rock blocks displacing laterally toward the west. During detachment, the rock blocks separated from each other by a few to many meters, forming streets and alleys. During movement, the blocks rotated laterally in both clockwise and counterclockwise directions. 3. A seismic event may have triggered slope movement, resulting in the formation of Rock City. The geomorphic setting resulting from Pond Drain carving a valley through the Tuscarora Sandstone ridges appears to have provided a gap in which collapsing rock could settle. ACKNOWLEDGMENTS The authors would like to thank the three anonymous reviewers whose constructive comments greatly helped improve the quality of this article. REFERENCES ANDRUS, R. D. AND YOUD, T. L., 1987, Subsurface Investigation of a Liquefaction Induced Lateral Spread, Thousand Springs Valley, Idaho: Miscellaneous Paper GL-87-8, Waterways Experiment Station, U.S. Army Corps of Engineers, Vicksburg, MS, 106 p. BOLLINGER, G. A. AND WHEELER, R. L., 1988, The Giles County, Virginia, Seismic Zone- Seismological Results and Geological Interpretations: U.S. Geological Survey Professional Paper 1355, 85 p. BUTTS, C., 1940, Geology of the Appalachian Valley in Virginia, Part 1: Virginia Geological Survey Bulletin, No. 52, 568 p. CAWLEY, J. C., 1999, A Re-Evaluation of Mountain Lake, Giles County, Virginia: Lake Origins, History and Environmental Systems: PhD Dissertation, Virginia Polytechnic Institute and State University, 118 p. CAWLEY, J. C.; PARKER, B. C.; AND PERREN, L. J., 2001, New observations on the geomorphology and origins of Mountain Lake, Virginia: Earth Surface Processes Landforms, Vol. 26, pp. 429–440. CRUDEN, D. M. AND VARNES, D. J., 1996, Landslide types and processes. In Turner, A. K. and Schuster, R. L. (Editors), Landslides—Investigation and Mitigation: Special Report 247, Transportation Research Board, National Research Council: National Academy Press, Washington, DC, pp. 37–75. EASTERBROOK, D. J., 1999, Surface Processes and Landforms, 2nd ed.: Prentice-Hall, Upper Saddle River, NJ. 546 p. ECKROADE, W. M., 1962, Geology of the Butt Mountain Area, Giles County, Virginia: Master’s Thesis, Virginia Polytechnic Institute, 64 p. FERGUSON, F. F.; STIREWALT, M. A.; BROWN, T. D.; AND HAYES, W. J., JR., 1939, Studies on the turbellarian fauna of the Mountain Lake Biological Station. I. Ecology and Distribution: Journal Elisha Mitchell Scientific Society, Vol. 55, pp. 274–288. FREEMAN, J.; CROOK, E. C.; AND WATTS, C. F., 2012, Seismic refraction survey of landslide colluvium and lacustrine sediments at Mountain Lake, Virginia (abstract): Geological Society America Abstracts Programs, Vol. 44, No. 7, p. 390.

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 93–111

Origins of Rock City and Mountain Lake HOLDEN, R. J., 1938, Geology of Mountain Lake, Virginia: Virginia Journal Science, 73 p. HUTCHINSON, G. E., 1957, A Treatise on Limnology, Vol. 1, Geography, Physics and Chemistry: John Wiley and Sons, New York. 1015 p. HUTCHINSON, G. E. AND PICKFORD, G. E., 1932, Limnological observations of Mountain Lake, Virginia: International Revenue GesamtenHydrobiologie Hydrographie, Vol. 27, pp. 252–264. JANSONS, M.; PARKER, B. C.; AND JACOB, E. W., 2004, Subterranean loss and gain of water in Mountain Lake, Virginia: A hydrologic model: Virginia Journal Science, Vol. 55, No. 3, pp. 107–113. JOYCE, W. L., 2012, Examining Pathways for Water Loss from Mountain Lake, Giles County, V.A.: Master’s Thesis, Virginia Polytechnic Institute, 44 p. KEEFER, D. K., 1984, Landslides caused by earthquakes: Geological Society America Bulletin, Vol. 95, No. 4, pp. 406–431. MARLAND, F. C., 1967, The History of Mountain Lake, Giles County, Virginia: An Interpretation Based on Paleolimnology: Ph.D. Dissertation, Virginia Polytechnic Institute and State University, 129 p. MILLS, H. H., 1981, Boulder deposits and the retreat of mountain slopes, or, “Gully Gravure” Revisited: Journal Geology, Vol. 89, pp. 649–660. MILLS, H. H., 1988, Surficial Geology and Geomorphology of the Mountain Lake Area, Giles County, Virginia, Including Sedimentological Studies of Colluvium and Boulder Streams: U.S. Geological Survey Professional Paper 1469, 57 p. MILLS, H. H., 1989, Hollow form as a function of boulder size in the Valley and Ridge province, southwestern Virginia: Geology, Vol. 17, pp. 595–598. MILLS, H. H., 1990, Thickness and character of regolith on mountain slopes in the vicinity of Mountain Lake, Virginia, as indicated by seismic refraction, and implications for hillslope evolution: Geomorphology, Vol. 3, pp. 143–157. PARKER, B. C.; WOLFE, H. E.; AND HOWARD, R. V., 1975, On the origin and history of Mountain Lake, Virginia: Southeastern Geology, Vol. 16, No. 4, pp. 213–226. ROCSCIENCE, 2012, Graphical and Statistical Analysis of Orientation Data: Electronic document, available at http://www. rocscience.com/products/1/Dips

ROGERS, W. B., 1884, Report on geological reconnaissance of the state of Virginia, made under the appointment of the Board of Public Works, 1835. Reprinted in Rogers, W. B., A Reprint of Annual Reports and Other Papers on the Geology of the Virginias: D. Appleton Company, New York, pp. 109–110. RONINGEN, J. M., 2011, Hydrogeologic Controls on Lake Level at Mountain Lake, Virginia: Master’s Thesis, Virginia Polytechnic Institute, 92 p. SCHULTZ, A. P.; STANLEY, C. B.; GATHRIGHT, T. M., II; RADER, E. K.; BARTHOLOMEW, M. J.; LEWIS, S. E.; AND EVANS, N. H., 1986, Geologic Map of Giles County, Virginia: Virginia Division of Mineral Resources, Publication 69. SHARP, H. S., 1933, The origin of Mountain Lake, Virginia: Journal Geology, Vol. 41, pp. 636–641. SOETERS, R. AND VAN WESTEN, C. J., 1996, Slope instability recognition, analysis, and zonation. In Turner, A. K. and Schuster, R. L. (Editors), Landslides—Investigation and Mitigation: Special Report 247, Transportation Research Board, National Research Council, National Academy Press, Washington, DC, pp. 129–177. U.S. Geological Survey (USGS), 2011, Electronic document, available at http://earthquake.usgs.gov/earthquakes/events/2011 virginia/overview.php] VARNES, D. J., 1978, Slope movement—Types and processes. In Schuster, R. L. and Krizek, R. J. (Editors), Landslides— Analysis and Control: Special Report 176, Transportation Research Board, National Research Council, Washington, DC, pp. 12–33. WATTS, C. F., 2013, Mountain Lake local geology and bathymetry overlays: KML files, Radford University https://sites.google. com/a/vt.edu/radford-geology—mountain-lake-files/googleearth-files WATTS, C. F.; GILLIAM, D.; HROVATIC, M.; AND HONG, H., 2012, ROCKPACK III for Windows: Rock Slope Stability Computerized Analysis Package, Part One—Stereonet Analysis: RockWare, Inc., Earth Science and GIS software, Golden, CO 80401. WILLIAMS, P., 2003, Dolines. In Gunn, J. (Editor), Encyclopedia of Caves and Karst Science: Fitzroy Dearborn, London, U.K., pp. 304–310. WYLLIE, D. C. AND MAH, C. W., 2004, Rock Slope Engineering: Civil and Mining, 4th ed.: Spon Press/Taylor and Francis Group, London, U.K. 432 p.

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 93–111

111

Discovering and Characterizing Abandoned Waste Disposal Sites Using LIDAR and Aerial Photography ANDREW DE WET1 Department of Earth & Environment, Franklin and Marshall College, Lancaster, PA 17603

Key Terms: LIDAR, Historical Aerial Photographs, Remote Sensing, Waste Disposal Sites, Landfills, GIS, Site Investigations

ABSTRACT Currently there are 1,908 active landfills across the United States; however, in the 1970s, prior to the modern era of sanitary landfills, there may have been as many as 100,000 landfills and dumps. Most of these landfills and dumps were unregulated and were aban‐ doned, and details about their locations and char‐ acteristics are poorly documented. The Abandoned Landfill Inventory produced by the Pennsylvania Department of Environmental Protection lists 2,620 waste facilities in the state, of which 1,309 are described as “landfills” or “abandoned landfills.” There are 157 reported facilities in Lancaster County, PA. Seventeen are “landfills” or “abandoned landfills”; however, it is clear that most of the landfills and dumps that existed in Lancaster are not listed. A 1971 report documented approximately 23 land disposal sites, including eight landfills (one of which was a licensed sanitary landfill), 15 dumps, approximately 44 identified informal dumps, and perhaps 200 to 300 additional unidentified informal dumps. This study uses LIDAR and historical aerial photography integrated into a GIS database to investigate three dump sites in Lancaster County, PA. The sites had varying degrees of available information about their location and extent. The techniques discussed here can be used to investigate other ‘known’ and possible abandoned sites in order to significantly increase the robustness of the available data about these sites, leading to better monitoring and even remediation in extreme cases. As land-use pressures increase with an expanding popu‐ lation abandoned landfill sites need to be avoided or used in appropriate ways.

Corresponding author. adewet@FandM.edu

INTRODUCTION Abandoned Waste Disposal Sites According to the most recent (2012) data there are 1,908 active Municipal Solid Waste landfills in the United States (EPA, 2014). The total number of landfills has been declining for decades; however, the total capacity has increased by 65 percent since 1980 and is currently around 251 million tons per year (Center for Sustainable Systems, 2014; EPA, 2014). Landfills have the potential to seriously affect environmental and human health (Herndon et al., 1990; Suflita et al., 1992), but landfill technology has improved significantly in recent decades, and most landfills in the United States are now well documented and monitored by various state and federal agencies (EPA, 2014). Potentially more troubling is that prior to the modern era of regulated sanitary landfills, municipal and other waste was disposed of in thousands of open dumps, haphazard (also referred to as promiscuous) dumps, and unregulated landfills across the United States. Precise numbers are unavailable, but estimates suggest that there are up to 100,000 dumps and landfills in these categories (Suflita et al., 1992). This number does not include the countless small farm and household dumps that have existed since colonial times. This study’s focus is on using remote sensing techniques to discover and characterize abandoned waste sites by examining three sites in Lancaster County, PA. Lancaster County has a long history of human occupation. Prior to colonial times the area was occupied by Native Americans, primarily from the Susquehannock Group (Kent, 1984; Wallace, 1989). In the early 1700s the area was settled and largely deforested by Europeans. Lancaster County was founded in 1729, and the City was founded in 1730. The population grew rapidly, particularly from the 1950s on, and is currently over 530,000 (US Census Bureau). The population density is 550 per square mile, double the Pennsylvania average (US Census Data for 2010). The county is also an important agricultural area, with nearly 6,000 farms and a growing Amish community (Lancaster Farmland Trust, 2015). While the county is only one of 67 counties in Pennsylvania, according to a Lancaster County Planning Commission

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 113–130

113

de Wet

report in 1971 (Lancaster County Board of Commissioners, 1971) there were approximately 23 waste disposal sites, including eight landfills (including one licensed sanitary landfill), 15 dumps, approximately 44 identified haphazard dumps and perhaps 200 to 300 additional unidentified haphazard dumps in the county at the time. In most cases the report does not give precise locations for these waste disposal sites. More recently the Pennsylvania Department of Environmental Protection (PADEP) has produced an inventory of abandoned landfills in Pennsylvania, part of the Abandoned Landfill Inventory (ALI) Project, which collects geospatial and descriptive data for closed and abandoned landfills throughout the state of Pennsylvania (www.psda.psu. edu). ALI site locations were determined using historic records such as microfiche, index cards, topographic maps, and staff personal files. These data were then compiled into site lists and incorporated into a GIS database (“Municipal Waste Operations”) available from the Pennsylvania Spatial Data Access (PASDA) on the Pennsylvania Geospatial Data Clearinghouse Web site (www.pasda.psu.edu). Of the 2,620 facilities listed, 1,309 are described as “landfills” or “abandoned landfills.” There are 157 reported locations in Lancaster County. Seventeen are “landfills” or “abandoned landfills”; however, none of the landfills examined in this study are in the ALI database. But many of these former waste disposal sites are known to local Lancaster County officials and residents or are recorded in various unpublished documents and reports held by local government agencies. There is concern that as time passes much of this local knowledge will be lost. This also raises the question of how many other former dumps/ landfills remain to be re-discovered in the future. In 1965 the US Congress passed the Solid Waste Disposal Act, and in the early 1970s many waste disposal sites were closed because of increased public awareness of the environmental problems associated with unregulated waste disposal and because more strenuous legislative and oversight requirements were being implemented (Lancaster County Board of Commissioners, 1987). Lancaster County is typical of counties across the United States and elsewhere. Unidentified former waste disposal sites are a national and international issue that warrants attention for human and environmental health and safety reasons. In addition, increased land use pressure in urban and suburban areas may lead to re-use of these locations without an understanding of their past history, with potential negative health and environmental consequences.

114

Study Locations This study demonstrates how a combination of LIDAR and historical aerial photographs provides a powerful method for identifying abandoned waste disposal sites and determining detailed information about each site’s history and specific characteristics. Three sites in Lancaster County were investigated using this combined technique; two sites are fairly welldocumented former landfills; the other site was ‘discovered’ and characterized in this study (Figure 1). All three locations were field checked. Using this technique numerous other sites in Lancaster County have been identified as potential waste disposal sites that should be investigated further. The waste disposal locations discussed here were active between the 1940s and the late 1960s. Local government reports, unpublished documents, and “anecdotal knowledge” provided varying amounts of information about their existence and approximate location. The first site—Baker Woodlands/Spalding Conservancy—has previously been studied using a variety of techniques including near-surface geophysics (de Wet et al., 1998, 1999). Approximately 100 acres of this area was recently designated as a nature conservancy by Franklin & Marshall College (Spalding Conservancy). This site includes several landfills that were active during the 1950s and 1960s (de Wet et al., 1998, 1999) and received Lancaster City municipal waste, debris from an urban renewal project in Lancaster City, and waste from several local manufacturing companies. Because the location of former waste deposits on the site are already fairly well defined it is an excellent place to test the value of using newly available LIDAR data combined with historical aerial photography to discover and characterize abandoned landfills. Between 1962 and 1968, after the Baker Woodlands landfills were closed, Lancaster City municipal and demolition waste went to another site, now located in the County Park (Lancaster County Commissioners, 1973). This location is partially documented in local government reports but was delineated in detail here using LIDAR data and historical aerial imagery and is the second location investigated in this study (referred to as the County Park site). The third site, the South Duke Street landfills, is substantially older; it was active during the 1940s and is located very close to Lancaster City within the floodplain of the Conestoga River. Little is known about this landfill, but field observations suggest that it also received primarily municipal waste. After the 1960s, Lancaster City waste was disposed of in the Manor Township landfill, located close to the currently active Frey Farm sanitary landfill (Figure 1). Presently, most Lancaster County household waste

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 113–130

Abandoned Waste Disposal Sites

Figure 1. Map showing the three sites studied here and the location of the municipal waste operations in Lancaster County included in the Abandoned Landfill Inventory (ALI) produced by the Pennsylvania Department of Environmental Protection. There are 157 reported facilities in Lancaster County, PA, of which 17 are “landfills” or “abandoned landfills.” Study site 1 is the Baker Woodlands/Spalding Conservancy site, study site 2 is the Lancaster County Park site, study site 3 is the South Duke Street site, and study site 4 is Frey Farm landfill, the current sanitary landfill used for most of the disposal of most of the municipal waste generated in Lancaster County (principally in the form of ash produced by the Resource Recovery Facility located along the Susquehanna River in the northwest part of the county).

is incinerated, and the ash and other municipal waste are buried in the Frey Farm sanitary landfill. Topographic Information Using LIDAR Since landfills typically modify the land surface in ways that are distinguishable from natural processes,

high-resolution spatial information can be used to identify and characterize specific features. In some cases, landfills simply change the topography through the addition of material on top of the original natural surface. The resulting topographic features are usually distinct from naturally produced landforms. In other cases, landfills occupy areas that were previously

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 113–130

115

de Wet

disrupted by human activities such as quarrying or surface mining, and in these situations landfill detection can be challenging if the natural grade of the landscape was reestablished when the landfill closed. Most landfills and dumps used prior to 1970, however, were merely abandoned with little or no effort to return the landscape to its natural state. Even landfills that were ‘graded’ prior to closure likely leave a telltale topographic signature that should be evident in detailed topographic data sets. While digital topographic data such as digital elevation models (DEMs) derived from Shuttle Radar Topography Mission and Advanced Spaceborne Thermal Emission and Reflection Mission have been available for years and cover most of the United States, these data sets are not of sufficiently high resolution to be helpful in identifying the relatively small topographic changes associated with landfills. Recently, however, high-resolution LIDAR data with typical horizontal resolutions of ,1 m and vertical resolutions of ,0.5 m are becoming available for large parts of the United States. According to an inventory of known high-resolution digital elevation data sources conducted as part of the National Enhanced Elevation Assessment (NEEA) in the summer of 2011, LIDAR data have been collected over 28 percent of the conterminous United States and Hawaii (http://nationalmap.gov/3DEP/neea.html). In the United States, elevation data are available through The National Map (nationalmap.gov), and while lower resolution data are available across the United States, high-resolution (1-m) bare-earth DEM data will be populated as new data is acquired in 2015 and beyond (nationalmap.gov/elevation.html). In many cases individual states have partial or complete high-resolution (,1-m) LIDAR coverage that is publically available. For example, in Pennsylvania, data are available from the Pennsylvania Spatial Data Access (PASDA) Web site (www.pasda.psu.edu) and in Maryland from the GeoSpatial Data Center (http://dnrweb. dnr.state.md.us/gis/data/index.asp). These highresolution elevation data offer a way to observe and characterize landscape changes, and despite some limitations, these data provide detailed topographic information even in vegetated areas. This is particularly important since many old landfills were left to re-vegetate and may now be forested or brush covered. Historical Aerial Photography Numerous human activities modify topography, so more than just LIDAR information is needed to rediscover and characterize former waste disposal sites. Current and historical aerial photography provides this necessary additional information. Current aerial photography includes very high-resolution normal

116

color and color infrared (CIR) imagery that provides detailed information about recent changes in vegetation and the land surface, which in combination with topographic data is useful for detecting former landfills. Older aerial photography is usually black and white and varies in spatial resolution. In many parts of the United States, aerial photography dates back into the 1930s and occasionally even into the 1920s. Since the 1940s virtually every part of the U.S. land surface has been photographed, with updates every several years. Historical aerial photography to detect and assess former landfills and hazardous waste sites is an established technique (Erb et al., 1981; Lyon 1987; Pope et al., 1996; de Wet et al., 1998, 1999; Biotto et al., 2009; and Silvestri and Omri, 2009), but there are limitations to it, especially because a wide variety of other human activities can disrupt the earth’s surface and be confused with landfills. Since landfills operate over the timescale of years and produce significant surface disruption and modification, combining aerial photography with co-registered high-resolution LIDAR data and its derivatives yields enhanced detection and characterization of landfills. METHODS LIDAR and Aerial Photography LIDAR data for the three sites investigated here were obtained from two different sources—the U.S. Geological Survey (USGS) and the Pennsylvania Department of Conservation and Natural Resources (PA DCNR). LIDAR data used for the analysis of the Baker Woodland/Spalding Conservancy site were obtained from the USGS. The data were acquired in 2004 by AERO-METRIC, Inc., using an airborne OPTECH ALTM 30/70 sensor. USGS post-processed the data with Realm software and used TerraScan software. AERO-METRIC provided 15 ground truth points for quality assurance and quality control checks against global positioning system (GPS) Rapid-Static survey methods. LIDAR returns were filtered and classified to produce a last-return bare-earth data set. The bare-earth returns have a vertical root mean square error of better than 15 cm relative to North American Vertical Datum of 1988 and a nominal horizontal spacing of better than 2.0 m based on Universal Transverse Mercator (UTM) coordinate system related to the North American Datum of 1983. In this study, the LIDAR bare-earth returns were imported into an ArcGIS Terrain model and triangulated. The triangulated data were rasterized to a ground resolution of 1 m for further processing and visualization. The LIDAR data for sites 2 and 3 were

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 113–130

Abandoned Waste Disposal Sites

Figure 2. WDOW LIDAR map and topographic profiles of study site 1 (Baker Woodlands/Spalding Conservancy) and surrounding area. The properties owned by Franklin & Marshall College are shown. The western property has been designated by the college as a conservancy (Spalding Conservancy). It is presently mostly wooded and includes much of the area affected by the brick-making activities (brick kilns, clay pits, etc.) and later landfills. The eastern property comprises mainly sports fields but includes some woods that were also affected by clay mining and landfills.

downloaded from the PASDA Web site (www.pasda. psu.edu) and were produced by the PAMAP Program, PA DCNR, Bureau of Topographic and Geologic Survey. The data were collected in 2006 and are available in a processed bare-earth surface DEM format using the Pennsylvania State Plane coordinate system and have a horizontal ground resolution of ,1 m and a typical vertical resolution of less than 1 m. The LIDAR data were processed and displayed using different rendering techniques, including basic gray-scale or color-coded schemes. These images form, in essence, a highly detailed and accurate topographic map of the earth’s surface. This type of map

was combined with additional spatial data for interpretation and visualization purposes. For this study, hillshading, slope, and aspect pro‐ cessing techniques were employed to highlight topography. Hillshading, or illuminating a topographic surface from one direction, emphasizes features oriented at right angles to the illumination direction and de-emphasizes features parallel to the illumination direction. An illumination angle of 45u above horizontal was used in this study. To compensate for any loss of features due to one-direction illumination, composite multiple hillshading was performed using Spatial Analyst using Mul‐ tiple Direction Oblique Weighted (MDOW) analysis

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 113–130

117

de Wet Table 1. Details of the aerial photographs used for each study site. Site 1, Spalding Conservancy

2, County Park

3, South Duke Street

Number ahg102193 ahg102194 AHG-3D-22 GS-OY 1 59 ahg_6r_18 ahg_7r_44 ahg_7r_45 ahg_7r_46 AHG-6EE-280 ahg_3mm_19 173-189 42071-178-57 27002360PAS AHG-3D-22 ahg_2r_64 AHG-6EE-242 ahg_3mm_103 173-191 42071-178-57 26002370PAS 18TUK885295 18TUK885310 18TUK900295 18TUK900310 ahg102101 ahg102102 ahg102103 ahg102139 ahg102140 AHG-3D-22 AHG-3D-23 GS-OY 1 59 ahg_2r_64 26002370PAS 18TUK900310

Date 4/29/1940 4/29/1940 6/17/1947 4/21/1951 11/11/1957 6/7/1958 06/071958 06/071958 5/23/1964 7/5/1971 10/27/1974 10/29/1978 2007 6/17/1947 9/27/1957 5/23/1964 7/5/1971 10/27/1974 10/29/1978 2007 2012 2012 2012 2012 4/29/1940 4/29/1940 4/29/1940 4/29/1940 4/29/1940 4/17/1947 4/17/1947 4/21/1951 9/27/1957 2007 2012

Source 1

Penn Pilot Penn Pilot1 F&M College F&M College Penn Pilot2 F&M College F&M College F&M College F&M College Penn Pilot3 F&M College F&M College4 PASDA5 F&M College Penn Pilot2 F&M College Penn Pilot3 F&M College F&M College4 PASDA5 PASDA6 PASDA6 PASDA6 PASDA6 Penn Pilot1 Penn Pilot1 Penn Pilot1 Penn Pilot1 Penn Pilot1 F&M College F&M College F&M College Penn Pilot2 PASDA5 PASDA6

Scale

Digital Resolution (ft)

1:20,000

0.686 (2.25) 0.686 (2.25) 0.896 (2.94)

1:20,000

0.677 (2.22) 0.643 (2.11) 0.643 (2.11) 0.643 (2.11) 0.24 (0.8) 0.67 (2.19)

Coord. None

None

1:20,000 — 1:20,000 1:20,000

— — — — — 1:20,000 1:20,000 1:20,000 1:20,000 1:20,000

1:20,000 — —

0.3 (0.98) 0.896 (2.94) 0.677 (2.22) 0.25 (0.8) 0.67 (2.19)

0.3 (0.98) 0.3 (0.98) 0.3 (0.98) 0.3 (0.98) 0.3 (0.98) 0.69 (2.25) 0.69 (2.25) 0.69 (2.25) 0.69 (2.25) 0.69 (2.25) 0.90 (2.94) 0.90 (2.94) 0.68 (2.22) 0.3 (0.98) 0.3 (0.98)

None PA st pl7 None None None None None None PA st pl7 PA st pl7

None

None None None PA st pl7 PA st pl7

Produced by: U.S. Department of Agriculture (USDA) Agricultural Adjustment Administration Northeast Division. USDA Commodity Stabilization Service. 3 USDA Agricultural Stabilization and Conservation Service. 4 USDA. 5 PAMAP Program, PA Department of Conservation and Natural Resources, Bureau of Topographic and Geologic Survey. 6 Lancaster County. 7 PA st pl: Pennsylvania State Plane. 2

(Mark, 1992) (Figure 2). Using ArcMap, a hillshade grid (shade_flt or shade_int) was placed above the bare-earth grid in the layers tree. The hillshade layer was then displayed using a gray-scale or monochromatic gradi‐ ent color scheme, made 50 percent transparent, and stretched to ‘minimum-maximum.’ The bare-earth elevation data were displayed using a different color, showing where land use/land cover relates to elevation. The bare-earth elevation was also stretched using ‘minimum-maximum.’ The raster shade_flt is kept as a floating point raster to allow for certain operations, such as histogram stretching or lightening in ArcGIS or Photoshop, which required additional gray-scale bit-depth. A significant advantage of this technique is that it integrates complex information into one layer.

118

Slope and aspect were derived using standard techniques in ArcGIS, and the resulting data were combined with other spatial data, such as aerial photographs. The LIDAR data set forms the threedimensional (3-D) visualization base for the study sites. Detailed 3-D characterization of each site was obtained by combining the LIDAR-derived DEM with historical aerial photography and other spatial data sets. In this study, historical aerial photographs were downloaded from the Penn Pilot Web site or scanned from photographs archived in the Department of Earth & Environment at Franklin & Marshall College (Table 1). The scanned historical photographs typically had a pixel resolution of ,2 m and were

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 113–130

Abandoned Waste Disposal Sites

georeferenced using standard methods in ArcGIS. Recent aerial photographs (PAMAP_cycle2; and Lancaster County, PA RGB and CIR Orthoimages) were obtained from the PASDA Web site and integrated into the ArcGIS database. The PAMAP_cycle2 orthorectified digital raster images were taken in 2007 and have a horizontal ground resolution of ,0.5 m (1 ft) and use the Pennsylvania State Plane South coordinate system (Table 1). The Lancaster County, PA, True Color (RGB) and Color InfraRed (CIR) Orthoimages were taken in 2012 and have a resolution of 0.076 m and are available in either UTM or PA State Plane South coordinates (Table 1).

pits. The volumes of the landfills were then calculated based on the aerial extent and the estimated average depths. Bedrock geology, soil type, and drainage patterns were examined to determine potential envi‐ ronmental impacts of landfill leachate. Historical documents such as newspapers and magazine articles provide rich details about a site’s past, although they generally only refer to major events or projects and so were used to supplement other data sources. Remote-sensing can potentially provide detailed information about waste disposal sites, but on-site field observations, mapping, near-surface geophysics, and sampling, where possible, will always be needed to compliment remotely derived information.

Database and Abandoned Landfill Determination Combining LIDAR and historical aerial photographs allows for planiform analysis of former landfills, leading to detailed spatial and temporal information about their location and extent. This forms the backbone of the GIS database used in this study, but the data set also includes roads, topographic contours, geology, soils, land parcels, and hydrology (all available from the PASDA Web site). Spatial data were then analyzed either individually or in combination with other data sets in the GIS. For example, initial inspection of the hillshaded LIDAR DEM in the target areas revealed many raised, relatively flat areas surrounded by slopes or peculiar hummocky terrain strongly suggestive of an anthropogenically modified surface. Viewing the slope derived from the LIDAR DEM further emphasized these areas. In other cases, hummocky terrain suggestive of dumping was observed. These data were then overlaid on recent aerial photographs to exclude areas that have residential or commercial development in order to focus on potential landfill sites. It was assumed that it was un‐ likely that new housing or commercial developments were established on abandoned landfills. Georeferenced historical aerial photographs of potential landfill sites were studied. Aerial photographs taken before, during, and after the landfills were active were used to map and determine the timing and extent of each landfill. Where available, stereo pairs of historical aerial photographs were examined manually and were used in conjunction with the digital data for estimating landfill boundaries and to provide semi-quantitative topographic information. Aerial extent and cross-sectional profiles were measured using standard techniques in ArcGIS. Landfill depth or thickness was estimated either by extrapolating the pre-landfill surface from the surrounding area, in the case of landfills in which material was just added to the surface, or depth estimates from stereo aerial photos, where material was dumped into excavated

CASE STUDY 1: BAKER WOODLANDS/ SPALDING CONSERVANCY Background The Baker Woodlands site was chosen because it has been well studied, including analysis of aerial photos, field observations, mapping, and near-surface geophysics (Figures 1 and 2) (de Wet et al., 1998, 1999; De Wet and Sternberg, 1999). Prior studies have focused on the history of commercial brick making (Horning, 1992), farming and other land-use history (de Wet and Sternberg, 1999; de Wet et al., 1999, 2000; and Carlson et al., 2000), and the site’s ecological recovery (de Wet et al., 1998). These studies revealed a complex land-use history dating back to European colonization of the area in the 1700s. A railway defines the northern boundary of the site, and it is part of the original Philadelphia and Columbia Railroad that was completed in 1834 (Wilson, 1985). Substantial land impacts occurred after the construction of a brick factory in 1920 (Horning, 1992). A large portion of the site was graded for the construction of the brick-making facilities, and between 1920 and the 1940s clay was excavated, creating pits covering 10 acres that were probably up to 10 m (30 ft) deep (de Wet et al., 1999). Many of these open pits were subsequently used as landfills in the 1950s and 1960s. After the brick factory closed in the late 1970s, the buildings were demolished, with the debris left largely in place. Most of the site has been allowed to re-vegetate, thus obscuring the surface topography and restricting near-surface geophysical surveys to narrow transects cut in swathes through the woods and scrub. The vegetation now consists of native and numerous non-native invasive species such as Norway maple (Acer platanoides), Ailanthus “tree-of-heaven” (Ailanthus altissima), English ivy (Hedera helix), and Multifora rose (Rosa multiflora) (de Wet et al., 1998).

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 113–130

119

de Wet

A large part of the original Baker Woodlands was recently designated a conservation area for recreation and long-term ecological and environmental research (Spalding Conservancy). Detailed understanding of the siteâ&#x20AC;&#x2122;s land-use history is important in shaping the conservation plan. As a result of the conservancy designation, clear-cutting transects for geophysical surveys are less likely to be permitted, meaning that non-destructive techniques such as LIDAR will become even more important in the future. Conservation land restoration plans include actively managing woodland areas, enhancing and expanding an existing wetland, restoring and preserving historical structures on site, construction of a trail system, and adding several deer exclosures. Results The LIDAR maps show elevation variations across the site and adjacent areas (Figure 2). The effect of extensive and complex alterations to the topography related to decades, and in some cases, centuries, of land-use change is clearly evident. The central and south portions are generally suburban development and associated infrastructure. Most obvious is the topographic patchwork of modifications for houses and streets. The northern area is dominated by major roads, a railway line, sports fields, and commercial and industrial buildings and parking lots. The flat sports fields and the flat interpolated area of commercial and industrial building sites are clearly evident. Located between these areas is the Spalding Conservancy, with its complex and highly variable topographic signature consisting of smooth regions interspersed with hummocky terrain and areas of parallel ridges and depressions on the scale of 1 to 10 m. Figure 2 shows a bare-ground MDOW layer (50 percent transparent) overlain on the LIDAR topographic layer with railways, streams, property boundaries of the Baker Woodlands/Spalding Conservancy, and the landfill locations. The combination of MDOW bare ground on LIDAR topography revealed subtle variations in topography. Overlaying the MDOW layer (50 percent transparent) on the topographic layer showed multiple shadows and revealed new details, especially in the low-lying areas adjacent to the Little Conestoga Creek and within the residential area south of the Spalding Conservancy. LIDAR data were used to derive a slope grid, in which the steepness of the slope is represented by increasing shades of gray. Slope maps were combined with color-coded topographic results to emphasize slope variations. Landfills in particular are often characterized by broad flat areas bounded by steep slopes;

120

thus, slope maps are particularly useful in defining the edges and size of former landfills, as seen in Figure 2. High-resolution cross sections extracted from the LIDAR data reveal topographic variations across the site, illustrating its complex past (Figures 2 and 3). Historical aerial photographs from 1936 show that farming had given way to the brick works, and photographs from the 1940s show numerous buildings and excavations on the site. Photographs from the 1950s and 1960s show major site disruptions when large areas were de-vegetated or covered with debris associated with landfill activity. The Landfills Historical documents (Horning, 1992), field observations (de Wet et al., 1999), and aerial photographs indicate that the brick workâ&#x20AC;&#x2122;s clay pits were converted to landfills during the late 1950s and early 1960s (de Wet et al., 1998). The landfill areas are apparent in the LIDAR maps and photographs (Figures 2 and 3). Two landfills are located within the present Spalding Conservancy; one landfill is located north of the Conservancy (north of the railway line), and there are several landfills northwest of the Conservancy (not shown on these maps) (Figure 3). All of the former landfills appear as positive topographic features, and they have a distinctive LIDAR texture because debris was not graded or smoothed after being dumped into the former clay pits. This produced small-scale, hummocky surfaces evident in the LIDAR results. The contrast between the disrupted landfill surfaces and the relatively undisturbed farm fields is particularly clear along the boundary of the landfill north of the railway and outside of the Spalding Conservancy (Figure 3). This landfill boundary is particularly well known because it was excavated in 2012 to stabilize the ground for the construction of a new rail yard (Figure 4). Slight topographic variation in each landfill coincides with different episodes of landfill activity, as evident in aerial photographs (Figure 3) and near-surface geophysical data (de Wet et al., 1999). Prior to its excavation, the northern landfill covered an area of 33,400 m2 (8.25 acres) and had an average thickness of 4 m, resulting in a volume of 133,600 m3 (174,742 yd3) (Figure 3 and Table 2). The two landfills located in the Spalding Conservancy covered areas of 10,904 m2 (2.69 acres) and 47,378 m2 (11.7 acres), and assuming an average thickness of 4 m, these landfills have volumes of approximately 40,000 m3 (52,318 yd3) and 190,000 m3 (248,510 yd3), respectively. Leachate-contaminated seeps occur on the downslope sides of both landfills and discharge into the adjacent wetlands before flowing into the Little Conestoga

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 113â&#x20AC;&#x201C;130

Abandoned Waste Disposal Sites

Figure 3. (a) Location map; (b) WDOW shaded relief LIDAR map showing parts of the site affected by the landfills and the location of the clay pits, lime kiln, and associated limestone quarry; (c) 1964 aerial photograph showing the main landfill area, location of topographic profile A-B, and the location of currently active seeps; (d) topographic profile A-B showing the location of the landfill and the seeps downslope of the landfill.

Creek (Figure 3). The wetlands appear to act as a sink for much of the metal content of the leachate (Wilson et al., 2006). Visualization of the landfills is enhanced by using the LIDAR results as an elevation base and by

rendering the historical aerial photographs in 3-D on top of it. Figure 5 shows the 1964 historical aerial photograph rendered in 3-D using the LIDAR data as the elevation base. Vertical exaggeration is 106. Multiple layers were combined and rendered in 3-D

Table 2. Details of the landfills/dumps at each study site. Site 1, 1, 1, 2, 3, 3,

Dates Active

Spalding Conservancy (north) Spalding Conservancy (west) Spalding Conservancy South (east) County Park South Duke Street (north) South Duke Street (south)

Late 1950s–1962 Late 1950s–1962 Late 1950s–1962 1962–1968 Mid-1940s–early 1950-s Mid-1940s–early 1950-s

Approximate Dimensions (m) 365 150 135 860 162 185

6 90 6 4 6 80 6 4? 6 360 6 4 6 200 6 5 6 105 6 6 6 45 6 5

Area (m2)

Area (acres)

Volume (m3)

33,400 10,904 47,378 187,761 13,063 7,161

8.25 2.69 11.71 46.40 3.23 1.77

133,600 43,616 189,512 938,805 78,378 35,805

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 113–130

121

de Wet

lows and topographic landfill highs are readily characterized using this representation method. Clay Excavations

Figure 4. Field photograph from study site 1 showing the excavation of the landﬁll located north of the railway. Up to 4 m of relatively unconsolidated landﬁll debris was removed and replaced by stable material in order to construct a railyard.

to enhance the overall interpretation. Alternatively, multiple layers can be rendered in 3-D and then displaced vertically to allow several layers to be visualized in 3-D simultaneously. Clay pit topographic

Clay pit distribution was mapped and characterized based on the LIDAR images (Figure 2, cross sections A-B and C-D). Excavations extend to the southern edge of the site and follow the orientation of the regional geology toward the east. Some of the excavated areas were later filled with landfill debris, as discussed above, but several others were abandoned after the clay was mined out, and the LIDAR data clearly show these excavated areas as linear and arcuate pits and benches (Figure 2). After the 1940s, clay excavation moved north of the railway line and west of the Little Conestoga Creek. Geological Features—Regional Structure Cambro-Ordovician limestone with thin graphitic or micaceous beds, the Conestoga Formation, constitutes the local bedrock (Meisler and Becher, 1971; de Wet et al., 1999). Bedrock influences the ENEWSW orientation of numerous small streams, while

Figure 5. Three-dimensional view of the site looking toward the ENE. The 1964 aerial photograph is draped over the LIDAR data (106 vertical exaggeration). The landﬁlls are clearly visible and correspond to relative topographic highs. In 2013 the wetland area was expanded (light green) from the original palustrine wetlands (dark green) that existed between the main landﬁll and the Little Conestoga Creek in the foreground.

122

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 113–130

Abandoned Waste Disposal Sites

the major drainage (Little Conestoga Creek) cuts across the geological structure, flowing south to the Conestoga River (Figure 2). As revealed in the LIDAR results, bedrock controls the distribution of clay pits and, therefore, landfill locations. It also directs groundwater flow and hence the distribution of landfill leachate seeps. A previously unknown structural feature in the southern residential area became apparent with the LIDAR results (Figure 2). The feature occurs along a ridge that separates two minor drainages and is oriented N70uE. It parallels the general orientation of the bedrock but is difficult to detect in the field because of the density of trees and houses. Close examination of the 1940 aerial photographs, which record pre-suburban development, reveals a very subtle break in slope that reflects the structural lineament revealed by LIDAR. Field observations of numerous vein quartz float blocks in the area suggest that the lineament is related to fractures filled by resistant quartz veins that produce a topographic high. The LIDAR results also delineated a former limestone quarry in a wooded part of the adjacent neighborhood. Other Features The oldest constructed features recognized on the site are a lime kiln and several small associated limestone quarries. These features are associated with the limestone ridge that parallels the regional geological structure (Figure 2; cross section A-B). This area was the only part of the site not affected by either brick making or subsequent activities. Mature trees and relatively undisturbed soils occur along the ridge. Debris from the demolition of the brickworks buildings is visible in the LIDAR maps (Figure 3), and foundations and walls are traced out. Beehive kilns for baking bricks are also visible, based on the dis‐ tribution of their demolition debris. Documenting historical structures is important for site characterization, and in this case, the Spalding Conservancy management plan may include removal of building debris, making this study’s maps particularly important in terms of historical knowledge preservation and re‐ moval methods. The site’s original hydrology has been significantly modified by land use over time. Seeps associated with the landfills drain leachate with elevated iron content that is naturally attenuated by wetlands downslope (de Wet et al., 2000). LIDAR data revealed new details of the site’s hydrology and clarified relationships between the drainage areas and wetlands (Figures 3 and 5). The exceptional accuracy of the LIDAR made it possible to map surface water flow from the seeps near the landfill area, through the

wetlands, to a discharge point into the Little Conestoga Creek. This information is guiding ongoing monitoring of the leachate produced by the landfills. Numerous enigmatic features visible in the 1940 (and later) aerial photographs of the wetland areas are now recognized as pits (partially in-filled) and mounds through this study’s LIDAR results. These features are previously unknown remnants of the 1920–1940 clay mining operations but have been evaluated and incorporated into the Conservancy’s wetland expansion project (Figure 5). New results from this study have significantly increased our confidence in determining the former landfill dimensions. This is particularly important as we quantify current environmental conditions and make recommendations for ongoing site management at the Conservancy. CASE STUDY 2: LANCASTER COUNTY PARK Background This site is located within a large meander in the Conestoga River and is bounded on the west and north by the Conestoga River and to the southeast by Mill Creek (Figure 6). Most of this area is currently a county park, but between 1962 and 1968 much of it was used as a landfill for Lancaster City and the surrounding townships. There are a number of county documents and maps for this site, and historical aerial photographs show that before the early 1960s, flatter parts of the area were farmland and steep slopes were wooded. The underlying bedrock is Cambro-Ordovician limestone (the Conestoga Formation, as at site 1), with a regional strike and dip such that groundwater moving through the landfill flows downslope east and west along fractures and karst cavities toward the Conestoga River and Mill Creek. Since the land‐ fill is unlined, leachate seeps surrounding the landfill became active soon after the landfill closed. In the early 1970s a leachate report was produced that included several proposals for remediation (Lancaster County Commissioners, 1973), but it is unclear if any of the plan was implemented since leachate continues to discharge into the river and stream and periodically raises public concern (Rutter, 2008). Recent analyses note the presence of iron, nickel, mercury, zinc, arsenic, chloroethane, and benzene, but the landfill has been removed from a list of possible Superfund cleanup sites, suggesting that solute concentrations are low (Rutter, 2008).

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 113–130

123

de Wet

Figure 6. Maps and topographic profiles of study site 2 Lancaster County Park landfill, (a) slope map derived from the LIDAR data with the outline of the landfill. Topographic profile A-A9 is located north of the landfill, B-B9 shows the northern part of the landfill and the location of the seeps at the base of the slope, and C-C9 shows the southern part of the landfill. The thickness of the landfill was determined by a combination of extrapolating the topography from outside of the landfill, changes in the slope, and observations from stereo-pairs of historical aerial photographs. (b) Colorized topography derived from LIDAR data overlain on the 1971 aerial photograph with the location of the landfill. The photograph, taken shortly after the closure of the landfill, clearly shows the maximum extent of the landfill.

Results: Location and Extent of the Landfill The landfill’s outline is defined by a relatively flat area bounded by regular steep sloping sides that contrast with natural slopes in the surrounding area

124

(Figure 6). The slope map derived from the LIDAR results clearly distinguishes the relatively flat upper surface of the landfill from the surrounding topography. Its steep side slopes formed as debris accumulated over the original hillsides. The landfill comprises three

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 113–130

Abandoned Waste Disposal Sites

sections—a smaller, lower northern section and two larger upper sections. The overall landfill area is 187,761 m3 (,46 acres) and varies up to 10 m deep along the western edge (Table 2). Assuming an average depth of 5 m, the volume of debris is almost 1,000,000 m3 (1,307,951 yd3). Historical photographs confirm that the landfill started after 1957 but before 1964. Photographs from 1971 and 1978 show some continued disruption of the landfill’s surface; however, stereo-pair photograph examination indicates that while new material was not added after 1971, some reworking of the landfill surface occurred. Maps of the approximate extent of the landfill are available in local documents (Lancaster County Commissioners, 1973) and confirm these LIDAR and photography results in terms of landfill area. The local maps, however, do not include any information about the landfill’s depth or in-fill history. This study shows that in 1964 the southwestern part of the landfill was active while material was being excavated from the northern section of the landfill. The excavated northern section was later backfilled with debris, resulting in a significantly thicker layer of waste material than would have been expected by simply comparing the pre-1962 topography with the current topography. The older southwest part of the landfill comprises refuse principally hauled from the city and surrounding areas, whereas the later northern part likely includes appreciable amounts of demolition debris from late1960s Lancaster City urban renewal projects. CASE STUDY 3: SOUTH DUKE STREET Background The South Duke Street site is the oldest and least well-documented location examined (Figure 7). There were two dump sites located in the floodplain of the Conestoga River, only a few meters from the river channel. They were both active during the 1940s. Little information is available about the type of material in these dumps, but field observations indicate that much of the material was probably household waste, most likely from nearby Lancaster City.

begun at that time. The 1947 aerial photograph clearly shows debris being added at the northern site, covering an area of 100 m by 100 m (10,000 m2, 2.5 acres), and it extended to within 30 m of the Conestoga River. The southern site was also active at this time, and because the floodplain was narrower here, the debris was located very close to the river and extended for 120 m parallel to the river along the floodplain. By 1950, there is no evidence of active dumping at the southern site, but there is evidence for continued activity at the northern site. The area of exposed ground in 1950 is very similar to the final ‘footprint’ for both dumps, as determined from the LIDAR results. Ultimately, the northern fan-shaped dump covered approximately 13,000 m2 (3.2 acres), and the southern dump covered 7,200 m2 (1.8 acres) (Table 2). By estimating the original topography of this area and comparing it to the LIDAR topography it is possible to determine the approximate volume of material (Figure 7). The depth of the northern dump varied from a few meters to over 15 m (average, ,6 m), resulting in an approximate volume of 78,000 m3 (102,020 yd3). The southern dump was much smaller and varied in depth from a few meters to 8 m (average, ,5 m), resulting in a volume of around 36,000 m3 (47,086 yd3). Both dumps are adjacent to the Conestoga River and are located in floodplain sediment; therefore, leachate and even solid waste entered the river in the past and may still possibly do so. In the past, waterways and wetlands were considered a convenient way to dispose of waste, but we now understand their vulnerabilities. It makes sense to look along waterways close to urban areas for abandoned waste disposal sites. Using the combination of LIDAR and historical aerial photographs several other sites along the Conestoga River that exhibit modification from their natural state and may be old landfills have been identified. Preliminary field observation confirms significant debris at these locations, but more detailed work is required to determine the nature of the material and whether it poses any danger to human health or the environment. DISCUSSION

Results The LIDAR data show several locations in which material appears to have been added to the floodplain along the Conestoga River (Figure 7). These areas are recognized by a relatively flat upper surface, steep bounding slopes, and planiform shapes. The oldest aerial photographs available (1940) indicate some activity (bare ground) at two of these locations, but it is unclear whether any dumping of material had

There are a number of techniques to identify former landfills and dumps but they all have drawbacks. For example, topographic data such as 10-m DEMs or 7.5-minute quadrangle topographic maps usually do not offer sufficient vertical or horizontal resolution to be useful for detailed land-use analysis. Field mapping may miss subtle changes in terrain or be challenging as a result of vegetation or lack of site access. Nearsurface geophysics requires equipment, time on the

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 113–130

125

de Wet

Figure 7. Maps and topographic profiles of study site 3—South Duke Street landfills, (a) colorized topographic map derived from the LIDAR data with the outline of the landfills and location of the topographic profiles. The estimated thickness of the landfills was determined by a combination of extrapolating the topography from outside of the landfills, changes in the slope, and observations from stereo-pairs of historical aerial photographs. (b) Colorized topography derived from LIDAR data overlain on the 1947 aerial photograph with the location of the landfills. The photograph was taken during active dumping at the landfills. Activity at the site was not evident in the 1950 aerial photographs.

ground, and possible vegetation clearing for transect lines. Although aerial photography provides information about land-use change through time, it is generally limited to the years after 1930, and there are technical shortcomings inherent in older photographic systems. For these reasons, the combined method used

126

in this study creates a robust model that is widely applicable to other waste disposal and land-use problems. LIDAR data are becoming widely available as a result of private, state, and federal efforts (for example, Myers, 2009), and historical aerial photographic coverage of the United States is generally good.

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 113–130

Abandoned Waste Disposal Sites

New Techniques, New Information LIDAR has been used to understand a wide variety of natural processes where geomorphic features are important, such as tectonics, landslides, glaciers, streams, and coastal processes as well as ecological studies, such as forest characterization, wetland mapping, and biomass determination (Lyon and Greene, 1992; Lefsky et al., 2002; Mitasova et al., 2004; Robertson et al., 2004; Drake et al., 2003; Arrowsmith and Zielke, 2009; and Haneberg et al., 2009). LIDAR is also being used for risk assessment and understanding human-environment interactions, including coastal erosion, volcanic hazards, flood risk mapping, landuse, and infrastructure mapping (Hodgson et al., 2003; Neelz et al., 2006; Rufin-Soler et al., 2008; Bisson et al., 2009; Tralli et al., 2009; and Zhou and Xie, 2009). LIDAR is now recognized as a key procedure for documenting natural land-surface change. In many cases, human impact on the environment might be suspected but the exact nature of the activity might not be obvious. As demonstrated in this study, the combination of LIDAR with aerial photography offers an effective method with which to distinguish natural from anthropogenically altered terrains and to locate and characterize long-abandoned sites, such as the South Duke Street dumps. By combining LIDAR data and its derivatives with other data sets, such as historical aerial photography, geophysical surveys, and field observations, as described here, detailed temporally and spatially complete interpretations of three former waste disposal locations were produced. According to PADEP there are 17 potential former landfills or dumps in Lancaster County, but the report did not include the three locations documented here. This strongly suggests that many more dumps remain to be discovered in the area and that this is an issue that affects counties, cities, and countries globally. Even if some minor dumps are never identified, hopefully they will be the dumps with the smallest environmental impact. Systematic use of a LIDAR and aerial photography combination, particularly if used within a GIS database, will permit agencies to locate and characterize major abandoned waste disposal sites with significant efficiency. While former waste disposal sites are one example in which the combination of LIDAR and historical aerial photographs becomes a powerful tool for documenting and characterizing past land use, there are a myriad of other applications; for example, former abandoned industrial sites in urban areas remain an issue despite the many successes of the Brownfields Program (Greenberg and Hollander, 2006). Many of these sites are conveniently located and could be returned to productive use either as new industrial

sites, as development opportunities, or as public parks and green spaces. Returning these sites to productive use requires thorough site characterization to determine the nature of past activity and whether any environmental concerns or hazards exist. The ability to virtually view a site in 3-D by combining LIDAR with a wide variety of current and historical data sets is useful for analytical purposes but also can help municipal planners and others to visualize site changes through time. For example, by rendering historical aerial photographs in 3-D using LIDAR results as a base, it is now straightforward to see how current topographic high areas correspond to landfills that were active in the late 1950s to early 1960s at the Baker Woodlands/Spalding Conservancy site. This effectively takes the viewer back in time to see the landscape as it appeared in the pastâ&#x20AC;&#x201D;an especially powerful visualization tool for students, environmental scientists, urban planners, and other constituents (Figure 5). Significance: Long-Term Implications of Abandoned Waste Disposal Sites Waste disposal sites may have an impact on human health, on the environment, on site engineering, and/or on local aesthetics. Human health issues may arise from leachate contamination of surface or groundwater (Suflita et al., 1992), physical harm from exposed debris, or airborne contaminates, contributing to asthma and other respiratory problems (EPA, 2008). In almost all cases, waste disposal sites prior to the 1970s were unlined, and few were capped with anything more than a thin soil cover. Rainwater percolates easily through them, carrying soluble material into local surface and/or groundwater systems (Cummins, 1968). In limestone karst terrains, such as the Conestoga Formation bedrock underlying the three Lancaster County dumps documented here, leachate may accumulate and travel significant distances underground. At the County Park waste disposal site documented in this study, leachate seeps drain along the perimeter of the former landfill and in some place cross public walking trails (Lancaster County Commissioners, 1973; Rutter, 2008). Exposed debris may cause physical harm, particularly to children and pets, if rusting metal, sharp plastics, and large objects (such as old refrigerators) are accessible on the ground surface. Unstable footing due to irregular ground settling may lead to injury if a site is open for recreational use. Asbestos was widely used in ceiling and flooring materials in the early 20th century, and demolition debris from that era may contain these materials. As the binding agents decay, asbestos and other particulates may be released into the air as the waste disposal site erodes (EPA, 2008).

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 113â&#x20AC;&#x201C;130

127

de Wet

Old waste disposal sites are likely to have received a wide variety of materials because there were no restrictions on what could be disposed of at these sites prior to the stricter regulations introduced after the 1970s. There is no information available about the type of material deposited in the Duke Street dump (site 3), and it is unclear whether the material disposed of in the 1940s was more or less harmful than typical material disposed of in the 1950s and 1960s. It is likely that there were many cases between the 1940s and 1960s of hazardous material disposal in unregulated municipal waste sites (Colten and Skinner, 1996). The Spalding Conservancy site received a significant quantity of ceiling and floor tile material as waste, much of which is now exposed on the land surface (de Wet et al., 1999). The age of the landfill not only determines the content but also relates to change with time. For example, over time the quantity and composition of leachate may change, as different materials decay at different rates. This influences the composition of leachate draining from a site (Department of the Environment, 1990; Marsh and Garnham, 1996; and Metcalfe and Rochelle, 1999). Study of leachate from the County Park landfills shows a change in leachate composition over time (Rutter, 2008). The environment is affected by the presence of a landfill or dump because of the physical disruption to the local area while the site is active (i.e., excavations, road building, and debris accumulation) and, subsequently, with land subsidence and poor soil development, leading to scrubby vegetation, often dominated by invasive species and thus generally poor wildlife habitat. Over time, habitat may improve somewhat as the ecosystem recovers, but, as noted for human health, significant wildlife hazards remain. Former waste disposal sites may shed trash material into the environment for decades as they erode. This affects an areaâ&#x20AC;&#x2122;s aesthetic value, affecting property values and quality of life for nearby residents. Compaction and decomposition of material within waste disposal sites makes them inherently unstable and, therefore, unsuitable for most development projects. Cavities may form where material readily decays, roofed by less soluble or compactable material that could give way with increased overburden. If a location is chosen for development but is atop a former waste disposal site that is difficult to detect from the surface, such as the three locations documented here, serious construction problems will arise. Making sound land-use decisions for safety, environmental concerns, municipal growth, and aesthetics requires an understanding of the previous land use. A database in which former landfills and dumps are clearly delineated is an important aspect of municipal planning that needs to be updated as new information becomes available. The combination of

128

LIDAR and aerial photography is an efficient method of obtaining critical land-use data that should be used to guide future land-use planning. CONCLUSIONS LIDAR is a valuable addition to the tools needed for accurate site assessment. First, used directly, it provides spatially complete, high vertical- and horizontalresolution information about the topography of a site. This data set can then be used to derive numerous other useful data sets, including slope, aspect, and shaded relief (particularly using MDOW analysis). Creative symbolization of the topographic data can reveal information such as subtle changes in topography that provides clues to natural and anthropogenic influences on the site. Second, when integrated with other historical and current spatial data sets, such as historical aerial photographs and geophysical surveys, LIDAR provides a powerful way to interpret and understand a siteâ&#x20AC;&#x2122;s land-use history. LIDAR is particularly useful in areas in which highresolution data are needed but are difficult, prohibited, or expensive to obtain. Sites with significant obstructions, including buildings and vegetation, may prevent or restrict the use of other mapping techniques that use instruments such as electronic distance meters or GPS. LIDAR is useful for sites that have restricted access because of hazards, or in sensitive areas such as wetlands or other special ecological areas. In these circumstances, using remotely sensed data such as LIDAR has significant advantages. Finally, LIDAR provides a high-resolution base for the 3-D rendering of other data sets, such as historical and recent aerial photographs. Visualization in 3-D can provide unique insight into the history of an area and provide a powerful way to analyze and visualize natural and anthropogenic change over time. The ability to seamlessly transition between different data sets is one of the powerful attributes of integrating information in a GIS environment, and the resultant data reveals new insights into situations that could directly affect human health, environmental standards, and future development opportunities. ACKNOWLEDGMENTS Mike Rahnis provided technical assistance with the processing of the LIDAR data. Chris Williams and Carol de Wet provided helpful comments, and Steve Sylvester provided local government documents and personal knowledge of sites across Lancaster County.

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 113â&#x20AC;&#x201C;130

Abandoned Waste Disposal Sites

REFERENCES ARROWSMITH, J. R. AND ZIELKE, O., 2009, Tectonic geomorphology of the San Andreas Fault zone from high resolution topography: An example from the Cholame Segment Source: Geomorphology, Vol. 113, Nos. 1–2, pp. 70–81. BIOTTO, G.; SILVESTRI, S.; GOBBO, L.; FURLAN, E.; VALENTI S.; AND ROSSELLI, R., 2009, GIS, multi-criteria and multi-factor spatial analysis for the probability assessment of the existence of illegal landfills: International Journal Geographical Information Science, Vol. 23, No. 10, pp. 1233–1244. BISSON, M.; FORNACIAI, A.; BEHNCKE, B.; AND NERI, M., 2009, LiDAR-based digital terrain analysis of an area exposed to the risk of lava flow invasion: The Zafferana Etnea territory, Mt. Etna (Italy): Natural Hazards, Vol. 50, No. 2, pp. 321–334. CARLSON, E.; DE WET, A.; CAMPBELL, S.; AND WILLIAMS, A., 2000, Geophysics used to investigate the structure of a landfill: Abstracts Programs—Geological Society America, Vol. 32, No. 1. CENTER FOR SUSTAINABLE SYSTEMS, UNIVERSITY OF MICHIGAN, 2014, Municipal Solid Waste Factsheet: Publication No. CSS04-15, 2 p. COLTEN, C. E. AND SKINNER, P. N., 1996, The Road to Love Canal: Managing Industrial waste Before the EPA: University of Texas Press, Austin, TX. 217 p. CUMMINS, R. L., 1968, Effects of Land Disposal of Solid Wastes on Water Quality, Cincinnati, Ohio: U.S. Department of Health, Education and Welfare, SW-2ts, 29 p. DEPARTMENT OF THE ENVIRONMENT, 1990, Landfilling Wastes, Technical Memorandum for the Disposal of Wastes on Landfill Sites: Waste Management Paper 26 (HMSO, London). DE WET, A. P.; RICHARDSON, J.; AND OLYMPIA, C., 1998, Interactions of land-use history and current ecology in a recovering “urban wildland”: Urban Ecosystems, Vol. 2, pp. 237–262. DE WET, A. P. AND STERNBERG, R. S., 1999, Geophysical investigation of an abandoned landfill: Proceedings SAGEEP, March 1999, pp. 143–152. DE WET, A. P.; STERNBERG, R.; AND WINICK, J., 1999, Interpreting land-use history by integrating aerial photographs, near-surface geophysics, and field observations into a digital database: Environmental Engineering Geoscience, Vol. 5, No. 2, pp. 235–254. DE WET, A. P.; WEAVER, S.; AND ACHEAMPONG, S., 2000, Hydrology, geochemistry and geophysics of the Baker Woodlands Environmental Research Site, Lancaster, PA: Proceedings 13th KECK Geology Consortium Symposium, pp. 245–249. DRAKE, J. B.; DUBAYAH, R. O.; HOFTON, M.; KNOX, R. G.; CLARK, D. B.; CONDIT, R.; AND BLAIR, J. B., 2003, Above-ground biomass estimation in closed canopy neotropical forests using LIDAR remote sensing: Factors affecting the generality of relationships: Global Ecology Biogeography Letters, Vol. 12, No. 2, pp. 147–159. ENVIRONMENTAL PROTECTION AGENCY (EPA), 2008, Framework for Investigating Asbestos-Contaminated Superfund Sites: OSWER Directive 9200.0-68, September 2008, 71 p. ERB, T. L.; PHILIPSON, W. R.; TANG, W. T.; AND LIANG, T., 1981, Analysis of landfills with historic air photos: Photogrammetric Engineering Remote Sensing, Vol. 47, pp. 1363–1369. GREENBERG, M. R. AND HOLLANDER, J., 2006, The Environmental Protection Agency’s Brownfields Pilot Program: American Journal Public Health, Vol. 96, No. 2, pp. 277–281. HANEBERG, W. C.; COLE, W. F.; AND KASALI, G., 2009, Highresolution LIDAR-based landslide hazard mapping and modeling, UCSF Parnassus Campus, San Francisco, USA: Bulletin Engineering Geology Environment, Vol. 68, No. 2, pp. 263–276. HERNDON, R. C.; MOERLINS, J. E.; LAMBOU, V. W.; AND GEBHARD, R. L., 1990, Proximity of Pennsylvania Sanitary Landfills to

Wetlands and Deepwater Habitats: EPA/600/S4-89/047 May 1990, 4 p. HODGSON, M. E.; JENSEN, J. R.; TULLIS, J. A.; RIORDAN, K. D.; AND ARCHER, C. M., 2003, Synergistic use of LIDAR and color aerial photography for mapping urban parcel imperviousness: Photogrammetric Engineering Remote Sensing, Vol. 69, No. 9, pp. 973–980. HORNING, R. A., II, 1992, The Lancaster Brick Company 1919– 1979: Journal Lancaster Historical Society, Vol. 94, pp. 2–32. KENT, B. C., 1984, Susquehanna’s Indians. Anthropological Series no. 6: Pennsylvania Historical and Museum Commission, Harrisburg, PA. LANCASTER COUNTY COMMISSIONERS, 1973, County Park Leachate Study: Huth Engineers, Inc., County of Lancaster, PA, 64 p. LANCASTER COUNTY BOARD OF COMMISSIONERS, 1971, Solid Waste Management Plan 1971: Lancaster County Planning Commission, County of Lancaster, PA. 132 p. LANCASTER COUNTY BOARD OF COMMISSIONERS, 1987, Lancaster County Solid Waste Management Plan 1986: Lancaster County Planning Commission, County of Lancaster, PA. 206 p. LANCASTER FARMLAND TRUST, 2015, Lancaster Farmland Trust: Electronic document, available at http://www.lancasterfarmlandtrust. org/heritage/farming-lancaster.html LEFSKY, M. A.; COHEN, W. B.; HARDING, D. J.; PARKER, G. G.; ACKER, S. A.; AND GOWER, S. T., 2002, Lidar remote sensing of above-ground biomass in three biomes: Global Ecology Biogeography Letters, Vol. 11. No. 5, pp. 393–399. LYON, J. G., 1987, Use of maps, aerial photographs, and other remote sensor data for practical evaluations of hazardous waste sites: Photogrammetric Engineering Remote Sensing, Vol. 53, pp. 515–519. LYON, J. G. AND GREENE, R. G., 1992, Use of aerial photographs to measure the historical areal extent of Lake Erie coastal wetlands: Photogrammetric Engineering Remote Sensing, Vol. 58, pp. 1355–1360. MARK, R., 1992, Multidirectional, Oblique-Weighted, Shaded-Relief Image of the Island of Hawaii: U.S. Geological Survey, OF-92422, 5 p. MARSH, A. H. AND GARNHAM, A., 1996, Investigation, hazard assessment and remediation of existing landfills: Geological Society, London, Engineering Geology Special Publications, Vol. 11, pp. 3–7. MEISLER, H. AND BECHER, A. E., 1971, Hydrogeology of the Carbonate Rocks of the Lancaster 15-Minute Quadrangle, Southeastern Pennsylvania, Groundwater Report W26: Pennsylvania Geological Survey, Harrisburg, PA. METCALFE, R. AND ROCHELLE, C. (Editors), 1999, Chemical containment of waste in the geosphere: Geological Society, London, Special Publications, Vol. 157, pp. 7–16. MITASOVA, H.; DRAKE, T. G.; BERNSTEIN, D.; AND HARMON, R. S., 2004, Quantifying rapid changes in coastal topography using modern mapping techniques and geographic information system: Environmental Engineering Geoscience, Vol. 10, No. 1, pp. 1–11. MYERS, M. D., 2009, USGS: Science at the intersection of land and ocean: Sea Technology, Vol. 50, No. 1, pp. 18–21. NEELZ, S.; PENDER, G.; VILLANUEVA, I.; WRIGHT, N. G.; WILSON, M.; BATES, P.; MASON, D.; AND WHITLOW, C., 2006, Using remotely sensed data to support flood modeling: Proceedings Institution Civil Engineers: Water Management, Vol. 159, No. 1, pp. 35–43. POPE, P.; VAN EECKHOUT, E.; AND ROFER, C., 1996, Waste site characterization through digital analysis of historical aerial photographs: Photogrammetric Engineering Remote Sensing, Vol. 62, No. 12, pp. 1387–1394.

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 113–130

129

de Wet ROBERTSON, V. W.; WHITMAN, D.; ZHANG, K.; AND LEATHERMAN, S. P., 2004, Mapping shoreline position using airborne laser altimetry: Journal Coastal Research, Vol. 20, No. 3, pp. 884–892. RUFIN-SOLER, C.; GARDEL, A.; AND HEQUETTE, A., 2008, Assessing the vulnerability of coastal lowlands to marine flooding using LiDAR data, Sangatte coastal dunes, northern France: Zeitschrift Geomorphologie, Vol. 52, No. 3, pp. 195–211. RUTTER, J., 2008, Gunk fouls trail at park. Lancaster Newspapers, June 15, 2008: Electronic document, available at http://Lancaster online.com SILVESTRI, S. AND OMRI, M., 2009, A method for the remote sensing identification of uncontrolled landfills: Formulation and validation: International Journal Remote Sensing, Vol. 29, No. 4, pp. 975–989. SUFLITA, J. M.; GERBA, C. P.; HAM, R. K.; PALMISANO, A. C.; RATHJE, W. L.; AND ROBINSON, J. A., 1992, The world’s largest landfill—A multidisciplinary investigation: Environmental Science Technology, Vol. 26, No. 8, pp. 1486–1495.

130

TRALLI, D. M.; BLOM, R. G.; ZLOTNICKI, V.; DONNELLAN, A.; AND EVANS, D. L., 2009, Satellite remote sensing of earthquake, volcano, flood, landslide and coastal inundation hazards: ISPRS Journal Photogrammetry Remote Sensing, Vol. 59, No. 4, pp. 185–198. WALLACE, P. A. W., 1989, Indians in Pennsylvania, Anthropological Series no. 5: Pennsylvania Historical and Museum Commission, Harrisburg, PA. WILSON, M. D.; DE WET, A. P.; AND DAVIES, G. M., 2006, Wetland sedimentary record as an indicator of the land use at a former industrial site: Geological Society America Abstracts Programs, Vol. 38, No. 2, p. 80. WILSON, W. H., 1985, The Columbia-Philadelphia Railroad and its Successor: American Canal and Transportation Center, York, PA. 72 p. ZHOU, G. AND XIE, M., 2009, Coastal 3-D morphological change analysis using LiDAR series data: A case study of Assateague Island National Seashore: Journal Coastal Research, Vol. 25, No. 2, pp. 435–447.

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 113–130

Identification of Wall Tension Fractures Caused by Earthquakes, Blasting, and Pile Driving JEFFREY A. JOHNSON1 3830 Valley Centre Drive, No. 705-804, San Diego, CA 92130

ALAN “BOB” MUTCHNICK GMU Geotechnical, Inc., 23241 Arroyo Vista, Rancho Santa Margarita, CA 92688 bmutchnick@gmugeo.com

Key Terms: Structure Survey, Crack Mapping, Site Investigations

ABSTRACT Determining whether an earthquake, blasting, or pile driving caused non-structural “cosmetic” wall cracks or extended pre-existing fractures is often based on judgment: The structure presumably vibrated; therefore, new cracks formed, and existing fractures were extended. The proposed two-part, primarily post-event, fracture mapping and analysis of intensity data are designed to test the static, as opposed to vibratory, cause of wall fractures, assuming: (1) tension fractures form perpendicular to the principal tensile stress (σ3) direction; (2) during structure vibrations, the local stress field rotates relative to the static stress field; (3) stress reversals, common during earthquakes, can cause tension fractures at all corners of rectangular doors and windows; (4) progressively younger en echelon tension fractures, originating at the corner of a rectangular wall opening, trend toward horizontal due to the transient rotation of the principal stresses, whereas fractures from non-vibrational causes trend toward vertical; (5) crack extension occurs when two fractures are linked by hook-shaped fractures; and (6) site-specific intensity observations and local intensity data provide qualitative data regarding the timing and amplitude of wall strains and the potential for co-seismic damage. Co-seismic tension fractures can therefore be differentiated from pre-existing cracks, suggesting pre-blasting and piledriving crack mapping, building condition surveys, and co-seismic ground and structure motion recordings may not be required: (1) if the building has no prior exposure to potentially damaging vibrations; and (2) a reconnaissance inspection indicates there are no preexisting horizontal or sub-horizontal wall fractures.

Corresponding author phone: 858-243-4438; email: jallenjohn@aol.com

INTRODUCTION The purpose of this study is to present the criteria and tests that form the basis of a two-part analysis that can be used to distinguish between wall cracks caused by vibrations from small or distant earthquakes, pile driving, blasting events, and other construction activities such as demolition and dynamic ground improvements, from pre-existing fractures caused by non-seismic processes (Oriard, 1999; Audell, 2004). Part one consists of a reconnaissance structure condition survey and an exterior site review followed by a post-event fracture analysis. Although important, the proposed part one analysis is not dependent on knowledge of the cause(s) of pre-existing fractures and data from co-seismic ground motion monitoring, provided the structure(s) has not experienced potentially damaging vibrations, and sub-horizontal to horizontal wall fractures are not observed during the pre-event reconnaissance inspection. The post-event part one data collection is limited to the orientation of tension fractures, angular changes between successive en echelon tension fractures, and evidence of fracture linkage. Part two includes collecting and documenting site-specific intensity observations and regional intensity data obtained from the U.S. Geological Survey (USGS) or the California Geological Survey (CGS). The collection of intensity data (e.g., the degree of shaking at a specified place [Richter, 1958]) is important because the data support the timing and the relative degree of shaking and therefore the opportunity for co-seismic wall tension fracture formation and the extension of pre-existing cracks. Emphasis is placed on wall tension fractures to determine vibratory damage because: (1) they form perpendicular to the principal tensile stress (σ3), thereby giving an indication of the local stress field in the wall at the time of their formation (Scholz, 2002; Gudmundsson, 2011); (2) during three-dimensional (3D) structure vibrations, σ3 is generally not horizontal (Gudmundsson, 2011); (3) tension fractures often

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 131–139

131

Johnson and Mutchnick

(2004) discuss non-seismic crack-formation processes in walls covered with lath and plaster, drywall, and stucco. BACKGROUND Fractures

Figure 1. Exterior stucco fractures, most likely caused by differential settlement, observed during a post-pile-driving site inspection. (a) Approximately six tension cracks linked by hookshaped fractures. Tension crack 1 is the oldest. (b) Angular relationship between progressively younger fractures trending toward vertical indicates fractures caused by a non-seismic process.

form when shear fractures do not because wall coverings, such as lath and plaster, drywall, and stucco, are weak in tension compared to their shear and compression strengths; and (4) it is unlikely cracking of floor slabs, foundations, driveways, sidewalks, patios, and pool decks that are in contact with the ground can be used to demonstrate vibratory damages because co-seismic damages of this type are “inconsequential, except perhaps in the near fault region” (Bolt et al., 2004), where vibratory damage is identifiable. The reader is referred to Griffith (1924), Scholz (2002), and Gudmundsson (2011) for a general review of fracture mechanics. Other methods that are used to distinguish between pre-existing cracks and fractures caused by mine, quarry, and construction blasting and pile driving have been documented since the 1920s and are discussed by Siskind et al. (1980), Stagg et al. (1984), the Transportation Research Board (1997), and Oriard (1999). Oriard (1999) and Audell

132

For the purposes of this analysis, it was assumed a wall includes two of the three mutually perpendicular principal compressive stresses, σ1 and σ3, where σ1 . σ3. For a vertical wall at rest, σ1 is approximately vertical, and σ3 is approximately horizontal. The assumption is consistent with Anderson’s (1951) theory of faulting, where the three principal stresses are arranged so that two are parallel and one is perpendicular to a relatively level ground surface. For purposes of this analysis, σ2 can be ignored, and the type (i.e., tension or shear) and orientation of a fracture(s) are dependent on the static or seismic orientation of principal stresses σ1 and σ3. A crack or fracture is a discontinuity with no tensile strength. There are two types of fractures, extension, which is the result of tensile stress, and shear, which is the result of either shear or compressive stress. Only tension fractures (i.e., an extension fracture that forms when σ3 is negative; Gudmundsson, 2011) are relevant to this discussion. Walls are not homogeneous or isotropic structures. However, in a vertical wall that lacks stress concentrations, such as doors and windows, σ1 is approximately vertical, and σ3 is approximately horizontal, and tension fractures will form perpendicular to σ3. Nonvertical tension fractures indicate rotation of the principal stresses or local stress field in the wall potentially due to either permanent or seasonal differential foundation movement or transient vibratory ground motions. Tension fractures are therefore important diagnostic discontinuities because they indicate the local stress field at the time of their formation. Discontinuities such as wall openings, often referred to as stress raisers, increase the stress at corners of the openings, alter the orientation of the stress trajectories or principal stress axes, in the adjoining wall, and therefore change the direction of fracturing if sufficient additional stresses are added from vibrational and non-vibrational causes. Static tension fractures, nearest to the corners of doors and windows, are therefore, although not always (Audell, 2004), inclined rather than vertical (Figures 1a and 2a). However, as distance from a stress raiser increases, successively formed tension fractures will trend toward vertical (Figures 1b and 2b). An integral part of conducting post-event crack mapping is determining if pre-existing fractures were

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 131–139

Identification of Wall Tension Fractures

Figure 2. Interior drywall fractures, most likely caused by differential settlement, observed during a post-blasting site inspection. (a) Approximately four tension cracks linked by hook-shaped fractures. Tension crack 1 is the oldest. (b) Angular relationship between progressively younger fractures trending toward vertical indicates fractures caused by a non-seismic process.

extended or increased in length. Crack extension is herein defined as the formation of a new fracture, roughly in the same linear alignment but not connected, that eventually links by hook-shaped tension fractures

with the older fracture (Figures 1a and 2a). Hookshaped tension fractures are characteristic of the linkage process and are caused by curved stress trajectories between the offset and initially unconnected fractures. Once hook-shaped fractures are evident, fracture extension should be assumed because the fractures are most likely acting as a single, hard linked structure. For example, fractures 3 and 4 (Figure 1a) exhibit hookshaped fractures without a clear visual connection. It is unlikely fractures 5 and 6 would have formed without a hard link between fractures 3 and 4. Stress reversals during an earthquake cause sideto-side, as well as up-and-down structure vibrations and can induce “X” wall tension cracks (Figure 3). “X” cracks are important because they are “characteristic of co-seismic structural damage” (Stienbrugge, 1970) and demonstrate σ3 is generally not horizontal while the structure is vibrating. Small or distant earthquakes, blasting, and pile driving generally will not result in structural damage such as “X” wall cracks, but they can cause cracking at and near the corners of rectangular doors and windows (Figure 4). One indication of damaging vibrations and stress reversals is a set of tension fractures that occur at all corners of a rectangular wall opening (Figures 3c and 4). In addition, fractures that extend due to seismic shaking will trend toward the horizontal (Figure 5a and 5b). Overprinting, a type of crack extension, occurs when en echelon fractures are caused by two or more processes (Audell, 2004). The residence shown on Figure 6a was inspected and photographed after a number of quarry blasts. The results of the inspection suggested fractures 1 to 4 were not caused by the blasting, whereas the fracture number 5 was. Additional fractures that are also most likely the result of blasting are highlighted in white, to improve visibility (Figure 6b). A second example of overprinting was observed on a wall of a residence in Los Angeles, CA, affected by a landslide and not seismic shaking (Figure 7). The slope supporting the residence failed and accelerated during the winter rains of 1993, most likely causing fractures 3 and 4. Fractures 1 and 2 were approximately vertical and located at only one corner of the window (Figure 7b), suggesting they were the result of a non-seismic process. Intensity Intensity is an observational measure of the co-seismic damage to a structure and the degree of shaking at the site (Richter, 1958). The first descriptive intensity scale intended for general use was the Rossi-Forel scale (De Rossi, 1883). In 1902, Mercalli (1902) introduced a scale that reduced some of the problems of

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 131–139

133

Johnson and Mutchnick

Figure 4. (a and c) Interior drywall fractures at both corners of a door observed during a post-blasting site inspection. (b) Overprinting suspected based on increase in fracture angles for one set of linked cracks (30u to 56u) and ,28u angle for isolated fractures with no clear evidence of linkage. (d) Overprinting suspected based on similar fracture angles (i.e., 25u shown in d and 28u to 30u on b), asymmetric number of fractures, reduction in fracture angles from 15u to approximately horizontal, and no clear evidence of linkage. Lack of evidence of linkage between suspected seismic fractures suggests cracking due to one or possibly two blasts.

Figure 3. “X” wall fractures. (a) Five-story structure in the city of Managua after the M 6.2 1972 Managua, Nicaragua, earthquake. (b) Low-rise building, San Fernando Valley, CA, after M 6.7 1994 Northridge, CA, earthquake. (c) Fractures on four corners of wall opening, M 6.7 1994 Northridge, CA, earthquake.

assigning Rossi-Forel intensities. In 1931, Wood and Neumann (1931) proposed the modified Mercalli intensity (MMI) scale and an abridged version designed for use by observers with varied experience levels (Neumann, 1954). MMI is a progressive scale ranging from not felt (MMI I) to total damage (MMI XII). MMI is important because of its extensive field testing and updating over the past 80+ years, and it includes the “felt” intensities suitable for supporting or eliminating the potential for co-seismic fracture formation and extension. The 1931 MMI scale, with some modifications, is still in use today (Stover and Coffman, 1993; Dewey et al., 1995). Selected wording from the 1931 unabridged MMI scale for estimating MMI IV, V, and VI earthquake

134

intensities, which includes the formation of fractures, is listed in Table 1. For comparison, criteria used to estimate MMI following the 1994 Northridge, CA, earthquake (Dewey et al., 1995) are also listed. An important difference between the unabridged 1931 wording describing MMI V and that used by Dewey et al. (1995) was the addition of “hairline cracks in interior walls.” In the 1931 abridged version, “a few instances of cracked plaster” was included in MMI V (Wood and Neumann, 1931, pp. 279–280). Dewey et al. (1995) did not specify the type of wall covering, even though the results of blasting studies indicate plaster is more prone to cracking than drywall or stucco. In the meizoseismal region, where structural damage is common, MMI is determined primarily by field observations. At greater distances, MMI was, until recently, based in part on questionnaires mailed to post offices and other government facilities by the U.S. Coast and Geodetic Survey and more recently by the USGS. The USGS appears to have started the practice following the 1886 Charleston, SC, earthquake (Talwani, 2014). Stover and Coffman (1993) amended MMI to account for post-1931 field experience, recent modifications to the USGS mailed questionnaires, and the poor reliability of the subjective effects on people in assigning MMI. The U.S. Bureau of Mines (Siskind et al., 1980, 1993) and the Transportation Research Board (1997) also concluded that human perception

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 131–139

Identification of Wall Tension Fractures

Figure 5. (a) Interior wall fractures extending from top-right corner of a door. Un-numbered en echelon tension fractures are most likely the result of differential settlement. Numbered and linked fractures are likely due to vibrations from a blast. (b) Trend toward horizontal of numbered fractures shown on Figure 5a supports origin due to blast vibrations.

of vibrations is not an accurate gauge of the damage potential of the vibrations. Dengler and Dewey (1998) conducted a telephone survey following the M 6.7 1994 Northridge, CA,

earthquake, obtaining building-specific intensity information from a relatively large number of locations compared to responses obtained from the mailers. Results of their survey were compared with the

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 131â&#x20AC;&#x201C;139

135

Johnson and Mutchnick

Figure 6. Fractures 1, 2, 3, and 4 are most likely the result of differential settlement. Fracture 5 (a) and highlighted fractures (b) are relatively low angle and unlinked, suggesting they formed due to structure vibrations resulting from possibly one blast that contained asymmetric stress reversals (evidence of fractures due to structure vibrations was not observed at the opposite corner of the door).

USGS MMI data, and an algorithm to estimate what was called the “community decimal intensity” (CDI) was developed. CDI estimates are similar but are not identical to MMI (Dengler and Dewey, 1998). The USGS adopted the Dengler and Dewey (1998) CDI algorithm and a form of their telephone questionnaire for use on the Internet (Wald et al, 1999a, 1999c). The program that became known as “Did you feel it?” (DYFI) has three main sections, including the two-part earthquake effects section. Part one includes observations related to sounds, the performance of exterior walls and fences, and the movement of objects within the structure. Part two includes observed building damage from “hairline cracks in walls” to “building permanently shifted over foundation.” According to the USGS web page, as the data are collected, CDI values are rounded to integers, average values are computed for individual ZIP codes, and then those values are converted to MMI

136

Figure 7. Linked and relatively low-angle fractures 3 and 4 were the result of accelerated movement of a 1993 landslide and not structure vibrations. Seismic exposure occurred prior to the construction of the residence and a year later during the 1994 Northridge, CA, earthquake. Fractures 1 and 2 may have been the result of differential settlement. It is unclear if the differential settlement was related to the slope movement.

intensities. CDI data complement site-specific MMI observations and are of particular importance if sitespecific data do not exist or cannot be obtained. ShakeMaps, posted online by the USGS and the CGS, are a valuable, rapidly available source of ground motion and instrumental intensity data (Wald et al., 1999b, 1999c). Instrumental intensity ShakeMaps are based on correlations of MMI with peak ground motion parameters (Wald et al., 1999b; Caprio et al., 2015) and can also be used to supplement site-specific MMI data. Siskind et al. (1980) published a correlation between blast recordings and damage to structures. Blasts generally contain a high concentration of energy that can cause damage to nearby structures. In addition, air blasts (air overpressures), also caused by blasting, can fracture windows and crack walls. Table 1 contains a portion of the Siskind et al. (1980) damage classification system. Although numerous studies have pointed out blasts are not earthquakes, our preliminary

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 131–139

Identification of Wall Tension Fractures Table 1. Comparison of modified Mercalli intensity and the blasting damage classification. Rating

Modified Mercalli Intensity

Wood and Neumann (1931) Creaking of walls, frame Rattling of dishes, windows, doors

Swinging of hanging objects Cracked windows (generally not) Overturned small or unstable objects Moved small objects, furnishings Swinging of hanging objects, doors Broke dishes to some extent

A few instances of cracked plaster Some windows broke out Fall of plaster in small amount Cracked plaster, especially fine cracks chimneys

Blasting Damage Classification

Dewey et al. (1995) Walls creaked loudly Buildings shaken moderately to strongly

A few windows cracked A few small objects overturned and fallen Hanging pictures tilted, out of place, or fallen

Hairline cracks in interior walls Some windows broken out A few instances of fallen plaster or damaged

un-reinforced masonry chimneys Overturned furniture Moved moderately heavy furnishings Fall of knick-knacks, books, pictures Broke dishes in considerable quantity

Siskind et al. (1980) Threshold Loosening of paint; small plaster cracks at joints between construction elements

Lengthening of old cracks Minor Hairline to 3 mm cracks Loosening and falling of plaster Fall of loose mortar Cracks in masonry around openings near partitions

Light furniture overturned Moderately heavy furniture displaced Many small objects overturned and fallen Many glassware items or dishes broken Large cracks in interior walls

analysis suggests “threshold” damage ranges from MMI IV to V, and “minor” damage ranges from MMI V to VI. The Transportation Research Board (1997) concluded vibrations due to pile installation could cause wall cracking at distances generally limited to the length of the pile. In extreme cases, differential settlement and resultant foundation and building damage, due to either compaction or consolidation of sandy sediments, have been recorded at distances of 400 m (Transportation Research Board, 1997). ANALYSIS The objective of the proposed two-part analysis is to determine if a small or distant earthquake, blasting, or pile driving caused or extended non-structural “cosmetic” wall fractures. Part one consists of pre-event data collection, which includes documenting the history of exposure to potentially damaging earthquakes and other sources of ground vibrations and, in the case of blasting and pile driving, conducting a reconnaissance structure(s) inspection to determine if existing wall fractures include horizontal or approximately horizontal tension cracks and/or fractures at all corners of at least one rectangular wall opening. If the answer is yes to either of these, detailed pre-event crack mapping or a structure condition survey should be considered. Post-event part-one analysis consists of conducting a detailed analysis of wall fractures

focusing on stress raisers to determine if: (1) there are tension fractures at or near all corners of one or more of the rectangular wall openings; (2) as the distance from a wall opening increases, successive tension fractures trend toward horizontal; and (3) there is evidence of hook-shaped linkage fractures between successive horizontal-trending tension fractures. If the answer is no to all of the above, the probability is low that an earthquake, pile driving, or blasting caused tension fractures or extended existing cracks. Part-two site-specific intensity data should be collected as soon after the event as possible from occupants or by inspection. Six questions, based on Table 1, for estimating a site-specific MMI are listed in Table 2. For example, was wall creaking noticed, did wall hangings move or fall, did any items fall from shelves, and were any windows cracked? Although wall creaking by itself (MMI IV) is not sufficient to determine whether or not co-seismic wall cracks developed, the absence of creaking would likely preclude fracture formation. Minimal wall movement is also suggested if hanging pictures and mirrors were undisturbed and items did not fall from shelves or furniture. A cracked window (MMI V) supports fracture formation at the corners of the window and potentially at other wall openings including doors. A word of caution: Humans can detect vibrations well below levels that can cause damage (Richter, 1958; Siskind et al., 1980, 1993), and the USGS, the Transportation Research Board (1997), and Oriard

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 131–139

137

Johnson and Mutchnick Table 2. A list of questions for estimating the potential for fracture formation and extension. MMI Did the walls creak loudly? Did you hear rattling of dishes, windows, or doors or see hanging objects swing? Were any small objects overturned or did you see doors swing? Any hanging pictures tilted, out of place, or fallen? Any cracked windows? Any light furniture overturned or heavy furniture displaced?

IV IV V V V VI

(1999) concluded that human perception of vibrations is not an accurate gauge of the damage potential of the vibrations. However, the implied or expressed correlation between the human perception of vibrations and building damage suggests a potential need for public outreach and pre-event structure condition surveys independent of the technical need (Oriard, 1999).

CONCLUSIONS The type of data collected during the proposed twopart analysis can be used to test if non-structural wall fractures were caused by seismic shaking, potentially eliminating the need for detailed pre-blasting and pile driving crack mapping and instrumental recordings of ground or structure motions if, during part one, it is determined the subject structure(s) has not been exposed to seismic shaking and if non-seismic processes have not caused horizontal or sub-horizontal wall tension fractures. The part-one post-event crack analysis focuses on evidence of: (1) tension fractures at all corners of rectangular wall openings; (2) successive tension fractures trending toward the horizontal as distance from the wall opening increases; and (3) fracture extension or hook-shaped tension fractures between successive tension fractures. Part-two site-specific intensity observation questions (Table 2) are designed to determine if potentially damaging structure vibrations equaled or exceeded MMI IV (Table 1). It is unlikely for vibratory tension fractures to have formed or existing fractures to have been extended if MMI was #IV. ACKNOWLEDGMENTS We would like to acknowledge and thank the three reviewers, who provided valuable insight and constructive suggestions and comments.

138

REFERENCES ANDERSON, E. M., 1951, The Dynamics of Faulting and Dyke Formation with Applications to Britain, 2nd ed.: Oliver and Boyd, Edinburgh, U.K. AUDELL, H. S., 2004, Field Guide to Crack Patterns in Buildings, a Guide to Residential Building Cracks Caused by Geologic Hazards: Special Publication No. 16, Association of Engineering Geologists, Denver, CO. Southern California Section, 71 p. BOLT, B., SOMERVILLE, P., ABRAHAMSON, N., ZERVA, A., 2004, Workshop Proceedings: Effects of Earthquake-induced transient ground surface deformation on at-grade improvements, CUREE Publication No. EDA-04, 41 p. CAPRIO, M.; TARIGAN, B.; WORDEN, C. B.; WIEMER, S.; AND WALD, D. J., 2015, Ground motion to intensity conversion equations (GMICEs): A global relationship and evaluation of regional dependency: Bulletin Seismological Society America, Vol. 105, No. 3, pp. 1476–1490. DENGLER, L. A. AND DEWEY, J. W., 1998, An intensity survey of households affected by the Northridge, California, earthquake of 17 January, 1994: Bulletin Seismological Society America, Vol. 88, No. 2, pp. 441–462. DE ROSSI, M. S., 1883, Programma dell'osservatorio ed archive centrale geodinamico: Bollettino del Vulcanismo Italiano, Vol. 10, pp. 3–124. DEWEY, J. W.; REAGOR, B. G.; DENGLER, L.; AND MOLEY, K., 1995, Intensity Distribution and Isoseismal Maps for the Northridge, California, Earthquake of January 17, 1994: U.S. Geological Survey Open-File Report 95-92, 35 p. GRIFFITH, A. A., 1924, Theory of rupture. In Biezeno, C. B. and Burgers, J. M. (Editors), Proceedings of the First International Congress on Applied Mechanics: Delft, Netherlands, Waltman, pp. 55–63. GUDMUNDSSON, A., 2011, Rock Fractures in Geological Processes: Cambridge University Press, Cambridge, U.K., 578 p. MERCALLI, G., 1902, Sulle modificazioni proposte alla scala sismica De Rossi-Forel: Bollettino della Società Sismologica Italiana, Vol. 8, pp. 184–191. NEUMANN, F., 1954, Earthquake Intensity and Related Ground Motion: University of Washington Press, Seattle, WA, 77 p. ORIARD, L. L., 1999, The Effects of Vibrations and Environmental Forces, a guide for the investigation of structures: International Society of Explosive Engineers, Solon, OH, 284 p. POLLARD, D. D. AND FLETCHER, R. C., 2006, Fundamentals of Structural Geology: Cambridge University Press, Cambridge, U.K., 500 p. RICHTER, C. F., 1958, Elementary Seismology: W. H. Freeman and Company, San Francisco, CA, 768 p. SCHOLZ, C. H., 2002, The Mechanics of Earthquakes and Faulting, 2nd ed.: Cambridge University Press, Cambridge, U.K., 404 p. SISKIND, D. E.; CRUM, S. V.; AND PLIS, M. N., 1993, Blast Vibrations and Other Potential Cause of Damage in Homes near a Large Surface Coal Mine in Indiana: U.S. Bureau of Mines Report of Investigations 9455. SISKIND, D. E.; STAGG, M. S.; KOPP, J. W.; AND DOWDING, C. H., 1980, Structure Response and Damage Produced by Ground Vibration from Surface Mine Blasting: U.S. Bureau of Mines Report of Investigations 8507. STAGG, M. S.; SISKIND, D. E.; STEVENS, M. G.; AND DOWDING, C. H., 1984, Effects of Repeated Blasting on a Wood-Frame House: U. S. Bureau of Mines Report of Investigations 8896, 82 p. STIENBRUGGE, K. V., 1970, Earthquake damage structural performance in the United States. In Wiegel, R. L. (Editor), Earthquake Engineering: Prentice-Hall, Inc., Ch. 9, Englewood Cliffs, NJ, pp. 167–226.

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 131–139

Identification of Wall Tension Fractures STOVER, C. W. AND COFFMAN, J. L., 1993, Seismicity of the United States, 1568–1989 (Revised): U.S. Geological Survey Professional Paper 1527, 418 p. TALWANI, P., 2014, The impact of the early studies following the 1886 Charleston earthquake on the nascent science of seismology: Seismological Research Letters, Vol. 85, No. 6, pp. 1366–1372. TRANSPORTATION RESEARCH BOARD, 1997, Dynamic Effects of Pile Installations on Adjacent Structures: National Cooperative Highway Research Program Synthesis 253. WALD, D. J.; DENGLER, Q. L.; AND DEWEY, J., 1999a, Utilization of the Internet for rapid community intensity maps: Seismological Research Letters, Vol. 70, pp. 680–697.

WALD, D. J.; QUITORIANO, V., HEATON, T. H.; AND KANAMORI, H., 1999b, Relationships between peak ground acceleration, peak ground velocity and modified Mercalli intensity in California: Earthquake Spectra, Vol. 15, No. 3, pp. 557–564. WALD, D. J.; QUITORIANO, V.; HEATON, T. H.; KANAMORI, H.; SCRIVNER, C. W.; AND WORDEN, C. B., 1999c, TriNet “ShakeMaps”; rapid generation of peak ground motion and intensity maps for earthquakes in Southern California: Earthquake Spectra, Vol. 15, No. 3, pp. 537–555. WOOD, H. O. AND NEUMANN, F., 1931, Modified Mercalli intensity scale of 1931: Bulletin Seismological Society America, Vol. 21, pp. 277–283.

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 131–139

139

Geologic and Geotechnical Factors Controlling Incipient Slope Instability at a Gravel Quarry, Livermore Basin, California PHILIP L. JOHNSON1 PATRICK O. SHIRES TIMOTHY P. SNEDDON Cotton, Shires and Associates, 330 Village Lane, Los Gatos, CA 95030

Key Terms: Slope Instability, Gravel Quarry, Landslide, Stratigraphy

ABSTRACT Mine pit slopes at Arroyo del Valle Quarry in the Livermore Basin of northern California expose late Quaternary sandy gravel. In the deep subsurface, the gravel unconformably overlies gently folded lacustrine sediments. Within the lacustrine sediments, a bed of sheared, unoxidized clay overlies a marl bed, forming a distinctive marker bed couplet. Structure contours on this marker bed show an anticline and a syncline in the vicinity of the quarry. These northwest-striking folds are aligned parallel to regional Quaternary fold and thrust belt structures. Where the marker bed dips toward the pit, slope inclinometers consistently deflected toward the pit at the depth of the unoxi‐ dized clay. Where the marker bed dips away from the quarry pit, no slope inclinometer deflections were recorded. Thus, slope instability was controlled by the site stratigraphy, geologic structure, and the location of quarry slopes relative to that geologic structure. High pore-water pressures within the unoxidized clay also contributed to slope instability. Shearing of the unoxidized clay occurred prior to excavation of the quarry pit, and the resulting low residual strength of this high-plasticity clay made it particularly vulner‐ able to incipient landsliding when lateral confinement was removed during excavation of the quarry pit. Analysis of the critical region between the quarry pit and the anticline axis showed that the static factor of safety remained below 1.5. Seismic displacement analyses indicated that moderate to large displace‐ ments would be anticipated. Thus, depressurization wells and an earth-fill buttress were designed and implemented to mitigate deep-seated slope instability.

Corresponding author email: pjohnson@cottonshires.com.

INTRODUCTION Mine slope stability has been studied extensively at sites at which mine slopes expose fractured rock, and the application of rock mechanics to the study of mine slopes has aided in the design of stable slopes for long-term mine reclamation and short-term slope stability during mining operations (Hoek and Karzulovik, 2000; Wyllie and Mah, 2004). In some cases, rock slopes in open pit mines have experienced large, fast-moving failures (Pankow et al., 2014) that present significant challenges to mining operations. By contrast, the stability of gravel quarry slopes that expose Quaternary sediments has received less attention. However, as mining of aggregate resources from basins within large metropolitan areas continues and urban development encloses the mined lands, the stability of gravel quarry slopes has become a significant concern (Doughton, 2009). Modern mining methods have allowed extraction of aggregate and other resources to significant depths within sedimentary basins, and dewatering systems have allowed mining of aggregate well below the groundwater table. Thus, relatively deep quarry excavations may be found locally within populated regions. Where the mine pits expose coarse granular materials, pit slope stability should be a simple function of slope angle, slope height, and (relatively high) material strengths. However, where the geologic conditions are more complex, slope stability may be more challenging to achieve. Arroyo del Valle Quarry is a gravel quarry located in the Livermore Basin of northern California (Figure 1). Sand and gravel have been mined extensively in the Livermore Basin since the early 20th century for use as aggregate (Goldman, 1964; Dupras, 1999). The quarry is the easternmost of a series of gravel pits within the southern portion of the basin. Once mining of an individual pit is completed and dewatering systems are shut down, the pits are allowed to flood slowly by groundwater seepage, becoming artificial

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 141–155

141

Johnson, Shires, and Sneddon

Figure 1. Location map for the Livermore Basin and surrounding uplifts.

lakes. Suburban residential development within the Livermore Basin has grown extensively since the 1980s and has approached the margins of the former gravel pits. At Arroyo del Valle Quarry, residential development extends up to the northeast boundary of the quarry property, currently within approximately 125 ft of the quarry pit. The Arroyo del Valle Quarry is located within a geologically complex setting that may appear to be relatively simple at the surface and within the shallow subsurface. The mine pit slopes expose dense sandy gravel that appears to remain stable at slope inclinations of 2H:1V (26.6u) or greater. However, slope inclinometers installed around the mine pit have recorded deflections at depth, below the gravel exposed in the quarry pit. In this article, the unique site stratigraphy, geologic structure, and hydrogeology found in the subsurface at Arroyo del Valle Quarry and the impact of this unique geology on the stability of the quarry slopes are described. In addition, analyses of static and seismic slope stability and the design of mitigation measures that were implemented to achieve long-term stability of the former quarry pit slopes are described.

GEOLOGIC SETTING The roughly east-westâ&#x20AC;&#x201C;trending late Quaternary Livermore Basin (Figure 2) is filled with non-marine clastic sediments (Barlock, 1989; Helley and Graymer, 1997). During Miocene to early Pleistocene time, the ancestral Livermore Basin (Unruh et al., 1997) extended beyond the limits of the late Quaternary basin, and upper Miocene to lower Pleistocene sedimentary rocks of the Sycamore, Tassajara, and Livermore formations are exposed in the hills that surround the late Quaternary basin (Andersen et al., 1995). The sediments that fill the late Quaternary Livermore Basin consist of Pleistocene to Holocene alluvial fan,

142

terrace, and floodplain deposits that overlie the Livermore Formation (Helley and Graymer, 1997). Early work by the California Department of Water Resources (CDWR, 1974) included characterization of the alluvial aquifers of the late Quaternary Livermore Basin. These aquifers contain a large volume of coarse-grained sediment that was derived from the Diablo Range to the south and transported by alluvial processes northward into the basin. Fine-grained lacustrine sediments form aquitards that cap the aquifer units. Ehman et al. (2004) interpreted electric logs and cuttings logs from 37 water wells in the central portion of the basin (north of Arroyo del Valle Quarry) and used sequence stratigraphic methods to characterize the aquifer stratigraphy and to identify several depositional sequences that are bounded by unconformities. Crane (1995, 2007) mapped the northern, eastern, and western boundaries of the late Quaternary basin as faults that bound the hills that surround the basin. He interpreted the northern boundary of the basin as a southwest vergent thrust fault that bounds the Mount Diablo region (Figure 2), while complex thrust and strike slip faulting characterize the hills to the east and west. Thus, the late Quaternary Livermore Basin is flanked by faulted uplifts that expose tilted and folded Miocene to lower Pleistocene rocks. Sawyer and Unruh (2004) identified a southwest vergent fold and thrust belt, the Mount Diablo Fold and Thrust Belt (MDFTB), that encompasses the Mount Diablo region and the Livermore Basin. Compressional deformation of the MDFTB has been attributed to a restraining left stepover between the right lateral Greenvillle and Concord faults (Unruh and Lettis, 1998). Unruh et al. (2007) view the Livermore Basin as deformed by active folding and southwest vergent thrust faulting behind the leading edge of the MDFTB, which extends as far southwest as the Verona and Williams faults (Figure 2). Several northwest striking and actively growing folds were identified within the basin and surrounding hills (Sawyer and Unruh, 2004). Thus, compressional tectonics of the MDFTB appear to be responsible for local deformation of Livermore Basin sediments. GEOLOGY OF THE ARROYO DEL VALLE QUARRY AREA Site Geomorphology The geomorphology of the Arroyo del Valle Quarry site and surrounding area was mapped using stereo pairs of historic aerial photographs that pre-date the excavation of the quarry (Figure 3). The geomorphic

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 141â&#x20AC;&#x201C;155

Arroyo Del Valle Quarry

Figure 2. Simpliﬁed map of the Livermore Basin and major mapped faults, modiﬁed from Wagner et al. (1990). The labeled primary stream channels within the basin are Arroyo del Valle (AV) and Arroyo Mocho (AM).

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 141–155

143

Johnson, Shires, and Sneddon

Figure 3. Photogeologic map showing the geomorphic setting of the Arroyo del Valle Quarry area, based upon interpretation of stereo pairs of historic aerial photographs. The polygon in the upper left portion of the map shows the area of Figure 7.

setting of the southern Livermore Basin was shaped by fluvial deposition along Arroyo del Valle, a low-sinuosity (braided) stream. The stream channel is flanked by uplifts to the northeast and southwest. To the northeast, a subdued intrabasinal uplift is centered on the Livermore anticline that may be associated with a blind reverse or thrust fault (Sawyer and Unruh, 2004). To the southwest, tilted beds of the Livermore Formation flank the southwest margin of the basin. Between these two uplifts, Arroyo del Valle flows toward the northwest. A series of fluvial terraces ascend from the channel of Arroyo del Valle (youngest to oldest) and up the flanking uplifts northeast and southwest of the channel (Figure 3). The Arroyo del

144

Valle Quarry study area is located within the channel and adjacent alluvial terraces. Site Stratigraphy The quarry cut slopes expose only the shallowest portion of the basin stratigraphy. As a result of the lack of surface exposure, the primary tool for characterization of the deeper stratigraphy was subsurface exploration. Thus, 39 continuously cored borings and two large-diameter bucket auger borings were drilled for this study. The core samples were logged in detail, sedimentary characteristics were described, and marker beds were identified. Correlations between borings

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 141â&#x20AC;&#x201C;155

Arroyo Del Valle Quarry

Figure 5. A ﬁne-grained bed that pinches out laterally (exposed in the quarry pit wall). Figure 4. Sandy gravel over a ﬁne-grained bed exposed in the shallow portion of the quarry pit wall.

were added to detailed geologic cross sections, and structure contour maps were compiled using marker bed elevations. Upper Sandy Gravel Deposits The quarry pit slopes expose the upper sandy gravel deposits. These poorly sorted, coarse sediments consist of rounded to subangular cobbles, pebbles, and very coarse to fine sand with local boulders and minor low-plasticity silt and clay. Where exposed in quarry pit walls, local horizontal stratification and pebble imbrication are visible. Individual beds of sandy gravel have sharp, irregular lower contacts with welldeveloped erosional relief; this is particularly noticeable where the sandy gravel overlies local fine-grained beds (Figure 4). The sandy gravel unit varies in thickness from approximately 90 to 100 ft (27–30 m). The youngest gravel deposits exposed in the area of the historic stream channel are mostly gray, while older beds are oxidized to yellowish brown (Munsell color 10YR). Interbedded with the sandy gravel are local beds of yellowish brown (10YR) clayey silt to silty clay that are generally less than 6 ft (2 m) in thickness (Figure 5). Based upon subsurface exploration, local fine-grained beds within the sandy gravel are mostly discontinuous, and where correlation of individual beds can be accomplished between boreholes, they appear to be horizontal or subhorizontal. Lower Fine-Grained Deposits In borings surrounding the quarry pit, fine-grained deposits were encountered below the upper sandy gravel deposits. The fine-grained strata consist of distinct beds of oxidized silty clay, unoxidized clay, marl, silty fine sand, and sandy silt. The oxidized

clay displays variable coloration from light yellowish brown to strong brown (Munsell colors range from 2.5Y to 7.5YR). Though mostly lacking in stratification, the oxidized clay is locally laminated. The results of Atterberg Limits testing of the oxidized clay indicate that the average liquid limit and plasticity index are 37 and 15 (respectively), which correlate with low-plasticity clay. The color and texture of the oxidized clay strongly contrast with those of the underlying unoxidized clay. The unoxidized clay is greenish gray to olive gray (5G to 5Y), with local lamination and significantly lower silt content (,10 percent silt) than the overlying oxidized clay. The results of laboratory testing of the unoxidized clay indicate that the average liquid limit and plasticity index are 75 and 48 (respectively), which places this clay in the high-plasticity range. In a few of the cores drilled northeast of the quarry pit, small pelecypod and gastropod shells and shell fragments were identified within the unoxidized clay. This stratigraphic unit displays evidence of intense shearing that includes numerous highly polished surfaces and development of clay gouge (Figure 6). The unoxidized clay overlies a white to light gray marl that is highly reactive to dilute hydrochloric acid. Below the marl is fine silty sand to sandy silt with interbedded silty clay; these strata are more permeable than the overlying unoxidized clay and marl. The sheared, unoxidized clay and underlying marl form a distinctive and laterally persistent stratigraphic couplet. This couplet was traced across the quarry site and adjacent area to the northeast and is designated as the unoxidized clay-marl marker bed couplet. Interpretation of Sedimentary Depositional Environments Upper Sandy Gravel Deposits The sandy gravel to gravelly sand encountered in the upper 100 ft (30 m) in the study area appears to

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 141–155

145

Johnson, Shires, and Sneddon

Figure 6. A core sample exposing a polished surface within the sheared unoxidized clay.

have been deposited in a braided stream environment similar to modern Arroyo del Valle. The predominantly coarse texture of this deposit is consistent with a high-velocity flow regime, and the horizontal stratification and pebble imbrication are consistent with deposition within a braided stream environment (Bridge and Lunt, 2006; Miall, 2010). Local discontinuous fine-grained beds within the upper sandy gravel were deposited at low flow velocity in an interchannel floodplain setting. Lower Fine-Grained Deposits The lower fine-grained deposits are interpreted as lacustrine sediments. The primary constituents are locally laminated clay with low silt content, marl, and silty clay. These sediments overlie silty fine sand and sandy silt. Though there is no single characteristic that defines a lacustrine depositional environment, typical indications include abundant fine-grained deposits, laterally continuous thin beds, lamination, freshwater fauna, organic-rich sediments, and evaporite or carbonate mineralogy (Picard and High, 1972, 1981; Platt and Wright, 1991; and Carroll and Bohacs, 1999). The laterally persistent fine-grained deposits, local lamination, carbonate-rich sediments, and local molluscan fauna of the lower fine-grained sedimentary strata are consistent with a lacustrine setting. Interpretation of Geologic Structure and Stratigraphy The study of geologic structure of the lower finegrained strata has been greatly aided by the recognition of the unoxidized clay-marl marker bed couplet. Through detailed logging of core borings and careful

146

correlation of the marker bed couplet between these borings, structure contours were drawn on the top of the marker bed couplet (Figure 7). The structure contour map shows a northwest-plunging anticline in the northern portion of the study area and a northwestplunging syncline in the southern portion. These folds align well with the orientation of other folds and thrust faults in the MDFTB. This implies that folding of the lacustrine sediments is related to northeast-southwest shortening within the MDFTB. Though the lower fine-grained sedimentary deposits had not been recognized at the time of quarry excavation, the pit was excavated above the southwest-dipping limb of the anticline (Figure 7). A cross section across the quarry and northeast flank of the quarry shows the site stratigraphy and geologic structure (Figure 8). In the southwest limb of the anticline, the maker bed couplet strikes roughly 303u to 313u (N47W to N57W) and dips approximately 2u to the southwest. The dip of the northeast limb is approximately 4u to the northeast. By contrast, the overlying upper sandy gravel deposits do not appear to be folded. Based upon this angular discordance and the apparent truncation of the lower fine-grained strata in the northeastern portion of the study area, it appears that an unconformity separates the upper sandy gravel from the lower fine-grained deposits. Thus, we designate them as separate stratigraphic sequences. Though the exact ages of these two sequences is poorly constrained, based upon the topographic setting of these deposits and the contrast with the steeply tilted and folded Livermore Formation beds that are exposed in the nearby hills south of the basin, the lower fine-grained strata appear to be Pleistocene in age, and the upper sandy gravel strata appear to be late Pleistocene to Holocene in age. HYDROGEOLOGY OF THE ARROYO DEL VALLE QUARRY AREA Based upon subsurface exploration, three separate aquifers were identified at the Arroyo del Valle Quarry site. The shallowest aquifer is within the upper portion of the sandy gravel and above the discontinuous finegrained beds. Below the discontinuous fine-grained beds is a semi-confined aquifer within the lower portion of the sandy gravel. The oxidized clay, unoxidized clay, and marl form an aquitard below the sandy gravel. The beds of sand and sandy silt below the unoxidized clay and marl constitute a confined aquifer. Within the study area, these three aquifer units are designated as the upper, middle, and lower aquifers (Figure 9). Though deeper aquifers may be present below the designated lower aquifer, this study did not include characterization of deeper aquifer stratigraphy.

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 141â&#x20AC;&#x201C;155

Arroyo Del Valle Quarry

Figure 7. Site map with structure contours drawn on the top of the unoxidized clay-marl marker bed couplet. The cross-section line indicates the location of the geologic cross section (Figures 8, 9, and 13).

Long-term monitoring of vibrating wire piezometers installed in the lower aquifer indicates that piezometric surface elevations typically ranged from 75 to 102 ft (23–31 m) above the top of the lower aquifer. By contrast, the middle aquifer water levels have remained lower than those in the lower aquifer. Thus, the piezometer data indicate an upward gradient from the lower aquifer to the middle aquifer (Figure 9). Pore pressure measurements from piezometers installed in the unoxidized clay are intermediate between those of the middle and lower aquifers. This upward pore pressure gradient has caused elevated pore pressures within the intervening unoxidized clay interval. SLOPE INSTABILITY The instability of the quarry pit slopes is not readily apparent at the ground surface. As a result of the relatively small magnitude of displacement, roughly 4 to 5 in. (10–12 cm), typical landslide-related landforms (scarps, grabens, and hummocky topography) have not developed. The initial evidence for instability of

the quarry slopes came from linear cracks that formed in roadways located approximately 125 ft (38 m) northeast of the quarry pit (Figure 10). Following the initial observation of pavement distress in 2001, multiple slope inclinometers were installed around the quarry pit from 2002 to 2006. Based upon detailed logging of continuous core samples from the slope inclinometer borings, discrete (landslide-type) deflections in the slope inclinometers were consistently found to correspond to the depth of the sheared unoxidized clay. This is shown in a data plot from a slope inclinometer located near the northeast flank of the quarry pit combined with the stratigraphic column from the same slope inclinometer boring (Figure 11). Even where the quarry pit walls were locally steeper than 2H:1V (26.6u), the slope inclinometer casings deflected only within the unoxidized clay and not within the upper sandy gravel sequence. Thus, the site stratigraphy exerts strong control on the instability of the quarry slopes. Slope instability is also controlled by geologic structure. As shown in Figure 7, the slope inclinometer

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 141–155

147

Johnson, Shires, and Sneddon

Figure 8. Geologic cross section across the quarry site and adjacent area northeast of the quarry.

displacement vectors are consistently oriented in a down-dip and downslope direction. At the northeast margin of the quarry, in the area between the anticline axis (Figure 7) and syncline that underlies the pit floor, slope inclinometer monitoring data show deflections in a down-dip direction, toward the pit. We refer to this critical sliding block where the sheared unoxidized clay dips toward the quarry pit as the Northeast Block. The anticline axis that forms the updip margin of the Northeast Block acts as a natural barrier to retrogressive movement farther to the northeast. Across the anticline axis, the sheared unoxidized clay bed dips away from the pit slope, and the slope inclinometers on this fold limb have not deflected. The monitoring results from a representative slope inclinometer located on the upper bench on the northeast side of the quarry provide a useful record of displacement of the northeast block over a period of several years (Figure 12), although it does not include the estimated several inches of movement that occurred prior to slope inclinometer installation. For the purposes of this study, the rates of displacement were averaged and annualized to allow meaningful comparison of monitoring periods of varied durations. Between January 2004 and April 2005, slope

148

inclinometer monitoring results showed that the displacement rates averaged 0.26 in. (6.6 mm) per year. Between April and September of 2005, the displacement rate increased to an average of 1.0 in. (25.4 mm) per year. Between September 2005 and September 2006, an average displacement rate of 0.14 in. (3.5 mm) per year was recorded. Thus, prior to implementation of mitigation measures designed to address slope instability at the quarry pit, the rate of movement within the Northeast Block had already slowed. The recent displacement that occurred on the northeast flank of the quarry pit would be insufficient in magnitude to produce the well-developed polished

Figure 9. Hydrogeologic cross section showing the piezometric surfaces of the middle and lower aquifers within the Northeast Block. U, upper aquifer; M, middle aquifer; L, lower aquifer.

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 141â&#x20AC;&#x201C;155

Arroyo Del Valle Quarry

Figure 10. Map of ground cracks that were the ďŹ rst indication of incipient slope instability related to the quarry pit slope.

surfaces and clay gouge observed in the unoxidized clay. In addition, the unoxidized clay was encountered in an intensely sheared state in borings throughout the study area, including the fold limb that dips away from the quarry pit. Therefore, it appears that the unoxidized clay was sheared prior to excavation of the quarry pit. Based upon the slope inclinometer data, it is clear that the inclinometer deflections were a response to stress release and removal of lateral confinement due to quarry pit excavation. This stress release triggered downslope movement in the weak unoxidized clay bed where it dips toward the quarry pit excavation.

The high pore-water pressures within the unoxidized clay also contributed to instability of the Northeast Block. The movement of the Northeast Block that followed pit excavation appears to represent incipient landsliding because it did not develop into a larger magnitude failure. GEOTECHNICAL ANALYSIS AND DESIGN OF MITIGATION MEASURES Limit equilibrium and finite element slope stability analyses were performed to assess the instability and factor of safety (FS) of the Northeast Block. The

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 141â&#x20AC;&#x201C;155

149

Johnson, Shires, and Sneddon

Figure 11. Slope inclinometer data plot and stratigraphic column from a slope inclinometer located near the northeast ďŹ&#x201A;ank of the quarry pit.

analyses indicated that the shear strength of the sandy gravel was incrementally mobilized as displacements occurred within the sheared unoxidized clay. This explains the decline in slope inclinometer deflection over time. Under static conditions, the slopes should experience little displacement once the gravel strength has been fully mobilized. However, under seismic conditions, the Northeast Block would be expected to reactivate, potentially leading to much larger displacements. Consequently, implementation of mitigation measures was necessary to provide for the long-term stability of the Northeast Block during future seismic events. Though several methods of mitigation were considered, an earth-fill buttress repair was chosen for its simplicity and reliability. Limit equilibrium slope stability analyses showed that the pressure exerted by

150

Figure 12. Graph of cumulative displacement over time at the unoxidized clay interval from a representative slope inclinometer on the northeast ďŹ&#x201A;ank of the quarry pit.

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 141â&#x20AC;&#x201C;155

Arroyo Del Valle Quarry

water within the quarry pit acted to resist movement of the Northeast Block. If the pit were drained to allow placement of an earth-fill buttress, displacements within the Northeast Block would likely accelerate until the gravel strength was fully mobilized. Thus, mitigation measures designed to stabilize the Northeast Block had to maintain short-term stability during placement of the buttress. A two-dimensional limit equilibrium slope stability analysis was performed on a representative cross section using Spencer’s method (Wright, 1975) to evaluate the stability of the Northeast Block during placement of the buttress and to calculate the potential seismic displacements after completion of the buttress. The key parameters for this slope stability analysis included topographic profile, rupture surface geometry, material shear strengths, unit weights, and piezometric surface elevations. The material shear strengths were determined by laboratory testing (consolidated-undrained triaxial compression and torsional ring shear testing) and back-calculation analyses. The shear strengths for the sandy gravels in particular were difficult to estimate because of difficulties in obtaining undisturbed samples of the sandy gravel material for triaxial shear strength testing due to the large clast size. Therefore, in addition to triaxial compression testing, back-calculation analyses of the sandy gravel material strengths were performed on areas of the quarry where the sandy gravel was exposed in near-vertical cut faces of approximately 35 ft in height. Several small failures had occurred in these cut faces, and based on these analyses the laboratory-derived peak shear strength was actually lower than the back-calculated shear strengths. Correlations between gravel particle size and friction angle were also considered in developing the shear strengths. Based on these sources, representative shear strengths for the sandy gravel material were selected as shown in Table 1. The material unit weights were determined by laboratory testing of undisturbed samples. The shear strengths and unit weights are compiled in Table 1. The topographic profile and subsurface geometry were taken from a representative geologic cross section (Figure 8), and the piezometric surface elevations were derived from monitoring of piezometers within the study area. In addition to the stratigraphic units previously described, two additional materials were placed as fill for the earth-fill buttress: K-in.– (13-mm–) diameter pea gravel and compacted pit run fill. The pea gravel was produced by processing of material excavated from one of the other nearby gravel quarries. The pit run material consisted of sandy gravel excavated from a borrow area at the southwest margin of the

Table 1. Summary of static material properties. Unit Weight (pcf)

Cohesion (psf)

Friction Angle (u)

Sandy gravel Clay beds within sandy gravel Oxidized clay Sheared unoxidized clay Unoxidized clay gouge (residual shear strength) Lower confined aquifer Compacted pea gravel fill

139 127 130 121 121

200 1,500 1,000 0 0

133 134

1,300 0

Compacted pit run fill

145

100

45 27 24 11 Non-linear (,6) 28 Non-linear (,40 to 51) 39

Material

quarry pit and placed as engineered fill without additional processing. The piezometric surface elevations were determined by long-term monitoring of staged vibrating wire piezometers that were installed within the Northeast Block and elsewhere around the study area. The piezometer sensors were divided into three groups: “upper sensors,” “mid-sensors,” or “lower sensors,” based on the aquifer stratigraphy and depth of sensor placement (Figure 9). The upper aquifer and discontinuous finegrained beds within the gravel were assigned the “upper sensor” piezometric surface. The middle aquifer was assigned the “mid-sensor” piezometric surface. The unoxidized clay and lower aquifer were assigned the “lower sensor” piezometric surface. For design of the buttress, the highest recorded levels for each sensor group were used to estimate the piezometric surface for each hydro-stratigraphic unit. The initial analyses indicated that the calculated FS values ranged from 1.22 to 1.70 (depending on the rupture surface analyzed) using peak shear strengths for most materials and residual shear strength for rupture surface gouge, as shown in Table 1. The unexpected finding that FS remained above 1.0 while active landslide-type movement was apparently underway is best explained by the high strength of the sandy gravel material. The slope movement response to the quarry excavation was concentrated on the unoxidized clay, while the relatively strong sandy gravel near the toe of the slope provided shearing resistance to this deep movement in the unoxidized clay. Because the amount of movement in the unoxidized clay was relatively small, displacement did not reach the point at which the peak shear strength of the sandy gravel near the toe of the slope was fully mobilized. Therefore, when using peak strength values for the sandy gravel in the slope stability analyses, the overall static FS of the slope (including the impact of the resisting sandy gravel) remained above unity, although local discrete

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 141–155

151

Johnson, Shires, and Sneddon

Figure 13. The geologic cross section annotated to show the phases of grading during placement of the earth-ďŹ ll buttress. The depressurization wells are screened in the lower aquifer. U, upper aquifer; M, middle aquifer; L, lower aquifer. The dashed horizontal line represents the elevation of the lower aquifer piezometric surface after installation of the depressurization wells.

shearing movement in the unoxidized clay occurred in response to the excavation. Once sufficient displacement within the unoxidized clay had occurred, the shearing resistance of the sandy gravel was partially mobilized, and displacement within the unoxidized clay slowed. These analyses expose an oversimplifying assumption typically used for limit equilibrium slope stability analysis, that the FS is uniform along all portions of the rupture surface analyzed whether the rupture surface passes through weak clay at residual strength or strong gravel with no previous history of shearing. In addition to the static slope stability analyses that were conducted, seismic displacements were also estimated using a Newmark-type sliding block analysis (Jibson and Jibson, 2003). Three active faults (the Calaveras, Greenville, and Hayward faults, located 9.3 to 19.1 km from the site) were identified as most likely to affect the site, with moment magnitudes ranging from 6.6 to 7.1. A design target response spectrum was developed for the site based upon multiple attenuation relationships with modifications for nearfault effects. Seven strong motion records (Pacific Earthquake Engineering Research Center, 2000) were selected that were representative of the expected ground motion at the site and scaled as necessary to obtain a relatively close fit between the average response spectrum of the seven motions and the target design response spectrum. One-dimensional equivalent-linear seismic response analysis (SHAKE2000, 2000) was performed on three representative stratigraphic columns to obtain horizontal equivalent acceleration time histories at the depth of sliding, which were then used in conjunction with yield accelerations from pseudo-static slope stability analyses to estimate seismic displacements (Koragappa et al., 2004). Based upon these analyses, the movement of the Northeast Block resulted from stress relief upon excavation of the free face at the northeast margin of the quarry pit. The previously sheared clay that was at or near residual shear strength prior to quarry pit excavation and the elevated pore pressures within the critical sliding surface were the primary factors resulting in the movement detected initially as ground cracks in

152

nearby roads and later by the slope inclinometers. Using partially mobilized upper gravel strengths, piezometric influence from the lower aquifer, and potential movement that extends northeastward to the anticline axis, our analysis showed that an industry-accepted static FS of greater than 1.5 and seismic displacements of less than 15 cm (6 in.) could be achieved by partial filling of the quarry pit with engineered fill to an elevation of 390 ft (119 m) above sea level (Figure 13). IMPLEMENTATION OF MITIGATION MEASURES Between September 2006 and May 2007, 40 depressurization wells were installed on the northeast flank of the quarry to relieve pore-water pressures within the lower aquifer. Based upon the site hydrogeology, lowering of the piezometric surface within the lower aquifer was expected to reduce pore pressures within the overlying unoxidized clay and, in turn, to improve the static FS of the Northeast Block. Immediately upon installation, the depressurization wells flowed at the surface under artesian conditions. As anticipated, the artesian flow resulted in lowering of the piezometric surface of the lower aquifer to roughly the elevation of the well discharge pipes. The piezometric surface at a representative piezometer within the Northeast Block (Figure 14) shows this initial drop. Following installation of the depressurization wells, the displacement rates in slope inclinometer SI2 decreased to an average of 0.04 in. (1 mm) per year between September 2006 and May 2008 (Figure 12). During May 2008, submersible pumps with float-activated controllers were installed in 10 of the wells. Within 9 days of pump activation, the elevation of the piezometric surface within the lower aquifer (Figure 14) had dropped by 43 ft (13 m). Within 3 months after pump activation, the piezometric surface elevation had dropped an additional 20 ft (6 m). Following activation of the well pumps and prior to placement of the fill buttress, the quarry pit (artificial lake) was partially drained. Between June 2008 and August 2008, the water surface elevation of the lake

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 141â&#x20AC;&#x201C;155

Arroyo Del Valle Quarry

Figure 14. Piezometric surface elevation data from a representative piezometer installed in the conďŹ ned lower aquifer on the northeast side of the quarry pit. (a) Initial drop in piezometric surface during installation of depressurization wells; (b) brief pumping test on a single depressurization well; (c) well discharge to the quarry pit maintains a constant piezometric surface elevation of approximately 407 ft (the elevation of the well discharge lines); (d) signiďŹ cant lowering of the piezometric surface during pumping of 10 depressurization wells; (e) return to equilibrium conditions following cessation of well pumping; (f) slight rise of the piezometric surface as the lake surface elevation increased and the well discharge lines were inundated.

was lowered approximately 13 ft (4 m) by pumping. During this period of time, some slope inclinometers within the Northeast Block showed a modest acceleration in deflection rate. Between June 2008 and November 2008, the average slope inclinometer displacement rate was 0.18 in. (4.5 mm) per year (Figure 12). Thus, even partial draining of the lake did affect slope stability, as predicted. In order to temporarily stabilize the Northeast Block before the quarry pit was fully drained, a temporary buttress was placed on the lake bottom. The temporary buttress consisted of pea gravel fill that was dropped through the water column and onto the lake bottom using a floating conveyor belt system. Figure 13 shows the buttress in cross-section view. Prior to construction, analysis of the loose dumped pea gravel fill indicated that it could be susceptible to liquefaction associated with strong seismic ground motions, so it was necessary to compact the pea gravel fill to provide long-term stability under future seismic loading conditions. Once the lake was drained and the pea gravel fill could be dewatered, it was excavated in slots oriented perpendicular to the slope and not more than 100 ft (30 m) wide in order to preserve the buttressing effect. The pea gravel was placed back

into the slots in lifts and compacted to 95 percent relative compaction. After recompaction was completed, the elevation of the top of the pea gravel fill was 364.5 ft (111.1 m). Following placement and recompaction of the pea gravel fill, drainage of the pit continued, and pit run material was excavated from a shallow borrow area southwest of the pit and placed as compacted fill over the pea gravel using conventional grading methods. As placement of the fill buttress progressed, the slope inclinometer displacement rate slowed to 0.045 in. (0.11 mm) per year between November 2008 and April 2009. The completed fill buttress reached the final elevation of 390 ft (118.9 m) during April 2009. Subsequent monitoring of site slope inclinometers and piezometers from 2009 through 2014 has shown no evidence of an increased rate of movement (Figure 12) even though the piezometric surface of the lower aquifer has risen since pumping of the depressurization wells ceased (Figure 14). CONCLUSIONS The unique stratigraphy, geologic structure, and hydrogeology at the Arroyo del Valle Quarry site

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 141â&#x20AC;&#x201C;155

153

Johnson, Shires, and Sneddon

and surrounding area were key factors that led to incipient failure of the northeast slope of the quarry pit (the Northeast Block). The highly sheared unoxidized clay found in the subsurface was weakened to residual shear strength by other geologic processes prior to excavation of the quarry pit, and this very low shear strength made it vulnerable to instability once the quarry pit was excavated. The slope inclinometers deflected only within the unoxidized clay interval; thus, the site stratigraphy strongly controlled incipient landslide movement. Slope inclinometer displacements occurred only where the unoxidized clay dips toward the quarry pit. Therefore, geologic structure also controlled incipient landsliding. Elevated pore-water pressures within the unoxidized clay appear to be related to the high pore pressures within the underlying lower aquifer and the upward hydraulic gradient between the lower and middle aquifers. Thus, the site hydrogeology also contributed to the instability of the Northeast Block. Based upon our analyses, we conclude that incipient landsliding resulted from the site geologic and hydrogeologic conditions described above combined with stress relief related to excavation of the quarry pit. Initial deflection within the weak unoxidized clay eventually led to partial mobilization of the strength of the overlying gravel and a decline in slope inclinometer deflection rate. However, the static FS remained below the industry-accepted minimum value of 1.5, and unacceptably large displacements were anticipated under seismic loading conditions. Thus, it was necessary to design and implement mitigation measures to address the static and seismic stability of the Northeast Block. Because the pressure exerted by the water that filled the former quarry pit provided some counterbalance to the forces driving slope movement, the lake could not be drained without triggering significant acceleration in slope movement. Initially, the pore pressures within the unoxidized clay were lowered using depressurization wells screened in the lower aquifer combined with pumping of those wells. Then a temporary buttress of pea gravel was placed on the floor of the former quarry pit before the lake was completely drained. Once drained, the pea gravel was removed in slots, compacted as it was placed back into the slots, and capped with additional engineered fill that was placed using conventional grading methods. Limit equilibrium analyses show that the completed buttress should provide a FS of greater than 1.5 under static conditions, even as pore pressures within the unoxidized clay rise. Displacements that could result from strong seismic shaking were calculated at less than 6 in. (15 cm) with the earth-fill buttress in place.

154

Five years of post-construction monitoring have shown no evidence of renewed slope inclinometer deflection within the Northeast Block or elsewhere around the quarry site. ACKNOWLEDGMENTS This project was completed with the review, assistance, and guidance of professors Jonathan Bray, Scott Kiefer, and Matthew Mauldon. Geotechnical laboratory testing was provided by Cooper Testing Labs, Inc. The authors wish to thank Dale Marcum, Joe Durdella, Jason Nichols, Jonathan Sleeper, Jamie Smith, and many others at Cotton Shires and Associates, Inc., for their assistance during completion of this project. We wish to thank Michael Hart, Douglas M. Yadon, and an anonymous reviewer for their helpful comments that improved this manuscript. REFERENCES ANDERSEN, D. W.; ISAACSON, K. A.; AND BARLOCK, V. E., 1995, Neogene evolution of the Livermore Basin within the California Coast Ranges. In Fritsche, A. E. (Editor), Cenozoic Paleogeography of the Western United States II: Pacific Section SEPM Publication 75, pp. 151–161. BARLOCK, V., 1989, Sedimentology of the Livermore Gravels (Miocene-Pleistocene), Southern Livermore Valley, California: U.S. Geological Survey Open File Report 89-131, 93 p. BRIDGE, J. S. AND LUNT, I. A., 2006, Depositional models of braided rivers. In Sambrook Smith, G. H.; Best, J. L.; Bristow, C. S.; and Petts, G. E. (Editors), Braided Rivers: International Association of Sedimentologists Special Publication No. 36, pp. 11–50. CALIFORNIA DEPARTMENT OF WATER RESOURCES (CDWR), 1974, Evaluation of Groundwater Resources, Livermore and Sunol Valleys: CDWR Bulletin 118-2, 153 p. CARROLL, A. R. AND BOHACS, K. M., 1999, Stratigraphic classification of ancient lakes: Balancing tectonic and climatic controls: Geology, Vol. 27, No. 2, pp. 99–102. CRANE, R., 1995, Geology of the Mount Diablo region and East Bay hills. In Sangines, E. M.; Andersen, D. W.; and Buising, A. V. (Editors), Recent Geologic Studies in the San Francisco Bay Area: Pacific Section SEPM Special Publication, Vol. 76, pp. 87–114. CRANE, R., 2007, Geology of Central California: Pacific Section American Association of Petroleum Geologists, CD-ROM No. 4, 237 p. DOUGHTON, S., 2009, State warned gravel pit of slope instability 4 years before landslide: Seattle Times, October 16, 2009. DUPRAS, D., 1999, Representative industrial mineral mines of the San Francisco Bay region: Sand and gravel, crushed rock, and limestone. In Wagner, D. L. and Graham, S. A. (Editors), Geologic Field Trips in Northern California: California Division of Mines and Geology Special Publication 119, pp. 235–245. EHMAN, K. D.; FIGUERS, S.; AND BLAKE, R., 2004, Preliminary stratigraphic evaluation, west side of the main basin, LivermoreAmador groundwater basin: unpublished consultant’s report to Zone 7 Water Agency, 25 p.

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 141–155

Arroyo Del Valle Quarry GOLDMAN, H. B., 1964, Sand and Gravel in California, and Inventory of Deposits, Part B Central California: California Division of Mines and Geology Bulletin 180-B, 58 p. HELLEY, E. J. AND GRAYMER, R. W., 1997, Quaternary Geology of Alameda County and Surrounding Areas, California: U.S. Geological Survey Open File Report 97-97, 1:100,000 scale. HOEK, E. AND KARZULOVIC, A., 2000, Rock mass properties for surface mines. In Hurstrulid, W. A.; McCarter, M. K.; and Van Zyl, D. J. A. (Editors), Slope Stability in Surface Mining: SME, Littleton, CO, pp. 59–70. JIBSON, R. W. AND JIBSON, M. W., 2003, Java Programs for Using Newmark’s Method and Simplified Decoupled Analysis to Model Slope Performance During Earthquakes: U.S. Geological Survey, Open File Report 2003-005, CD-ROM. KORAGAPPA, N.; BRAY, J. D.; MEYER, D.; AND TRAVASAROU, T., 2004, Seismic slope stability of a landfill on young bay mud: Proceedings Waste Tech Landfill Technology Conference, Dallas, TX, May 17–19, 2004, 22 p. MIALL, A., 2010, Alluvial deposits. In James, N. P. and Dalrymple, R. W. (Editors), Facies Models 4: Geological Society of Canada, GEOText 6, pP. 105137. PACIFIC EARTHQUAKE ENGINEERING RESEARCH CENTER, 2000, Ground Motion Database: Electronic document, available at http://peer.berkeley.edu PANKOW, K. L.; MOORE, J. R.; HALE, J. M.; KOPER, K. D.; KUBACKI, T.; WHIDDEN, K. M.; AND MCCARTER, M. K., 2014, Massive landslide at Utah copper mine generates wealth of geophysical data: GSA Today, Vol. 24, pp. 4–9. PICARD, M. D. AND HIGH, L. R., 1972, Criteria for recognizing lacustrine rocks. In Rigby, J. K. and Hamblin, W. K. (Editors), Recognition of Ancient Sedimentary Environments: SEPM Special Publication, No. 16, pp. 108–145. PICARD, M. D. AND HIGH, L. R., 1981, Physical stratigraphy of ancient lacustrine deposits. In Ethridge, F. G. and Flores, R. M. (Editors), Recent and Ancient Nonmarine Depositional

Environments: SEPM Special Publication, No. 31, pp. 108–145. PLATT, N. H. AND WRIGHT, V. P., 1991, Lacustrine carbonates: Facies models, facies distributions and hydrocarbon aspects. In Anadon, P.; Cabrera, L.; and Kelts, K. (Editors), Lacustrine Facies Analysis: International Association of Sedimentologists Special Publication, No. 13, pp. 57–74. SAWYER, T. L. AND UNRUH, J. R., 2004, Characterization of Late Quaternary Deformation of the Mt. Diablo Fold-and-Thrust Belt, Eastern San Francisco Bay Area, California: EOS Transactions, Vol. 85, Fall Meeting Supplement, Abstract T32C-02. SHAKE2000, 2000, A Computer Program for the 1-D Analysis of Geotechnical Earthquake Engineering Problems: Gustavo A. Ordonez, GeoMotions, LLC, Lacy, WA. UNRUH, J. R.; DUMITRU, T. A.; AND SAWYER, T. L., 2007, Coupling of early Tertiary extension in the Great Valley forearc with blueschist exhumation in the underlying Franciscan accretionary wedge at Mount Diablo, California: Geological Society America Bulletin, Vol. 119, pp. 1347–1367. UNRUH, J. R. AND LETTIS, W. R., 1998, Kinematics of compressional deformation in the eastern San Francisco Bay region, California: Geology, Vol. 26, pp. 19–22. UNRUH, J. R.; SAWYER, T. L.; KNUDSEN, K. L.; AND LETTIS, W. R., 1997, Assessment of Blind Seismogenic Sources, Livermore Valley, Eastern San Francisco Bay Region, Final Technical Report: U.S. Geological Survey NEHRP Award No. 1434-95-G-2611, 88 p. WAGNER, D. L.; BORTUGNO, E. J.; AND MCJUNKIN, R. D., 1990, Geologic Map of the San Francisco–San Jose Quadrangle: California Division of Mines and Geology, Regional Geologic Map Series, Map No. 5A, 5 sheets, 1:250,000. WRIGHT, S. G., 1975, Evaluation of Slope Stability Procedures: American Society of Civil Engineers, preprint 2616, National Convention, Denver, CO, 28 p. WYLLIE, D. C. AND MAH, C. W., 2004, Rock Slope Engineering: Civil and Mining: Taylor & Francis, New York, NY, 456 p.

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 141–155

155

Factors Affecting Failure by Internal Erosion of Geosynthetic Clay Liners Used in Freshwater Reservoirs HAKKI O. OZHAN1 Department of Civil Engineering, Istanbul Kemerburgaz University, Mahmutbey Dilmenler Cad, No. 26, Bagcilar-Istanbul, Turkey

EROL GULER2 Department of Civil Engineering, Bogazici University, 34342 Bebek-Istanbul, Turkey

Key Terms: Internal Erosion, Geosynthetic Clay Liner, Freshwater Reservoir, Permittivity

ABSTRACT Geosynthetic clay liners (GCLs) are often used as lining materials for freshwater reservoirs. To irrigate agricultural land without depleting groundwater, surface water is stored in these artificial ponds. In this study, hydraulic conductivity tests were performed on GCLs placed in flexible-wall permeameters under hydraulic heads of up to 50 m in order to investigate the risk of internal erosion. In these tests, base pedestals made of Plexiglas with uniform circular voids were placed beneath the GCLs instead of a typical gravel subgrade. The voids in the base pedestal represented the voids between uniform rounded gravel particles. Different types of GCLs were tested. GCL-1 was reinforced using needle-punching technology, whereas GCL-2, GCL-3, and GCL-4 were un-reinforced GCLs that were assembled in the laboratory. We investigated the effects on internal erosion of the void size in the subbase; the geotextile component that was in contact with the subbase; the bentonite component; and the manufac‐ turing process of the GCLs. Test results indicated that internal erosion was directly related to the void diameter of the base pedestal. The resistance of the needlepunched GCL to internal erosion was better than that of the un-reinforced GCLs. The degree of internal erosion was also related to the engineering properties of the geotextile in contact with the base pedestal. Higher tensile strength of the GCL reduced the possible potential for internal erosion within it. The type of bentonite did not have a significant effect on internal erosion.

Corresponding author email: hakki.ozhan@kemerburgaz.edu.tr

INTRODUCTION A GCL is a barrier material that consists of a thin layer of bentonite (5–15 mm) sandwiched between two geotextiles and/or glued to a geomembrane by using a water-soluble, non-polluting adhesive (Bouazza, 2002; Koerner, 2005;). Geotextile components of GCLs are bonded to the bentonite on both sides with needlepunching, stitch-bonding, adhesives, or just by attaching the geotextiles to the wet bentonite. This attachment is maintained by placing the bentonite on top of the lower geotextile, then wetting the bentonite, and finally laying the upper geotextile on top of the bentonite layer (Ozhan, 2011). According to the manufacturing process, GCLs are classified into two groups: reinforced GCLs and un-reinforced GCLs. Reinforcement is maintained by either needle-punching or stitch-bonding. Needle-punched GCLs are reinforced by sewing the top geotextile through the bentonite into the bottom geotextile (von Maubeuge and Heerten, 1994). Bonding is maintained by punching the polypropylene fibers between the top and bottom geotextiles, which provides higher internal shear strength (Bouazza, 2002). Alternatively, in stitch-bonding, the GCL is reinforced with parallel rows of sewn yarn (Bouazza, 2002). In an unreinforced GCL, the high cohesion capability of the wetted bentonite particles provides the bond between the bentonite and the upper and lower geotextiles (Ozhan, 2011), or an adhesive is added to the bentonite to increase its bonding capacity to the geotextiles (Bouazza, 2002). Reinforced GCLs have greater shear strength and stronger bonding capability between the bentonite and the geotextiles than un-reinforced GCLs. The manufacturing process of GCLs (reinforced or un-reinforced), the manufacturing of the geotextile components of GCLs (woven or non-woven), and the composition and form of the bentonite used in GCLs (sodium bentonite or calcium bentonite; granular or powdered) are the main parameters used to compare different GCLs. GCLs are widely preferred for environmental protection barriers due to their low hydraulic conductivity

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 157–169

157

Ozhan and Guler

(,10−10 m/s), low cost, and ease of installation in both cover systems and composite bottom liners (Benson and Meer, 2009). GCLs are used either as part of composite liners at the bottom of landfills, canals, storage tanks, and surface impoundments (Hornsey et al., 2010; Kang and Shackelford, 2010; and Rowe and Abdelatty, 2012) or as the sole lining material in freshwater reservoirs. The collection of surface water into these GCL-lined ponds necessitates only a small investment and reduces the demand on groundwater. GCL can be used to overlie a wide range of soils from clay to coarse-grained gravel (Ozhan and Guler, 2013). High water levels have to be taken into consideration when designing any reservoir. As the depth of water increases, the hydraulic head and, consequently, potential for the seepage flow increase, which might cause the bentonite to erode through the geotextile. This process is known as bentonite extrusion. After the threshold hydraulic head is exceeded, bentonite particles can start to flow through openings and damaged zones within the geotextile (McCook, 2007). If this occurs, it will cause the hydraulic conductivity of the GCL to significantly increase and impair the hydraulic barrier capability of the GCL (Ozhan and Guler, 2013). This process is called internal erosion (McCook, 2007). According to Li (2008), grain size distribution, grain shape, and loading conditions influence the degree of internal erosion. The combined properties of the geotextile and the bentonite determine the hydraulic head at which internal erosion begins. However, when internal erosion begins, the combined structure begins to change, and excessive deformations begin to occur in the carrier geotextile, and, consequently, more bentonite particles start to flow. At the end of this process, a total failure in terms of hydraulic conductivity is observed (Ozhan and Guler, 2013). Piping is another process that causes hydraulic failure of earth materials. Piping is induced by erosion of particles that results in a continuous pipe through the material (Jacobson, 2013). Piping initiates at zones of concentrated leakage through the openings of the material, whereas internal erosion is caused by flow through the openings of the material (Fell et al., 2003; McCook, 2007). Based upon this definition, hydraulic failure of GCLs due to significant amount of bentonite loss through the openings of the geotextile is caused by internal erosion. Internal erosion also occurs in dams, embankments, and other structures that consist of earth materials. During internal erosion, the repulsive forces between soil particles become greater than the attractive forces, leading to deflocculation and dispersion of the soil (Burns and Ghataora, 2007). The seepage forces

158

caused by high hydraulic gradients initiate the detachment of soil particles (Greene et al., 2010; Muresan et al., 2011; Benahmed and Bonelli, 2012; and Baena and Toledo, 2014). During the initiation phase, fine particles erode slowly within the matrix of coarser particles (Chang and Zhang, 2013). After the onset of internal erosion, the flow increases and causes a sudden increase in the hydraulic conductivity of the soil, and an increase in the displacement and flow of soil particles. Finally, the soil skeleton becomes unstable and collapses (Bendahmane et al., 2006; Chang and Zhang, 2013; Rodriguez et al., 2014; and Zhang et al., 2015).

BACKGROUND Fox et al. (2000) conducted hydraulic conductivity tests in flexible-wall permeameters on both needlepunched and adhesive-bonded GCLs to study the effects of the particle size of soil cover and the rate of loading. The GCLs tested were placed beneath gravels with a particle diameter varying from 12.7 to 50.8 mm. Test results indicated that bentonite extrusion in‐ creased with increasing cover soil particle size and rate of loading. The confinement provided by the needle-punching process caused the GCLs to have less bentonite extrusion and subsequent variability in thickness than the adhesive-bonded GCLs. Fox et al. (2000) further stated that although bentonite displacement was observed for some of the GCLs, internal erosion was not detected under hydraulic heads of up to 0.60 m. The effect of the engineering characteristics of gravel subgrade on the hydraulic performance of both needle-punched and adhesive-bonded GCLs was investigated by Shan and Chen (2003). One of the subgrade materials was uniformly graded, angular crushed gravel with diameters ranging from 25.4 to 50.8 mm, while the other subgrade material was uniformly graded, rounded gravel and cobble with diameters ranging from 50.8 to 76.2 mm. According to the test results, a hydraulic head of 0.69 m on the GCLs did not cause internal erosion. However, larger particle sizes and increased angularity of the subgrade caused more bentonite loss from the GCL than smaller particle sizes and more rounded shapes of the subgrade (Shan and Chen, 2003). Results of laboratory tests performed by Fox et al. (2000) and Shan and Chen (2003) indicated that more bentonite displacement took place in adhesivebonded GCLs than in needle-punched GCLs. However, internal erosion causing hydraulic failure of the GCLs was not observed in these studies, and this was most likely due to the very low hydraulic heads.

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 157–169

Barriers in Freshwater Reservoirs

Rowe and Orsini (2003) conducted hydraulic conductivity tests on GCLs using a rigid-wall permeameter in order to examine the effect of subgrade type on internal erosion. The subgrades placed beneath the GCLs were sand, 6 mm gravel, and geonet. A geonet is an open grid-like geosynthetic material consisting of two sets of parallel, polymeric ribs, which conveys liquids or gases. The primary function of a geonet is drainage. Most of the GCLs placed over gravel or geonet experienced internal erosion under hydraulic heads ranging from 8 to 90 m. However, when sand subgrade was used, there was no internal erosion under hydraulic heads of up to 90 m. Rowe et al. (2014a) investigated the effects of different factors such as flow rate and slope angle on downslope erosion of bentonite from needle-punched GCLs. In this case, erosion occurred along the downward gradient within the liner. According to the results, the flow rate and slope did not significantly affect the time needed to initiate erosion. However, after the erosion holes were formed, higher flow rates and steeper slopes caused more bentonite particles to be eroded. Apart from the laboratory studies considering internal erosion of GCLs, the literature provides few field studies where internal erosion is associated with GCLs. Stam (2000) reported a case study where a significant amount of leakage was observed through the GCL lining of a reservoir. Internal erosion occurred through the non-woven geotextile of the GCL into the coarse sand subgrade. Orsini and Rowe (2001) indicated that internal erosion took place through the GCLs that were in contact with a coarse gravel subgrade under hydraulic heads greater than 10 m. Rowe et al. (2014b) and Brachman et al. (2014) also performed field tests to investigate the effects of GCL type on down-slope erosion. According to the test results, erosion did not occur for the GCLs with a polypropylene coating facing up or the GCLs with polyacrylamide-based polymer-enhanced bentonite even after 15 months of exposure. However, erosion began after just 6 months of exposure when the GCLs were not coated or polymer enhanced (Brachman et al., 2014). According to the results, additional needle-punching of the GCLs did not reduce the risk of down-slope erosion. Furthermore, granularity of the bentonite had almost no effect on the development of down-slope erosion (Rowe et al., 2014b). The laboratory tests and case studies where internal erosion of GCLs was investigated in the field suggest the factors that affect the amount of bentonite extrusion, and therefore internal erosion, include coarseness of the subgrade, height of the hydraulic head, type of geotextile and bentonite used in the GCLs, and whether the GCLs were needle-punched.

The objective of our research was to investigate the performance of different GCLs against internal erosion by placing the GCLs over coarse subgrade materials under high hydraulic heads, which could simulate possible field conditions for GCL usage in lining reservoirs. For this purpose, the effects of the manufacturing process, the engineering properties of the geotextiles, and the type of bentonite on internal erosion were investigated. In the tests, a base pedestal made of Plexiglas with uniform circular voids was used to simulate natural gravel as a subgrade beneath the GCL. In this way, it was possible to avoid effects of randomness of the void size of natural earth materials. The diameters of the voids were selected as 20, 15, 10, and 5 mm. It was shown by Ozhan and Guler (2013) that a perforated base pedestal with uniform circular voids successfully simulated rounded coarsegrained gravel in terms of internal erosion. Both GCLs reinforced by needle-punching and un-reinforced GCLs placed over base pedestals with circular voids of different sizes were tested under hydraulic heads up to 50 m. MATERIALS Four different GCLs (one reinforced and three unreinforced) were tested in this study. The first GCL, Bentomat SS100 (CETCO, 2007), designated as GCL-1, was a reinforced GCL consisting of a layer of granular sodium bentonite between a woven slitfilm polypropylene geotextile (108 g/m2) and a nonwoven needle-punched polypropylene geotextile (203 g/m2). To provide reinforcement, polypropylene fibers from the non-woven geotextile were needle-punched through the bentonite to the woven geotextile. The engineering properties of GCL-1 are listed in Table 1 (CETCO, 2007). GCL-2 was assembled in the laboratory using the same sodium bentonite as GCL-1, sandwiched between the same woven and non-woven geotextiles used in GCL-1 without needle-punching the components. GCL-3 was assembled in the laboratory using calcium bentonite sandwiched between the same woven and non-woven geotextiles of which GCL-1 was composed. Again, the components were not needle-punched. GCL-4 was assembled using sodium bentonite sandwiched between the same non-woven geotextile used in GCL-1 and a woven geotextile with a relatively higher tensile strength than the one used in GCL-1, without needle-punching the components. The non-woven geotextile designated as N1 that was used in all of the GCLs was a polypropylene, staplefiber, needle-punched, non-woven geotextile. The woven geotextile that was used in GCL-1, GCL-2, and GCL-3 specimens was designated as W1, while

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 157â&#x20AC;&#x201C;169

159

Ozhan and Guler Table 1. Engineering properties of GCL-1. Property Bentonite swell index Bentonite mass per unit area GCL tensile strength GCL tensile elongation GCL index flux GCL permeability

Test Method

Value

ASTM D5890 (2011) ASTM D5993 (2009) EN ISO 10319 (2008) EN ISO 10319 (2008) ASTM D5887 (2009) ASTM D5887 (2009)

min. 24 mL/2 g min. 4.8 kg/m2 8 kN/m 15% 2 6 10−9–2 6 10−10 m3/m2/s 1 6 10−11–1 6 10−12 m/s

the woven geotextile used in GCL-4 was designated as W2. Both of the woven geotextiles were polypropylene, slit-film geotextiles. The non-woven geotextile that was used as a filter beneath the perforated base pedestal was designated as NF. The engineering properties of the geotextiles used in this study are listed in Table 2. The sodium bentonite in GCL-1, GCL-2, and GCL4 was a naturally occurring, granular, high-swelling Wyoming sodium bentonite, whereas the calcium bentonite in GCL-3 was a naturally occurring, powdered calcium bentonite from Ankara, Turkey. The swell index, liquid limit, and plastic limit were 28 mL/2 g (ASTM D5890, 2011), 344 percent, and 36 percent (ASTM D4318, 2010) for the sodium bentonites, and 20 mL/2 g (ASTM D5890, 2011), 141 percent, and 41 percent (ASTM D4318, 2010) for the calcium bentonite, respectively. SPECIMEN PREPARATION A 100-mm-diameter template was placed on specimen GCL-1 with the woven geotextile side facing up. Then, the outer perimeter was wetted with de-ionized water to prevent bentonite loss during cutting. After the bentonite was saturated, the specimen was cut with a utility knife (Fox et al., 2000; Rowe and Orsini, 2003). Preparation of specimens GCL-2, GCL-3, and GCL-4 was different from that of GCL-1. Both non-woven and woven geotextiles with a diameter

of 100 mm were cut separately with scissors, and the bentonite was placed between the geotextiles. Then, the granular sodium bentonite was wetted with de-ionized water before placing it between the geotextiles. The wetting of the bentonite enabled it to bond with the upper and lower geotextiles. For specimen GCL-3, the calcium bentonite used was originally in powder form and was pre-moistened to cause expansion before being placed between the geotextiles (Ozhan, 2011). METHODOLOGY Void Size of the Base Pedestal Used Beneath the GCL In order to standardize the tests and eliminate the variability of the natural gravel subbase, a perforated base pedestal with uniform circular voids was used instead of a natural subgrade. The base pedestal was made of Plexiglas that had uniform circular voids. The diameter of the circular voids was varied to simulate different types of subgrade soils. Using the trigonometric assumption presented by Ozhan and Guler (2013), the maximum void diameter was calculated as approximately 20, 15, 10, and 5 mm for a uniform grain size of 50, 37.5, 25, and 12.5 mm, respectively. The pedestal with hole diameters of 20, 15, 10, and 5 mm had 4, 6, 12, and 44 holes, respectively (Figure 1a, b, c, and d). The ratio of the area of voids to the total area was between 11 and 16 percent. GCL specimens were placed over these base pedestals during the tests. Ozhan and Guler (2013) showed that the

Table 2. Engineering properties of the geotextile components of the GCLs.

Property Mass per area (g/m2) Thickness (mm) Wide-width tensile strength (kN/m) Wide-width elongation (%) Apparent opening size (mm)

160

Test Method

Non-Woven Geotextile (N1)

Woven Geotextile (W1)

Woven Geotextile (W2)

Non-Woven Geotextile as a Filter (Nf)

ASTM D 5261 ASTM D 5199 ASTM D 4595 ASTM D 4595 ASTM D 4751

203 2.0 15.4 45 0.212

108 0.4 12.2 10 0.425

210 0.77 44 15 0.380

120 1.1 9 50 0.08

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 157–169

Barriers in Freshwater Reservoirs

Figure 1. Perforated base pedestals with: (a) 20 mm, (b) 15 mm, (c) 10 mm, and (d) 5 mm hole diameters.

resistance against internal erosion of GCLs placed over granular soils could be successfully simulated by perforated base pedestals with 10-mm- and 15-mmdiameter voids.

Test Procedure After specimen preparation, hydraulic conductivity tests were performed using a standard flexible-wall permeameter and the constant-head method (ASTM D5887, 2009). The thickness of a GCL specimen varies throughout the surface area, which makes it hard to construct a constant thickness (Fox et al., 2000). Due to thickness variation, it is possible to measure the permeation through the GCL specimen by not taking the thickness parameter into account. For this reason, permittivity (Ψ) values were measured in these tests. The use of Ψ is more accurate than the use of permeability (k) for thin, compressible geosynthetic products such as GCLs because there is no need to measure the thickness of the GCL specimen (Koerner, 2005). Permittivity (Ψ) is the ratio of the coefficient of permeability (k) to the average thickness (L) of the soil. Permittivity (Ψ) is expressed in Eq. 1 as follows: W¼

DQ ; A Dh Dt

ð1Þ

where Ψ (1/T) is the permittivity, ΔQ (L3) is the average of inflow and outflow amounts for a given time interval, A (L2) is the cross-sectional area of the GCL specimen, Δh (L) is the hydraulic head difference, and Δt (T) is the interval of time over which the flow ΔQ occurs. The configuration of the test setup is shown in Figure 2. From top to bottom, the setup consisted of a rigid top cap, porous stone, filter paper, GCL specimen, perforated base pedestal, non-woven geotextile filter (NF), and a rigid bottom cap. A latex membrane was used to prevent side leakage. More details of the test setup and the specimen preparation in the permeameter are reported in Ozhan (2011). The GCL-1, GCL-2, and GCL-3 specimens were tested with both the woven and the non-woven geotextile facing the perforated base pedestal. However, only the woven geotextile component of the specimen GCL-4 was tested. After the GCL specimen was fully saturated and consolidated, permeation was initiated by raising the pressure at the top of the specimen to 530 kPa while the effluent pressure was kept constant at 515 kPa. As a result, downward flow was produced through the GCL specimen with a pressure difference of 15 kPa (ASTM D5887, 2009). The pressure measured at the location where the water entered at top of the GCL specimen was called the influent pressure, whereas the

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 157–169

161

Ozhan and Guler

Figure 2. Test setup conﬁguration.

pressure measured at the location where the water left the bottom of the GCL specimen was called the effluent pressure. When the flow became steady or the increase in permittivity was less than one order of magnitude, the head was increased to 5 m by 1 m increments at 10 minute intervals. To increase the hydraulic head, both the cell pressure and the influent pressure were kept constant, while the effluent pressure was decreased to the desired value. At each 5 m increment, the head was kept constant for 12 days. A duration of 12 days was chosen because a significant increase in permittivity, up to three orders of magnitude, was obtained within approximately 12 days for all of the GCLs that experienced internal erosion (Ozhan, 2011; Ozhan and Guler, 2013). Therefore, it was concluded that the test duration was satisfactory for comparison reasons. When internal erosion occurred, the hydraulic conductivity test was terminated; otherwise, the hydraulic head was increased to 50 m in 5 m increments (Ozhan, 2011; Ozhan and Guler, 2013). In this study, the maximum applied hydraulic head was chosen as 50 m because water depths can be as high as 40–50 m in freshwater reservoirs (Zohary and Ostrovsky, 2010). Consequently, a hydraulic head of 50 m was considered sufficient to simulate the worstcase scenario. Moreover, a 50 m hydraulic head was used to investigate internal erosion of the GCLs in most of the hydraulic conductivity tests performed by Dickinson and Brachman (2010) and Rowe and Orsini (2003). The GCL specimens tested in this study were designated according to the geotextile type in contact with the base pedestal (W for woven or NW for nonwoven), and the void diameter of the base pedestal (D2 for 2 cm, D1.5 for 1.5 cm, D1 for 1 cm, or D0.5 for 0.5 cm). For example, a GCL-1 specimen tested with a non-woven geotextile over the base pedestal

162

with a void diameter of 1.5 cm is designated as GCL1-NW-D1.5. Tests were performed on the woven and non-woven geotextiles of GCL-1, GCL-2, and GCL3 using void diameters of 0.5, 1, 1.5, and 2 cm, and on the higher-tensile-strength woven geotextile of GCL-4 using void diameters of 1, 1.5, and 2 cm. TEST RESULTS The results of the hydraulic conductivity tests and the measured permittivities of the GCLs are listed in Table 3. The hydraulic heads that caused internal erosion and the confining, influent, and effluent pressures are also listed in Table 3. When exposed to a high hydraulic head, at first, the measured permittivity of the GCL decreased very slowly. After a while, some bentonite particles began to erode from the stretched lower geotextile, usually causing an increase of up to one order of magnitude in permittivity at the beginning of internal erosion. Higher deformation of the lower geotextile component of the GCL was detected when compared with the deformation of the upper geotextile. As time passed, more bentonite particles were extruded through the geotextile. Finally, an increase up to three orders of magnitude in permittivity was measured for all of the GCLs that experienced hydraulic failure. This behavior is similar to the test results reported by Rowe and Orsini (2003), where hydraulic conductivity of the GCL decreased almost one order of magnitude during the first 175 hours of permeation under a 10 m hydraulic head. Then, the hydraulic conductivity began to increase very slowly during the next 155 hours of permeation. Finally, hydraulic conductivity increased drastically within a couple of hours. The measured increase in hydraulic conductivity that

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 157–169

Barriers in Freshwater Reservoirs Table 3. Permittivity test results.

Specimen Name

Hydraulic Head at Failure (m)

GCL-1-NW-D2 GCL-1-W-D2 GCL-1-NW-D1.5 GCL-1-W-D1.5 GCL-1-NW-D1 GCL-1-W-D1 GCL-1-NW-D0.5 GCL-1-W-D0.5 GCL-2-NW-D2 GCL-2-W-D2 GCL-2-NW-D1.5 GCL-2-W-D1.5 GCL-2-NW-D1 GCL-2-W-D1 GCL-2-NW-D0.5 GCL-2-W-D0.5 GCL-3-NW-D2 GCL-3-W-D2 GCL-3-NW-D1.5 GCL-3-W-D1.5 GCL-3-NW-D1 GCL-3-W-D1 GCL-3-NW-D0.5 GCL-3-W-D0.5 GCL-4-W-D2 GCL-4-W-D1.5 GCL-4-W-D1

35 30 No failure 45 No failure No failure No failure No failure 20 10 35 15 50 30 No failure No failure 20 5 35 15 50 25 No failure No failure No failure No failure No failure

Confining Pressure (KPa)

Influent Pressure (KPa)

Effluent Pressure (KPa)

Permittivity at 50 m Hydraulic Head (1/s)

Permittivity Before Internal Erosion (1/s)

550 550 550 550 550 550 550 550 550 550 550 550 550 550 550 550 550 550 550 550 550 550 550 550 550 550 550

530 530 530 530 530 530 530 530 530 530 530 530 530 530 530 530 530 530 530 530 530 530 530 530 530 530 530

186.7 235.7 39.5 88.6 39.5 39.5 39.5 39.5 333.8 431.9 186.7 382.9 39.5 235.7 39.5 39.5 333.8 481 186.7 382.9 39.5 284.8 39.5 39.5 39.5 39.5 39.5

— — 3.04 6 10−10 — 9.95 6 10−11 1.32 6 10−10 4.29 6 10−11 5.12 6 10−11 — — — — — — 3.97 6 10−11 1.08 6 10−10 — — — — — — 6.77 6 10−10 8.69 6 10−10 2.05 6 10−10 8.92 6 10−11 6.53 6 10−11

5.46610−10 8.39 6 10−10 — 4.92 6 10−10 — — — — 1.25 6 10−9 1.67 6 10−9 3.85 6 10−10 3.50 6 10−9 1.32 6 10−10 5.59 6 10−10 — — 1.11 6 10−8 9.72 6 10−9 8.61 6 10−9 1.13 6 10−8 1.18 6 10−9 5.54 6 10−9 — — — — —

caused the failure of the GCL was more than two orders of magnitude (Rowe and Orsini, 2003). Based upon the hydraulic conductivity test results of specimens GCL-1, three of the eight tested GCL-1 specimens experienced internal erosion as shown in Figure 3: GCL-1-W-D2 failed under 30 m of hydraulic head, GCL-1-NW-D2 failed under 35 m of hydraulic head, and GCL-1-W-D1.5 failed under 45 m of hydraulic head. The measured permittivity of these three specimens was more than four orders of magnitude higher at the end of the tests. However, internal erosion did not occur for GCL-1-NW-D1.5, GCL1-W-D1, GCL-1-NW-D1, GCL-1-W-D0.5, or GCL1-NW-D0.5, even under 50 m of hydraulic head. In the hydraulic conductivity test results of GCL1-NW-D2, the permittivity began to increase after almost 175 hours, as shown in Figure 3. The rate of increase in permittivity remained constant for approximately 15 hours. Then, the rate of increase increased and remained almost constant for another 30 hours. Subsequently, a sudden and much higher increase in permittivity was measured within 1 to 2 hours. This sudden increase was higher than two and a half orders of magnitude. The permittivity (Ψ) was 2.30 6 10−9 1/

s at the beginning of the test, increasing to 2.29 6 10−5 1/s at the end of the test. Based upon the hydraulic conductivity test results of specimens GCL-2, six of the eight GCL-2 specimens experienced internal erosion, as shown in Figure 4. The measured permittivity of these specimens was four to five orders of magnitude higher at the end of the tests. Internal erosion did not occur for GCL-2W-D0.5 and GCL-2-NW-D0.5, even under 50 m of hydraulic head. Based upon the hydraulic conductivity test results of specimens GCL-3, six of the eight GCL-3 specimens experienced internal erosion, and the permittivity values increased by several orders of magnitude, as shown in Figure 5. Again, internal erosion did not occur in the case of the smallest void diameter (0.5 cm) when testing the woven and non-woven geotextiles up to a hydraulic head of 50 m. None of the three GCL-4 specimens experienced internal erosion, even under 50 m of hydraulic head. The measured permittivity of these specimens was approximately half an order of magnitude lower after almost 250 hours, as shown in Figure 6. The decrease in permittivity was attributed to the increase in

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 157–169

163

Ozhan and Guler

Figure 3. Permittivity (Ψ) vs. elapsed time (t) graph for GCL-1 (the values given in parentheses are the hydraulic heads).

seepage stresses and decrease in void ratio of the bentonite when the hydraulic head was increased. The zones of the GCL facing the voids on the perforated base pedestal exhibited a convex geometry and, therefore, were called convex zones, as shown in Figure 7a and b. The zones where the GCL was in contact with the solid parts of the perforated base pedestal were called flat zones. As shown in Figure 7a, a large amount of bentonite was eroded through the openings of the woven geotextile. Most of the high increase in

measured permittivity was due to the internal erosion that occurred through the convex zones of the GCL. The slit-film tapes of the woven geotextile in the convex zones of GCL-1-W-D2 were also damaged and torn, as shown in Figure 7a. Similar behavior was observed in the non-woven geotextiles (Figure 7b). This shows that the GCL failed in the convex zones, where internal erosion was the most severe. Specimens GCL-4-W-D1.5 and GCL-2-NW-D0.5 did not experience internal erosion. The appearance

Figure 4. Permittivity (Ψ) vs. elapsed time (t) graph for GCL-2 (the values given in parentheses are the hydraulic heads).

164

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 157–169

Barriers in Freshwater Reservoirs

Figure 5. Permittivity (Ψ) vs. elapsed time (t) graph for GCL-3 (the values given in parentheses are the hydraulic heads).

of these specimens after the termination of the hydraulic conductivity tests is shown in Figure 8a and b, and according to Figure 8a and b, the depth of the convex zones was smaller, and the slit-film tapes on the woven geotextile were not damaged in either.

using four uniform hole sizes. Also, the effects of the manufacturing process of the GCLs, engineering properties of the geotextile components of the GCLs in contact with the subbase, and the type of bentonite used in the GCLs were investigated. Based upon the test results, the effects of these different parameters on internal erosion of the GCLs are discussed below.

PRACTICAL CONSIDERATIONS Ozhan and Guler (2013) have shown that perforated base pedestals could simulate rounded uniform gravel subgrades almost perfectly for testing internal erosion. In their study, only two different perforated base pedestals with uniform 15 mm and 10 mm holes were used. In our study, the effect of the void size of the subbase material on internal erosion was modelled by

Effect of the Void Size of the Perforated Base Pedestal A question that must be addressed is the loss of bentonite in the convex zones. Test results show that the failure of the carrier geotextile did not occur immediately after the application of the hydraulic head. The carrier geotextile was the geotextile component of the

Figure 6. Permittivity (Ψ) vs. elapsed time (t) graph for GCL-4 (the values given in parentheses are the hydraulic heads).

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 157–169

165

Ozhan and Guler

Figure 7. (a) Woven geotextile in contact with a perforated base pedestal of 2-cm void size. (b) Non-woven geotextile in contact with a perforated base pedestal of 1.5 cm void size.

GCL placed at the bottom during the manufacturing process. Bentonite was sandwiched between the lower, carrier geotextile and the upper, cover geotextile. The permittivity of the GCL started to increase slowly and after 25 to 30 hours, increasing approximately one order of magnitude. Our interpretation is that this increase in permittivity occurred because the bentonite particles began to erode. As a consequence of this loss of bentonite particles, the configuration of the GCL changed, and excessive deformation occurred in the convex zones of the geotextile. This caused an increase in the rate of permittivity and, eventually, complete failure. Our opinion is that the failure was initiated by internal erosion, which caused the initial loss of bentonite particles, and the final failure occurred due to the failure of the carrier geotextile. There is a direct relation between internal erosion and the void size of the subbase material underlying

166

Figure 8. (a) Woven geotextile in contact with a perforated base pedestal of 1.5 cm void size. (b) Non-woven geotextile in contact with a perforated base pedestal of 0.5 cm void size.

a GCL. When the void size increased, the depth of the convex zones also increased. As a result, a greater amount of bentonite could erode through the openings and damaged regions in the convex zones, and internal erosion could occur under lower hydraulic heads. GCLs tested over the perforated base pedestal with a uniform void diameter of 5 mm did not experience internal erosion, even under a hydraulic head of 50 m. This indicates that internal erosion becomes less critical as the grain size of the natural soil, and its void size, decreases. In this study, it was shown that placing GCLs over gravel particles with void diameters of 5 mm or smaller would not cause internal erosion even under hydraulic heads of 50 m. Thus, it can be concluded that if the subbase upon which the GCL is placed consists of coarser particles, internal erosion can be prevented

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 157â&#x20AC;&#x201C;169

Barriers in Freshwater Reservoirs

Figure 9. Hydraulic head at failure and void diameter of perforated base pedestal.

by placing a sand layer having a nominal thickness of 5 cm above the gravel subgrade. In order to prevent the loss of sand particles into the voids in the gravel, either an intermediate geotextile or a sand-mat (sand sandwiched between geotextiles) can be placed over the gravel subbase. Effect of the Manufacturing Process of GCLs The performance of GCL-1 against internal erosion was better than that of specimens GCL-2 and GCL-3. Specimens GCL-1-W-D2, GCL-1-NW-D2, and GCL1-W-D1.5 were the only GCL-1 specimens that experienced internal erosion, whereas all of the GCL-2 and GCL-3 specimens tested over the perforated base pedestal with uniform void diameters of 20, 15, and 10 mm experienced internal erosion. Hydraulic heads that caused internal erosion on GCL-1 specimens were also higher than those on GCL-2 and GCL-3 specimens when the perforated base pedestal beneath the GCLs had the same void size. The results of the permittivity tests in which GCL-1, GCL-2, and GCL-3 specimens experienced internal erosion are summaâ&#x20AC;? rized in Figure 9. As shown in Figure 9, the needlepunched specimens (GCL-1) performed better, indicating that it contributed to the prevention of internal erosion under high hydraulic heads. Effect of the Geotextile Component of GCLs The geotextile type played a significant role in internal erosion. Among the tests conducted with the woven geotextile side facing the perforated base pedestal, GCL-4 performed the best. The main reason for this was the higher tensile strength of the woven

geotextile component (W2), which prevented the formation of defects on the slit-film tapes. Bulging of the carrier geotextile into the voids, which caused the formation of the convex zones, might have contributed to the failure of the GCL. Sample W2, with a higher tensile strength, seemed to perform the best and appeared to have the least amount of bulging into the voids simulated by the base pedestal. As shown in Figure 9, GCL-1, GCL-2, and GCL-3 specimens tested with their woven geotextile component (W1) over the perforated base pedestal experienced internal erosion under lower hydraulic heads than the GCL specimens tested with non-woven geotextiles. It can be speculated that one of the reasons for the better performance of the non-woven geotextile (N1) than the woven geotextile (W1) was the capability of N1 to deform extensively without losing its mechanical properties. Probably, the rigid slit-film tapes of the woven geotextile (W1) got damaged and caused a significant increase in the opening size in the convex zones more easily than those of the non-woven geotextile. The apparent opening size of the geotextile in contact with the perforated base pedestal can also influence internal erosion. The woven geotextile (W1) had an opening size of 0.425 mm, and the non-woven geotextile (N1) had an opening size of 0.212 mm (ASTM D4751, 2012). This indicates that using geotextiles with smaller opening sizes can be beneficial. The comparison between the behavior of the GCLs with N1 and W2 facing the perforated base pedestal might be considered to contradict the above statement, because W2 performed much better. However, our opinion is that the higher tensile strength of W2 prevented the openings of the geotextile from enlarging. Effect of the Bentonite Component of GCLs Although calcium bentonite has a higher hydraulic conductivity than sodium bentonite (Gleason et al., 1997), both GCL-2 and GCL-3 specimens performed similarly against internal erosion. This indicates that the bentonite type does not influence internal erosion significantly for the un-reinforced GCLs tested in this study. Furthermore, at the beginning of the tests, a decrease in permittivity was measured for all of the GCL specimens composed of either sodium bentonite or calcium bentonite that experienced internal erosion. This decrease was attributed to the increase in seepage stresses caused by the increase in hydraulic head, which resulted in a decrease in the void ratio of the bentonite (Rowe and Orsini, 2003). However, to come to a decisive conclusion, further studies with permeation periods of several years should be performed.

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 157â&#x20AC;&#x201C;169

167

Ozhan and Guler

CONCLUSIONS The diameter of the voids of the perforated base pedestal plays a significant role in internal erosion. As the void size increased, internal erosion occurred under lower hydraulic heads. For the test configurations chosen and the GCLs used in this study, no internal erosion occurred for the base pedestal with a void diameter of 5 mm. The manufacturing process (reinforced with needlepunching versus un-reinforced) plays a significant role in internal erosion. The performance of the GCLs that were manufactured by needle-punching was better than that of the GCLs that were assembled in the laboratory without needle-punching, indicating that a proper bonding of the two geotextile components through needle-punching was beneficial. The engineering properties of the geotextile components also play a significant role in internal erosion. The GCL with the woven geotextile (W2) facing the perforated base pedestal performed the best. This was due to the high tensile strength of W2, which helped to prevent excessive deformation. However, the performance of the non-woven geotextile (N1) was better than that of the woven geotextile (W1) against internal erosion. This result was interpreted as being due to the capacity of the non-woven geotextile to deform extensively without significantly changing its mechanical properties, e.g., the apparent opening size. The smaller initial opening size of the non-woven geotextile was also instrumental in preventing internal erosion. The type of bentonite (sodium versus calcium bentonite) used in the un-reinforced GCLs did not significantly affect internal erosion. In conclusion, GCLs have been shown to be used successfully as a barrier if the above conditions are followed and the same materials are used. ACKNOWLEDGMENTS The authors are appreciative of the financial support provided by the Scientific Research Project Foundation of Turkey (Project No. 07HA401). The authors also sincerely thank Dr. Nigel Webb and Dr. Susan English of the Colloid Environmental Technologies Company (CETCO) for providing GCL-1 specimens and the geotextile and bentonite components of GCL-1 specimens in order to use them for assembling specimens GCL-2, GCL-3, and GCL-4. REFERENCES ASTM STANDARD D4318, 2010, Standard Test Methods for Liquid Limit, Plastic Limit, and Plasticity Index of Soils: ASTM International, West Conshohocken, PA.

168

ASTM STANDARD D4595, 2011, Standard Test Method for Tensile Properties of Geotextiles by the Wide-Width Strip Method: ASTM International, West Conshohocken, PA. ASTM STANDARD D4751, 2012, Standard Test Method for Determining Apparent Opening Size of a Geotextile: ASTM International, West Conshohocken, PA. ASTM STANDARD D5199, 2012, Standard Test Method for Measuring the Nominal Thickness of Geosynthetics: ASTM International, West Conshohocken, PA. ASTM STANDARD D5261, 2010, Standard Test Method for Measuring Mass per Unit Area of Geotextiles: ASTM International, West Conshohocken, PA. ASTM STANDARD D5887, 2009, Standard Test Method for Measurement of Index Flux through Saturated Geosynthetic Clay Liner Specimens Using a Flexible-Wall Permeameter: ASTM International, West Conshohocken, PA. ASTM STANDARD D5890, 2011, Standard Test Method for Swell Index of Clay Mineral Component of Geosynthetic Clay Liners: West Conshohocken, PA. ASTM STANDARD D5993, 2009, Standard Test Method for Measuring Mass per Unit of Geosynthetic Clay Liners: ASTM International, West Conshohocken, PA. BAENA, C. M. AND TOLEDO, M. A., 2014, An experimentally verified model of the seepage progress due to dissolution of soluble particles in foundations subject to intergranular flow: Environmental Earth Sciences, Vol. 72, pp. 3369–3382. BENAHMED, N. AND BONELLI, S., 2012, Investigating concentrated leak erosion behaviour of cohesive soils by performing hole erosion tests: European Journal of Environmental and Civil Engineering, Vol. 16, No. 1, pp. 43–58. BENDAHMANE, F.; MAROT, D.; ROSQUOET, F.; AND ALEXIS, A., 2006, Characterization of internal erosion in sand kaolin soils: Revue Européenne de Génie Civil, Vol. 10, No. 4, pp. 505–520. BENSON, C. H. AND MEER, S. R., 2009, Relative abundance of monovalent and divalent cations and the impact of desiccation on geosynthetic clay liners: Journal of Geotechnical and Geoenvironmental Engineering, Vol. 135, No. 3, pp. 349–358. BOUAZZA, A., 2002, Geosynthetic clay liners: Geotextiles and Geomembranes, Vol. 20, pp. 3–17. BRACHMAN, R. W. I.; RENTZ, A.; ROWE, R. K.; AND TAKE, W. A., 2014, Classification and quantification of downslope erosion from a GCL when covered only by a black geomembrane: Canadian Geotechnical Journal, Vol. 52, No. 4, pp. 395–412, doi: 10.1139/cgj-2014-0241. BURNS, B. AND GHATAORA, G. S., 2007, Internal erosion of kaolin. In Puppala, A. J.; Hudyma, N.; and Likos, W. J. (Editors), GSP 162, Problematic Soils and Rocks and In Situ Characterization, GeoDenver 2007: American Society of Civil Engineers, Denver, CO, pp. 1–8. CETCO, 2007, BENTOMAT SS100 Certified Properties Technical Data Sheet: CETCO, Hoffman Estates, IL. CHANG, D. S. AND ZHANG, L. M., 2013, Critical hydraulic gradients of internal erosion under complex stress states: Journal of Geotechnical and Geoenvironmental Engineering, Vol. 139, No. 9, pp. 1454–1467. DICKINSON, S. AND BRACHMAN, R. W. I., 2010, Permeability and internal erosion of a GCL beneath coarse gravel: Geosynthetics International, Vol. 17, No. 3, pp. 112–123. EN ISO 10319, 2008, Geosynthetics—Wide-Width Tensile Test, 2nd ed.: International Organization for Standardization, Geneva, Switzerland. FELL, R.; WAN, C. F.; CYGANIEWICZ, J.; AND FOSTER, M., 2003, Time for development of internal erosion and piping in embankment dams: Journal of Geotechnical and Geoenvironmental Engineering, Vol. 129, No. 4, pp. 307–314.

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 157–169

Barriers in Freshwater Reservoirs FOX, P. J.; DE BATTISTA, D. J.; AND MAST, D. G., 2000, Hydraulic performance of geosynthetic clay liners under gravel cover soils: Geotextiles and Geomembranes, Vol. 18, pp. 179–201. GLEASON, M. H.; DANIEL, D. H.; AND EYKHOLT, G. R., 1997, Calcium and sodium bentonite for hydraulic containment applications: Journal of Geotechnical and Geoenvironmental Engineering, Vol. 123, No. 5, pp. 438–445. GREENE, B. H.; CROCK, J.; MOSKOVITZ, L.; AND PREMOZIC, J. W., 2010, Evaluation of seepage, internal erosion, and remedial alternatives for East Branch Dam, Elk County, Pennsylvania: Environmental & Engineering Geoscience, Vol. 16, August 2010, pp. 229–243. HORNSEY, W. P.; SCHEIRS, J.; GATES, W. P.; AND BOUAZZA, A., 2010, The impact of mining solutions/liquors on geosynthetics: Geotextiles and Geomembranes, Vol. 28, No. 2, pp. 191–198. JACOBSON, T., 2013, An Analysis on Soil Properties on Predicting Critical Hydraulic Gradients for Piping Progression in Sandy Soils: M.Sc. Thesis, Utah State University, Logan, UT, 38 p. KANG, J. B. AND SHACKELFORD, C. D., 2010, Consolidation of a geosynthetic clay liner under isotropic states of stress: Journal of Geotechnical and Geoenvironmental Engineering, Vol. 136, No. 1, pp. 253–259. KOERNER, R. M., 2005, Designing with Geosynthetics, 5th ed.: Prentice Hall, Upper Saddle River, NJ, 816 p. LI, M., 2008, Seepage Induced Instability in Widely Graded Soils: Ph.D. Thesis, University of British Colombia, Vancouver, B. C., Canada, 300 p. MCCOOK, D. K., 2007, Filter Diaphragms: Part 628 Dams— National Engineering Handbook: U.S. Department of Agriculture, National Resources Conservation Service, Washington, D.C. 58 p. MURESAN, B.; SAIYOURI, N.; GUEFRECH, A.; AND HICHER, P. Y., 2011, Internal erosion of chemically reinforced granular materials: A granulometric approach: Journal of Hydrology, Vol. 411, pp. 178–184. ORSINI, C. AND ROWE, R. K., 2001, Testing procedure and results for the study of internal erosion of geosynthetic clay liners: In Proceedings of the Geosynthetics 2001: Industrial Fabrics Association International, Portland, OR, pp. 189–201. OZHAN, H. O., 2011, Internal Erosion of Geosynthetic Clay Liners under High Hydraulic Heads: Ph.D. Thesis, Bogazici University, Istanbul, Turkey, 222 p. OZHAN, H. O. AND GULER, E., 2013, Use of perforated base pedestal to simulate the gravel subbase in evaluating the internal

erosion of geosynthetic clay liners: ASTM Geotechnical Testing Journal, Vol. 36, No. 3, pp. 418–428. RODRIGUEZ, V.; GUTIERREZ, F.; GREEN, A. G.; CARBONEL, D.; HORSTMEYER, H.; AND SCHMELZBACH, C., 2014, Characterizing sagging and collapse sinkholes in a mantled karst by means of ground penetrating radar (GPR): Environmental & Engineering Geoscience, Vol. 20, May 2014, pp. 109–132. ROWE, R. K., AND ABDELATTY, K., 2012, Effect of a calcium-rich soil on the performance of an overlying GCL: Journal of Geotechnical and Geoenvironmental Engineering, Vol. 138, No. 4, pp. 423–431. ROWE, R. K.; ASHE, L. E.; TAKE, W. A.; AND BRACHMAN, R. W. I., 2014a, Factors affecting the down-slope erosion of bentonite in a GCL: Geotextiles and Geomembranes, Vol. 42, No. 5, pp. 445–456, doi:10.1016/j.geotexmem.2014.07.002. ROWE, R. K.; BRACHMAN, R. W. I.; TAKE, W. A.; ASHE, L. E.; AND RENTZ, A., 2014b, Field and laboratory observations of downslope bentonite migration in exposed composite liners. In Bouazza, A.; Brown, B.; and Yuen. S (Editors), 7th International Conference on Environmental Geotechnics: Curran Associates, Inc. Melbourne, Australia, pp. 1–18. ROWE, R. K. AND ORSINI, C., 2003, Effect of GCL and subgrade type on internal erosion in GCLs under high gradients: Geotextiles and Geomembranes, Vol. 21, No. 1, pp. 1–24. SHAN, H. Y. AND CHEN, R. H., 2003, Effect of gravel subgrade on hydraulic performance of geosynthetic clay liner: Geotextiles and Geomembranes, Vol. 21, No. 6, pp. 339–354. STAM, T. G., 2000, Geosynthetic clay liner field performance. In Proceedings of the 14th GRI Conference (Hot Topics in Geosynthetics): Geosynthetic Institute, Las Vegas, NV, pp. 242–254. VON MAUBEUGE, K. P. AND HEERTEN, G., 1994, Needle-punched geosynthetic clay liners. In Proceedings of the 8th GRI Conference: Geosynthetic Research Institute, Philadelphia, PA, pp. 129–207. ZHANG, X.; DONG, J.; HUANG, Z.; NIE, X.; WONG, H.; WANG, J.; SUN, W.; LEO, C.; AND WANG, J., 2015, Study of soil structures strength and stiffness loss based on thermodynamics and continuum mechanics: Environmental Earth Sciences, Vol. 73, No. 8, pp. 4143–4149. ZOHARY, T. AND OSTROVSKY, I., 2010, Ecological impacts of excessive water level fluctuations in stratified freshwater lakes: Inland Waters, Vol. 1, pp. 47–59.

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 157–169

169

Book Review Geomodels in Engineering Geology—An Introduction

(Peter Fookes, Geoff Pettifer, and Tony Waltham) Review by: Richard Jackson 11 Venus Crescent, Geofirma Engineering Ltd., Heidelberg, Ontario N0B 2M1, Canada

This is an unusual and welcome contribution to the engineering geology literature. The preamble outlines the concept of a “geomodel” and its depiction in simplified block diagrams is useful for the engineering geologist and geotechnical engineer. The geomodel approach was developed in the UK because of the popularity of Fookes’ Glossop Lecture before the Engineering Geology Group of the Geological Society of London, which was subsequently published as Fookes (1997). It is written in the British tradition of engineering geomorphology. The purpose of the book is “to help engineers visualize the three-dimensional geology and to act as a quick introduction to new or unfamiliar ground or environments for geologists and engineers.” The block diagrams are to be employed as “springboards” in the design and construction of geotechnical models, which, perhaps, indicates the authors’ desire to get the engineering geologists on site well before geoengineers like me develop too many conclusions! My colleague, Rob Sengebush of INTERA, Inc., in Albuquerque, NM, developed one such geomodel of a Pleistocene alluvial-fan aquifer cutting through brackish Eocene sediments in the heart of San Diego, CA, which we published in this journal recently (Sengebush et al., 2015, Figure 5). However, this volume is unusual in several ways. The book itself is laid out in landscape, not portrait, style to accommodate the numerous illustrations. It is divided into five parts: 1) underlying factors: climate and geology; 2) near-surface ground changes; 3) basic geological environments influencing engi‐ neering; 4) ground investigations; and 5) case histories and some basic ground characteristics and properties. An appendix, which defines the various geotechnical problems associated with different types of soils, a bibliography of textbooks, and a list of the locations of the numerous photographs complete the text.

Each of the parts (chapters really) comprises several sections, each of which contains a block diagram drawn by one of the authors (Pettifer), text, and summary tables, followed by numerous color photographs from across the world. Thus, Part 3 begins with a section led by a block diagram of glacial environments, which is supplemented by three pages of text and tables on “glacial landforms” and “engineering in glacial environments.” The section is completed by a dozen photographs of glaciers, moraines, drumlins, and tills—adequate for Brits perhaps but not enough, I think, for most Canadians, Scandinavians, and many Americans who live their lives on these materials. The format is instructive, although I would prefer having fewer and larger photographs so that the details are more evident. This same format is followed in the other sections of Part 3, i.e., periglacial environments; temperate environments such as the Mediterranean; relict periglacial environments in southern Britain; hot desert environments; savanna environments; hot, wet tropical environments; and mountain environments. The photographs are excellent, and many of the block diagrams are superb. One might have hoped that the authors would have digested some modern hydrogeological thinking because the section on groundwater and “permeability” is poor. Nearly 40 years after Freeze and Cherry (1979) laid out the basic principles of modern hydrogeology, the authors show little appreciation of shallow groundwater flow systems and their relation to the topographic features. Reading about the “coefficient of permeability,” a term dead here since the 1970s, rather than “hydraulic conductivity” is disappointing. This is a very British book written for a British audience and might not be appreciated by many in North America. Although probably one quarter of the photographs are from the Americas, the authors have not cited the North American scientific literature to any significant degree—just 3 of ,70 references are by North American authors or institutions. This is disappointing when they have cast their photographic

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 171–172

171

Book Review

net so wide and it undermines the adoption here of this useful book. Nevertheless, for those of us with a working interest in geomodels, this is a welcome monograph illustrated with some truly splendid block diagrams. Hopefully, this book will serve to keep alive this pictorial tradition. REFERENCES FOOKES, P., 1997, First Glossop Lecture: Geology for engineers: The geological model, prediction and performance: Quarterly Journal of Engineering Geology, Vol. 30, pp. 293–424.

172

FOOKES, P.; PETTIFER, G.; AND WALTHAM, T., 2015, Geomodels in Engineering Geology—An Introduction: Whittles Publishing, Taylor & Francis, Caithness, Scotland, UK. Available in North America through CRC Press, Boca Raton, Florida, 176 p. ISBN 9781498740043, Paperback US$70. https://www.crcpress.com/Geomodels-in-EngineeringGeology-An-Introduction/Fookes-Pettifer-Waltham/ 9781498740043 FREEZE, R. A. AND CHERRY, J. A., 1979, Groundwater: PrenticeHall Inc., Englewood Cliffs, NJ. SENGEBUSH, R. M.; HEAGLE, D. J.; AND JACKSON, R. E., 2015, The late Quaternary history and groundwater quality of a coastal aquifer, San Diego, California: Environmental & Engineering Geoscience, Vol. XVIII, No. 4, pp. 249–275.

Environmental & Engineering Geoscience, Vol. XXII, No. 2, May 2016, pp. 171–172