Environmental & Engineering Geoscience FEBRUARY 2016
VOLUME XXII, NUMBER 1
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 Photo of the tunnel at the Yucca Mountain Nuclear waste disposal site. Photo by John Stuckless, USGS. See article on page 1.
Environmental & Engineering Geoscience Volume 22, Number 1, February 2016 Table of Contents
1
The Road to Yucca Mountain—Evolution of Nuclear Waste Disposal in the United States John S. Stuckless and Robert A. Levich
27
A Comparison of Surface- and Standard Penetration Test-Derived Shear-Wave Velocity Peter J. Hutchinson and Maggie H. Beird
37
Reliable Soil Property Maps over Large Areas: A Case Study in Central Italy Giulia Fanelli, Diana Salciarini, and Claudio Tamagnini
53
Investigation of Air Bulging Beneath Geomembranes Used as a Liner for the Datun Reservoir Xue-shan Cao, Jun-ping Yuan, Zong-ze Yin, Gui-lin He, and Ying-hao Liu
67
Assessment of the CO2 Storage Potential in the Deep Saline Formation of Offshore Bohai Basin, China Guanbao Li, Chuanlin Huo, Tianyun Su, and Baohua Liu
81
The Integration of Data Review, Remote Sensing and Ground Survey for a Regional-Level Karst Assessment Robert Kenneth Denton Jr., Ashley Hogan, and Ronald Drew Thomas
The Road to Yucca Mountain—Evolution of Nuclear Waste Disposal in the United States JOHN S. STUCKLESS U.S. Geological Survey, Emeritus, Denver Federal Center, MS 908, Denver, CO 80225
ROBERT A. LEVICH U.S. Department of Energy, Retired, 405 Norwood Lane, Las Vegas, NV 89107
Key Terms: Hazardous Waste, Waste, Nuclear, Geopolitical
ABSTRACT The generation of electricity by nuclear power and the manufacturing of atomic weapons have created a large amount of spent nuclear fuel and high-level radioactive waste. There is a world-wide consensus that the best way to protect mankind and the environment is to dispose of this waste in a deep geologic repository. Initial efforts focused on salt as the best medium for disposal, but the heat generated by the radioactive waste led many earth scientists to examine other rock types. In 1976, the director of the U.S. Geological Survey (USGS) wrote to the U.S. Energy Research and Development Administration (ERDA), predecessor agency of the U.S. Department of Energy (DOE), suggesting that there were several favorable environments at the Nevada Test Site (NTS), and that the USGS already had extensive background information on the NTS. Later, in a series of communications and one publication, the USGS espoused the favorability of the thick unsaturated zone. After the passage of the Nuclear Waste Policy Act (1982), the DOE compiled a list of nine favorable sites and settled on three to be characterized. In 1987, as the costs of characterizing three sites ballooned, Congress amended the Nuclear Waste Policy Act directing the DOE to focus only on Yucca Mountain in Nevada, with the proviso that if anything unfavorable was discovered, work would stop immediately. The U.S. DOE, the U.S. DOE national laboratories, and the USGS developed more than 100 detailed plans to study various earth-science aspects of Yucca Mountain and the surrounding area, as well as materials studies and engineering projects needed for a mined geologic repository. The work, which cost more than 10 billion dollars and required hundreds of man-years of work, culminated in a license application submitted to the U.S. Nuclear Regulatory Commission (NRC) in 2008.
INTRODUCTION Since the dawn of the Atomic Age more than 70 years ago, the operation of nuclear power plants as well as the development and manufacture of nuclear weapons have resulted in considerable amounts of radioactive industrial by-products. These by-products are commonly referred to as nuclear or radioactive waste. Some of the components in this waste are in low concentrations and consist of radioisotopes with short half-lives. These “low-level” wastes can be disposed of in shallow surface trenches. Other types of radioactive wastes (including transuranic waste) contain radioisotopes in higher concentrations, some of which have half-lives of thousands of years or longer. In order to protect public health and safety, as well as Earth’s environment, these “high-level” wastes must be disposed of in a permanent manner, such that they can be isolated for tens to hundreds of thousands of years. World-wide, there has been a long-standing scientific consensus (National Academy of Sciences, 1957, 1990; Nuclear Energy Agency, 1995) that the best method for permanent disposal of high-level nuclear waste (HLW) is deep geologic disposal. A good analogy as to why deep geologic disposal is considered a strategy that can protect public health and safety and the environment is the comparison of geologic repositories to mineral deposits, especially metallic ore deposits. In concept, deep geologic disposal is capable of isolating nuclear waste from the surface environment in a manner similar to ore deposits. During much of geo‐ logic time, metallic ore deposits have been formed and isolated in diverse geologic host rocks and under a variety of geologic conditions. These deposits have remained isolated for millions to billions of years until exposed at Earth’s surface by tectonic activity and (or) erosion. Ore deposits that are not exposed at the surface are generally very difficult to discover, in part because the ore metals generally do not migrate far from the original site of deposition. Thus, experience
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 1–25
1
Stuckless and Levich
with ore deposits leads to the conclusion that various geologic media and tectonic settings are capable of successfully isolating nuclear waste and protecting health, safety, and the environment. The field of geochronology provides another set of analogues showing long-term isolation of radioactive material from the environment. Most geochronom‐ eters measure a radioactive parent isotope present in a rock or minerals separated from a rock and the stable daughter isotope. If neither parent nor daughter isotope has moved, the age of the material can be calculated using a known decay constant (cf. Faure, 1986). Independent evidence for ages and use of multiple decay systems that have differing half-lives and geochemical properties show closed-system behavior to be very common. A final argument for geologic disposal of radioactive waste can be seen in the long-term immobility of radioactive elements in whole-rock systems. Evaluation of the uranium-thorium-lead isotope systematics for granite from the Granite Mountains of Wyoming showed that these elements remained immobile from 2,600 Ma until about 1,400 Ma, when the granite was subjected to an ill-defined metamorphic event that mobilized 10 to 45 percent of the uranium (Stuckless and Nkomo, 1978). The uranium then remained immobile until the Laramide event (about 60 Ma), which brought the granite close enough to the surface to allow groundwater to remove as much as 90 percent of the uranium (Stuckless and Nkomo, 1978). Analyses of the uranium series decay chain show that mobilization of uranium continued until at least the last 300,000 years. Other studies have also shown long-term immobility of uranium and its radioactive daughter products in granite (Nkomo et al., 1978; Stuckless et al., 1981, 1982, 1985; and Smellie and Stuckless, 1985). SOURCES OF MATERIALS CONSIDERED FOR DISPOSAL Nuclear waste destined for disposal in a geologic repository consists of spent nuclear fuel (SNF) from commercial, research, and naval reactors, and vitrified re-processed HLW that mainly originates from production of nuclear weapons (Figure 1). The largest quantity of waste destined for disposal in a geologic repository is commercial SNF, which consists of fuel assemblies from civilian nuclear power plants containing enriched uranium fuel pellets that have completed their useful life producing electricity. SNF is characterized by intense penetrating radiation and by high heat generation. Approximately 20 percent of our nation’s electricity has been produced at 74 sites in 33 states that host 118 commercial nuclear power reactors, more than 100 of which are still in
2
operation. In 2002, more than 40,000 metric tons of heavy metal (MTHM) were stored in 39 states at 72 commercial reactor sites and storage sites. It is estimated that if the existing nuclear power plants continue to operate for their license periods of 40 years, they would generate about 87,000 MTHM of SNF. If each were granted an additional 10 years of operation by extending their licenses, they could produce a total of about 105,000 MTHM of SNF (U.S. DOE, 2002). The isotopic composition of uranium occurring in nature is 0.7 percent fissionable 235U, 99.3 percent 238 U, and a negligible amount of 234U. Commercial reactors use uranium fuel pellets that have been enriched to between 3 and 5 percent 235U. After undergoing fission in a reactor, the now “spent” fuel contains approximately 95 percent 238U, 1 percent 235U, 1 percent plutonium, and 3 percent other elements, including activation products in the transuranic series (e.g., neptunium, americium, and curium) and a number of fission products (e.g., isotopes of zirconium, cesium, and some rare earth elements) (Brookins, 1985; U.S. DOE, 1998). Material referred to as HLW is next in abundance to SNF and is the radioactive by-product resulting from the re-processing of commercial SNF. Like SNF, HLW is characterized by intense penetrating radiation; however, because of its isotopic composition, it has a somewhat lower heat-generation rate. High-level liquid waste from re-processing of spent fuel or other nuclear material is mixed with a variety of chemicals (e.g., borates) and then is heated to high temperatures and fused into solid vitrified waste products, such as borosilicate glass. As most of the uranium and plutonium has been removed, the residues consist mainly of fission products and other actinides. However, reprocessing fails to separate a small amount of plutonium and other transuranic elements. Early plans indicated that between 4,000 and 5,000 MTHM of HLW would be sent to the first geologic repository (the license application lists 4,667 MTHM; U.S. DOE, 2008). A third source of waste is U.S. DOE SNF, which consists of fuel removed from naval reactors (those aboard U.S. Navy ships) plus irradiated fuel from weapons production and research reactors, both domestic and foreign. Approximately 2,000–2,500 MTHM of U.S. DOE SNF are destined for the repository (the license application lists 2,333 MTHM; U.S. DOE, 2008). Surplus weapons plutonium is a possible fourth source of waste that may be disposed of in a repository. Some of the plutonium could be placed in the repository as immobilized plutonium ceramic. Alternatively, plutonium can be formulated into a mixed
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 1–25
Deep Geologic Disposal of Nuclear Waste
Figure 1. U.S. Department of Energy illustration showing various types of spent nuclear fuel and high-level radioactive waste (U.S. DOE, 2001).
oxide (MOX 5 uranium plus plutonium) or MOX fuel that could be used in commercial reactors. Transuranic elements (TRU) are those elements that have an atomic number greater than uranium (element 92). TRU wastes are the by-products from fuel-assembly and weapons-manufacturing operations. TRU wastes actually contain some minor amounts of fission products, but the inclusion of the alpha-emitting and long-lived isotopes causes them to be treated as a separate class of waste. This is true even when their activity is less than 10 rem/hr, equivalent to low-level wastes. Most TRU wastes generate very little heat, and most can be handled directly (contact-handled waste). Some transuranic wastes, however, contain higher levels of radioactivity and will require robotic handling (remote-handled waste). The United States is disposing of its TRU wastes in a bedded salt formation that underlies the Waste Isolation Pilot Plant (WIPP) near Carlsbad in southeastern New Mexico (Figure 2).
In 1980, Congress authorized WIPP for disposal of transuranic waste. The WIPP repository lies 650 m below the ground surface. Construction began in 1981, but it was delayed several times until 1984. The first disposal of nuclear wastes at WIPP took place in 1999. During the first 2 years of operation, several thousand shipments of contact-handled TRU wastes were disposed of in the WIPP repository (Patterson and Nelson, 2001). By February 2, 2014, the number of shipments had grown to 11,866, with 14,193,873 miles of loaded trucks (Patterson and Gross, 2014). In 2001, the radioactive wastes described above and destined for disposal in a repository were located at 129 sites in 39 states across the United States (U.S. DOE, 2001; Figure 3). These sites include operating commercial reactors, shutdown commercial reactors, commercial SNF pool storage (away from reactor), commercial dry storage, and U.S. DOE–owned waste. In fact, there were an estimated 161 million Americans
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 1–25
3
Stuckless and Levich
Figure 2. Photograph of surface facilities at the Waste Isolation Pilot Plant (WIPP) site and a truck loaded with transuranic waste.
living within 75 miles (120 km) of these sites (U.S. DOE, 2002). HISTORY OF NUCLEAR WASTE MANAGEMENT Early Search for a Site The problem of managing HLW dates back to 1944, when the first storage tanks for liquid waste were constructed at Hanford, WA (see Congress of the United States, Office of Technology Assessment, 1985, Appendix A-2 which summarizes the major events in the history of waste management from 1944 to 1981). The Atomic Energy Act of 1954 assigned the responsibility for managing SNF from civilian reactors to the U.S. Atomic Energy Commission (AEC) and established the federal responsibility for SNF and HLW. This act also permitted private industry to construct and operate nuclear reactors for generating electricity.
4
In 1955, the AEC requested the National Academy of Science (NAS) to study the problem of disposal of nuclear waste. In 1957, the NAS National Research Council reported that, “… radioactive waste can be disposed of safely in a variety of ways and at a large number of sites in the United States.” The NAS also indicated that, “… the most promising method of disposal of high-level waste … is in salt deposits” (National Academy of Sciences, 1957, p. 3–4). This NAS recommendation was based on disposal of liquid wastes from re-processing, and the assumption of self-sealing properties of salt and fact that the presence of salt was thought to be an indication that there was no water pres‐ ent to dissolve the salt. The NAS also recommended the development of a process to solidify liquid wastes. In 1958, the AEC commissioned the U.S. Geological Survey (USGS) to review U.S. salt deposits. In 1962, the USGS published a report that identified salt deposits in 24 of the 50 states, noted the lower cost of mining salt as opposed to other rock types, and included a section on the plastic flow of salt
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 1–25
Deep Geologic Disposal of Nuclear Waste
Figure 3. Map of the contiguous United States showing the location of 126 sites where high-level radioactive waste is currently stored (modified from U.S. DOE, 2001).
deposits (Pierce and Rich, 1962). From 1962 until 1972, investigations were focused on salt formations as potential host rocks for disposal of radioactive waste. In the late 1960s, the AEC set up a 19 month demonstration project in a salt mine beneath Lyons, KS, using SNF. However, technical problems were uncovered, namely, the discovery of old abandoned wells drilled through the deposit, and these raised serious concerns about the potential suitability of the site. The president of the American Salt Corporation, which operated the adjacent mine, reported to the Kansas Geological Survey that in 1965 they had pumped about 170,000 gallons (643,520 L) of water into their mine as part of an attempt to solution mine their deposit, but they were unable to recover any of the water. W. W. Hambleton, the director of the Kansas Geological Survey, characterized the salt deposits around Lyons, KS as, “like a piece of Swiss cheese and (that) the potential for entrance and circulation of fluids is great” (Walker, 2007, p. 283). By 1972, in addition to the potential technical prob‐ lems, there was intense local opposition to developing the abandoned mine as a geologic repository for nuclear waste. These issues resulted in the cancellation of the project and led the AEC to reevaluate its strategy for geologic disposal. However, the AEC’s focus
remained on salt as the geologic medium of choice for waste disposal (Congress of the United States, Office of Technology Assessment, 1985). Late in 1972, the AEC asked the USGS to evaluate the geohydrologic possibilities of placing HLW in geologic formations, principally other than salt. The final report (Ekren et al., 1974) cited 30 previous reports on the subject and concluded by listing several optimal considerations for a site: (1) hydrologic isolation was paramount, requiring lowpermeability rock and a virtually fault-free site; (2) low seismic risk; (3) low possibility of flooding by rising sea level; (4) low potential hazard for surface water or groundwater in glacial or rainy climates; and (5) low potential for exhumation by erosion. One specific recommendation reads, “The Basin and Range province of the western United States, particularly the Great Basin exclusive of seismic-risk zone 3, [a zone corresponding approximately to the east and west province margins where extension is most active] appears to have potential for mined chambers above the deep water tables in tuff, shale, or argillite” (Ekren et al., 1974, p. 2). The body of the report provides
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 1–25
5
Stuckless and Levich
Figure 4. Diagram illustrating alternatives to geologic disposal in repository considered by the Atomic Energy Commission.
several examples of favorable geologic features at the Nevada Test Site (NTS). While the USGS concentrated its efforts on a mined geologic repository in the contiguous United States, the AEC broadened the scientific basis for its program by examining other means of disposal and waste management. Between 1972 and 1974, the AEC identified concepts and began work on alternative methods of disposal. A variety of concepts was considered, including several methods of deep geologic disposal on land, disposal under the ocean, in polar ice sheets, in space, and via transmutation (Figure 4). These concepts are summarized in Volume 1 of the U.S. DOE’s Final Environmental Impact Statement for the Management of Commercially Generated Radioactive Waste (U.S. DOE, 1980). However, when this evaluation was completed, disposal in a mined geologic repository within the contiguous United States remained the preferred option, and U.S. DOE documented this in a record of decision published in the Federal Register on May 14, 1981 (Federal Register, 1981). Sub-seabed disposal consists of the burial of nuclear waste beneath the floor of the ocean. SNF and (or) vitrified HLW would be sealed inside specially
6
designed canisters and buried within the deep-sea sediments of the abyssal plain in a tectonically stable area, far from plate boundaries. Emplacement in a midplate region would isolate the waste from future tectonic activity for millions of years, and the fine oceanic clays would help isolate the waste from the biosphere and protect marine life. Transport of radionuclides from the buried canisters to the biosphere would be a slow process. Oceanic currents in the thousands of meters of seawater lying above the seafloor would disperse and dilute any radionuclides that leaked from the canisters and migrated upward through the sediments to the ocean floor. Disadvantages included the difficulty of placing waste containers beneath the ocean floor to ensure containment until the waste decayed to acceptably low levels, international sea treaties, and regulations relating to performance confirmation and retrievability. The AEC initiated a program to study sub-seabed disposal in 1973. It was carried out for a number of years, but it was effectively halted in 1988 due to lack of appropriated funds (Nadis, 1996). The Office of Subseabed Disposal Research was established within the U.S. DOE by the Nuclear
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 1–25
Deep Geologic Disposal of Nuclear Waste
Waste Policy Amendments Act and was officially closed in 1996 (Nadis, 1996). The concept of island geologic disposal is the isolation of HLW in a deep geologic repository beneath an island, particularly one that is uninhabited, lacks natural resources, and lies in a remote part of the ocean. Islands have relatively simple hydrologic systems, and leakage of radioactive waters from the repository could be easily detected. An important siting attribute for island disposal is hydrologic separation from the populated continents, minimizing potential radiological health effects. The primary advantage of island disposal is the isolation of the island’s hydrologic system by the surrounding ocean, which would dilute any leakage of radioactive particles transported by groundwater. Drawbacks of this concept include ocean transport of nuclear waste in adverse weather conditions and waste handling and (or) transport through highly populated port cities. In addition, many islands lie at or near tectonic plate boundaries or were formed by mantle plumes (hot spots), and frequent and intense seismic or volcanic activity is not uncommon. Furthermore, groundwater-saturated volcanic rocks are both porous and permeable and may provide pathways for corrosive freshwater and (or) marine waters to interact with waste canisters and for radionuclides to leak into the surrounding environment. Ice-sheet disposal is the concept of isolating nuclear waste beneath the ice sheets covering Greenland or Antarctica. Several concepts have been developed for ice-sheet waste emplacement. (1) The canister containing the nuclear waste is emplaced in a shallow hole, and the heat emitted due to radioactive decay causes the canister to melt its way to the bottom of the ice sheet. (2) Similar to (1), except that an anchor cable is attached to the canister, limiting its descent to a selected depth and permitting retrieval. (3) Nuclear waste is placed in large storage units on the surface of the ice, and the emitted heat causes the storage units to melt their way to the bottom of the ice sheet. The Greenland and Antarctic ice sheets lie in uninhabited areas, are thousands of meters thick, are uniform in extent, and have been stable for long time periods. At depth, ice behaves as a plastic and flows to seal fractures and close openings. The slow movement of the ice caps would ensure isolation of nuclear waste for long time periods. Disadvantages include uncertainties of global and regional climate evolution and the stability of ice sheets in the future, the potential high costs, and safety concerns due to remote locations and adverse weather conditions, and inadequate understanding of ice dynamics and engineering for an ice repository. It was a concern that heat from the nuclear waste might accelerate melting of ice sheets, climate change, and (or) sea-level rise and would
provide pathways for transport of radionuclides to the accessible environment (cf. U.S. NRC, 1983, 10CFR60.2). In addition, the Antarctic Treaty of 1959 specifically prohibited the disposal of nuclear waste in Antarctica. Meeting regulatory requirements for performance confirmation and retrievability were additional concerns. Disposal of nuclear waste in very deep holes requires the placement of waste canisters in drill holes up to 10,000 m deep, well below circulating groundwater and far beneath the accessible environment. At this depth, the waste would effectively be isolated and would not threaten human health. The host rock at these depths would have to retain its character and structural stability under the heat and radiation conditions introduced by the waste. However, the hydrolog‐ ic characteristics and in situ mechanical properties of rocks under high pressure and thermal loads at these depths are not fully understood. In addition, there is much technical uncertainty associated with drilling large holes to these depths, and waste emplaced in such configurations would be extremely difficult to monitor and retrieve. The proposal for rock-melt disposal would have emplaced nuclear waste in liquid or slurry form into a deep underground drill hole or rock opening. Radioactivity from the waste would heat the liquid/slurry and evaporate the contained water. The heat generated by radioactive decay would continue to increase the temperature, and the surrounding rock would melt. In turn, this molten rock would slowly dissolve the waste, forming a molten solution of waste and rock. Then, over about 1,000 years, the waste-rock melt would eventually cool, solidifying the radioactive material in a relatively insoluble form far beneath the surface. This solidified “waste rock” was expected to be very resistant to leaching and provide long-term containment of the radionuclides in the waste. As this would eliminate the need for considerable mining and construction activity, substantial cost savings would accrue. A primary disadvantage of rock-melt disposal is the long time period for complete solidification. Furthermore, there is insufficient understanding of heat transfer and phase change phenomena in rock, and this is necessary to establish the stability of the molten rock matrix and the engineered emplacement methods. In addition, it is unlikely that the waste could be monitored or that it could be retrieved. Disposal by deep-well injection envisioned a similar mechanism, but without melting the host rocks. Liquid or slurried nuclear waste would be injected into a deep geologic formation, which was capped by a layer of impermeable rock. Pressurized pumping would be used to inject the waste into porous or hydro-fractured rock at depths of 1,000 to 5,000 m. In theory, the waste
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 1–25
7
Stuckless and Levich
would progressively disperse through the host rock and would be isolated from the biosphere by impermeable overlying strata. Well injection is a working technology that is commercially used in a number of applications, but it generally involves the injection of relatively harmless solutions. However, deep well injection would require the chemical or mechanical processing of spent fuel before disposal. Moreover, deep well injection was a commonly practiced disposal method in the former Soviet Union, and it has resulted in radioactive contamination of several areas (Chapman, 2004). The U.S. DOE and the National Aeronautics and Space Administration (NASA) have studied several alternative concepts for disposal of nuclear waste in space. These include: (1) transport to and injection of nuclear waste into the sun; (2) emplacement of waste on Earth’s moon; and (3) emplacement of re-processed waste into a circular solar orbit, midway between Earth and Venus. Disadvantages of space disposal include potential failure of a space launch and the concurrent inability for waste packages to contain waste in the event of the catastrophic failure that would result. In addition, the volume and tonnage of SNF and HLW are great, and the costs of launching the expected tonnage of nuclear waste into space would be extremely high. In 1982, U.S. DOE and NASA agreed to discontinue further study of the extraterrestrial nuclear waste disposal concept. Long-term surface storage was considered as an option or possible adjunct to permanent disposal. After SNF assemblies have been removed from nuclear reactors, they are generally stored in waterfilled pools that cool the spent fuel rods and shield workers from radiation. Spent fuel can also be kept in dry storage casks at reactor sites or at centralized localities, but this storage is considered a temporary measure and requires constant monitoring and security. France’s Commissariat à l’Energie Atomique (CEA), the French Atomic Energy Commission, defines long-term storage as storage for a period of at least 50 years, but no more than 300 years. Reprocessing, partitioning, and transmutation were all considered as adjuncts to permanent disposal. Re-processing is a chemical process in which SNF is dissolved, and the uranium, plutonium, and other actinides are separated from the waste and fission prod‐ ucts that were produced in nuclear reactors. However, it is a costly and complex process, and it requires the development and operation of a major infrastructure. Re-processing can successfully recover fissile material (uranium and plutonium) for recycling as MOX fuel, and it can separate out several of the waste products. However, re-processing SNF was thought to be much more costly than direct disposal. Re-processed HLW is generally consolidated in vitrified form, which still
8
must be permanently isolated from the biosphere. The United Kingdom, France, and Japan were long committed to re-processing all or most of their SNF prior to disposal. Although France remains committed to re-processing, recently the United Kingdom has vacillated between re-processing and direct disposal of SNF, while Japan is now reconsidering its preferred method of disposal. The national disposal strategies of Canada, Sweden, and Finland call for direct disposal of SNF. During President Carter’s administration (1977–1980), the United States established the policy of not re-processing commercial SNF. In the United States, most re-processed waste originates from defense programs; however, some of the vitrified HLW that will be disposed of in the repos‐ itory is a product from civilian reactors. Partitioning and transmutation of radioactive waste are methods of waste treatment that can be used in conjunction with disposal. SNF must be re-processed in advance of using this method. Re-processing separates uranium and plutonium from the waste products. The remaining HLW is partitioned into actinide waste and fission product. The liquid containing the fission products is concentrated, solidified, and sent to a geologic repository for disposal. The actinide waste is combined with uranium (or uranium plus plutonium) and fabricated into mixed oxide or MOX fuel rods and re-inserted into a reactor. In the reactor, approximately 5 to 7 percent of the recycled waste actinides are transmuted to stable or short-lived isotopes, and these are separated out during the next recycle step for disposal in a repository. Numerous recycles would result in the nearly complete transmutation of the waste actinides; however, additional waste streams are generated during each recycle. Furthermore, transmutation does not reduce the quantities of long-lived fission products such as Tc99 and I129, and these must be sent for geologic disposal. Nuclear Waste Disposal Outside the United States Other nations that use nuclear power to generate electricity have faced the same decisions concerning disposal of long-lived nuclear wastes as the United States, and the U.S. DOE has worked closely with their counterparts in other nations to ensure that all possible options for safe disposal of waste are considered. All but one of the foreign nations has opted for deep geologic disposal. In the Netherlands, a national law forbids disposal of “valuable resources,” and SNF has been defined as such. Therefore, the Dutch determined to pursue temporary surface storage and decide on a more permanent solution in the future. Aside from the United States, all other countries with HLW are only considering repository sites in
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 1–25
Deep Geologic Disposal of Nuclear Waste
saturated rocks. Local geologic and hydrologic factors have strongly influenced national decisions on which geologic media are most suitable to investigate as potential repository host rock. Canada, Finland, and Sweden are studying the crystalline rocks of the Canadian and Fennoscandian Shields, respectively. France, Belgium, and Switzerland are currently examining sites underlain by a variety of argillaceous rocks. Some countries, including Spain, Japan, Russia, Argentina, and the United Kingdom, have examined or are examining a variety of potential rock types but have made no final decision concerning the repository host rock. Although Germany spent more than 20 years examining final disposal at the Gorleben Salt Dome in Lower Saxony, a change in government placed all geologic media back into the mix of available options. China has selected saturated crystalline rocks in the Gobi Desert of arid Gansu Province in northwest China as its planned repository host rock. The relatively high water table underlying the Gobi Desert precludes an unsaturated site for China’s repository. Finland has selected a single site adjacent to a nuclear power facility on the Baltic coast of southwest Finland for a repository site, and Sweden has selected and is studying a site along its Baltic coast that is adjacent to other nuclear power facilities. France’s underground laboratory is located in marly argillites of Jurassic age in the eastern part of the Paris Basin. The French authorities wanted to select a site for a second underground research laboratory, preferably in granites; however, they were unsuccessful in locating a site that did not face strong local opposition. Belgium is testing the Boom Clay of Tertiary age at an underground research laboratory in northeast Belgium, and the Swiss are studying the Opalinus Clay at a facility in the Jura Mountains of northwest Switzerland. The Swiss do not intend to locate a repository at or near the research site, but they are considering possible sites underlain by the same clay formation to the east, in northern Switzerland, south of Germany’s Black Forest region. The Swiss have also developed a test facility in crystalline rocks in the Swiss Alps, but studies have indicated that the tectonically stable areas underlain by crystalline rocks are volumetrically insufficient for siting a repository. Broader Search for Geologic Disposal Sites (1976–1982) In 1974, the AEC was divided into two separate agencies. The regulatory function became the U.S. Nuclear Regulatory Commission (NRC), a separate and independent agency to regulate nuclear issues. The remainder of the AEC was joined with parts of
other federal departments and agencies responsible for different aspects of energy programs to create the U.S. Energy Research and Development Administration (ERDA). In 1976, ERDA created the National Waste Terminal Storage (NWTS) Program, which is the direct predecessor to the program of U.S. DOE’s Office of Civilian Radioactive Waste Management (OCRWM). The NWTS expanded the program to study other rock types and planned to develop six repositories, two in salt formations and four in other host rocks, e.g., shale, basalt, or crystalline rocks. The number of repositories was based on predictions of future growth of nuclear power and the desire to distribute the burden of disposal among several states. ERDA began investigations in a number of states, and also began a search for potential sites on federal land, especially where previous activities had been conducted using radioactive materials, including the NTS and the Hanford Reservation in Washington. As early as 1957, the NAS considered crystalline rocks as potential repository host rocks (NAS, 1957). The Nuclear Waste Policy Act (NWPA) of 1982 required the siting and construction of two repositories. The areas of interest for new sites for the second repository were in crystalline rocks. In late 1982, U.S. DOE created the Crystalline Repository Project Office (CRPO), which surveyed crystalline rocks throughout the conterminous 48 states as potential host formations for the second repository, and recommended that U.S. DOE concentrate on the vast areas of crystalline rocks exposed at the surface or buried beneath glacial debris of Pleistocene age. Areas included the Appalachian Mountains from Georgia to Maine and the North American Precambrian Shield in the upper Midwest states of Minnesota, Wisconsin, and Michigan (OCRD, 1983). Between 1983 and 1986, CRPO conducted a literature study of the Appalachians and the Precambrian Shield and identified 235 crystalline rock bodies underlying 17 states. U.S. DOE compiled the published geologic, environmental, and socioeconomic data for these rock bodies and developed this information into specific criteria. These criteria were divided into disqualifying conditions and favorable or potentially adverse regional screening variables, and these were used to screen the crystalline rocks. Twelve areas in seven states (Georgia, North Carolina, Virginia, New Hampshire, Maine, Minnesota, and Wisconsin) were recommended for further study and to be considered for future site selection (U.S. DOE, 1986b). However, as part of the NWPA as Amended (1987), Congress prohibited U.S. DOE from further studies of crystalline rocks to determine their suitability to host a repos‐ itory. U.S. DOE was asked to report to Congress on
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 1–25
9
Stuckless and Levich
the need for a second repository during a 3 year period beginning January 2007. The NAS (1957) considered claystones or shales as appropriate media for geologic disposal of nuclear waste. Although the greater part of the contiguous United States is underlain by sedimentary rocks, only a few attempts have been made to examine the potential for these rocks to host a repository for HLW. The AEC examined argillaceous rocks as early as 1971, and funded several studies by USGS scientists to conduct literature reviews of the properties of shale (Merewether et al., 1973) and particular shale formations. Shurr (1977) reviewed the Pierre Shale in the northern Great Plains. Dames and Moore (1978) also studied the rock properties of shale for U.S. DOE. In the late 1980s, Oak Ridge National Laboratory (ORNL) provided U.S. DOE with reports on sedimentary rocks as potential host rocks for a geologic repository and provided generic comparisons of five sedimentary rocks: sandstone, shale, carbonates, chalk, and anhydrite (Jacobs, 1989). Shale was selected as the most favorable of the sedimentary rocks (Croff et al., 1985). Among the advantages of argilla‐ ceous rocks is their low in situ permeability, the lack of open fractures at depth, and their ability to sorb a number of different cations. However, they commonly have complex mineralogy and typically contain considerable quantities of water. Application of heat to argillaceous rocks may cause desiccation, phase changes in the mineralogy, shrinkage, and fracturing rocks (Croff et al., 1985). Throughout the 1970s and 1980s, the USGS was tasked by Congress to study and comment on the problem of disposal of HLW. In 1976, Dr. Vincent McKelvey, director of the USGS, wrote to Richard Roberts, assistant administrator for nuclear energy at ERDA, and suggested the NTS in southern Nevada as a potential site for a HLW repository. McKelvey based his suggestion on the remoteness of the NTS and the varied geologic environments, and that the USGS already had 400 man-years of data collection and interpretation for the NTS. Dr. McKelvey included a table of assets and limitations of the NTS. As a result of studies requested by Congress, a committee of senior USGS scientists published recommendations in Circular 779 (Bredehoeft et al., 1978) and concluded the following: (1) Salt deposit sites were less than ideal for a retrievable system of waste disposal in a geologic medium. (2) Systematic examination of media other than salt should continue. (3) Major studies of groundwater flow and transport would be needed, especially in fractured rock.
10
(4) More tools should be developed to evaluate potential repositories (e.g., methods of dating old groundwater). (5) More research was needed on the extent to which the repository itself could localize escape of radionuclides to the environment. (6) Uncertainties in earth-science predictions should be recognized as well as importance of a multiple barrier approach for radionuclide containment. In 1977, ERDA was combined with additional federal energy programs and agencies to create the U.S. DOE, and in 1978, U.S. DOE established the Nevada Nuclear Waste Storage Investigations (NNWSI) proj‐ ect. The NTS had been recommended earlier by the USGS and was now selected by U.S. DOE because of: (1) its generally favorable geologic and hydrologic conditions; (2) its prior use for nuclear testing; and (3) the ability for U.S. DOE to conduct work on federal land. The exploration and evaluation of the NTS included geologic, geophysical, hydrologic, tectonic, seismic, and volcanic studies, and these added to more than 30 years of previous NTS studies related to nuclear testing by the USGS and U.S. DOE national laboratories. Initially, 9 rock types and 15 alternative locations on or near the NTS were identified as potentially suitable for a repository (U.S. DOE, 1986a). The Climax Granite of Cretaceous age and argillites of the Eleana Formation of Devonian and Mississippian age were among the first rock units investigated on the NTS. Concurrently, geophysical and drilling studies attempted to locate potentially suitable granitic rocks at Timber Mountain and Twinridge Hill. Coincident studies included developing a regional hydrologic model and the evaluation of potential disruptive events. Some of the earliest work included establishing a seismic monitoring network to determine whether a nuclear waste repository would be compatible with the nuclear weapons tests at the NTS and to measure natural seismicity to evaluate tectonic activity. In 1978, because of nuclear weapons testing activities in the northern and eastern parts of the NTS, NNWSI was required to focus its studies on the southwestern part of the NTS. Investigations examined four geologic media: granite, argillite, tuff, and alluvium. Studies involving alluvium were eventually halted due to its low thermal conductivity. Two boreholes were drilled and cored to depths of 760 m. One borehole, located in the Calico Hills, was preceded by geophysical studies that suggested the presence of a granitic body underlying the hydrothermally altered surface rocks of the Calico Hills Formation; however, this borehole failed to encounter intrusive rock. The second borehole cored and tested the tuffs underlying
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 1–25
Deep Geologic Disposal of Nuclear Waste
Yucca Mountain. In both cases, the target zone of investigation was below the water table. In addition, NNWSI continued or initiated geologic, hydrologic, and geophysical studies, conducted an inventory of exposed granitic rocks in southern Nevada, and began laboratory studies of sorption and permeability to determine the ability of various rocks to retard radionuclides. In 1979, NNWSI completed the Eleana full-scale heater test and modeled the thermal and mechanical data from the test. The NNWSI also completed several hydrologic studies of the NTS area to determine distribution and movement of groundwater, and a major paleo-hydrologic study concluded that during the time of Wisconsin glaciation, the water table at the NTS was not likely more than 60 m above its present level on the east side of the test site (Winograd and Doty, 1980). The USGS reviewed the data available concerning the geology, hydrology, and geophysics for Wahmonie, Calico Hills/Topopah Wash, and Yucca Mountain and recommended the tuffs underlying Yucca Mountain as the most promising potential repository host rock. In mid-1979, NNWSI determined to focus future efforts on studying and characterizing the thick tuff units at Yucca Mountain and determined that it would not initiate future studies in other rock units or areas of the NTS. NNWSI formed the Site Evaluation Steering Committee, which established the Site Evaluation Working Group, which was chartered to establish siting criteria, evaluate and screen potential repository sites, and recommend site locations to the Steering Committee (Sinnock and Fernandez, 1984). During 1979, construction began on the Spent Fuel Test (SFT) facility, which was excavated in the Climax stock (Figure 5). The SFT was located 145 m above the water table in the northeastern part of the NTS and at a depth of 420 m in porphyritic quartz monzonite of Cretaceous age. The purpose of the SFT was to demonstrate that by use of existing engineering technology, it was feasible to transport, encapsulate, emplace, and retrieve SNF at a mined facility, and to store it safely underground for several years. The test also provided information concerning the effect of simultaneous thermal loading and nuclear radiation on granitic rock in comparison to the effects of heat alone. The experiment also tested a preliminary repository design (vertical emplacement in shallow boreholes) and demonstrated the successful handling of HLW using a remote-controlled underground transfer vehicle. The SFT facility consisted of three parallel horizontal drifts, each 64 m long. The two side drifts were constructed first and instrumented to provide rock stress and displacement measurements during the mining of
the central drift. After completion of the central drift, 17 vertical boreholes were drilled 3 m apart into the drift floor. Eleven boreholes were fitted with a steel liner, and a canister 1.17–1.5 m in diameter and 4.7 m in length was lowered into each lined borehole. Eleven SNF assemblies, which had been out of the reactor core for 2.5 years, were safely transported to the NTS from Florida and were encapsulated in stainless-steel canisters inside a hot cell located on the NTS. The canisters containing the spent fuel were transported to the Climax facility and remotely lowered into the SFT. These 11 canisters and six others containing only electric heaters simulating the thermal output of spent fuel were placed in the boreholes using a remote-controlled vehicle, and a concrete floor plug was placed on top of each canister. In each of the side drifts, 10 electrical heaters were placed 6 m apart and powered to simulate the thermal loading of a large repository containing thousands of spent fuel canisters. The SFT-Climax emplacement phase started in May 1980, and the test continued for almost 3 years until April 1983, when the canisters were removed. Additional studies, including coring and sampling of the host rock to examine thermal and radiation effects, lasted until the spring of 1984 (LLNL, 1984, 1985; Patrick, 1986). During 1980, NNWSI prepared an evaluation for U.S. DOE headquarters concerning the suitability of the NTS for a repository. This report concluded that no discoveries to date precluded a repository from being located at the NTS; however, it did not recommend a specific preferred site. Subsequently, further investigations concentrated on the layered tuffs underlying Yucca Mountain, and NNWSI cored, logged, and conducted hydrologic tests on stratigraphic borehole USW G-l, which penetrated 1,830 m of competent, partially welded and moderately welded tuffs. Other NNWSI investigations included geologic mapping and studies of alluvial deposits, groundwater recharge, geophysics, and paleo-climate. Two new tasks on rock mechanics and radionuclide migration in subsurface environments compared tuffs and gra‐ nitic rocks. The Site Evaluation Working Group prepared an interim status report on the suitability of tuffs to host a repository for HLW and provided it to the National Academy of Sciences, Committee on Radioactive Waste Management for review. In 1981, the Site Evaluation Working Group developed, evaluated, and began implementing a screening methodology for locating repository sites on the southwestern NTS and adjoining areas. Technical experts identified the attributes and rationale for assessing site suitability and rated their relative importance as screening tools. Preliminary results confirmed Yucca
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 1–25
11
Stuckless and Levich
Figure 5. Photograph of spent fuel test at the Climax stock showing central drift and vertical emplacement chambers.
Mountain as one of the better locations. Among the reasons for the selection of Yucca Mountain was its isolated location on the southwestern edge of the NTS, far away from nuclear weapons tests, the thick and virtually undisturbed layers of welded, partially welded, and non-welded tuffs extending into the deep subsurface, the results of hydrologic testing of a deep borehole that extended 1,830 m into the layered tuffs that underlay Yucca Mountain, and generally favorable results from preliminary geologic mapping, and volcanic, mineralogic, petrographic, geochemical, geophysical, seismic, groundwater recharge, and paleo-climate studies. NNWSI continued to evaluate the technical acceptability of Yucca Mountain as a potential repository site and characterized the area by geologic, geophysical, geochemical, hydrologic, volcanic, seismic, and environmental investigations (Sinnock and Fernandez, 1984). Borehole USW H-1 was started in 1980 near the center of the Yucca Mountain block, was completed to 1,830 m, and was cored, logged, and hydrologically tested. Mineralogic and petrographic studies of core samples from drill-hole USW G-l identified zeolites and other alteration products in the welded tuffs.
12
A second drill hole (USW G-2) was started to study stratigraphy and was completed to a depth of 1,830 m. Several other boreholes were completed, including USW VH-1, which was drilled in Crater Flat west of Yucca Mountain to study hydrology and relatively recent volcanology (Figure 6). Basaltic volcanic centers were evaluated, and trenches were excavated to study the age and activity of faults in Crater Flat. During 1981, NNWSI also studied radionuclide migration processes, initiated a program to design and develop a waste package for a repository in tuff, and conducted prelim� inary design studies for an exploratory shaft at Yucca Mountain for in situ characterization studies of potential repository horizons. In 1978 and again in 1979, an Interagency Review Group (1978, 1979) recommended that the United States should develop at least two repositories in different regions of the country. The report concluded that candidate sites should be located in a variety of host rocks, and the development of repositories should proceed in a technically conservative step-by-step manner. Partly in response to this recommendation, the Office of Nuclear Waste Management of U.S. DOE and the USGS of the U.S. Department of the
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 1–25
Deep Geologic Disposal of Nuclear Waste
Figure 6. Aerial view of Yucca Mountain looking southeast. The drill pad is UZ-6, sixth of 25 holes drilled speciďŹ cally to investigate and characterize the unsaturated zone.
Interior (DOI) jointly released a draft plan for disposal of radioactive waste in a mined repository (U.S. DOE and U.S. DOI, 1980). The report was written by 17 scientists from five organizations and concluded that there was a need to direct research away from generic plans to characterization of four or five specific sites. Furthermore, they recommended that detailed study plans be prepared for each site and concluded that research and development at the current level should be sufficient to resolve major technical issues within the next 10 years. In 1980, President Jimmy Carter announced a comprehensive waste-management program that included an effective role for state and local governments. In 1981, in its Record of Decision for the Environmental Impact Statement, U.S. DOE selected geologic repositories as the preferred disposal method (U.S. DOE, 1980). During 1981 and 1982, the NWTS program developed criteria for mined geologic repositories for disposing of nuclear waste consistent with the listed criteria then available from existing sources: IAEA (1977), NAS (1978), U.S. DOE (1980), and U.S. NRC (1980). The NWPA of 1982 established the current geologic disposal program, including a comprehensive national policy for management and disposal of SNF and HLW. It remains, with amendments, the statutory
framework for the U.S. HLW disposal program, and it partitioned the responsibility for waste disposal among U.S. DOE, NRC, and the Environmental Protection Agency (EPA). U.S. DOE was given the responsibility to implement the NWPA, NRC to develop the implementing regulations, and EPA to develop the standards that repositories must meet to assure public and worker health and safety. The NRC issued its general regulation for disposal of SNF and HLW, 10 CFR 60, in two parts, in 1980. (This regulation had been developed for saturated zone repositories; the NRC amended Part 60 in 1985 to include criteria that were applicable to repositories in the unsaturated zone.) U.S. DOE issued its general siting guidelines in 1984 (10 CFR 960; U.S. DOE, 1984). The NWPA also set up the nuclear waste fund, which was to be funded by nuclear electricity generation at a rate of 1 mil per kilowatt-hour. The EPA issued its radiation protection standards for geologic disposal of radioactive wastes (including transuranic waste) in 1985 (40 CFR 191; U.S. EPA, 1985). A Refined Search for a Site The NWPA created the OCRWM to manage the program. The NWPA of 1982 developed a process
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 1–25
13
Stuckless and Levich
Figure 7. Diagram showing the various candidate sites for a repository and the time line ending with the mandate to study only Yucca Mountain, Nevada.
for site selection for the first repository, which required U.S. DOE to nominate at least five sites and screen down to three. The U.S. DOE OCRWM identified nine potentially acceptable sites, including Yucca Mountain (Figure 7). The nine sites had three different rock types: salt, basalt, and tuff. The National Academy of Sciences–National Research Council had first recommended salt as a potentially suitable host rock for geologic disposal of nuclear waste (NAS, 1957). During the 1960s and 1970s, the contiguous United States was searched for salt deposits, and four large areas were identified. These areas were: (1) bedded salt in the Salina Basin of Michigan, Ohio, Pennsylvania, and New York; (2) salt domes in the Gulf Coastal Plain of Mississippi, Louisiana, and Texas; (3) bedded salt in the Permian Basin of Kansas, Oklahoma, Texas, and New Mexico; and (4) bedded salt in the Paradox Basin of Utah, Colorado, Arizona, and New Mexico (U.S. DOE, 1986a). During screening of the salt deposits, the Salina Basin was deferred from consideration due to unfavorable characteristics, including high population density and abundant natural resources. Following regional screening, seven potentially acceptable salt sites were selected from the other three regions (U.S. DOE, 1986a). Three of the sites evaluated were in salt domes, and these included Richton Dome and Cypress Creek Dome in Mississippi and the Vacherie Dome in Louisiana. The four remaining sites were in bedded
14
salt: Davis Canyon and Lavender Canyon in Utah’s Paradox Basin, and Deaf Smith County and Swisher County in the Permian Basin of West Texas. This list was subsequently pared down to one salt dome and two bedded salt deposits. U.S. DOE selected Richton Dome in southeast Mississippi as the most favorable salt dome. It was an elliptical northwest-trending salt dome on the edge of the Gulf Coast geosyncline. Anhydrite capped the dome and partially draped its flanks. The top of the cap rock was about 250 m below the surface and was overlain by strata of Miocene age consisting of silts and clays interbedded with limestones and sands. Although the Richton Dome was in an area of low seismicity, two parallel subsurface faults were within 8 km of the dome, and another small fault intersected its northwestern edge. The climate was sub-tropical, and flooding had occurred in the vicinity. Three principal aquifers flowed to the south and southwest away from the dome and discharged in the Gulf of Mexico. Several small oil and gas fields were located within 16 km of Richton Dome; however, the potential for additional hydrocarbon deposits seemed low. Also the salt was judged not to be an economic resource under then current economic conditions (U.S. DOE, 1986c). The Davis Canyon site in San Juan County, southeast Utah, was considered the more favorable of the two Paradox Basin sites, even though it was located within 0.3 km (0.2 miles) of the boundary of
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 1–25
Deep Geologic Disposal of Nuclear Waste
Canyonlands National Park. The proposed host rock was a salt bed, about 60 m thick, within the Paradox Formation of Pennsylvanian age about 900 m underground. The rock strata beneath the site included siltstones, sandstones, bedded salt, and limestones overlying crystalline basement rock. The proposed host rock was one of 29 evaporite layers that formed the Paradox Formation. No faults were known at the site; however, several faults were within 16 km of Davis Canyon. Salt dissolution features occurred 19 and 23 km, respectively, north and southwest of Davis Canyon. Surface water was saline and not potable. The proposed host unit was located within an aquitard with water-bearing units both above and below. Potentiometric data suggested a downward hydraulic gradient from the upper to the lower water-bearing units; however, it was possible that they were not be connected. The groundwater flow direction in the upper unit was north, and that in the lower unit was west- southwest. Uranium and vanadium had been produced from several localities in the vicinity of the site (U.S. DOE, 1986f). The Deaf Smith County site lay in the Texas Panhandle and in the Southern High Plains physiographic sub-province. The site lay in the Palo Duro structural basin within the larger West Texas Permian Basin. A sequence of sandstone, limestone, dolomite, shale, arkose, anhydrite, and bedded salt overlay a basement consisting of metamorphic and igneous rocks. The repository horizon was about 750 m below the ground surface, and the host rock was a bedded salt about 50 m thick that lay in the Permian San Andres Formation. The host rock contained many impurities and was interbedded with shale and anhydrite. Active dissolution fronts were identified 100 km to the east, 30 km to the north, and 170 km to the west. The area was tectonically stable, and no known active faults were identified at the surface. Although oil and gas were produced in the Palo Duro Basin, none was produced in Deaf Smith County. The unconfined High Plains Aquifer, which is within the Ogallala Formation and Dockum Group, lay above the proposed repository host unit. The host unit was located in the middle of an aquitard consisting of shales and evapo‐ rites, 1,500 to 1,800 m thick. Beneath the aquitard containing the host rock, there was a brine aquifer (U.S. DOE, 1986e). The defense-waste management program had been investigating plateau basalts in the Pasco Basin of south-central Washington from 1972 to 1976. Eventually, five sites were evaluated, and one was selected by the Basalt Waste Isolation Project (U.S. DOE, 1986d). The layered basalts of Miocene age were located in the Cold Creek Syncline of the Columbia Plateau on the Hanford Nuclear Reservation in
south-central Washington. The basalt flows in this area were nearly horizontal and had a less disturbed structure than those of other areas of the Columbia Plateau. More than 50 thick basalt flows had been identified in the Pasco Basin of south-central Washington with a combined thickness of about 5,000 m. The selected candidate horizon, the Cohassett basalt flow, was between 75 and 80 m thick in the reference repository location and lay at a depth of 700 m. The deep waters that underlay the Hanford Site were chemically reducing, which would be expected to retard the transport of key radionuclides. At the reference repository location, groundwaters occurred in an unconfined aquifer in near-surface sediments and in a number of confined aquifers in the basalts. However, the highly fractured nature of the Columbia Plateau basalts raised serious concerns relative to hydraulic conductivity and pathways to the accessible environment. The hydrology of the area was complex due to permeable contact zones between basalt and sediment layers plus structural and stratigraphic discontinuities. The location of the groundwater discharge area was unknown, and the high temperature (,60uC.) at the reference repository level was an issue related to pre-closure construction and safety. The Yucca Mountain site lay in the southern part of the Great Basin on federal land, on and adjacent to the Nevada National Security Site, formerly known as the Nevada Test Site, in Nye County, Nevada. The site and regional geology have been described in Stuckless and Levich (2009). Yucca Mountain was formed by layered welded and nonwelded silicic tuffs, deposited during late Miocene time. A number of faults, several of which had moved during the last 2 million years, offset the tuffs in the vicinity of Yucca Mountain. West of Yucca Mountain, several basalt flows and cinder cones were deposited between 3 million and 70,000 years ago. The hydrologic system was characterized by low precipitation, a deep water table, intermittent surface water, and closed topographic and hydro‐ logic basins (Stuckless, 2012). Groundwater was recharged by the slow infiltration of rain, snowmelt, and the surface water (Stuckless and Dudley, 2002). The first rock units at Yucca Mountain that were considered as a repository host were saturated tuffs lying beneath the water table. However, Dr. Isaac Winograd of the USGS suggested that the thick unsaturated zones of the deserts in the southwestern United States would be suitable for the disposal of nuclear waste. In February 1982, U.S. DOE received a letter from the USGS, signed by John Robertson, chief of the USGS Office of Hazardous Waste Hydrology; Gary Dixon, acting USGS program coordinator of the Nevada Nuclear Waste Site
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 1–25
15
Stuckless and Levich
Investigation (NNWSI) program; and William Wilson, chief of the Nuclear Hydrology Program. This letter indicated that the USGS believed that the thick unsaturated zone (400–600 m) at Yucca Mountain might offer considerable advantages for disposal of radioactive waste. The letter further stated that the basis for disposal in the unsaturated zone is that only a very small amount of water would reach the repository. U.S. DOE could design a repository that would permit this water to pass through the repository into the permeable rocks below and would have only minimal contact with waste containers. In addition, the operation of a repository in the unsaturated zone would permit easier access for monitoring and retrieval of the waste. This recommendation was subsequently published with more detail (Roseboom, 1983). Focus on Yucca Mountain The Yucca Mountain site has been studied for more than 30 years by hundreds of scientists from the USGS, the U.S. DOE, the U.S. DOE national laboratories, associated universities, and private contractors. The work performed has been by unbiased world-class scientific studies of the site and at a cost of more than 10 billion dollars (Voegele and Vieth, in press). It has produced new innovative technologies. During 1982, NNWSI cored, logged, and hydrologically tested six major drill holes at Yucca Mountain and began three more. Groundwater was sampled and characterized, and drill core samples underwent physical, chemical, and mechanical tests, including geochemistry, mineralogy, petrology, and rock physics. These data were used to develop a three-dimensional (3-D) model of Yucca Mountain, defining major litho-stratigraphic and hydro-stratigraphic units. By mid-1982, NNWSI examined the strata underlying Yucca Mountain using the collected information, the 3-D model, and considerations of repository performance, safety, and economics. The project compared and evaluated physical properties of several rock units and selected a promising horizon lying within the welded Topopah Spring Member of the Paintbrush Tuff for further study (Johnstone et al., 1984; U.S. DOE, 1986a). At Yucca Mountain, the Topopah Spring Member is at least 300 m thick and consists of a zeolitized lower zone, upper and lower vitric zones, and a thick interior of devitrified tuff with varied degrees of welding and alteration. The selected unit lies approximately 370 m below the surface and about 170 m above the water table, and it is underlain by the highly sorptive and zeolitized non-welded tuff of the Calico Hills Formation. NNWSI installed a computerized seismic monitoring and recording system to replace the previous network
16
and continued studies of the geology, hydrology, geophysics, and geochemistry of the Yucca Mountain site. These studies included tuff-water interaction experiments, digitizing the topography of the potential repository block, stability studies of candidate materials for shaft and borehole sealing, and a radionuclide migration field experiment. Other work included mining of experimental facilities in G-Tunnel at Rainier Mesa, design development for an exploratory shaft, and operation of the SFT at the Climax facility. NNWSI also began three additional major new tasks, including waste package development, repository conceptual design, and repository performance assessment. The NWPA of 1982 prohibited shaft excavation at any site prior to official nomination unless such activities were already in progress. In 1983, NNWSI drilled borehole UE-25p#1, east of Yucca Mountain, at a locale that geophysical data had indicated the depth to Paleozoic bedrock was at its shallowest. Borehole UE-25p#1 penetrated the entire sequence of tuffs of Tertiary age and reached carbonate rocks of Paleozoic age at a depth of 1,250 m (Carr et al., 1986). However, at other localities on the NTS, no Paleozoic rocks were encountered in boreholes to depths of 1,830 m (Keefer et al., 2007). The NWPA of 1982 required the U.S. DOE to develop a detailed site characterization plan (SCP), and in 1988, The Yucca Mountain Project completed the final SCP for Yucca Mountain (U.S. DOE, 1988). The SCP outlined studies including: regional and site geology; volcanic stratigraphy; climate and paleo-climate; erosion; unsaturated zone hydrology; saturated zone hydrology; mineralogy and petrology; rock and fluid geochemistry; fracture fillings characterization; rock mechanics; thermal testing; coupled processes testing; radionuclide transport; tectonics; seismic and volcanic risk hazard; geophysics; natural resources evaluation; and geosphere modeling. All of these subjects were defined in greater detail in approximately 100 study plans. The SCP also proposed excavating a small underground test facility accessed by two vertical shafts. Before starting underground testing at Yucca Mountain that might adversely affect the site, the U.S. DOE conducted extensive prototype testing in tunnels originally constructed as part of the weapons testing program. Between 1962 and 1971, the G-Tunnel Complex was constructed at Rainier Mesa in the north-central NTS. In 1971, Sandia National Laboratories (SNL) began using the tunnel complex as an underground research facility for a variety of technical concerns. SNL initiated the development of the G-Tunnel Underground Facility, which provided access to the Grouse Canyon Tuff, a welded tuff that had properties similar to the rocks of Yucca
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 1–25
Deep Geologic Disposal of Nuclear Waste
Mountain, for NNWSI-related tests. This facility provided a setting where project scientists could develop and test instruments and design and conduct prototype tests for characterizing unsaturated tuff. Participants in the G-Tunnel testing in addition to SNL included Lawrence Livermore National Laboratory (LLNL), Los Alamos National Laboratory (LANL), the USGS, and the Bureau of Reclamation (USBR). Various studies were proposed for G-Tunnel. Some were completed, and others were partially completed before the G-Tunnel Complex was closed down in 1990 as a cost savings and in anticipation of having a tunnel at Yucca Mountain. SNL conducted geomechanical studies at G-Tunnel that included field measurements (e.g., in situ stress measurements in welded tuffs). SNL also conducted smalldiameter heater experiments to investigate thermal and hydrothermal phenomena in welded and non-welded tuffs, and a heated block experiment to collect data on thermo-mechanical properties. In addition, SNL excavated demonstration drifts in welded tuffs using an alpine miner (also called a road header) and performed pressurized slot testing to investigate mechanical properties. SNL and USBR planned joint studies of blast effects on repository host rocks and planned tests related to design of controlled blasting in tuffs. LLNL conducted experiments related to the waste package environment. In addition, LLNL studied properties related to hydrothermal phenomena. LANL completed studies of special drilling tech‐ niques, notably air coring, and developed and validated methods for collecting samples for mineralogy and petrology studies. LANL also conducted an in situ diffusion experiment to study geochemical processes. The USGS conducted a variety of tests at G-Tunnel related to the future Exploratory Studies Facility (ESF). Hydrologic properties were studied using chemical tracers. Other studies tested design and function of borehole instrumentation, measured hydrologic properties by cross-hole testing, and a thermal stress test studied thermo-mechanical properties. Other USGS prototype tests included drift-wall mapping using photogrammetric techniques, intact fracture tests to study flow mechanisms, and wet and dry drilling to test special drilling operations and measure hydrologic properties. The USGS planned additional prototype tests including an infiltrometer test to measure fluid flow properties, a bulk permeability test to study hydrologic phenomena, and dry rubble coring and a perched water test to examine hydrologic phenomena, but these were never conducted (Zimmerman, 1988). In 1987, Congress amended the NWPA of 1982. The Nuclear Waste Policy Amendments Act (NWPAA) of 1987 designated Yucca Mountain as the only site to be characterized with the proviso that the U.S. DOE stop
work immediately if any findings indicated that the site was unsuitable. The amendment also created the Nuclear Waste Technical Review Board (NWTRB), which was to judge the technical and scientific work of the U.S. DOE and to report twice a year to Congress and the Secretary of Energy. The 11 member board was to be nominated by the National Academy of Sciences and appointed by the President of the United States. In 1990, the NWTRB recommended that the Yucca Mountain Project review alternatives for configuration and construction of the ESF. The NRC also objected to the ESF plan, citing an inadequacy of the designcontrol process presented in the SCP. A formal decision methodology was developed to compare and evaluate the benefits of alternative configurations for the ESF by construction of vertical shafts and (or) ramps. A selected team used this formalized methodology to develop and compare 34 alternatives for design and construction. The selected design provided a roughly “horseshoe” configuration (Figure 8). Two gently inclined passages, the northwest to southeast “North Ramp” and the east-west “South Ramp”, would connect the north-south main tunnel to the surface. The report concluded that this particular configuration would provide project scientists with their best opportunity to investigate and evaluate the critical features of the Yucca Mountain site (Stevens and Costin, 1991; U.S. DOE, 1992). Construction of the Yucca Mountain ESF began in late 1992 by conventional drill and blast mining of a short, 60 m starter tunnel at the North Portal Pad. This served as the launch chamber for a 7.6-mdiameter tunnel boring machine (TBM). The TBM started excavation of the main 7.78 km tunnel loop in September 1994 (Figure 9), and it emerged at the South Portal in April 1997. Most tests in the ESF were conducted in eight alcoves and five niches excavated into the walls of the tunnels (U.S. DOE, 1998, p. 1-8/9). The NWTRB also suggested that the project construct a second tunnel to provide better exposure of all the units that would be host horizons for the nuclear waste. During 1997–1998, the project constructed the cross-drift tunnel, originally named the “Enhanced Characterization of the Repository Block” (ECRB) tunnel. The cross-drift was excavated using a 5-m-diameter TBM. Throughout the period of site characterization, the U.S. DOE received oversight from not only the NWTRB, but also the NRC, the State of Nevada, and Nye County (the Nevada jurisdiction in which the Yucca Mountain site is located). The state and local governments were funded through the NWPA for the oversight function. In addition, the U.S. DOE sought independent expert advice for three studies.
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 1–25
17
Stuckless and Levich
Figure 8. Block diagram view of the Exploratory Studies Facility, cross drift, and location of experiments.
The first study was in response to a question of whether or not calcite-silica deposits might represent a time when the water table stood much higher than the proposed repository horizon. The U.S. DOE’s internal team of scientists had concluded that the deposits were pedogenic in origin (Stuckless et al., 1992). The NAS assembled a team of 17 experts to evaluate the two alternatives. They concluded that there was no evidence to support the idea that the groundwater table had risen hundreds of meters in the past and that the evidence supported a surficial origin (NAS, 1992). Similarly, U.S. DOE asked the NAS to assemble a team to review its work on surficial characteristics, pre-closure hydrology, and erosion. The team of seven experts was given a technical basis report on the subjects mentioned and was asked to review it. The U.S. DOE requested the experts to evaluate the technical adequacy of the data, to suggest alternative interpretations of the data, where appropriate, and to suggest ways to improve the report. The committee presented their findings and suggestions to the NWTRB and published them in a book (NAS, 1995). The committee found the work to be technically adequate but added several recommendations for areas of improvement. The U.S. DOE assembled a team of independent experts in volcanology to evaluate the volcanic hazard
18
and possible consequences of magma entering the repository. A second team of experts in seismology and earthquake hazard was assembled. These teams produced probabilistic analyses, which U.S. DOE used in their license application to the NRC (U.S. DOE, 2008, Chapters 2 & 1, respectively). Viability Assessment Congress directed U.S. DOE (by the Energy and Water Development Appropriations Act of 1995) to provide a viability assessment of the Yucca Mountain site to Congress and the president. The Viability Assessment (VA) is the U.S. DOE report that responded to that directive (U.S. DOE, 1998). The VA was issued in 1998 and consists of five volumes: (1) Introduction and Site Characteristics; (2) Preliminary Design Concept for the Repository and Waste Package; (3) Total System Performance Assessment; (4) License Application Plan and Costs; and (5) Costs to Operate the Re� pository. It included the results of 20 years of scientific studies and design development. The VA summarized field investigations, laboratory tests, models, analyses, and engineering. It also identified major uncertainties related to analyses and designs that resulted from variability in the natural system at Yucca Mountain. The VA identified work necessary to complete the Site Recommendation (SR) the License Application (LA) and
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 1–25
Deep Geologic Disposal of Nuclear Waste
Figure 9. Photograph inside Exploratory Studies Facility.
provided cost estimates for completing the LA and constructing and operating the repository (U.S. DOE, 1998; Dyer and Voegele, 2001). As part of the VA, the U.S. DOE provided a Total System Performance Assessment (TSPA), which developed process models using the most current science and design. The results of individual process models, and abstractions of results from other models were linked together to construct a total system numeric model (Figure 10). This numeric model described the natural and engineered components of the repository system (including the biosphere) that would be important to assessing the overall performance of the total system. The TSPA model was used to estimate the releases of various radionuclides from a geologic repository over thousands of years and under a range of potential future conditions, and the consequent range of potential radiological impacts on the groundwater resource and to individual members of the public (U.S. DOE, 1998, p. 1–20). The TSPA examined the safety of the repository after it was closed permanently. A second model (pre-closure safety assessment) evaluated safety before closure (U.S. DOE, 2008, Chapter 1).
The U.S. DOE requested an international review of its TSPA in terms of its adequacy to support a site recommendation. A team of 10 experts concluded, “While presenting room for improvement, the TSPASR methodology is soundly based and has been implemented in a competent manner. Moreover, the modelling incorporates many conservatisms, including the extent to which water is able to contact the waste packages, the performance of engineered barriers and retardation provided by the geosphere. Overall, the IRT (International Review Team) considers that the implemented performance assessment approach provides an adequate basis for supporting a statement on likely compliance within the regulatory period of 10,000 years and, accordingly, for the site recommendation decision.” (OECD/NEA and IAEA, 2002, p. 63). To ensure pre-closure radiological safety, U.S. DOE intended to use a combination of prevention activities and plans for mitigation of unlikely events that could cause excess exposure to radiation. The strategy for pre-closure radiological safety would include primary safety features determined by analyzing design basis events and defense-in-depth features to provide
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 1–25
19
Stuckless and Levich
Figure 10. A representation of the development of Total System Performance Assessment (TSPA).
additional safety margins and reduce the overall safety risk. Other factors that would have been studied and evaluated for effects on pre-closure safety include the risk and consequence of seismic activity, tornadoes, and the potential for flooding. (U.S. DOE, 1998, p. 1–26). OFFICIAL RECOMMENDATION OF YUCCA MOUNTAIN The NWPA of 1982 specified a process for the recommendation and approval of a site for developing a repository. In February 2002, pursuant to the NWPA, the Secretary of Energy (Spencer Abraham) provided a comprehensive statement of the basis for his site recommendation to the President (George W. Bush).
20
Based on the secretary’s recommendation, on February 15, 2002, President Bush sent a letter to the Speaker of the House of Representatives and the President of the Senate stating that he considered the Yucca Mountain site qualified to apply for a construction authorization for a repository. Therefore, he recommended the Yucca Mountain site for a deep geologic repository. He noted that it is important for our national security and our energy future, as nuclear energy is the second largest source of U.S. electricity generation and must remain a major component of national energy policy in the future. On May 8, 2002, the House of Representatives overwhelmingly passed a resolution supporting the president’s recommendation. On July 9, 2002, the Senate concurred by voting 60-39 on a procedural motion preceding (by previous agreement) an unrecorded final voice vote approving
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 1–25
Deep Geologic Disposal of Nuclear Waste
the Yucca Mountain site for development as the nation’s first deep geologic repository for SNF and HLW. During the pre-closure period, a geologic repository must be designed to protect the health and safety of repository workers during repository operations, as well as the public in general, and provide for longterm monitoring. The period for pre-closure performance begins when nuclear waste is first emplaced in the repository and ends at the completion of the monitoring phase and the closure of the repository. NRC regulations are in place for operating nuclear facilities, and operators of nuclear facilities have a 50 year history of complying with the applicable standards, codes, and regulatory guides that can be applied directly to pre-closure operations in a geologic repository and its supporting facilities. The pre-closure radiological safety case would develop the repository system design and sub-system performance requirements. These would focus on high-priority features important to radiological safety and would include the evaluation of the consequence of radiological exposure both to workers and the public at the site boundary. For the post-closure period, the NRC must be able to conclude, on the basis of U.S. DOE’s demonstration, that there is reasonable expectation that nuclear waste can be disposed of safely for 10,000 years without posing unreasonable risk to the health and safety of the public. (The time requirement has since been extended to 1 million years.) A repository safety case was developed in accordance with NRC regulatory requirements and expectations that identified and addressed the uncertainties of assessing long-term repository performance. The safety strategy underlying the safety case was described in the Yucca Mountain Science and Engineering Report (U.S. DOE, 2001). It included: (1) assessment of expected post-closure performance; (2) design margin and defense in depth; (3) consideration of disruptive processes and events; (4) insights from natural and anthropogenic analogues; and (5) a performance confirmation plan. The strategy that formed the basis for the safety case described how a geologic repository located in the unsaturated zone beneath Yucca Mountain would safely isolate nuclear waste for thousands of years. It was based on a conceptual model that specified that the natural and engineered components would work in tandem to meet the post-closure performance objectives and ensure that radiation doses to nearby residents did not exceed the regulatory limits. The strategy integrated site information, design elements, and performance assessment outputs. It identified that the most important attributes of the repository safety case were: (1) limited water contacting waste
packages, (2) long waste package lifetime, (3) low rate of release of radionuclides from breached waste packages, (4) radionuclide concentration reduction during transport away from the waste package, and (5) low consequences of disruptive events. Each of these five main attributes was dependent on several factors important to repository performance, for which data would have to be collected and analyzed. The implementation of the safety strategy would result in a repository system that met post-closure performance criteria and accounted for disruptive processes and events. It would have included engineering features that enhanced repository performance and reduced the effect of uncertainties in the natural system by use of design margin and defense in depth, i.e., redundancies in the overall approach to safety. To the extent possible, performance assessment analyses are based on available data relating to processes and conditions that might occur in a repository system and the chances of their occurrence. Models use numerical simulations to estimate system performance and assess the site’s ability to contain and isolate wastes for many thousands of years, and they are also useful as a means to assess the relative contributions of a variety of factors. Due to spatial and temporal variability of site conditions and the complexity of the coupled physical and chemical processes that operate in a repository environment over time, the results from computational models are not a precise prediction of the actual performance of a repository. According to NRC’s regulation 10 CFR 63.101 (U.S. NRC, 2002), “Proof that the geologic repository will conform with the objectives for postclosure performance is not to be had in the ordinary sense of the word because of the uncertainties inherent in the understanding of the evolution of the geologic setting, biosphere, and the engineered barrier system. For such long-term performance, what is required is reasonable expectation, making allowance for the time period, hazards, and uncertainties involved …” (http://www. nrc.gov/reading-rm/doc-collections/cfr/part063/part0630101.html). Therefore, there are significant uncertainties in the results of performance assessments (U.S. DOE, 1998, p. 1-22/25; U.S. NRC, 2002). The TSPA must capture uncertainties in the understanding of processes, as well as uncertainties in the characteristics and properties of the natural and engineered systems. Generally, these uncertainties are specified as to the relative likelihood that model parameters fall within a range of possible values. This type of uncertainty assessment is typically evaluated by probabilistic sampling from probability-distribution functions that represent the likelihood of a given value for a parameter. Uncertainty also includes the assignment of ranges of probabilities of occurrence to unanticipated events, and
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 1–25
21
Stuckless and Levich
evaluating the range of likely repository behavioral impacts that may result from those initiating events (U.S. DOE, 1998, p. 1-20/22). The U.S. regulatory framework, put in place by the EPA (40 CFR 197) and the NRC (10 CFR 63), required compliance for 10,000 years (later changed to 1,000,000 years) along with post-closure standards for individual protection, groundwater protection, and human intrusions. In addition, the National Academy of Sciences (NAS, 1995) recommended a less rigorous calculation of likely performance for the individual protection standard carried out to the time of the peak dose, generally 200,000 to 800,000 years for Yucca Mountain, or the period of geologic stabil‐ ity, estimated by NAS for the Yucca Mountain area as being about a million years. The performance calculation for the Yucca Mountain Site Recommendation was based on the requirements contained in the NRC’s 10 CFR 63. 10 CFR Part 63.341 states, “… DOE must calculate the peak dose of the reasonably maximally exposed individual that would occur after 10,000 years following disposal but within the period of geologic stability. No regulatory standard applies to the results of this analysis; however, U.S. DOE must include the results and their bases in the environmental impact statement for Yucca Mountain as an indicator of long-term disposal system performance” (U.S. NRC, 2002). The longer-term calculation was published in the 2002 Environmental Impact Statement for the proposed repository at Yucca Mountain (U.S. DOE, 2002). The one-million-year time frame became mandatory when it was included in the ruling of a federal court in 2004. President Bush signed House Joint Resolution 87 on July 23, 2002, “Resolved by the Senate and House of Representatives of the United States of America in Congress assembled. That there hereby is approved the site at Yucca Mountain, Nevada, for a repository, with respect to which a notice of disapproval was submitted by the Governor of the State of Nevada on April 8, 2002.” The White House noted that: “Nuclear waste is now stored in 131 temporary locations in 39 states. The successful completion of the Yucca Mountain project will ensure our nation has a safe and secure underground facility that will store nuclear waste in a manner that protects our environment and our citizens” (Levich et al., 2002, p. 1–2). CONCLUDING REMARKS Settling on deep geologic disposal as the best means for isolating HLW from the environment and protecting life was a long and arduous process. Eventually, all nations that generate nuclear waste have come to the same conclusion, although the geologic medium of
22
choice varies from country to country. The choice of Yucca Mountain, NV, started in a methodical fashion, but in the end, Congress designated Yucca Mountain for characterization, because it seemed to have the fewest problems. Some have alleged that the science conducted at Yucca Mountain was influenced by politics. For example, Alison Macfarlane (then chairperson of the NRC) testified before the Senate subcommittee for Energy and Water Development in May of 2003 that, “First, we know the science done at Yucca Mountain is produced by scientists mindful of the political goal of the agencies they work for, and the work they produce is evaluated by managers trying to meet these goals.” She went on to say, “Now that we know that politics does indeed play a role in the science produced to uphold the Yucca Mountain site, we can ask the question, has politics limited some of the science done and questions asked about the site? I would argue the answer is yes” (cited by Alley and Alley, 2013, p. 230). This is contradicted by the fact that work for the project was conducted by numerous scientists, employed by numerous federal and non-federal agencies and institutions, and much of that work received extensive oversight and scrutiny from both federal and local organizations. Macfarlane also ascribed political motivation to management, which is unfounded and, in the case of the USGS, is explicitly incorrect. In June 1985, the director of the USGS wrote to U.S. DOE, “We are also prepared to defend our scientific investigations and findings.” The letter also states, “The USGS has a further obligation to the Government and the public to demonstrate impartiality in its statements. We must retain our position as a disinterested party.” Certainly, the scientists at LANL were not trying to slant their science in favor of a repository at Yucca Mountain either. When they reported apparent evidence of fast pathways for water from the surface to the repository horizon, they released the information immediately. These fast pathways would have been detrimental to U.S. DOE’s safety case. Later work by LLNL, the USGS, and the University of Nevada– Las Vegas failed to confirm the widespread existence of these fast pathways. Twice towards the end of the project, the U.S. DOE sought an opinion from the USGS. When asked for their appraisal of the VA, the USGS assembled a team of five senior scientists not associated with the Yucca Mountain Project. These scientists represented a range of earth science areas of specialties. They published their appraisal (Hanks et al., 1999, p. 16) stating that they agreed with the U.S. DOE, “that the site remains promising for the development of a geologic repository.” They also made five suggestions, which
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 1–25
Deep Geologic Disposal of Nuclear Waste
included preparing a semi-quantitative assessment in plain English, and the publication of the milestone reports and other gray literature. To a large extent, the U.S. DOE has complied with these suggestions. On August 28, 2001, The Office of Civilian Radioactive Waste Management asked the USGS to comment on the possible recommendation of Yucca Mountain by the secretary of energy. The response on October 4, 2001, was prefaced as, “The USGS is commenting within the scope of Earth science and is neutral regarding other information the Secretary may consider.” The letter states that, “the USGS believes that the scientific work performed to date supports the decision to recommend the site for the development as a nuclear waste repository.” The letter cites a number of positive attributes of the site as well as several recommendations. The letter closes with a request for a “stepwise decision-making process” such that “future decisionmakers [can] select alternative options, if necessary.” As noted in the introduction, the Yucca Mountain site has been studied for more than 30 years by hundreds of scientists from the USGS, the U.S. DOE, the U.S. DOE national laboratories, multiple universities, and private contractors. The work performed has been unbiased world-class scientific studies of the site, and it has produced some new innovative technologies. This was affirmed by Representative Paul Tonko, ranking member of the Subcommittee on Investigations and Oversight. During a congressional hearing, Representative Tonko noted that the decision to close Yucca Mountain, “was not driven by science, and it is fiction to pretend that it was.” He continued, “The truth is that the decision making process surrounding Yucca (Mountain) has always been political” (Showstack, 2011, p. 411). The NRC has finally released a five-volume evaluation of the LA and found nothing technical that would preclude approval of Yucca Mountain as a HLW repository. There are, however, some land access and water rights issues that would appear to require Congressional action. ACKNOWLEDGMENTS The authors received many helpful suggestions from Ike Winograd (USGS retired), Michael Voegele (Scientific Applications International Corporation (SAIC) retired), Robert Rechard (SNL), and James Paces (USGS). The authors also received helpful input from three reviewers chosen by the journal. REFERENCES ALLEY, W. M. AND ALLEY, R., 2013, Too Hot to Touch—The Problem of High-Level Radioactive Waste: Cambridge University Press, Cambridge, U.K., 370 p.
BREDEHOEFT, J. D.; ENGLAND, A. W.; STEWART, D. B.; TRASK, N. J.; AND WINOGRAD, I. J., 1978, Geologic Disposal of HighLevel Radioactive Wastes; Earth-Science Perspectives: U.S. Geological Survey Circular 779, 15 p. BROOKINS, D. G., 1985, Geochemical Aspects of Radioactive Waste Disposal: Springer-Verlag, Berlin, 347 p. CARR, M. D.; WADDELL, S. I.; VICK, G. S.; STOCK, J. M.; MONSEN, S. A.; HARRIS, A. G.; CORK, B. W.; AND BYERS, F. M., JR., 1986, Geology of Drill Hole UE25p#1—A Test Hole into Pre-Tertiary Rocks near Yucca Mountain, Southern Nevada: U.S. Geological Survey Open-File Report 86‐175, 87 p. CHAPMAN, N. A., 2004, Examining subsurface radionuclide movement—Natural and anthropogenic sources: EOS, Vol. 85, No. 4, pp. 387–389. CONGRESS OF THE UNITED STATES, OFFICE OF TECHNOLOGY ASSESSMENT, 1985, Managing the Nation’s Commercial High-Level Radioactive Waste: Office of Technology Assessment Report OTA-O-171, 348 p. CROFF, A. G.; LOMENICK, T. F.; LOWRIE, R. S.; AND STOW, S. H., 1985, Evaluation of Five Sedimentary Rocks Other Than Salt for HighLevel Waste Repository Siting Purposes: Oak Ridge National Laboratory–Martin Marietta Report ORNL1CF-85/2, 3 volumes. DAMES AND MOORE, 1978, Baseline Rock Properties—Shale: Office of Waste Isolation Report Y/OWVTM-366. DYER, J. R. AND VOEGELE, M. D., 2001, The Yucca Mountain site characterization project for the United States. In Witherspoon, P. A. and Bodvarsson, G. S. (Editors), Geological Challenges in Radioactive Waste Isolation: Earth Sciences Division, Lawrence Berkeley National Laboratory, University of California, Berkeley, CA, pp. 299–312. EKREN, E. B.; DINWIDDIE, G. A.; MYTTON, J. W.; THORDARSON, W.; WEIR, R. E., JR.; HINRICHS, E. N.; AND SCHRODER, L. J., 1974, Geologic and Hydrologic Considerations for Various Concepts of High-Level Radioactive Waste Disposal in Conterminous United States: U.S. Geological Survey Open-File Report 74‐158, 219 p. FAURE, G., 1986, Principles of Isotope Geology, 2nd ed.: John Wiley & Sons, New York, 589 p. FEDERAL REGISTER, 1981, Record of decision: Federal Register Notices, Vol. 46, No. 93, p. 26,677. HANKS, T. C.; WINOGRAD, I. J.; ANDERSON, R. E.; REILLY, T. E.; AND WEEKS, E. P., 1999, Yucca Mountain as a Radioactive Waste Repository—A Report to the Director of the U.S. Geo‐ logical Survey: U.S. Geological Survey Circular 1184, 19 p. INTERAGENCY REVIEW GROUP, 1978, Subgroup Report on Alternative Technology Strategies for the Isolation of Nuclear Waste: Interagency Review Group on Nuclear Waste Management Report TID-28818. INTERAGENCY REVIEW GROUP, 1979, Report to the President: Interagency Review Group on Nuclear Waste Management Report TID-29442, 149 p., 7 appendices. INTERNATIONAL ATOMIC ENERGY AGENCY (IAEA), 1977, Site Selection Factors for Repositories of Solid High-Level and Alpha-Bearing Wastes in Geologic Formations: IAEA, Vienna, Austria, Technical Report Series No. 177. JACOBS, G. K., 1989, The Sedimentary Rock Program: A Summary of the Geochemical Characteristics of Shale Important to Waste Isolation: Oak Ridge National Laboratory–Martin Marietta Report ORNL1TM-1098-1, 32 p. JOHNSTONE, J. K.; PETERS, R. R.; AND GNIRK, P. F., 1984, Unit Eval‐ uation at Yucca Mountain, Nevada Test Site, Summary Report and Recommendation: Sandia National Laboratories Report SAND83-0372, 26 p. KEEFER, W. R.; WHITNEY, J. W.; AND BUESCH, D. C., 2007, Geology of the Yucca Mountain site area. In Stuckless, J. S. and Levich, R. A. (Editors), The Geology and Climatology of
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 1–25
23
Stuckless and Levich Yucca Mountain and Vicinity, Southern Nevada and California: Geological Society of America Memoir 199, pp. 53–103. LAWRENCE LIVERMORE NATIONAL LABORATORY, 1984, Spent Fuel Test—Nevada Test Site: Lawrence Livermore National Laboratory Report LLL-TB-61, Energy Technologies, 4 p. LAWRENCE LIVERMORE NATIONAL LABORATORY, 1985, SFT-C Fact Sheet, Revision 4—5/30/85 (Unpublished): Lawrence Livermore National Laboratory, Livermore, CA, 11 p. LEVICH, R. A.; LINDEN, R. M.; PATTERSON, R. L.; PECK, J. H.; AND STUCKLESS, J. S., 2002, Geology of the Proposed Yucca Mountain Repository Site: Field Trip # l (J. H. Peck and R. A. Levich, co-leaders). In 2002 Joint Annual Meeting Field Trip Guide: Association of Environmental and Engineering Geologists–American Institute of Professional Geologists, Reno, NV, 37 p., 7 figs. MEREWETHER, E. A.; SHARPE, J. A.; GILL, J. R.; AND COOLEY, M. E., 1973, Shale, Mudstone and Claystone as Potential Host Rocks for Underground Emplacement of Waste: U.S. Geological Survey Open-File Report 73‐184, 44 p., 7 pl. NADIS, S., 1996, The sub-seabed solution: The Atlantic Monthly, Vol. 278, No. 4, pp. 28–39 (also available at https://www. theatlantic.com/past/docs/issues/96oct/seabed/seabed.htm). NATIONAL ACADEMY OF SCIENCES (NAS), 1957, The Disposal of Radioactive Waste On Land, Report of the Committee on Waste Disposal of the Division of Earth Sciences: National Research Council, National Academy of Sciences Publication 519, 142 p. NATIONAL ACADEMY OF SCIENCES (NAS), 1978, Geological Criteria for Repositories for High-Level Radioactive Wastes: National Research Council, National Academy Press, Washington, DC, 19 p. NATIONAL ACADEMY OF SCIENCES (NAS), 1990, Rethinking HighLevel Radioactive Waste Disposal: National Research Council, National Academy Press, Washington, DC, 38 p. NATIONAL ACADEMY OF SCIENCES (NAS), 1992, Ground Water at Yucca Mountain—How High Can it Rise: National Research Council, National Academy Press, Washington, DC, 231 p. NATIONAL ACADEMY OF SCIENCES (NAS), 1995, Review of the U.S. Department of Energy Technical Bases Report for Surface Characteristics, Preclosure Hydrology, and Erosion: National Research Council, National Academy Press, Washington, DC, 143 p. NKOMO, I. T.; STUCKLESS, J. S.; THADEN, R. E.; AND ROSHOLT, J. N., 1978, Petrology and uranium mobility of an early Precambrian granite from the Owl Creek Mountains, Wyoming. In Wyoming Geological Association Symposium and Guidebook for the Economic Geology of the Wind River Basin: Wyoming Geological Association, Casper, WY, pp. 335–348. NUCLEAR ENERGY AGENCY (NEA), 1995, The Environmental and Ethical Basis of Geological Disposal, a Collective Opinion of the NEA Radioactive Waste Management Committee: Organization for Economic Co-Operation and Development, Paris, 30 p. OFFICE OF CRYSTALLINE REPOSITORY DEVELOPMENT (OCRD), 1983, A National Survey of Crystalline Rocks and Recommendations of Regions to be Explored for High-Level Radioactive Waste Repository Sites: Office of Crystalline Repository Development (OCRD), Battelle Memorial Institute, OCRD-1, Columbus, OH, 110 p. ORGANISATION FOR ECONOMIC CO-OPERATION AND DEVELOPMENT (OECD)/NUCLEAR ENERGY AGENCY (NEA) AND IAEA, 2002, An International Peer Review of the Yucca Mountain Project TSPA-SR: IAEA and NEA, Vienna, Austria, 96 p. PATRICK, W. C., 1986, Spent Fuel Test—Climax: An Evaluation of the Technical Feasibility of Geologic Storage of Spent Nuclear Fuel in Granite, Final Report: Lawrence Livermore National Laboratory Report UCRL-53705, Livermore, CA, 228 p.
24
PATTERSON, R. L. AND GROSS, M., 2014, Compliance Recertification 2019 (CRA-2019) Big Changes Coming. In Proceedings, Waste Management Conference, March 2014: Phoenix, AZ. PATTERSON, R. L. AND NELSON, R. A., 2001, An update on the geological disposal of radioactive waste at the Waste Isolation Pilot Plant in southeastern New Mexico, U.S.A. In Witherspoon, P. A. and Bodvarsson, G. S. (Editors), Geological Challenges in Radioactive Isolation: Earth Sciences Division, Lawrence Berkeley National Laboratory, University of California, Berkeley, CA, pp. 313–318. PIERCE, W. G. AND RICH, E. I., 1962, Summary of Rock Salt Deposits in the United States as Possible Disposal Sites for Radioactive Waste: U.S. Geological Survey Bulletin 1148, 175 p., 6 pls. ROSEBOOM, E. H., JR., 1983, Disposal of High-Level Nuclear Waste Above the Water Table in Arid Regions: U.S. Geological Survey Circular 903, 21 p. SHOWSTACK, R., 2011, Yucca Mountain nuclear waste repository prompts heated Congressional hearing: EOS, Vol. 90, No. 46, pp. 410–411. SHURR, G. W., 1977, The Pierre Shale, Northern Great Plains; a Potential Isolation Medium for Radioactive Waste: U.S. Geological Survey Open-File Report 77‐776, 27 p. SINNOCK, S. AND FERNANDEZ, J. A., 1984, Location Performance Objectives for the NNWSI Area-to-Location Screening Activity: Sandia National Laboratories Report SAND82-0837, 165 p. SMELLIE, J. A. T. AND STUCKLESS, J. S., 1985, Element mobility studies of two drill cores from the Gotemar Granite (Krakemala Test Site), southeast Sweden: Chemical Geology, Vol. 51, pp. 55–78. STEVENS, A. L. AND COSTIN, L. S., 1990, Findings of the ESF Alternatives Study, an Executive Report: Sandia National Laboratories Report SAND90-3232, 68 p. STUCKLESS, J. S. (Editor), 2012, Hydrology and Geochemistry of Yucca Mountain and Vicinity, Southern Nevada and California: Geological Society of America Memoir 209, 393 p. STUCKLESS, J. S.; BUNTING, J. A.; AND NKOMO, I. T., 1981, U-Th-Pb systematics of some granitic rocks of the northeastern Yilgarn Block, Western Australia: Journal of the Geological Society of Australia, Vol. 28, pp. 365–375. STUCKLESS, J. S., AND DUDLEY, W. W., JR., 2002, The geohydrologic setting of Yucca Mountain, Nevada. In Gascoyne, M. and Peterman, Z. E. (Editors), Geochemistry of Yucca Mountain, Nevada; a Potential Site for a High-Level Nuclear Waste Re‐ pository: Applied Geochemistry, Vol. 17, No. 6, pp. 659–682. STUCKLESS, J. S.; HEDGE, C. E.; WORL, R. G.; SIMMONS, K. R.; NKOMO, I. T.; AND WENNER, D. B., 1985, Isotopic studies of the Late Archean plutonic rocks of the Wind River Range, Wyoming: Geologic Society of America Bulletin, Vol. 96, pp. 850–860. STUCKLESS, J. S. AND LEVICH, R. A. (Editors), 2009, The Geology and Climatology of Yucca Mountain and Vicinity, Southern Nevada and California: Geological Society of America Memoir 199, 205 p. STUCKLESS, J. S. AND NKOMO, I. T., 1978, Uranium-lead isotope systematics in a uraniferous, alkali-rich granite from the Granite Mountains, Wyoming—Implications for uranium source rocks: Economic Geology, Vol. 73, pp. 427–441. STUCKLESS, J. S.; PETERMAM, Z. E.; FORESTER, R. M.; WHELAN, J. F.; VANIMAN, D. T.; MARSHALL, D. B.; AND TAYLOR, E. M., 1992, Characterization of fault-filling deposits in the vicinity of Yucca Mountain, Nevada. In Proceedings of Waste Management ’92 Conference: Phoenix, AZ, pp. 929–935. STUCKLESS, J. S.; TROENG, B.; HEDGE, C. E.; NKOMO, I. T.; AND SIMMONS, K. R., 1982, Age of uranium mineralization at Lilljuthatten in Sweden and constraints on ore genesis: Sveriges Geolcgiska Undersokning, Ser. C, Vol. 789, 49 p.
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 1–25
Deep Geologic Disposal of Nuclear Waste U.S. DEPARTMENT OF ENERGY (U.S. DOE), 1980, Final Environmental Impact Statement Management of Commercially Generated Radioactive Waste: Office of Civilian Radioactive Waste Management (OCRWM), Report DOE/EIS-0046F, variously paginated. U.S. DOE, 1984, General Guidelines for the Recommendation of Sites for the Nuclear Waste Repositories, Final Siting Guidelines: U.S. DOE Report 10 CFR 960, 886 p. U.S. DOE, 1986a, Environmental Assessment, Yucca Mountain Site, Nevada Research and Development Area, Nevada: OCRWM Report DOE/RW-0073, 3 Volumes. U.S. DOE, 1986b, Area Recommendation Report for the Crystalline Repository Project: OCRWM Report DOE/CH-15(1), Volume 1. U.S. DOE, 1986c, Environmental Assessment Overview for Richton Dome, Mississippi: OCRWM Report DOE/RW-0078, 37 p. U.S. DOE, 1986d, Environmental Assessment Overview for Reference Repository Location, Hanford Site, Washington: OCRWM Report DOE/RW-0076, 34 p. U.S. DOE, 1986e, Environmental Assessment Overview for Deaf Smith County Site, Texas: OCRWM Report DOEIRW-0075, 36 p. U.S. DOE, 1986f, Environmental Assessment Overview for Davis Canyon Site, Utah: OCRWM Report DOE/RW-0077, 38 p. U.S. DOE, 1988, Site Characterization Plan, Yucca Mountain Site, Nevada Research and Development Area, Nevada: OCRWM, 9 Volumes, GPO, Washington, D.C. U.S. DOE, 1992, Management overview. In Lorenz, J. J. and Stephan, P. M. (Editors), Yucca Mountain Site Characterization Project Bibliography, 1990–1991: U.S. DOE Report DOE/ OSTI-3406 (DE92004874). U.S. DOE, 1998, Viability Assessment of a Repository at Yucca Mountain, Volume 1, Introduction and Site Characteristics: OCRWM, variously paginated, GPO, Washington, D.C. U.S. DOE, 2001, Yucca Mountain Science and Engineering Report: OCRWM, variously paginated, GPO, Washington, D.C. U.S. DOE, 2002, Final Environmental Impact Statement for a Geologic Repository for the Disposal of Spent Nuclear Fuel and High-Level Radioactive Waste at Yucca Mountain, Nye
County, Nevada: OCRWM, variously paginated, GPO, Washington, D.C. U.S. DOE, 2008, Yucca Mountain Repository SAR: U.S. DOE Report DOE/RW-0573, Revision 1, Table 1.5.1-1. U.S. DOE AND U.S. DEPARTMENT OF INTERIOR (U.S. DOI), 1980, Earth Science Technical Plan for Disposal of Radioactive Waste in a Mined Repository (Draft): DOE Office of Nuclear Waste Management and U.S. Geological Survey Report DOE/TIC-11033 (draft) and USGS (draft report). U.S. ENVIRONMENTAL PROTECTION AGENCY (U.S. EPA), 1985, Envi‐ ronmental Radiation Protection Standards for Management and Disposal of Spent Nuclear Fuel, High-Level and Trans‐ uranic Radioactive Wastes: U.S. EPA 40 CFR 191, variously paginated. U.S. NUCLEAR REGULATORY COMMISSION (NRC), 2002, Disposal of High-Level Radioactive Wastes in a Proposed Geologic Repository at Yucca Mountain, Nevada: U.S. NRC 10 CFR 63, not paginated. U.S. NRC, 1983, Disposal of High-Level Radioactive Wastes in Geologic Repositories: U.S. NRC 10 CFR60, GPO, Washington, D.C. U.S. NRC, 1980, Draft Licensing Regulations at (c.f. U.S. NRC, 1983, 10CFR60.2), GPO, Washington, D.C. VOEGELE, M. D. AND VIETH, D. L., in press, Waste of a Mountain: How Yucca Mountain Was Selected, Studied, and Dumped. WALKER, J. S., 2007, The controversy over the Lyons radioactive waste repository, 1970–1972: Kansas History—A Journal of the Central Plains, Vol. 27, pp. 266–285. WINOGRAD, I. J. AND DOTY, G. C., 1980, Paleohydrology of the Southern Great Basin. With Special Reference to Water Table Fluctuations beneath the Nevada Test Site during the Late(?) Pleistocene: U.S. Geological Survey Open-File Report 80‐ 569, 91 p. ZIMMERMAN, R. M., 1988, Background Material for Visitor Escorts to the G-Tunnel Underground Facility (Unpublished): U.S. DOE, OCRWM, NNWSI Geotechnical Projects Division, Sandia National Laboratories, 22 p.
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 1–25
25
A Comparison of Surface- and Standard Penetration Test-Derived Shear-Wave Velocity PETER J. HUTCHINSON1 MAGGIE H. BEIRD THG Geophysics, 4280 Old William Penn Highway, Murrysville, PA 15668
Key Terms: Geophysics, Geotechnical, Seismic, Engineering
ABSTRACT Shear wave velocity (Vs) values calculated from Standard Penetration Tests (SPTs) and from multichannel analysis of surface waves (MASW) are consistent in their prediction of Vs. This analysis was based upon SPT and MASW data from 10 localities representing four different depositional environments across North America. Homogeneous deposits tend to produce the closest agreement between MASW- and SPT-derived Vs. Poorly sorted deposits can predict less consistent agreement between MASW- and SPTderived Vs. Further, at depths of greater than 20 m below grade, based upon the testing geometry within this work, the prediction of Vs between the two methods can differ. INTRODUCTION Engineered structures rely on geotechnical studies of sheer wave velocity (Vs) data to properly calculate the design basis for construction. Recently, design basis in earthquake-active areas has driven many researchers to characterize dynamic soil properties (Hvorslev, 1949; Imai, 1977; Ohta and Goto, 1978; Fumal and Tinsley, 1985; Kayabali, 1996; Chien et al., 2000; Andrus et al., 2004; and Hanumantharao and Ramana; 2008). The application of Rayleigh (surface) wave analysis (SWA), introduced in late 1980s to evaluate the stiffness of near-surface material (i.e., Vs), is gaining popular acceptance in geotechnical studies. This article addresses one of the methods for calculating Vs from surface waves, the multichannel analysis of surface waves (MASW) method. Standard Penetration Test (SPT) data have been used for approximately 100 years and are considered the standard by which engineers can convert pene1
Corresponding author email: pjh@geo-image.com.
tration resistance to dynamic soil moduli (Ohsaki and Iwasaki, 1973; Ohta et al., 1978; and Sykora and Koester, 1988). Further, international standards have codified SPT data into the engineering design basis realm (for example, ASTM, 2008). SPTs are conducted using the shell and auger method pursuant to ASTM Standard Method D1586 (ASTM, 2008). SPT values are measured in 1.5-m depth intervals by connecting a split spoon sampler to drill rods. A 63.5-kg dead weight is dropped freely from a height 0.76 m and used to drive the split spoon 0.457 m into the subsurface. The number of blows for each 0.152 m of penetration of the split spoon sampler is recorded. The blows required to penetrate the initial 0.152 m of the split spoon are ignored as a result of the presence of possible loose material or cuttings. The SPT value or N-value is derived from the cumulative number of blows required to penetrate the remaining 0.305 m of the 0.457-m sampling interval. The energy generated by the hammer blow in the SPT test is principally shearing energy and can be used to model the shear strain modulus. This article will demonstrate that MASW data are comparable to SPT data, the engineering standard. One of the strongest aspects of MASW testing is that MASW data are digitally recorded. Consequently, MASW data are equally sensitive from the low velocity range (i.e., 10 m/s) to the high velocity range (i.e., .5,000 m/s), whereas SPT data are less sensitive in the low blow count range (for the predicted Vs of ,100 m/s) and saturate in the high blow count range (for the predicted Vs of .350 m/s). Furthermore, the strongest aspect of MASW testing is that it does not require a borehole, so cost per test is much less than that associated with SPT testing. Measurement of Vs, performed in situ, using geophysical methods can be one of the best methods for measuring the low strain shear modulus (Rollins et al., 1998a). Geophysical seismic methods are based on the velocity of propagation of a wave in an elastic body as a function of the modulus of elasticity, Poisson’s ratio, and density of the material (Hvorslev, 1949).
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 27–36
27
Hutchinson and Beird
Figure 1. An example of a MASW dispersion curve from south-central Puerto Rico of the fundamental mode of the frequency (Hz) versus phase velocity (m/s). Amplitude intensity is shown in color. The dotted white line represents the fundamental mode. Processed using KGS SurfSeis 3.06 computer program (KGS, 2010).
Within the past 10 years, however, SWA data have proven to offer a more effective and economical alternative to the use of SPT data for the prediction of soil dynamic properties. SWA is generated from a high-amplitude surface wave front initiated through induced elastic energy (i.e., hammer blow). Initially, non-intrusive surface wave method using a single pair of receivers, Spectral Analysis of Surface Waves (SASW), was introduced (Nazarian and Stokoe, 1984; Dennis et al., 1998). The SASW method produced results that were within 10–15 percent of measured Vs values (Nazarian and Stokoe, 1984; Stokoe et al., 1988; Dennis et al., 1998; and Brown et al., 2002). The use of multiple channels, as with the MASW method, records higher resolution, more consistent, more repeatable and higher amplitude energy than does the SASW method (Park et al., 1999). Furthermore, dispersion curves based upon MASW produce more readily interpreted data (Xia et al., 1999). MASW data are recorded using a series of lowfrequency geophones (4.5 Hz) in a linear array. A source (sledgehammer, propelled energy generator, etc.) is used to initiate the surface wavefront. The geophones receive the vertical component of the elliptical Rayleigh wave or “ground roll” and convert the mechanical movement to an electrical signal, which is recorded by the seismograph. The depth of surface wave penetration is limited to half of the surface wavelength, and the survey geometry must be chosen accordingly. However, there are limitations to the accuracy of the Vs values, especially in the depth ranges greater than 20 m (Hutchinson et al., 2008). The surface wave data are post-processed subsequent to the field data collection. Data are displayed in the frequency versus phase velocity format in the form of a dispersion curve. A fundamental mode is derived from the dispersion curve (phase velocity versus the frequency) and is inverted to produce a Vs-depth curve (Miller et al., 1999; for example, see Figure 1).
28
MASW analysis diverges from other seismic methods of Vs data collection and processing because MASW does not, in most cases, attenuate from velocity inversions with depth (Park et al., 1999; Miller et al., 2001; and Ivanov et al., 2008). The MASW method of measuring Vs is quite robust because this method is insensitive to the presence of subsurface utilities; to the occurrence of standing structures; to buried boulders, rocks, and concrete; and to conductive soil (Hutchinson and Beird, 2011). DISCUSSION Many authors have attempted to predict the Vs from SPT data. For example, the NovoSPT program provides more than 260 formulas and methods for correlating geotechnical engineering soil properties from SPT blow counts (N60 or N100) (NovoTech, 2010). The generic regression formula for the prediction of Vs from SPT data is: Vs ~AN B ,
ð1Þ
where shear wave velocity is a function of the blow count (N) raised to a power (B) times a constant (A). Hanumantharao and Ramana (2008) collected over 50 formulas within the literature that used this basic regression equation. Their work showed quite a variation in formulas from a low extreme (from Rollins et al., 1998b), Vs ~222:0 ðN Þ0:06 ,
ð2Þ
To a high extreme (from Jafari et al., 2002): Vs ~19:0 ðN Þ0:85 :
ð3Þ
Consequently, there is quite a bit of variability at the low and high range of N-values for predicting Vs (Table 1). Further, N-values of ,5 and .50 show
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 27–36
MASW- vs SPT-Derived S-Wave Velocity Table 1. Comparison of regression formulas for the derivation of Vs from SPT N-values using formulas 2 and 3, respectively. N 5 20 50
222.0 (N)0.06 75 242 528
Table 2. Comparison of Imai and Tonouchi (1982) and Inazaki (2006) formulas based upon the N-values provided in Table 1.
19.0 (N)0.85
N
91(N)0.337
99(N)0.3448
Average
244 266 280
5 20 50
156 250 340
173 280 384
164.5 265 362
Vs values in m/s.
Vs in m/s.
a large calculated Vs range between the two formulas and demonstrate the weaknesses with using the SPT method for the prediction of dynamic soil properties (Figure 2). Further, Inazaki (2006) states “… that it is hopeless to estimate S-wave velocity from generally incorrect [i.e., low or high] N-values.” Imai and Tonouchi (1982) and Inazaki (2006) provide regression equations that are more robust than most of the formulas reviewed. Imai and Tonouchi (1982) incorporate a large data set to derive their equation, whereas the Inazaki (2006) equation is based upon a strict data collection regiment. The Imai and Tonouchi (1982) equation is based upon 1,600 data sets from Japan: Vs ~91ðN Þ0:337 ;
ð4Þ
whereas, Inazaki (2006) showed results of 500 Nvalues: Vs ~99ðN Þ0:3448 :
ð5Þ
Both formulas appear to be robust in their approach to the prediction of Vs from N-values.
Figure 2. Continuous blow counts converted to Vs using regression equations from Imai and Tonouchi (1982) and Inazaki (2006). Note dashed straight line functions from N-value of one to six blows and from 30 to 50 blows.
The formulas from Imai and Tonouchi (1982) and Inazaki (2006) provide a narrow range of S-wave velocities for N-values. Not surprisingly, these equations provide similar results (Table 2). What is interesting is that the first six blow counts (N 5 6) account for Vs values from 100 m/s to 180 m/s, or 29 percent to 51 percent, respectively, of the predictive range of SPTs. The majority of engineering sites tested by the authors using the MASW method of Vs prediction have an engineering requirement of a Vs of 180 m/s as a minimum for an effective design basis, and therefore this value is referenced here only as a datum by which to further this discussion of SPT data. Consequently, an acceptable site, using only N-values, is one that has a design basis based upon six blow counts. Clearly, any errors in SPT data collection could compromise the analysis and conversion to Vs. Further, from an N-value of 30 and higher, the curve flattens to a straight line function (Figure 2). Consequently, the top 40 percent of the SPT curve (N 5 30 to 50) includes a predicted velocity range of only 50 m/s (300–350 m/s). As with any geophysical method of data collection there are weaknesses. When processing MASW data, the fundamental mode must be selected for the dispersion curve analysis to derive accurate shear wave velocities (Ivanov et al., 2008). Unfortunately, selecting a higher order mode as the fundamental mode for inversion will result in spurious results. This can occur, although rarely, when the fundamental mode blends with the subsequent modes; however, a fully trained interpreter can easily avoid this pitfall. A high-resolution dispersion curve will lead to a more accurate pick of frequencies and phase velocities. Acquisition factors influencing the dispersion curve resolution are number of channels, spread length, shot offset, and source (Park, 2014). During acquisition it is also beneficial to stack multiple records to diminish the contamination of the shot gather by unwanted data or noise. Hutchinson and Beird (2011) demonstrated that the active method of MASW data collection is only effective at predicting Vs to a depth of approximately 30 m below grade, consistent with the observations of
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 27–36
29
Hutchinson and Beird Table 3. A summary of the lithology listed by deposit type and location. Deposit
Locale
Lithology
Cultural
Western Pennsylvania Central Arizona South-central Kansas Central Saskatchewan Central Ontario North-central North Dakota South-central Puerto Rico Central Pennsylvania Southeastern Washington Texas Panhandle
Spoil below coal surface mine high wall Dynamically compacted construction and demolition debris Till Till; fluvial Till, fluvial, lacustrine clay Till, fluvial, lacustrine clay Fluvial, alluvial Fluvial Aeolian Aeolian, fluvial
Glacial
Fluvial Aeolian
Xia et al. (2002). Passive MASW methods of data collection have proven effective at the prediction of Vs to a depth of approximately 70 m below grade (Ivanov et al., 2008). ANALYSIS The data utilized within this article were derived from many studies that specifically addressed a depth range of at least 30 m below grade. Consequently, some of the disparities at depth can be attributed to lithologic attenuation of Raleigh waves, from too tight a survey geometry, and from the magnitude of the source. Four different depositional environments were used for this study. Changes in geometry and magnitude of the elastic shock wave will produce deeper data. The cultural deposit consists of anthropogenic fill, whereas the other three units—glacial, fluvial, and aeolian deposits—represent in situ conditions. Several examples from each deposit are used to compare and contrast the Vs values as derived from N-values and from MASW. The regression equations from Imai and Tonoughi (1982) and Inazaki (2006) were used to convert Nvalues to Vs for 10 sites in North America (Table 3). These two values for Vs were then averaged and compared to MASW-derived Vs for borings at two sites at 10 locales. Cultural Deposits Cultural deposits are an important part of urban development. As brownfields are brought back into use, on-site buried waste material can be a challenge for the developer. Construction and demolition debris (C&D) and mining spoil are common urban deposits encountered during development; however, municipal solid waste landfills are often used after closure as parks and other low-impact facilities (Hutchinson and Spieler, 1998; Hutchinson and Barta, 2003).
30
In general, C&D and mine spoil consist of a bimodal distribution of particle sizes. These deposits can contain cement or rock fragments that are several meters across mixed with sand-sized particles and multi-dimensional wood waste. Representative blow counts in C&D waste can be difficult to collect as buried concrete and brick can induce abnormally high blow counts. A well log, reviewed by the authors, documented the presence of a thick concrete slab at 15 m below grade with concomitant elevated SPT-generated Vs (Figure 3 A). MASW images the cultural waste as a unit and does not respond to discrete large objects in its analysis; consequently, no spike in the Vs is observed at 15 m below grade (Figure 3A). The Vs derived from the MASW is in general higher than that derived from SPT for mine spoil and C&D waste (see, for example, Figure 3A through C). The lower SPT values can be attributed to the split spoon tendency to find its way through the softer material (i.e., low N-values) until encountering rock or large refractory material. The MASW measurements in both coal mine spoil examples are roughly 300 m/s, showing that the readings represent inter-particulate contact (i.e., cement, rock, abandoned machinery, etc.) and may be more representative of the spoil dynamic properties. After dynamic compaction of C&D waste, SPTderived Vs values can be consistent with MASW data (Figure 3C and D). In this example, the pre-compaction SPT-derived Vs (Figure 3C) is lower than that derived from MASW measurements prior to dynamic compaction. After dynamic compaction the SPT- and MASW-derived Vs track closely until about 10 m below grade, where the two readings diverge (Figure 3D). Greater than 10 m below grade (Figure 3C), the SPT-derived Vs is consistently less than the MASW-derived Vs, suggesting that dynamic compaction impacted the material to a depth of 10 m below grade (Figure 3D).
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 27–36
MASW- vs SPT-Derived S-Wave Velocity
Figure 3. Velocity versus depth profiles for coal mine spoil from western Pennsylvania and C&D waste from central Arizona. Diagram C shows pre-compaction SPT-derived Vs with post-compaction MASW-derived Vs. Diagram D shows post-compaction SPT- and MASWderived Vs.
Glacial Deposits Glacial deposits consist of poorly sorted till, lacustrine clay, and fluvial sediments. Fine-grained aeolian and loess (i.e., wind-blown) deposits can also be glacially derived; however, this deposit is addressed separately as a result of the homogeneous nature of these sediments. Till with intercalated fluvial deposits are poorly sorted and can contain cobble or largesized fragments. In general, the Vs derived from SPT measurements are consistent with those derived from MASW measurements (Figure 4A through F). Several examples of divergence do exist between SPT- and MASW-derived Vs values, however (Figure 4B, C, and F). These differences fall into three categories based upon depth of the measurement (shallow and deep) or lithology. Lacustrine deposits contain a large percentage of water that could be suppressing the MASW-derived Vs measurement in comparison to the Vs derived from SPT: for example,
the lake clay found 4 to 7 m below grade (Figure 4B) and that found 8 to 9 m below grade (Figure 4D). Divergence also tends to occur at depths of greater than 10–15 m, where MASW can be weakest in its prediction of Vs based upon the project-derived MASW data collection methodology (Figure 4B, C, and F). The divergence typically shows the MASW prediction of Vs to be less than that derived by the SPT method (Figure 4C and F). However, MASW can predict a Vs greater than that predicted by SPT methods at depths of greater than 10 m (Figure 4B). This disparity can be attributed to the geometry of the survey (1.5-m geophone step-out distance), to the magnitude of initiated elastic shock wave (via 3.75-kg sledge hammer), and possibly due to lithologic contrasts. A larger step-out distance and heavier weight source would probably resolve this difference. In most cases, the Vs values derived from both methods in the shallow portion of each test site are consistent (Figure 4C). In this case, the disagreement
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 27–36
31
Hutchinson and Beird
Figure 4. Velocity versus depth profiles of glacial deposits from the United States (North Dakota) and Canada (Saskatchewan and Ontario).
between the two methods is probably related to poor SPT interval measurement (five blow counts) and weight on bit, as near-surface MASW-derived estimates of Vs are usually very reliable indicators of true Vs (Park, 2014). 32
Fluvial Deposits Fluvial deposits are poorly sorted and consist of sand- to cobble-sized particles intercalated with overbank clay and silt. In general, MASW-derived Vs values are equivalent to those derived from SPT
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 27–36
MASW- vs SPT-Derived S-Wave Velocity
Figure 5. Velocity versus depth profiles of fluvial deposits from central Pennsylvania, south-central Puerto Rico, and south-central Kansas.
measurements (Figure 5A through F). This equivalence can exist to a depth of 20 m below grade; however, agreement can be weak greater than 10 m below grade as a result of lithology, survey geometry, and magnitude of initiated elastic shock wave (Figure 5A and F). Any variation with the SPT- and MASW-derived Vs values can be attributed to large fragments within poorly sorted units. For example, large cobbles were encountered in an example from Puerto Rico (Figure 5C) between 4 and 6 m below grade. MASW-derived Vs did not reflect the increased Vs due to the bimodal distribution of the sediment size, measured as a unit and not as an individual particle.
Aeolian Deposits Aeolian deposits consist of very well-sorted windblown sand and silt. Typically the SPT- and the MASW-derived Vs values are very consistent (Figure 6A through D). Similar to the comparison for glacial deposits, there can be divergence between the two measurements in the shallow and deep portions of the site. The shallow divergence is typically not too great, with the MASW-derived Vs slightly higher than the SPT-derived Vs (Figure 6B and D). In addition, the deep divergence is not too great, with the MASW-predicted Vs slightly greater than the SPT-derived Vs (Figure 6D).
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 27–36
33
Hutchinson and Beird
Figure 6. Velocity versus depth profiles for aeolian deposits from the southeastern Washington State and the Panhandle of Texas.
Lithologic and clast size differences have a dramatic effect on the prediction consistency of the two methods. Borehole logs for Figure 6B and C indicate that these borings penetrated a fluvial unit at 3 m and 4.75 m, respectively. The lithologic change from wellsorted aeolian to poorly sorted fluvial deposit, with its concomitant coarse-sized fraction, attributed to the divergence in Vs measurements. Consequently, wellsorted and probably finer-grained deposits produce more consistent Vs predictions between the MASW and SPT methods. CONCLUSION SPT data conversion to Vs encompasses numerous formulas to express the relationship, many of which are not reasonable in predicting Vs. Furthermore, the N-values are an analog method of data collection,
34
through which continuous information is not collected. Low N-values (i.e., ,6) are very insensitive to Vs prediction where there is a significant change with every blow count. Moreover, N-values greater than 30 blow counts produce a narrow Vs range, leading to imprecise data. Large objects are not detected by MASW methods, yet the SPT method can produce spuriously high readings for the depth zone in which the large object is encountered. The MASW method continuously collects data, providing a digital representation of Vs throughout the entire depth of investigation, which, in the cases presented within this study, are limited to approximately 20 m below grade. MASW- and SPT-derived Vs values tend to diverge at approximately 20 m below grade; however, it is unclear which method produces more accurate readings. The limitation on depth of detection is due to the attenuation of
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 27–36
MASW- vs SPT-Derived S-Wave Velocity
Rayleigh waves with depth. The attenuation of Raleigh waves can be ameliorated by expanding the survey and increasing the source magnitude. However, lithology is also a factor in the attenuation of Raleigh waves, and there is no adjustment for this. Vs, derived from MASW testing, is more sensitive to subsurface lithologic changes and should be included within every soil dynamic geotechnical study of the subsurface. Homogeneous (and possibly isotropic) deposits within the sand-sized fraction of a deposit produce consistent Vs values between MASW- and SPT-derived methods. Shear wave velocity measurements from these two methods can show some disagreement in poorly sorted deposits. However, these differences are not that significant and can be minimized by the collection of more MASW tests in the area. These two test methods are complimentary, and while SPT is the “Engineering Standard,” 20–30 MASW tests can be collected for every boring. So in this regard, the two tests are complementary, and each contains strengths and weaknesses, so using both methods is recommended. ACKNOWLEDGMENTS The authors would like to thank all of the engineers and geologists who have contributed well logs for the MASW we have collected. In many cases these data were proprietary to the client and sometimes great effort was applied to obtain approval for the release of the corresponding well logs. We would like to thank the good people of RES-Americas, Pattern Energy, Terracon, CMT Laboratories, and Tulloch. Engineers and geologists who have helped collect well logs include Mohamed Nofal, Jordan Black, Michael Ehss, Richard Gray, Lloyd Pasley, Shad Hoover, and James Dermody, and for this we owe a big debt of gratitude. REFERENCES AMERICAN SOCIETY FOR TESTING AND MATERIALS (ASTM), 2008, Standard Test Method for Standard Penetration Test (SPT) and Split-Barrel Sampling of Soils: ASTM Standard Method D1586-08a: ASTM, West Conshohocken, PA. ANDRUS, R. D.; PIRATHEEPAN, P.; ELLIS, B.; ZHANG, J.; AND JUANG, H., 2004, Comparing liquefaction evaluation methods using penetration VS relationships: Soil Dynamics Earthquake Engineering, Vol. 24, pp. 713–721. BROWN, L. T.; BOORE, D. M.; AND STOKOE, K. H., 2002, Comparison of shear wave slowness profiles at 10 strong motion sites from noninvasive SASW measurements and measurements made in boreholes: Bulletin Seismologic Society America, Vol. 92, pp. 3116–3133. CHIEN, L. K.; LIN, M. C.; AND OH, Y. N., 2000, Shear wave velocity and SPT-N values of in-situ reclaimed soil in west Taiwan: Journal Geotechnical Engineering, Vol. 31, pp. 63–77.
DENNIS, R.; HILTUNEN, S. M.; AND WOODS, R. D., 1998, SASW and cross-hole test results compared. In Earthquake Engineering Soil Dynamics II, Recent Advances in Ground Motion Evaluation: Geotechnical Special Publication 20: American Society of Civil Engineers, Park City, UT, pp. 279–289. FUMAL, T. E. AND TINSLEY, J. C., 1985, Mapping Shear Wave Velocities of Near Surface Geological Materials: Predicting Aerial Limits of Earthquake Induced Landsliding: U.S. Geologic Survey Professional Paper 1360, pp. 127–150. HANUMANTHARAO, C. AND RAMANA, G. V., 2008, Dynamic soil properties for microzonation of Delhi, India: Journal Earth Science, Vol. 117, No. 2, pp. 719–730. HUTCHINSON, P. J. AND BARTA, L. S., 2003, Imaging your way to a better brownfield site. In RevTech: Cleaning Up Contaminated Properties for Reuse and Revitalization: Effective Technical Approaches and Tools Conference: Environmental Protection Agency, Pittsburgh, PA, 78 p. HUTCHINSON, P. J. AND BEIRD, M. H., 2011, A shear wave velocity comparison—MASW to SPT. In Proceedings of the 24th Symposium of the Application of Geophysics in Engineering and Environmental Problems (abs.): Environmental Engineering Geophysical Society, Charleston, SC. HUTCHINSON, P. J. AND SPIELER, R., 1998, Characterization of waste disposal facilities through geophysical methods: A case study from Boston, MA. In Ogunro, V. O. (Editor), 4th International Symposium on Environmental Geotechnology and Global Sustainable Development: International Society Environmental Geotechnology, Boston, MA, pp. 1197–1206. HUTCHINSON, P. J.; TESCHKE, B. J.; ZOLLINGER, K. M.; AND DEREUME, J. M., 2008, Field applicability of MASW data. In: Proceedings of the 21st Symposium of the Application of Geophysics in Engineering and Environmental Problems: Environmental Engineering Geophysical Society, Philadelphia, PA, pp. 1226–1231. HVORSLEV, M. J., 1949, Subsurface Exploration and Sampling of Soils for Civil Engineering Purposes: U.S. Waterways Experiment Station, Vicksburg, MS, 552 p. IMAI, T., 1977, P- and S-wave velocities of the ground in Japan. In Proceedings of the 9th International Conference Soil Mechanics and Foundation Engineering: International Society of Soil Mechanics Geotechnical Engineering, Tokyo, Japan, pp. 257–260. IMAI, T. AND TONOUCHI, K., 1982, Correlation of N-value with S-wave velocity. In Verruijt, A.; Beringen, A. H.; and de Leeuw, E. H. (Editors), Proceedings of the 2nd European Symposium on Penetration Testing: International Society Soil Mechanics Foundation, Amsterdam, The Netherlands, pp. 67–72. INAZAKI, T., 2006, Relationship between S-wave velocities and geotechnical properties of alluvial sediments. In Proceedings of the 19th Symposium on Application Geophysics Engineering to Environmental Problems: Environmental Engineering Geophysical Society, Seattle, WA, pp. 1075–1085. IVANOV, J.; PARK, C.; AND XIA, J., 2008, MASW/SurfSeis2 Workshop: Kansas Geological Survey, Lawrence, KS, 200 p. JAFARI, M. K.; SHAFIEE, A.; AND RAZMKHAH, A., 2002, Dynamic properties of fine grained soils in south of Tehran: Journal Seismologic Earthquake Engineering, Vol. 4, No. 1, pp. 25–35. KANSAS GEOLOGICAL SURVEY (KGS), 2010, SurfSeis: Seismic processing software, version 3.064: KGS, Lawrence, KS. KAYABALI, K., 1996, Soil liquefaction evaluation using shear wave velocity: Engineering Geology, Vol. 44, pp. 121–127. MILLER, R. D.; XIA, J.; PARK, C. B.; AND IVANOV, J. M., 1999, Multichannel analysis of surface waves to map bedrock: Leading Edge, Vol. 18, No. 12, pp. 97–173.
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 27–36
35
Hutchinson and Beird MILLER, R. D.; XIA, J.; PARK, C. B.; AND IVANOV, J. M., 2001, Shear wave velocity field to detect anomalies under asphalt. In Lodge, R. G. (Editor), 52nd Highway Geology Symposium: (abs): Maryland Geologic Survey, Baltimore, MD. NAZARIAN, S. AND STOKOE, K. H., 1984, In situ shear wave velocities from spectral analysis of surface waves. In Proceedings of the 8th World Conference on Earthquake Engineering: International Association for Earthquake Engineering, San Francisco, CA, pp. 31–38. NOVOTECH SOFTWARE LTD., 2010, NovoSPT: Proprietary SPT Correlation Computer Software: NovoTech Software, British Columbia, Canada. OHSAKI, Y. AND IWASAKI, R., 1973, Dynamic shear moduli and Poisson’s ratio of soil deposits: Soils Foundation, Vol. 13, pp. 61–73. OHTA, Y. AND GOTO, N., 1978, Empirical shear wave velocity equations in terms of characteristic soil indexes: Earthquake Engineering Structural Dynamics, Vol. 6, pp. 167–187. OHTA, Y.; GOTO, N.; KAGAMI, H.; AND SHIONO, K., 1978, Shear wave velocity measurement during a standard penetration test: Earthquake Engineering Structural Dynamics, Vol. 6, pp. 43–50. PARK, C. B., 2014, Symposium on the Application of Geophysics to Engineering and Environmental Problems, Boston, MA, March 20, 2014. PARK, C. B.; MILLER, R. D.; AND XIA, J., 1999, Multi-channel analysis of surface waves: Geophysics, Vol. 64, No. 3, pp. 800–808. ROLLINS, K. M.; DIEHL, N. B.; AND WEAVER, T. J., 1998a, Implications of Vs-BPT (N1)60 correlations for liquefaction
36
assessment in gravels. In Dakoulas, P.; Yegian, M.; and Holtz, R. D. (Editors), Geotechnical Earthquake Engineering and Soil Dynamics, Geotechnical Special Publication No. 75. American Society Civil Engineers, Reston, VA, Vol. 1, pp. 506–517. ROLLINS, K. M.; EVANS, M. D.; DIEHL, N. B.; AND DAILY, W. D., 1998b, Shear modulus and damping relationships for gravels: Journal Geotechnical Geoenvironmental Engineering, Vol. 124, No. 5, pp. 396–405. STOKOE, K. H.; NAZARIAN, S.; RIX, G. J.; SANCHEZ-SALINERO, I.; SHEU, J.; AND MOK, Y., 1988, In situ seismic testing of hard-to-sample soils by surface wave method. In Von Thun, J. L. (Editor), Earthquake Engineering and Soil Dynamics II, Recent Advances in Ground-Motion Evaluation: American Society of Civil Engineers, Park City, UT, pp. 264–278. SYKORA, D. W. AND KOESTER, J. P., 1988, Review of existing correlations between shear wave velocity or shear modulus and standard penetration resistance in soils. In Von Thun, J. L. (Editor), Earthquake Engineering and Soil Dynamics II Conference, Recent Advances in Ground-Motion Evaluation: American Society of Civil Engineers, Park City, UT, pp. 389–404. XIA, J.; MILLER, C. B.; AND PARK, J. A., 1999, Estimation of nearsurface shear-wave velocity by inversion of Rayleigh wave: Geophysics, Vol. 64, pp. 691–700. XIA, J.; MILLER, R. D.; PARK, C. B.; HUNTER, J. A.; HARRIS, J. B.; AND IVANOV, J., 2002, Comparing shear-wave velocity profiles inverted from multi-channel surface wave with borehole measurements: Soil Dynamics Earthquake Engineering, Vol. 22, pp. 181–190.
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 27–36
Reliable Soil Property Maps over Large Areas: A Case Study in Central Italy GIULIA FANELLI1 DIANA SALCIARINI2 CLAUDIO TAMAGNINI2 Department of Civil and Environmental Engineering, Via G. Duranti 93, 06125, Perugia, Italy
Key Terms: Georeferenced Database, Soil Strength Variability, Geostatistics, Kriging
ABSTRACT In this article we present results of a data collection project carried out to create a reliable georeferenced database to characterize the soil types constituting the near surface cover of study areas located in central Italy. The database includes the following features: i) the coordinates of the site investigations; ii) the geotechnical parameters; iii) the shear wave velocity in the upper 30 m of soil; iv) the type of test used to evaluate the geotechnical parameters and the shear wave velocity; and v) the general description of the stratigraphy. Preliminary analyses on the data were performed to determine the average values, the distribution of the measured data over the intervals, and the probability density function that best fits the measured values. Secondly, geostatistical analyses were done to assess the spatial correlation between data. Among the soils considered, only gravels show a low correspondence between the experimental variograms and the mathematical curve fitting them, while for all of the other soils, such agreement is high. Finally, two applications of the newly developed database are proposed. The first one is the development of continuous soil property maps for a selected study area, created from the information included in the database, which is constrained by a discrete amount of information. In the second application approximately the 80 percent of the measured data are considered to provide spatial predictions, which are tested with the remaining 20 percent of the data.
INTRODUCTION AND MOTIVATION In recent years, the use of physically based (PB) models to predict the occurrence of landslides has 1
Corresponding author email: giulia.fanelli@strutture.unipg.it. Emails: diana.salciarini@.unipg.it; claudio.tamagnini@.unipg.it
2
increased since reliable landslide hazard maps are needed to reduce damage and human losses. The PB approaches can be used to predict both rainfallinduced and earthquake-induced landslides with accuracy: successful examples of this application are presented in the works of Montgomery and Dietrich (1994), Jibson et al. (2000), Refice and Capolongo (2002), Salciarini et al. (2006), Saygili and Rathje (2009), Wang and Lin (2010), Salciarini et al. (2012), and Park et al. (2013). PB models are capable of handling all the physical processes occurring during a landslide event and therefore provide a substantial contribution to mapping hazardous areas. But, in practice, their application is still quite limited as a result of the difficulty inherent in collecting a representative number of data points, particularly over large areas. Indeed, one of the crucial factors that controls the accuracy of the PB predictions is the need for a detailed database of physical and mechanical properties of soils for the selected study area, which is rarely available. Although there is always a degree of uncertainty in the quality of a data set used for a model application (as a result of spatial variability, measurement and interpretation errors, model hypotheses or generalizations, etc.), it can be reduced by increasing the quality and the density of input data. In practice this is difficult since typically information (deriving from boreholes, surveys, or geophysical investigations) is widely spaced and randomly distributed over a large area (see, i.e., Park et al., 2013). In these cases the model results are affected by the paucity of information. In the literature, there are several examples of applications of the PB models for rainfall-induced and earthquake-induced landslide predictions to areas adequately characterized from the physical and mechanical point of view. For example, 1) Pack et al. (1998a) performed an analysis of the Kilpala watershed of northern Vancouver Island (British Columbia, Canada) through the SINMAP model (Pack et al., 1998b); 2) Simoni et al. (2008) applied GEOtop-FS in an alpine watershed in Italy; 3) Baum et al. (2010)
Journal of Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 37–52
37
Fanelli, Salciarini, and Tamagnini
produced landslides hazard maps for Seattle (WA, USA) using TRIGRS (Baum et al., 2008); and 4) in Italy, Salciarini et al. (2013) defined PB rainfall thresholds for early-warning systems, discussing the challenges in their application to a case study, while Salciarini and Tamagnini (2015a, 2015b) provided applications of PB models for assessing the critical rainfall thresholds over large areas. As for the evaluation of the seismic landslide hazard, Jibson et al. (2000), starting from an extensive geotechnical data set, obtained a probability curve relating predicted seismic displacement to probability of failure. In addition, Saygili and Rathje (2009) presented a probabilistic methodology for slope stability assessments after an earthquake, and Fanelli et al. (2015) developed a model to predict the spatial distribution of earthquake-induced landslides at a regional scale, based on the Newmark method. For Italy, a complete, systematically structured, and organized database that characterizes different soil types at large scale, providing information on their spatial distribution, is still not available. Some examples of well-documented investigations and data collections are available; however, they typically focus only on a single soil type, ignoring the distribution of the other soil types in the study area (Cotecchia and Chandler, 1995; Raspa et al., 2008). In the Umbria Region of central Italy particular morphological, geotechnical, and hydraulic conditions are very often predisposing for landslides and are relevant factors in the landscape classification for risk (see, e.g., Cardinali et al., 2002; Reichenbach et al., 2004; and Guzzetti et al., 2006). Despite the fact that the application of PB models would improve the current techniques used for management and territorial planning, their application has been slow because a reliable, complete, and well-organized database for the input data required by such models is still lacking. This work contributes to the creation of such a database. The database covers the entire Perugia Province and constitutes a high-resolution collection for reliable applications of PB models to predict rainfall- and earthquake-induced landslides. Once the database was created, the quality of the data was investigated to better understand its reliability and accuracy. To achieve this goal, statistical analyses were performed to quantify the distribution of the data (i.e., the histograms of the absolute frequency). Furthermore, to understand the level of spatial correlation between the data, geostatistical analyses were carried out to create the variograms. The variograms computed for each soil type form the basis for predicting the best estimation of the value of a parameter in a generic point in the study area, where such information is not currently available. To this aim, a well-known
38
interpolation technique, the Kriging Method, was used. This method provides a prediction of a function in a given point as a weighted average of known values of such function in the nearest points, deriving the weights from the variograms. DATA COLLECTION To include georeferenced data (physical and mechanical properties) of the soils in the analysis, a thorough procedure (conducted over 2 years) was undertaken to obtain information from the archives of the Structural Control and Civil Protection Office of the Perugia Province. The task was to verify reports containing geological and geotechnical information submitted in order to obtain a permit for building any structure. Approximately 8,000 records were thoroughly examined, and those with incomplete or unreliable data were excluded from the study. For example, when the datum was not specified or the method for evaluating the soil strength parameters was not clearly described, the record was not included in the database. The resulting database contains this information at about 3,000 data points covering the entire Perugia Province (Figure 1a and 1b). The data have the highest density in urban centers. Since the archived records were converted into digital data, each data point had to be geospatially rectified and manually entered into the georeferenced database. Each record holds the following information: a) coordinates of the site of investigation; in latitude and longitude with datum ED50 or WGS84; b) geotechnical parameters; (peak) effective cohesion c9 and (peak) friction angle Ď•9; c) shear wave velocity in the first 30 m of soil (Vs,30); d) type of investigation or test used to measure the geotechnical properties; e) type of geophysical method used to determine the Vs,30; and f) general description of the stratigraphy at the site. An example of records included in the database is shown in Table 1. The geotechnical parameters, effective cohesion c9 and effective friction angle Ď•9, characterize the shear strength of soils when drained. They can be evaluated through mechanical tests, such as laboratory tests or in situ investigations. These data provide the fundamental information for performing stability analyses. Shear wave velocity is a parameter for evaluating the dynamic behavior of soils under different seismic conditions. In addition, the shear wave velocity, averaged within the top 30 m of soil, is a parameter required by the European Code EC8 (CEN, 2003)
Journal of Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 37–52
Reliable Soil Property Maps over Large Areas
Figure 1. Maps showing (a) the Perugia Province of the Umbria Region in central Italy and (b) the distribution of the geological and geotechnical data collected in the Perugia Province territory.
and by the Italian Building Code (Ministero delle Infrastrutture, 2008) to classify sites according to the soil type. It is defined as follows: 30 Vs;30 ¼ P ; n hi vi
(1)
i¼1
where hi and vi denote the thickness and the shear wave velocity of the i-th layer existing in the top 30 m and n is the total number of layers. The shear wave velocity is a valuable indicator of the dynamic properties of the soil because of its relationship with the soil shear modulus G and the density (ρ), derived by G 5 ρV2s,30. In addition, the data allow calculating the stratigraphical amplification of the seismic waves moving from the bedrock to the surface. Shear wave velocity can be obtained from geophysical investigations, either as indirect surface tests or by down-hole logging.
For each data point, the method used to evaluate both the geotechnical parameters (cohesion and friction angle) and the shear wave velocity is included in “d” and “e” in the list above. This information is used in the statistical analyses to evaluate the recurrence of each type of method and to provide a first classification on the accuracy. The mechanical investigation performed and recorded in the database has been grouped into the following categories: i) penetration test, both Dynamic (Standard Penetration Tests, Dynamic Penetration Super Heavy) and Static (Cone Penetration Tests); ii) excavation and survey, such as borehole drilling and continuous core drilling; iii) Bieniawski (1989) classification; iv) laboratory test (direct shear, triaxial), and v) other. The recurrence of each category is summa‐ rized in Figure 2a. For most of the investigations (approximately 65 percent), penetrometric tests (dynamic and static) were conducted. For 16 percent of the tests, boreholes
Table 1. Example of a record included in the database. Latitude
Longitude
c9 (kPa)
ϕ9 (u)
Mechanical Test
Vs,30 (m/s)
Geophysical Test
Stratigraphy
12.168u
43.116u
10
30u
Lab test
300
MASW
Alternation of sands and silts
MASW 5 multi-channel analysis of surface waves.
Journal of Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 37–52
39
Fanelli, Salciarini, and Tamagnini
Figure 2. Frequency of tests used to evaluate (a) the geotechnical parameters and (b) the shear wave velocity of soil.
and excavations were conducted. Where rock was encountered, the RMR classification system (Bieniawski, 1989), based on multi-parametric indices, was applied (8 percent of the tests). For 3 percent of the tests, laboratory analyses (direct shear, triaxial) were performed. Finally, for 8 percent of the tests, other procedures, such as geological surveys, in situ testing, data from previous studies, or testing in neighboring areas, were used. The same procedure was used to classify the geophys‐ ical tests performed to characterize the seismicity of soils and to evaluate their shear wave velocities. The European Code EC8 (CEN, 2003) introduced the definition of the soil category to quantify the stratigraphical amplification. Such a parameter is derived from the average value of the shear wave velocity in the top 30 m of soil (Vs,30), and it is identified with letters ranging from “A” (no stratigraphical amplification) to “E” (highest stratigraphical amplification). The results of geophysical tests compiled in the database were grouped into the following categories: 1) 2) 3) 4)
surface test with survey; surface test with no survey; in-hole test; and other
The surface tests include several techniques, such as seismic refraction and reflection, spectral analysis of surface waves, multi-channel analysis of surface waves, refraction micro-tremor, and horizontal-to-vertical spectral ratio. The borehole tests include down-hole and cross-hole methods. The results of statistical anal‐ yses conducted on this data are presented in Figure 2b. The surface tests are the most widely used type of investigations (87 percent); in 73 percent of the tests, they were conducted with surveys that make the test results more reliable, while 14 percent of the surface tests were performed without surveys. Only 1 percent
40
of the investigations were done in a borehole, and 12 percent included other techniques, such as penetrometric tests or empirical methods. Finally, general stratigraphic descriptions were compiled at sites in which the soils were exposed in outcrop. These descriptions provided information about the soil in the upper 10 m. ANALYSIS OF THE RELIABILITY OF THE DATABASE In a large area such as the Perugia Province (covering 6,337 km2), there is a large number of soil types having different physical and mechanical properties. For this reason, the properties included in the database are assumed to be very different, depending upon the soil type they are describing. To investigate the reliability of the database most efficiently and thoroughly, the soil covers were grouped into a limited number of categories, each one characterized by similar physical and mechanical properties. Beginning with the digital geological map of the Umbria Region (scale 1:10,000) (Figure 3), which includes about 250 different geological units classified according to their lithology and relative age, it was possible to group the soils into a small number of categories (i.e., 16 lithologic classes). These lithologies are shown in Figure 3. In order to associate each record in the database to one of the 16 lithologic classes, the georeferenced data points were superimposed over the geological map. In this manner, the data points were split in the 16 lithological classes identified in the territory of the Umbria Region. The data in each of these classes were further analyzed to identify and exclude erroneous records. For example, when the stratigraphic description at one data point does not relate to the soil class to which it belongs, the record is excluded.
Journal of Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 37–52
Reliable Soil Property Maps over Large Areas
Figure 3. SimpliďŹ ed geological map of the Umbria Region, including the locations of the collected data.
A small number of the classes are represented by bedrock formations (i.e., limestone, marl limestone, conglomerate, flysh, marl, travertine). As is well known, the physical and mechanical properties of rocks and soils are very different (Jaeger et al., 2009; Bell, 2013). Furthermore, in the practical applications related to landslide identification, the failure mechanisms that can involve either rocks or loose soils are often very different (Bieniawski, 1989; Wyllie and Mah, 2004). For these reasons, these material types need to be treated separately. In this study, attention was focused on the characterization of soils which, compared to rocks, have generally the following features: 1) lower strength and higher deformability, producing a more critical behavior in the presence of certain triggering factors, such as storms or earthquakes; and 2) more reliable estimates of their physical and mechanical properties—derived from results of site investigations and lab tests. In this case study, rock type is usually identified through indirect methods, such as the Bieniawski (1989) classification, based on the estimation of an index obtained
by summing up the ratings of six different parameters of rock. Therefore, from this point onward, only soils that are classified as loose sediments are considered. This reduces the database to 10 soil classes (alluvium, terraced alluvium, clay, eluvial deposit, pyroclastic deposit, landslide deposit, gravel, silt, sand, and turbidite). The number of data points for each soil cover is listed in Table 2 (no points collected in the database fall into the pyroclastic deposit). Table 2. Number of data points for each soil cover. Soil Cover Type
No.
Alluvium Terraced alluvium Clay Eluvial deposit Landslide deposit Pyroclastic deposit Gravel Silt Sand Turbidite
848 364 154 160 204 0 126 149 235 483
Journal of Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 37–52
41
Fanelli, Salciarini, and Tamagnini
Figure 4. Scatter plots of (a) the friction angle and (b) the cohesion for the soil covers in the Perugia Province.
The soil strength parameters (cohesion c9 and friction angle ϕ9) at the locations included in the geotechnical database have been examined and distinguished on the basis of the soil classes to which they belong and their value. They are shown in Figure 4, with different colors representing the classes and with different sizes depending on the values. The scatter plot allows one to view, at a glance, how the classes of the data are distributed throughout the Perugia Province. The size of the circles provides a quantitative value of a given parameter. For the friction angle ϕ9 (Figure 4a), the plot shows modest differences, because generally the range of variation of ϕ9 is quite small, and all the lithologic classes are characterized within a narrow range. The differences in cohesion, represented by varying circles sizes, are much more evident (Figure 4b), meaning that this parameter has much greater variability. The strength parameters of the soil have been further analyzed in order to construct the absolute frequency histograms. The shape of the histograms allows one to understand the probability density function that best fits the experimental data. For each soil class, two absolute histogram distributions have been generated, one for the cohesion c9 and one for the friction angle ϕ9. For simplicity, only the histograms for alluvium, terraced alluvium, and turbidite are presented (Figure 5). Figure 5 (a to c) shows the histograms of the cohesion, plotted on the lognormal distributions. In general, a lognormal distribution can be useful for modeling positive correlated variables, such as cohesion, elastic modulus, etc. These variables may have a mode in the distribution just above zero, and the distribution tapers off gradually for larger values.
42
For cohesion, it is possible to recognize from the histograms that the mode is associated with the smallest values of c9, very close to zero. Beyond this mode, the distribution gradually decreases, tending toward zero as cohesion increases. Concerning the absolute frequency histograms of the effective friction angle, the mode is always far from zero, as expected, and in most cases corresponds to values of approximately 30u. Above and below 30u there is a lower probability of occurrence, represented by the lower frequencies, which are further away from the mode. In this study, the normal distribution is the best theoretical probability distribution representing the variation of the friction angle for the soil types analyzed in Figure 5d–f). A normal distribution with mean m and variance σ2 is a statistic distribution with a probability density function defined by 1 ðx mÞ2 : PðxÞ ¼ pffiffiffiffiffiffi exp 2r2 r 2p
(2)
A lognormal distribution, instead, is a statistic distribution of a random variable the logarithm of which has a normal distribution; its probability density function has the following expression: PðxÞ ¼
1 ðln x mÞ2 pffiffiffiffiffiffi exp ; x > 0: 2r2 x r 2p
(3)
There are only a few classes that exhibit irregular behavior (e.g., the eluvial deposit, the landslide deposit, and the turbidite) (Figure 5c). Each class is characterized by a decreasing frequency from the mode, except for a near-maximum frequency in cohesion at approximately 100 kPa. The histograms for
Journal of Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 37–52
Reliable Soil Property Maps over Large Areas
Figure 5. Absolute frequency histograms of the cohesion and of the friction angle for three soil classes: (a) and (d) alluvium, (b) and (e) terrraced alluvium, (c) and (f) turbidite.
the cohesion of sand (Figure 6a) and gravel (Figure 6b) classes also have a unique distribution not observable in the other soil classes: specifically, the frequency in the first interval is much higher than the following, suggesting that gravel and sand represent coarsegrained materials that have very low cohesion, often equal to zero. Greater values of c9 can be due to cementation phenomena. The properties of the soil covers have been further investigated using boxplots to visualize the degree of variance of the data. In Figure 7, each class is represented by a box on which the central mark represents the median and corresponds to the value separating the upper half of the data from the lower half. In other words, the top and bottom of the boxplot are aligned with the 25th and 75th percentiles (first and third quartiles), and the whiskers extend to the maximum and minimum data. Outliers are plotted individually with blue circles (Figure 7).
The values of the 25th and 75th percentiles of a sample depend on the number (n) of the elements of the sample and are computed as follows: I25 ¼
25ðn þ 1Þ ; 100
(4)
I75 ¼
75ðn þ 1Þ : 100
(5)
Figure 7a shows that the gravel and sand classes have very low cohesion (median very close to zero), while the other classes are more cohesive, with medi‐ ans ranging from 10 to 20 kPa, except for the turbidite class, which has a median greater than 20 kPa. This distribution can be expected for turbidite since these materials are deposited in a subaqueous environment and the material may be cemented to some degree. The most extended boxes are associated with the
Figure 6. Absolute frequency histograms of the cohesion for (a) sand and (b) gravel soil classes.
Journal of Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 37–52
43
Fanelli, Salciarini, and Tamagnini
Figure 7. Variability of (a) the cohesion and (b) the effective friction angle.
landslide deposit and the turbidite (i.e., the data range is larger): this result was expected because these materials are made up of a larger range of grain sizes. In fact, the distance between the 25th and the 75th percentile is consistent. Figure 7b shows the friction angle for the soil classes. Generally there is less variation in the data; in fact, the range is only a few degrees from the medi‐ an. Also, there is very little variability between the median values, between 27u and 30u.
GEOSTATISTICAL ANALYSIS: THE VARIOGRAM CREATION Geostatistics is a relatively recent discipline—it was developed in the 1950s by Danie Krige (Krige, 1951)— and is based on the observation that there exists a spatial correlation between the attributes of a spatial variable: the closer to each other the observations are, the more similar they will appear. Thus, in making predictions, the closest sample points will assume a greater weight than those located further apart. Compared to classical statistics that examine statistical distributions of a set of sampled data, geostatistics includes the evaluation of their spatial correlation. Geostatistics is fundamental in analyses in which there are sparse data from field observations and where predictions in unsampled locations are necessary. This is very common in the geosciences since, as is well known, soil strength parameters are spatially distributed (i.e., they assume values that vary continuously from point to point in the space). Geo‐ statistics allow the data analysis to be extrapolated by means of experimental variograms where data are sparse or not available.
44
For this reason, the data set of cohesion c9 and friction angle ϕ9 for each soil class has been further analyzed to obtain the empirical variogram. The experimental semi-variogram γ(h) was first introduced by Matheron (1963) as half the average squared difference between the paired data values: N ðhÞ 1 X ðzðxi Þ zðxi þ hÞÞ2 ; cðhÞ ¼ 2N ðhÞ i¼1
(6)
where N(h) is the set of all pairs of data whose locations are separated by a lag distance h and z(xi) and z(xi + h) are data values at spatial location xi and xi + h. Thus, the semi-variogram provides the variance of the difference between field values at two locations. It is logical to expect that values increase with respect to distance, because the further apart two data points are located, the less correlated they are. The variogram is twice the semi-variogram, and for simplicity we will refer to the variogram as if it was the semi-variogram. When a variogram is calculated, it is important that the data pairs considered for a lag distance h¯ (i.e., the distance at which the variogram is calculated) are at least 30 and that the maximum value of h¯ is less than L/2, where L is the maximum distance over the field of data. In this case a maximum lag distance of 20 km has been set, since analyzing the spatial correlation of geotechnical parameters at greater distances is not necessary. In order to provide a prediction of an attribute, the experimental variogram needs to be converted in a theoretical one, based on model distributions. The most commonly used theoretical variograms are the 1) spher‐ ical, 2) exponential, and 3) Gaussian. In this study, the
Journal of Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 37–52
Reliable Soil Property Maps over Large Areas
Figure 8. Variogram of the cohesion and friction angle for the alluvium soil class.
Figure 9. Variogram of the cohesion and friction angle for the clay soil class.
exponential model was used, as it is the best one for fitting the experimental variograms. The procedure to obtain the experimental variograms of cohesion and friction angle data was programmed into Matlab. Once the experimental variogram was computed, it was fitted using an exponential model. The variograms of only alluvium and clay are shown in Figures 8 and 9 as examples. The variograms have been calculated assuming that the spatial correlation is uniform in all directions: the variogram is isotropic and described as omni� directional. As expected, Figures 8 and 9 show globally increasing functions with a horizontal asymptote, confirming that the spatial correlation between the data points is high for short distances and decreases for greater distance. In general, the variograms are characterized by curves with very low mean-squared error with respect to the experimental data. In only a few cases were the experimental points characterized by a greater dispersion, even though the distribution is trending upward. In addition, experimental variograms of cohesion are less regular than for friction angle. This is due to the fact that there is a higher variability in the cohesion with respect to the friction angle. That can explain why the experimental variograms of the friction angle follow a trend that is easier to model by a mathematical curve. Among the analyzed classes, only gravel shows a very low correspondence between the experimental variograms and the exponential curve-fitting data points (Figure 10). This may be related to the small sample size, which is more influenced by each data point.
Once the theoretical variogram has been identified, it is possible to use it for performing simulations that will provide the best estimation of a parameter for which data are not available. To achieve this goal, Kriging was used; it is one of the most commonly used methods to make interpolations and provides a prediction of a function based on weighted averages of the closest points. This method derives the weights from the variogram, providing unbiased estimates with minimum variance. In the following paragraph, the variograms obtained for the soil classes will be tested in study areas to show the applications of the Kriging Method.
Figure 10. Variogram of the cohesion and friction angle for the gravel soil class.
Journal of Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 37–52
45
Fanelli, Salciarini, and Tamagnini
APPLICATION OF THE GEOSTATISTICS TO TWO SELECTED STUDY AREAS IN CENTRAL ITALY Study Area 1
Figure 11. Location of the two study areas in the Perugia Province of the Umbria Region.
In order to test the applicability of the variograms, two areas have been selected for further study (Figure 11). Study area 1 is about 8 km 6 4 km and is located in the very center of the Perugia Province. Study area 2 is 7 km 6 6 km and is located to the southeast. The two areas were chosen because there are a significant number of data points to use for spatial interpolations. The soils in study area 1 are characterized by a central zone of silt, bordered by areas of terraced alluvium, turbidite, gravel, and eluvial and landslide deposit (Figure 12). The remaining part of the area is covered by alluvium, except for a small area of silt and terraced alluvium in the northwest corner. The available data points primarily overlie the alluvium, with fewer located over terraced alluvium and silt classes. For this map, the spatial interpolation algorithm was computed in Matlab. The described experimental variograms were fitted using appropriate mathematical functions (exponential model). The modeling of an empirical variogram through the definition of a theoretical curve is necessary because the Kriging Method requires information about the relationship between
Figure 12. Distribution of soil classes in study area 1.
46
Journal of Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 37–52
Reliable Soil Property Maps over Large Areas
Figure 13. Map of the predicted cohesion in study area 1.
Figure 14. Map of the predicted friction angle in study area 1.
Journal of Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 37–52
47
Fanelli, Salciarini, and Tamagnini
Figure 15. Distribution of the soil classes in study area 2.
distance and variance γ(h) continuously. Then it is possible to apply Kriging and make predictions of a variable z in a point x0 using the weight, λi, provided by the variograms: zðx0 Þ ¼
N X
ki zðxi Þ;
(7)
i¼1
where z(xi) values are the observed values of z at points xi, i 5 1,…, N. This procedure has been applied for the shear strength parameters of the soil classes (i.e., cohesion and friction angle). The maps of the predicted values are shown in Figures 13 and 14. The range of the interpolated values of cohesion is approximately 2 to 80 kPa (Figure 14). The central part of study area 1, covered by silt, is characterized by high values of cohesion (approximately 70 kPa), decreasing northward, as well as to the east, into the alluvium. The cohesion of alluvium has a significant spatial variability that would not be possible to catch without the disposal of punctual data, which allow distinguishing local variations in soil properties’ attributes.
48
In this way, it is possible to differentiate alluvium characterized by lower values of cohesion in the western part of area 1 and by higher values in the eastern part. This difference would not be evident without sampled points in the study area. Otherwise an average value of c9 would be assigned to the alluvium. A similar trend was observed for the terraced alluvium. In fact, close low values of cohesion are modeled in the northwest corner, while in the northeast part two sampled points suggest a much lower value. Regarding the friction angle, Figure 14 shows the spatial variability in the study area 1, from 23u to 27u. In the central part of the area, the silt has an average value of about 30u, but this value increases toward the north, where one data point has a higher ϕ9. The alluvium class, which covers a very large part of the area, has values of ϕ9 that vary significantly away from the data points (Figure 14). The friction angle increases from the top to the bottom of the area on the left and increases from the south to the north (on the right). The turbidite and gravel are characterized by values of approximately 30u (Figure 14).
Journal of Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 37–52
Reliable Soil Property Maps over Large Areas Table 3. Values of cohesion and friction angle for the selected data points used to validate the interpolation procedure. Point No.
1
2
3
4
5
6
7
8
Cohesion (kPa) Friction angle (u)
2 —
2 —
60 —
5 —
— 28
— 18
— 25
— 31
*Dashes indicate no data.
Study Area 2 The same procedure was used in study area 2. The distribution of soil classes is shown in Figure 15. Most of study area 2 is covered by alluvium, except for the southwest quarter, which is covered by clay and turbidite. There are very small areas of eluvial and landslide deposit. Most of the data points are located over the alluvial and clay classes. In study area 2, four sample points were not used in the interpolation process in order to use them to verify the predictions provided by the Kriging procedure for the cohesion. In the same manner, four different data points were excluded in order to validate the predictions for the friction angle. The values for cohesion
and friction angle at these points are listed in Table 3, while their locations are shown in Figures 16 and 17. In Figure 16, the interpolated values of the cohesion are shown. There appears to be a significant variability in cohesion within the alluvium. The values are higher in the north and west and lower in the south and east. Greater cohesion values (between 40 and 60 kPa) are found in the southeast. The area covered by clay class has values of cohesion similar to the northern part of alluvium, but slightly greater (around 25–30 kPa). The turbidite class has a very high cohesion, below 40 kPa. Comparing the predicted values of cohesion with the measured values at data points 1–4, there is a very good agreement. For example, point 4, with a c9 of 5 kPa (see Table 3), is in a zone in which the value is predicted to range between 2 and 10 kPa. At point 3, cohesion was measured at 60 kPa (see Table 3) and falls in a zone in which the predicted values are between 50 and 60 kPa. The distribution of the friction angle values provided by the Kriging procedure is shown in Figure 17. The angles range from 18u to 33u. Based on the modeled friction angle, the map can be divided in two areas. The zone to the east has soils with very
Figure 16. Map of the predicted cohesion in study area 2.
Journal of Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 37–52
49
Fanelli, Salciarini, and Tamagnini
Figure 17. Map of the predicted friction angle in study area 2.
low friction angles, from 18u at the eastern border to 24u at its western edge. The western zone has angles that are less homogeneous, with angles to the north greater than 30u, and they decrease southward. Over the clay class, the friction angle ranges from 25u to 26u, while over the turbidite class the angles are the greatest (greater than 30u). The four data points (points 5–8, see Table 3) used to validate data confirm that the predicted and measured values are similar. For example, data point 6 has a measured angle of 18u and lies in an area in which the interpolated value is similar. Data point 5 has an angle of 28u, and the interpolated angles in the area range from 27u to 30u. These case studies show how important it is to have a robust database of information containing soil strength parameters. A homogeneous soil type can show local variations that can be evaluated only with an appropriate number of measurements. The Kriging Method allowed us to take into account the spatial variability of soils and to model the local variations.
50
CONCLUSIONS AND PERSPECTIVES In this article the creation of a database to characterize the soil properties in the Perugia Province of the Umbria Region (central Italy), starting from an extensive data collection, is described. The database was first analyzed to study its reliability, focusing only on the soil classes. In particular, the values of the soil strength parameters (c9 and ϕ9) were further interrogated to obtain their absolute frequency histograms and to identify the probability density function that best fits the experimental histograms (i.e., the lognormal distribution for the cohesion and the normal distribution for the friction angle). Secondly, geostatistical analyses were performed on the strength parameters for each soil class to quantify any spatial correlation. Experimental variograms were created and fitted via an exponential theoretical model. Such variograms allowed us to evaluate the spatial correlation of the soil classes in the study area, necessary for applying the interpolating tech‐
Journal of Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 37–52
Reliable Soil Property Maps over Large Areas
nique of Kriging. Finally, two case studies were chosen to test the applicability of the local variograms. We found the measured and predicted cohesion and friction angle to be similar. These results verified the predictions of the Kriging procedure. A database containing the physical and mechanical properties of the soil in a given study area has a great potential application in the use of PB models for landslide hazard predictions. In particular, the database for the characterization of the soils obtained in this study contains the stochastic information on the geologic and geotechnical properties that are necessary for applying probabilistic PB models. They constitute the most appropriate approach for developing realistic landslides hazard assessments, since stochastic maps of soil strength parameters are indeed more reliable than deterministic maps (i.e., with a fixed value of soil strength parameters for each soil class). The presence of an extensive database of geotechnical information in the Perugia Province allowed us to apply the procedure of spatial interpolation described in this article in study areas of the Umbria Region in central Italy; however, the procedure can be easily extended to any other geographical area provided that the zone is well characterized from a geotechnical point of view. ACKNOWLEDGMENTS This work was supported by Project PRIN 2010– 2011 (“La mitigazione del rischio da frana mediante interventi sostenibili”), funded by the Italian Ministry of University and Research, and Project “Ricerche applicate e servizi modellistico-informatici di early warning a scala territoriale per la prevenzione del rischio idrogeologico e sismico, mediante approcci fisicamente basati,” funded by the Civil Protection Department of the Umbria Region. The authors wish to acknowledge the technical support for the data collection provided by the Structural Control and Civil Protection Office of the Perugia Province. The authors express their gratitude to Dr. Nicola Berni, Dr. Francesco Ponziani, and Dr. Marco Stelluti for their important help in preparing and organizing the data. The valuable reviews provided by two anonymous reviewers and by Dr. Andrew Stumpf are gratefully acknowledged. REFERENCES BAUM, R. L.; GODT, J. W.; AND SAVAGE, W. Z., 2010, Estimating the timing and location of shallow rainfall‐induced landslides using a model for transient, unsaturated infiltration: Journal Geophysical Research: Earth Surface (2003–2012), 115 (F3). BAUM, R. L.; SAVAGE, W. Z.; AND GODT, J. W., 2008, TRIGRS—A Fortran Program for Transient Rainfall Infiltration and Grid-
Based Regional Slope-Stability Analysis, Version 2.0: U.S. Geological Survey, Open File Report. 2008-1159, 75 p. BELL, F. G., 1992, Engineering Properties of Soils and Rocks: Elsevier, Oxford, Great Britain. BIENIAWSKI, Z. T., 1989, Engineering Rock Mass Classifications: A Complete Manual for Engineers and Geologists in Mining, Civil, and Petroleum Engineering: John Wiley and Sons, New York, U.S.A. CARDINALI, M.; REICHENBACH, P.; GUZZETTI, F.; ARDIZZONE, F.; ANTONINI, G.; GALLI, M.; CACCIANO, M.; CASTELLANI, M.; AND SALVATI, P., 2002, A geomorphological approach to the estimation of landslide hazards and risks in Umbria, Central Italy: Natural Hazards Earth System Science, Vol. 2, No. 1/2, pp. 57–72. CEN, 2003, Eurocode 8: Design Provisions for Earthquake Resistance of Structures, Part 1.1: General Rules, Seismic Actions and Rules for Buildings. PrEN1998-1: European Committee for Standardisation TC250/SC8. COTECCHIA, F. AND CHANDLER, R. J., 1995, Geotechnical properties of the Pleistocene clays of the Pappadai Valley, Taranto, Italy: Quarterly Journal Engineering Geology Hydrogeology, Vol. 28, No. 1, pp. 5–22. FANELLI, G.; SALCIARINI, D.; TAMAGNINI, C.; PONZIANI, F.; STELLUTI, M.; AND BERNI, N., 2015, Mapping earthquake-induced landslide susceptibility in Central Italy. In Engineering Geology for Society and Territory, Vol. 2: Springer International Publishing, pp. 727–730. GUZZETTI, F.; GALLI, M.; REICHENBACH, P.; ARDIZZONE, F.; AND CARDINALI, M., 2006, Landslide hazard assessment in the Collazzone area, Umbria, Central Italy: Natural Hazards Earth System Science, Vol. 6, No. 1, pp. 115–131. JAEGER, J. C.; COOK, N. G.; AND ZIMMERMAN, R.W., 2007, Fundamentals of Rock Mechanics: John Wiley & Sons, Malden, U.S.A. JIBSON, R. W.; HARP, E. L.; AND MICHAEL, J. A., 2000, A method for producing digital probabilistic seismic landslide hazard maps: Engineering Geology, Vol. 58, No. 3, pp. 271–289. KRIGE, D. G., 1951, A statistical approach to some basic mine valuation problems on the Witwatersrand: Journal South African Institute Mining Metallurgy, Vol. 94, No. 3, pp. 95–112. MATHERON, G., 1963, Principles of geostatistics: Economic Geology, Vol. 58, No. 8, pp. 1246–1266. MINISTERO DELLE INFRASTRUTTURE, 2008, DM 14 Gennaio 2008: Norme Tecniche per le Costruzioni. MONTGOMERY, D. R. AND DIETRICH, W. E., 1994, A physically based model for the topographic control of shallow lands‐ liding: Water Resources Research, Vol. 30, No. 4, pp. 1153–1171. PACK, R. T.; TARBOTON, D. G.; AND GOODWIN, C. N., 1998a, The SINMAP approach to terrain stability mapping. In 8th Congress of the International Association of Engineering Geology, Vancouver, British Columbia, Canada: pp. 21–25. PACK, R. T.; TARBOTON, D. G.; AND GOODWIN, C. N., 1998b, Terrain Stability Mapping with SINMAP, Technical Description and Users Guide for Version 1.00: Salmon Arm: Terratech Consulting Ltd. PARK, H. J.; LEE, J. H.; AND WOO, I., 2013, Assessment of rainfallinduced shallow landslide susceptibility using a GIS-based probabilistic approach: Engineering Geology, Vol. 161, pp. 1–15. RASPA, G.; MOSCATELLI, M.; STIGLIANO, F.; PATERA, A.; MARCONI, F.; FOLLE, D.; VALLONE, R.; MANCINI, M.; CAVINATO, G. P.; MILLI, S.; AND COSTA, J. F. C. L., 2008, Geotechnical characterization of the upper Pleistocene-Holocene alluvial deposits of Roma (Italy) by means of multivariate geostatistics: Cross-validation results: Engineering Geology, Vol. 101, No. 3, pp. 251–268.
Journal of Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 37–52
51
Fanelli, Salciarini, and Tamagnini REFICE, A. AND CAPOLONGO, D., 2002, Probabilistic modeling of uncertainties in earthquake-induced landslide hazard assessment: Computers Geosciences, Vol. 28, No. 6, pp. 735–749. REICHENBACH, P.; GALLI, M.; CARDINALI, M.; GUZZETTI, F.; AND ARDIZZONE, F., 2004, Geomorphological mapping to assess landslide risk: Concepts, methods and applications in the Umbria region of central Italy: Landslide Hazard Risk, Chapter 15, pp. 429–468. SALCIARINI, D.; GODT, J. W.; SAVAGE, W. Z.; CONVERSINI, P.; BAUM, R. L.; AND MICHAEL, J. A., 2006, Modeling regional initiation of rainfall-induced shallow landslides in the eastern Umbria Region of central Italy: Landslides, Vol. 3, No. 3, pp. 181–194. SALCIARINI, D. AND TAMAGNINI, C., 2015a, Physically-based critical rainfall thresholds for unsaturated soil slopes. In Recent Advances in Modeling Landslides and Debris Flows: Springer International Publishing, pp. 253–264. SALCIARINI, D. AND TAMAGNINI, C., 2015b, Physically based rainfall thresholds for shallow landslide initiation at regional scales. In Engineering Geology for Society and Territory, Vol. 2: Springer International Publishing, pp. 1041–1044.
52
SALCIARINI, D.; TAMAGNINI, C.; CONVERSINI, P.; AND RAPINESI, S., 2012, Spatially distributed rainfall thresholds for the initiation of shallow landslides: Natural Hazards, Vol. 61, No. 1, pp. 229–245. SALCIARINI, D.; TAMAGNINI, C.; PONZIANI, F.; AND BERNI, N., 2013, Defining physically-based rainfall thresholds for early warning systems. In Landslide Science and Practice: Springer, Berlin, Germany, pp. 651–657. SAYGILI, G. AND RATHJE, E. M., 2009, Probabilistically based seismic landslide hazard maps: An application in Southern California: Engineering Geology, Vol. 109, No. 3, pp. 183–194. SIMONI, S.; ZANOTTI, F.; BERTOLDI, G.; AND RIGON, R., 2008, Modelling the probability of occurrence of shallow landslides and channelized debris flows using GEOtop‐FS: Hydrological Processes, Vol. 22, No. 4, pp. 532–545. WANG, K. L. AND LIN, M. L., 2010, Development of shallow seismic landslide potential map based on Newmark’s displacement: The case study of Chi-Chi earthquake, Taiwan: Environmental Earth Sciences, Vol. 60, No. 4, pp. 775–785. WYLLIE, D. C. AND MAH, C., 2004, Rock Slope Engineering: SPON Press, New York, U.S.A.
Journal of Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 37–52
Investigation of Air Bulging Beneath Geomembranes Used as a Liner for the Datun Reservoir XUE-SHAN CAO1 JUN-PING YUAN ZONG-ZE YIN Key Laboratory of Ministry of Education for Geomechanics and Embankment Engineering, Hohai University, 1st Xikang Road, Nanjing 210098, China; and the College of Civil and Transportation Engineering, Hohai University, 1st Xikang Road, Nanjing 210098, China
GUI-LIN HE YING-HAO LIU Shandong Survey and Design Institute of Water Conservancy, Jinan 250013, China
Key Terms: Geotechnical, Lined Reservoir, Pore-Air Pressure, Air Bulging, Simulation
geomembrane will result in air bulging. A comparison among French drains with spacings of 25 m and 50 m indicates that the original design comprising French drains spaced 50 m apart is reasonable and economical.
ABSTRACT Extensive and thick permeable strata are found beneath many reservoir sites. The leakage problem in these reservoirs can be effectively solved by constructing a full geomembrane liner design. However, applying this particular solution for the reservoir induces novel engineering problems. The air bulging produced beneath the geomembrane results in geomembrane rupture and causes a renewed leakage in reservoirs. The groundwater level may be relatively low during construction because of seasonal variation and dewatering. Geomembranes isolate pore air from the atmosphere, and a significant amount of air is trapped in the pores of the unsaturated soil above the groundwater level. The rise in the groundwater level or the compression of soil decreases pore-air volume and increases pore pressure, thereby causing air bulging to occur and prompting the ultimate destruction of the geomembrane. The presence of air bulging beneath the geomembrane is studied in the design process of the Datun Reservoir in Dezhou City, Shandong Province, China. Investigation results reveal that a rise in the groundwater level significantly influences the air bulging beneath the geomembrane. The French drain is set to prevent air bulging beneath the geomembrane from occurring. The 75-m French drain spacing is not reasonable because the air pressure beneath the 1
Corresponding author email: x.s.cao@163.com.
INTRODUCTION A large number of plain reservoirs have been constructed in China (Gu, 1994, 2009; Jiang et al., 2010; and Yuan et al., 2014). Beneath these reservoir sites, extensive and thick permeable strata can be observed. Accordingly, the problems of reservoir leakage and seepage are effectively solved by developing a full geomembrane liner design. This approach is also expected to greatly improve reservoir efficiency. Geomembranes are continuous flexible sheets manufactured from one or more synthetic materials. These geomembrane barriers are impermeable and are used as liners for fluid (Bathurst, 2007). Meanwhile, geotextile-geomembrane composites are manufactured from geomembranes and geotextiles. In particular, geotextiles can be laminated on one or both sides of a geomembrane and provide increased resistance to puncture, tear propagation, and friction related to sliding. Moreover, geotextiles can be equipped with tensile strength. In China, geomembranes have been used as a liner in many plain reservoirs. The Jashanzi Reservoir in Gansu Province, China (Yi and Chen, 1999), was constructed in 1991. To solve the reservoir leakage and seepage caused by the 100-m-thick sand-pebble layer and three surface troughs, a 500,000-m2 geomembrane was placed on the bottom sand-pebble layer of the Jashanzi reservoir (Gu, 1994).
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 53–66
53
Cao, Yuan, He, Liu, and Yin
A full geomembrane liner can solve the leakage problem in reservoirs without an anti-seepage soil foundation. However, new engineering problems in a reservoir site in China have emerged. In particular, air bulging developed beneath the geomembrane, leading to geomembrane rupture. This geomembrane rupture then led to a renewed leakage in the reservoir (Foose et al., 2001; Bouazzaa and Vangpaisalb, 2006). Geomembrane liners were designed for the Jindian Reservoir in Kunming Province, China. Consequently, the seepage control effect attained the expected target in the operation. However, “geomembrane rupture” and “soil collapse” occurred in the local area (Gong and Song, 1999). In 2003, the Shengli Reservoir in the Xinjiang Uygur Autonomous Region, China, was constructed with an area measuring 3,853,000 m2 (Shi and Li, 2005; Jiang et al., 2010). During the first impoundment of the Shengli Reservoir, 1,155 and 84 holes emerged in the smooth geomembrane and geotextile-geomembrane composites, respectively (Jiang et al., 2010). The permeability coefficient of this reservoir ranged from 2.17 3 1027 m/ s to 1.16 3 1024 m/s before the geomembrane was laid in the Shengli Reservoir and from 1 3 10211 m/s to 1 3 10210 m/s one year after the impoundment (2003) of the Shengli Reservoir with geomembrane liners. However, after another year of operation (2004), both the reservoir leakage and permeability increased, attaining 1 3 1029 m/s to 1 3 10210 m/s. The phenomenon revealed that the geomembrane was ruptured. The geomembrane was inspected and repaired in 2004, and the permeability of the Shengli Reservoir was reduced (Jiang et al., 2010). Technicians and engineers poorly comprehend the phenomenon of geomembrane bulging and rupture because of inadequate knowledge and the lack of detection methods related to such problems. For instance, a groundwater table is located 15–30 m and 30–50 m below the geomembranes in the Jashanzi and Shengli Reservoirs, respectively, and the foundations of these reservoirs consist mainly of sandy gravel. Thus, no drainage system was placed beneath the geomembrane in either reservoir (Gu, 2009). Geomembrane bulging and rupture are related to the property of unsaturated soil because the geomembrane isolates the pore air in the unsaturated soil from the atmosphere. Unsaturated soil consists of solid particles, pore water, and pore air. An increase in soil moisture content increases the volume of pore water, which then squeezes out the pore air. If the pore air is exposed to the atmosphere, the pore air is discharged and the pore-air pressure remains constant. Otherwise, the pore-air pressure increases, thereby inhibiting the infiltration of pore water (Dixon and Linden,
54
1972; Latifi et al., 1994). Soil compression reduces the closed pore volume and increases the pore-air pressure. The increased pore-water volume and reduced pore-air volume both produce excess porewater pressure and pore-air pressure. Accordingly, the pressure gradient causes the pore air to move. Air bulging develops beneath the geomembrane under three conditions: (1) Material conditions: deeper groundwater level corresponds to thicker unsaturated soil above the groundwater level. Pore air in the unsaturated soil is the first condition necessary for the occurrence of bulging beneath the geomembrane. (2) Boundary conditions: the area of the lined reservoir should be large, the ratio of the length or width to the thickness of the unsaturated soil should be large, and the lateral drainage path should be long, approximating an isolated system. (3) Induced conditions: when pore air is compressed, the pore-air volume decreases and the pore-air pressure increases. Either the rise in the groundwater level or the compression of soil decreases the pore-air volume, thus causing an increase in pore pressure, promoting the migration and accumulation of pore air and producing air bulging beneath the geomembrane (Li et al., 2013; Cao et al., 2015). This condition is significantly different from that associated with landfill gas (Li et al., 2013). Landfill gas is generated and migrated on post-closure (Liu et al., 2010; Aubertin et al., 2014). The gas, including methane, carbon dioxide, and other gases, is primarily generated from waste in the biochemical reaction (Behera et al., 2011). The gas beneath the geomembrane in the reservoir is aggregated and compacted in the unsaturated soil pore (Yuan et al., 2014). Gas is not produced but originates from the air. SITE CONDITION Background of the Project The Datun Reservoir in Shandong Province, China, is 13 km from Wucheng County and 25 km from Dezhou City. The total surface area of the reservoir ground plan is 6,489 km2. This reservoir is one of the regulating reservoirs in the South-North Water Transfer Project and is an important part of the Water Conveyance Project of Shandong Lubei. This reservoir primarily regulates the Eastern Line of the South-North Water Transfer Project of China and satisfies the industrial and urban water demands of Dezhou City and Wucheng County. Figure 1 illustrates that the embankment of the Datun Reservoir is roughly quadrilateral in plan view, with an embankment axis length of approximately 8,913.99 m. The reservoir area comprises low-
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 53–66
Bulging Geomembranes Used as Liner
Figure 1. Design scheme of the Datun Reservoir. (A) Datun Reservoir locations; (B) plan design of the Datun Reservoir; (C) section design of the Datun Reservoir; (D) section design of the No. 1 French drain; (E) section design of No. 2 French drain. Note 1. The height system in the figure is the Wusong elevation datum 1985; one unit is a meter, and the other unit is a millimeter. Note 2. The spacing of a French drain in the reservoir area is 50.0 m. The check valve is set on the intersection of French drains. Note 3. The No. 2 French drain connected the dam slope sand cushion to drains on the waste soil platform toe. Note 4. French drain is wrapped in geotextile of 300 g/m2. Note 5. Scale: 1:500.
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 53–66
55
Cao, Yuan, He, Liu, and Yin Table 1. Design parameter of the Datun Reservoir.
Maximum Minimum
Pool Elevation (m)
Storage Capacity (million m3)
29.80 21.00
52.09 7.45
lying, flat terrain that slopes gently from south to north. The natural ground surface elevation ranges from 21.4 m to 20.5 m (China Wusong Elevation System, 1985). Table 1 shows the design parameters of the Datun Reservoir. Geological and Hydrological Conditions The geological conditions of the Datun Reservoir engineering are complex. The Datun Reservoir exploration data indicate that Holocene strata are widely distributed across the area (Figure 2 and Table 2). The layers of alluvial soils above 17.20 to 23.50 m are unstable. The upper layer consists of sandy loam and fissured clay with sporadic loam and silt. The lower layer consists primarily of fissured clay, sandy loam, and loam and partially of alluvial silt and silty sand from the ancient river. The silty and fine sand of the alluvial swamp strata lie below the elevation of 17.20 m to 23.50 m. The degree of saturation of the shallow soils ranges from 75 percent
to 95 percent. The degree of saturation ranges from 86.9 percent to 94 percent at a depth of 5 m below ground level. The preliminary estimation indicates that the storage capacity of pore air in the unsaturated foundation soil in the Datun Reservoir is approximately 2.0 million m3 (Johnson, 1967) under extreme drought conditions. Pore water fills the sandy loam, fissured clay, silty sand, and fine sand in the area. The depth to groundwater is generally 1.10 m to 1.80 m below ground level. Local meteorological data specify that the average rainfall, the maximum annual precipitation, and the minimum annual rainfall in the vicinity are 558.8 mm, 1,047.7 mm (in 1964), and 297.0 mm (in 1968), respectively. The monsoon climate causes seasonal rainfall distribution throughout the year, with rains during the flood seasons (June to September) accounting for 76.95 percent of the annual rainfall. The 1995–2004 observation data obtained in Well No. 39 near the Datun Reservoir area (Figure 3) imply that the groundwater level varies between 3 m and 4 m within a year and between 5 m and 6 m for 2 years (Figure 3A); the rate of rise in the groundwater level is usually less than 1.00 m/mo, but the maximum rate can reach up to 1.94 m/mo (Figure 3B). This rise typically occurs between June and September but is also observed as early as May and as late as November.
Figure 2. Geologic profile across the Datun Reservoir.
56
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 53–66
Bulging Geomembranes Used as Liner Table 2. Soil properties in the Datun Reservoir for the Quaternary Yellow flood alluvial layer. Coefficient of Permeability (3 1025 m/s)
Void Ratio Propertya (1) (2) (3) (4) (5) (6) (7)
Loam Sandy loam Fissured clay Sandy loam Fissured clay Clay Sandy loam
Average
Maximum
Minimum
Average
Maximum
Minimum
Depth to bottom (m)
0.79 0.78 0.99 0.76 0.93 0.95 0.74
0.88 0.83 1.10 0.84 0.97 1.02 0.80
0.72 0.73 0.89 0.69 0.88 0.86 0.68
0.54 0.62 4.79 0.31 4.33 4.38 0.31
0.82 0.91 15.74 0.54 5.45 5.94 0.54
0.30 0.11 0.10 0.18 2.45 1.74 0.18
0–1.75 0.24–4.53 0.79–6.18 7.66–10.61 7.23–10.1 7.66–10.11 8.81–11.56
a
Numbers in parentheses correspond to the soil numbers in Figure 2.
Design and Construction Conditions A full geomembrane liner is designed to serve as the Datun Reservoir seepage control. The geotextilegeomembrane composite is composed of double-sided geotextiles (the upper and bottom layers are 200 and 300 g/m2, respectively) and one High Density Polyethylene (HDPE) membrane (0.5 mm). The permeability coefficient of the geomembrane is 5 3 10212 m/s, and the average thickness of the backfill above the geomembrane is 1.0 m. The geotextile-geomembrane composites (the upper and bottom layers are 300 and 200 g/m2, respectively) also cover the embankment slope.
The height of the fill embankment is 11.30 m. The embankment is a zoned compacted-earth fill structure, and the geotextile-geomembrane composites are laid after the completion of the embankment. Geomembrane sealing systems consist of (top to bottom) a 210-mm-thick pre-cast concrete slab, a 20– 40-mm gravel cushion with a thickness of 150 mm, a 5–20-mm gravel cushion with a thickness 150 mm, and a mixed sand mat layer with a thickness of 100 mm on the embankment slope. The load over the geomembrane on the upstream of the embankment slope is 14.2 kPa, and that of the 1.0-m-thick backfill is 20.0 kPa. The Datun Reservoir ground surface elevation is low, with an average elevation of approximately 20.8 m (Figure 1A), and the elevation of the excavated reservoir base is 19.0 m. In particular, the earth is excavated 100 m away from the embankment axis to protect the toe of the embankment (Figure 1C). Operating Conditions The total filling period of the Datun Reservoir for a year is 122 days. The first and second fillings both take 61 consecutive days from October to November and from April to May, respectively. The Datun Reservoir supplies 125.02 million m3 of water for 365 days. The analysis of the reservoir water and water supply process reveals that water loss and water shortage are prevented throughout the year. The average monthly pool elevations are shown in Table 3. AIR BULGING ANALYSIS METHODS OF UNSATURATED SOIL Unsaturated Soil Methods
Figure 3. Observation data related to the groundwater level in the Datun Reservoir area from 1995 to 2004. (A) The rise in the groundwater level; (B) the rate of rise in the groundwater level.
The formation of air bulging beneath the geomembrane is caused by the migration of pore water and pore air in the unsaturated soil under certain
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 53–66
57
Cao, Yuan, He, Liu, and Yin Table 3. Calculated monthly pool elevation and depth.
Cases Filling Draining
Calendar Month 10 11 12 1 2 3
Reservoir Water Elevation (m)
Reservoir Water Height (m)
Load over the Geomembrane at Analysis Point 1 (kPa)
Load over the Geomembrane at Analysis Point 2 (kPa)
25.42 29.58 27.43 25.30 23.36 21.00
6.22 10.38 8.23 6.10 4.16 1.80
56.2 97.8 76.3 55 35.6 20
74.2 115.8 94.3 73 53.6 30
4 5 6 7 8 9
conditions. The factors that cause the formation of air bulging beneath the geomembrane are complex. The lined reservoir has a huge area, and a large volume of air is trapped in the unsaturated soil beneath the geomembrane. The drainage path of excess air pressure is long. Therefore, unsaturated soil methods are proposed for the analysis of the occurrence of air bulging and geomembrane rupture. A great number of unsaturated soil methods (Chen et al., 1993; Fredlund and Rahardjo, 1993; Cao et al., 2007; and Cao and Yin, 2009) exist with which to explore the migration behavior of pore water. However, only a few of these methods have been employed in the analysis of air bulging and geomembrane rupture (Zhang et al., 2012). When the degree of saturation is high, pore water and pore air are assumed to form a mixed fluid that collectively moves. The continuity equation of this mixed fluid is established using the permeability coefficient and compressibility of the mixed fluid. The mixed fluid pressure, um, is a weighted average of the pore-water pressure, uw, and the pore-air pressure, ua, via the parameter x: um ~ð1{xÞua zxuw ~ua {xðua uw Þ
pressure, um, can be obtained (Cao et al., 2007; Cao and Yin, 2009) as follows: ( ) {½L T ½D ½L fwgz½L T fM gum ~f f g , ð4Þ T km 2 L 1 : Lum Lt fM g ½L fwg{ c + um z Bm Lt ~0 m
where {f} is the volume force, {w} is the displacement variable, cm is the mixed fluid unit weight, km is the mixed fluid permeability coefficient, and Bm is the volume-compressed modulus of mixed fluid. On the basis of the water and mixed fluid continuity equations, the water continuity equation (Cao et al., 2007; Cao and Yin, 2009) is kw km ð5Þ { +2 uw ~Sr { +2 um , cw cm where cm is the water unit weight and kw is the water permeability coefficient. Eq. 5 expresses that the volume change in pore water caused by drained water volume is equal to the product of the drained volume change in pore fluid and the degree of saturation. By combining um and uw, the air pressure can be obtained using Eq. 1.
ð1Þ Computation Element and Parameter
The magnitude of the parameter x is 1 and 0 for saturated soil and dry soil, respectively. Aitchison (1961) used Eq. 2 to obtain the effective stress parameter, x: x~
Sr : 0:4Sr z0:6
ð2Þ
Then, the effective stress equation (Bishop et al., 1960) can be modified as follows: fsg~fs’gzfM gum :
ð3Þ
Eq. 3 shows that the total stress is the sum of mixed fluid pressure and soil skeleton stress. On the basis of the pore-fluid continuity and mass balance equations, the displacement and mixed fluid
58
The design in Figure 1C exhibits that the ground elevation is 20.8 m near the reservoir embankment and 19.0 m in the middle area of the reservoir. Reservoir data show that the minimum elevation of the groundwater level, which is the saturated soil boundary, is 15.0 m. This boundary acts as an undrained boundary of airflow because the air permeability coefficient ranged from 10213 m/s to 10215 m/s (Fredlund and Rahardjo, 1993). In the reservoir area below the toe of the embankment, the model thickness ranges from 4 m to 5.8 m (Figure 4). The ground surface elevation is 20.8 m at Analysis Point 1 and 19.0 m at Analysis Point 2 (Figure 4). The model parameters (Table 4) include the Duncan model parameters obtained by the soil triaxial slow shear consolidation drain test and other
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 53–66
Bulging Geomembranes Used as Liner
Figure 4. Model grid of the Datun Reservoir and location of the analysis points.
a period of 30 days. The construction period of the waste soil platform is also 30 days. The groundwater level elevation is y 5 15.00 m at the time of embankment construction. Figure 5A indicates that the excess pore-air pressure in unsaturated soil is higher at the center of embankment than on both sides because the load is heavy and the drain distance is far from the center during construction. The geomembrane is installed on the slope of the embankment after the latter is completely assembled. Subsequently, the pore air beneath the geomembrane is isolated from the atmosphere. The effective stress of unsaturated soil increases with the load of the geomembrane sealing system, thereby compressing pore volume and increasing pore-air pressure. The pore-air pressure beneath the geomembrane increases. The pore-air pressure increases significantly in the reservoir area and insignificantly in the upstream slope of the embankment. By contrast, the pore-air pressure decreases obviously on the downstream slope of the embankment (Figure 5A and B). Pore air cannot be drained off and gathered beneath the geomembrane; hence, the pore-air pressure increases. However, the pore-water pressure is negative for unsaturated soils beneath the geomembrane. The increase in pore-air pressure is weakened by the negative pore-water pressure. The effective stresses share the load of the geomembrane sealing system, weakening the increase of the pore-fluid pressure beneath the geomembrane. In this event, the increase of pore-air pressure is exceedingly small. The pore-air pressure beneath the geomembrane is obviously lower than the load on the geomembrane. No air bulging occurs during the construction of the reservoir.
parameters acquired by the saturated soil permeability tests, the soil-water characteristic curve tests of unsaturated soil, and additional routine tests. Two points are designed to analyze the factors influencing air bulging beneath the geomembrane. Analysis Point 1 is located on the reservoir edges close to the embankment toe, at which point the geomembrane elevation is 20.8 m, the backfill thickness is 1.0 m, and the backfill soil surface elevation is 21.8 m, which is higher than the dead pool elevation at 21.00 m. Analysis Point 2 is situated on the central region, at which point the geomembrane elevation is 19.0 m, the backfill thickness is 1.0 m, and the backfill soil surface elevation is 20.0 m, which is lower than the dead pool elevation at 21.00 m. FACTORS INFLUENCING AIR BULGING BENEATH THE GEOMEMBRANE Construction Factors The pore-air pressure and pore-water pressure of the unsaturated soil increase with increasing height during embankment construction and gradually dissipate with time. The geomembrane is installed after the embankment is completely constructed, and the soil pores of the upstream slope and reservoir area are isolated from the atmosphere by the former. Thereafter, pore air is drained off with difficulty and is gathered under the geomembrane to increase the pore-air pressure. The embankment construction is simulated using the unsaturated soil consolidation method. We assume that seven layers of the embankment, with each layer having a height of 1.64 m, are placed over
Table 4. Model parameters of the Duncan E-m model.
Soil Loam Sandy loam Fissured clay Embankment soil
c (kN/m3)
W (u)
C (kPa)
Rf
k
N
G
F
D
Kur
Kws
E
sr0 srl (%) (%) l usb
kws (cm/s)
kd (cm/s)
19 19.1 18.5
27 10.5 0.93 155.4 0.603 0.274 20.17 1 311 0.464 0.794 75.5 32.5 4.33 0.946 255.9 0.571 0.271 20.18 1.8 350 0.535 0.784 86.7 24.9 19 0.902 174.3 0.622 0.311 20.14 1.5 349 4.14 0.988 85.0
15 15 15
1 30 5.37 3 1024 2 15 6.19 3 1024 1 30 4.79 3 1023
0.01 0.01 0.01
20.3
30
15
2 30 6.13 3 1026
0.01
19
0.95
300.0 0.6
0.6
0
0
400 0.005 0.633 90
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 53–66
59
Cao, Yuan, He, Liu, and Yin
Figure 5. Contour map of the pore-air pressure in the reservoir unsaturated soil. (A) After completion; (B) after laying the geomembrane and 1.0-m thickness of backfill; (C) filling the reservoir water level elevation to 29.80 m with 1.0-m thickness backfill and 15.0 m of the groundwater level elevation; (D) discharging the reservoir water level elevation from 29.80 m to 21.00 m with 1.0-m thickness of backfill and 15.0 m of the groundwater level elevation.
Reservoir Operation Factors The lined reservoir water over the geomembrane acts as a vertical load that compresses the foundation soil beneath the geomembrane liner, thereby increasing the effective stress, pore-water pressure, and pore-air pressure. The foundation of the unsaturated soil decompresses and rebounds when the reservoir water level decreases, thereby decreasing the effective stress, pore-water pressure, and pore-air pressure. However, the modulus of rebound is higher than that of the compression. Consequently, the rebound displacement is considerably less than that of the compression. The residual compression of pore volume then decreases the pore-gas volume because
60
of reservoir operation, thus increasing the pore-air pressure and prompting pore air to gather beneath the geomembrane. Several factors influence the occurrence of air bulging beneath the geomembrane. For instance, the compressibility of the soil under the geomembrane influences the residual amount of pore volume. The other two factors are the permeability and degree of saturation of the soil beneath the geomembrane. The excess pore-fluid pressure beneath the geomembrane dissipates quickly for soils with high permeability. Consequently, the effective stresses greatly increase with the operation of the reservoir, particularly in the case of the maximum pool elevation, thus leading to
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 53–66
Bulging Geomembranes Used as Liner
the high over-consolidation ratio of the soil beneath the geomembrane, the lower degree of saturation, and the higher pore-air pressure when dewatering to the dead pool elevation at 21.00 m. When the degree of saturation of the soil beneath the geomembrane is low, the ratio of the pore-air volume is high and the compressibility of unsaturated soils and the residual amount of pore volume increase. These occurrences increase effective stresses when impounding reservoirs and increase pore-air pressure when dewatering to the dead pool elevation at 21.00 m. Water filling at the Datun Reservoir began after the embankment was completely constructed and the geomembranes were placed. The reservoir filling plan states that the full storage level annually attains an elevation of 29.80 m from April to May and from September to October. The vertical compression produced by reservoir water weight increases the pore-air pressure beneath the geomembrane. Figure 5C shows the high pore-air pressure beneath the geomembrane when the reservoir water level increases to the designed full storage level elevation of 29.80 m. The pressure is approximately 61 kPa on the sides and 80 kPa at the reservoir center area; these values are both less than the 97.8-kPa and 115.8-kPa loads over the geomembrane, respectively. In this case, no air bulging occurs beneath the geomembrane. The water from the designed pool elevation of 29.80 m discharges toward the dead pool elevation of 21.00 m for nearly 4 months. Figure 5D shows the pore-air pressure beneath the geomembrane during the water discharge from 29.80 m to 21.00 m. In this case, the pressures are approximately 16 kPa on the sides and 26 kPa at the reservoir area, which are less than the load at 20 kPa on the sides and 30 kPa at the reservoir center area over the geomembrane, respectively. No air bulging occurs beneath the geomembrane. Figure 6A and B illustrate that the reservoir water weight load on the geomembrane linearly decreases with reservoir water elevation. A decrease in pore-air pressure beneath the geomembrane follows. At the initial dewatering time, the reduction of excess porefluid pressure is near that of the load on the geomembrane, and the pore-air pressure is close to the pore-water pressure because the degree of saturation of the soil beneath the geomembrane is nearly 100 percent. At the next dewatering time, the reduction of pore-air pressure during unloading is less than its increase during the loading of a similar load on the geomembrane because the rebound displacement is considerably less than that of the compression. Moreover, the degree of saturation of the soil decreases because of the reduction of pore-water pressure and the expansion of pore air, thereby
Figure 6. Relationship between the load on the geomembrane and the pore-air pressure beneath the geomembrane during filling and dewatering. (A) Analysis Point 1; (B) Analysis Point 2.
inducing a large difference between the pore-air pressure and the pore-water pressure. However, when rebound deformation is less than the compaction deformation and the pore-air pressure is higher than the pore-water pressure, the pore-air pressure beneath the geomembrane is obviously smaller than the load on the geomembrane. Therefore, no air bulging occurs because of the reservoir operation. This particular condition is related to the compressibility, permeability, and degree of saturation of the unsaturated soil beneath the geomembrane. The pore-air pressure beneath the geomembrane at Analysis Point 1 in Figure 6A is higher than that at Analysis Point 2 in Figure 6B for two reasons. First, the position of Analysis Point 1 is higher than that of Analysis Point 2. Second, the degree of saturation is lower because of the reservoir operation. Hence, the
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 53–66
61
Cao, Yuan, He, Liu, and Yin
air beneath the geomembrane accumulates easily and is influenced by the high pore-air pressure in the unsaturated embankment. Rise in the Groundwater Level Lined reservoirs are generally constructed during the drought season, when the groundwater level is lower and the unsaturated soil thickness is great. After the geomembrane is installed in the reservoir area, the geomembrane isolates pore air in the unsaturated soil from the atmosphere. During the rainy season, the groundwater level rises, and the pore water in the soil increases and squeezes out the pore air, thereby decreasing the pore-air volume and increasing the pore-air pressure. In this case, the degree of saturation also increases. A closed space in the foundation area of the plain reservoir on the vertical space distribution is formed after the geomembranes are installed because the latter are impermeable on the surface and because saturated soils and groundwater have low air permeability below. On the horizontal space distribution, some air drainage boundaries influence the air pressure beneath the geomembranes, such as the outside surface of embankment and the lateral boundary of the reservoir-embankment-foundation. The groundwater level observed between 1995 and 2004 in Well No. 39 (Figure 3A and B) signifies that September is the most critical month each year. During September, the pool elevation is 21.00 m (dead pool elevation), and a considerable rise in the groundwater level occurs at the Datun Reservoir. Figure 6 exhibits the condition of pore-air pressure beneath the geomembrane when the groundwater level rises and the pool elevation is 21.00 m (dead pool elevation). Influence of the Rise in the Groundwater Level The construction of reservoir sites requires the artificial lowering of the groundwater level. The groundwater level rises after construction. When water infiltrates into the vadose zone, soil air is displaced and may become compressed. Existing laboratory studies (Latifi et al., 1994) and field experiments (Dixon and Linden, 1972) have suggested that soil-air pressure increases. In unsaturated soil, volume changes during shearing and excess pore-air and pore-water pressures develop (Fredlund and Rahardjo, 1993). Figure 7A displays that while the Figure 7. Variation of pore-air pressure beneath the geomembrane when the groundwater level increases; (A) a 1.0-m-thick backfill, no reservoir water, and 15.0 m of the initial groundwater level elevation and 1.0 m/mo rate of rise in the initial groundwater level during construction; (B and C) a 1.0-m-thick backfill,
62
r the dead pool elevation, and 15.0 m of the initial groundwater level elevation and 1.0 m/mo rate of rise in the initial groundwater level during operation. See Figure 4 for the location of analysis points.
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 53–66
Bulging Geomembranes Used as Liner
load on the geomembrane is constant during construction, the pore-air pressure beneath the geomembrane increases linearly with the rise in the groundwater level. The load on the geomembrane of 1.0 m thickness of backfill is 20 kPa, with no water at the center and along the sides of the reservoir. The critical rises in the groundwater level are 2.5 m at Analysis Point 1 and 1.5 m at Analysis Point 2. Air bulging does not occur beneath the geomembrane when the rise in the groundwater level is less than the critical amount. By contrast, air bulging develops if the rise in the groundwater level is greater than the critical amount. The critical rise in the groundwater level at Analysis Point 1 is higher because the higher geomembrane surface leads to thick unsaturated soils and high pore-air volumes. However, engineering techniques can be used to directly prick the bulges and to repair the geomembrane after deflation because the backfill soil on the geomembrane is dry prior to reservoir impounding. The behavior of pore-air pressure beneath the geomembrane during reservoir operation reveals that its most critical case involves situations in which the reservoir water level decreases to the minimum pool elevation of 21.00 m because of the minimum load on the geomembrane and high pore-air pressure beneath the geomembrane. Figure 7B and C show that the critical rise in the groundwater level is 1.75 m. In particular, Figure 7B specifies that Analysis Point 1 is located on the reservoir edges, where the backfill surface elevation (21.80 m) is higher than the dead pool elevation (21.00 m) and where the load is 20 kPa. Figure 7C demonstrates that Analysis Point 2 is positioned on the central region, where the backfill surface elevation (20.0 m) is lower than the dead pool elevation (21.00 m) and where the load is 30 kPa. Influence of the Rate of Rise in the Groundwater Level The rise in the groundwater level influences the behavior of the pore-air pressure beneath the geomembrane and indicates that the critical rise in the groundwater level is from 1.5 m to 2.5 m during construction and 1.75 m during reservoir operation. Therefore, the rate of the rise in the groundwater level is analyzed with its 2.0-m increase (Figure 8). Figure 8 depicts that pore-air pressure is insignificantly influenced and reduced with an increase in the rate of rise in the groundwater level for two reasons. First, when the rise in the groundwater level is constant and the rate of rise is fast, the soil is easily saturated. This instance eventually leads to the insignificant characteristics of unsaturated soil, such as small suction. Second, after the reservoir is
completely constructed, the high pore-air pressure in the embankment fill and foundation influences the pore-air pressure beneath the geomembrane. The pore-air pressure beneath the geomembrane increases with time, leading to a slight difference in the pore-air pressure curve line between Analysis Points 1 and 2. DISCUSSION AND ANALYSIS OF MITIGATION MEASURES FOR AIR BULGING BENEATH THE GEOMEMBRANE Discussion The 1995–2004 groundwater observation data obtained in Well No. 39 near the Datun Reservoir signify that the highest groundwater level is y 5 21.26 m during flood season (from June to September) and y 5 20.58 m during non-flood season. The elevation of the geomembrane is 19.0 m at the middle reservoir. Air bulging may develop in the reservoir when the gas pressure beneath the geomembrane is higher than that of the load over the geomembrane. Therefore, the French drain and the check valve must be appropriately designed. In particular, the scheme must be designed for the most unfavorable situation (i.e., when the pool elevation of the reservoir is lowest and the load on the geomembrane is at its minimum). This situation occurs during flood season, during which the largest rise in the groundwater level is 6.26 m. The French drain is wrapped with coarse sand or gravel in geotextile to form a square with a side length of 300 mm (Figure 1E). The spacings of the French drains are designed to be 50 m in plan view. The first French drain is set at the toe of the slope along the reservoir circumference. The results of the gas permeability tests of soil (Fredlund and Rahardjo, 1993) reveal that the gas permeability coefficient and the soil water content (degree of saturation) are closely related. When soil is saturated, the gas permeability coefficient of the soil is small (10215 m/s) (Fredlund and Rahardjo, 1993); therefore, the saturated soil can be considered impermeable. The permeability coefficient of the French drain is large. On the basis of the relationship between the top and bottom surfaces of the French drain and the groundwater level, three cases are considered in the rising process of the groundwater level. Case 1 occurs when the groundwater level is lower than or equal to the bottom of the French drain; the French drain is fully effective. Case 2 occurs when the groundwater level is greater than the top of the French drain, the French drain soil is saturated, filled with water, and becomes an airtight passage; thus, the French drain fails to drain gas completely. Case 3 occurs between
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 53–66
63
Cao, Yuan, He, Liu, and Yin
Cases 1 and 2; the French drain is partially submerged and only partly effective. The spacing of the French drain must be satisfactory to dissipate air pressure beneath the geomembrane fully in Case 1 to ensure that no air bulging occurs. If the spacing of the French drains beneath the geomembrane is reasonable, no air bulging occurs during the rise in the groundwater level even in Cases 2 and 3. Therefore, the spacings of the French drain beneath the geomembrane are first analyzed and compared in Case 1, and then the reasonable spacing of the French drain is determined in Cases 2 and 3. Analysis of French Drain Spacing When the Groundwater Level Rises to the Geomembrane Surface The French drain spacings are 25 m, 50 m, and 75 m according to the original design scheme. The laying of French drain begins from the embankment toe and moves to the reservoir area. Case 1: If the groundwater level from the lowest position, y 5 15 m, increases to the geomembrane bottom surface, the rise is 3.7 m. In this case, no water exists in the medium coarse sand, and the water pressure is 0 kPa. The backfill on the geomembrane in the middle reservoir is 30 kPa because of the 1.0-mthick sand blanket and the dead pool elevation, which is 1.0 m deep in the reservoir center. Figure 9A and B show the relationships among air pressure beneath the geomembrane, French drain spacing, and the load over the geomembrane when the rise in the groundwater level is 3.7 m in Case 1. The 75-m spacing of the French drain is not reasonable because the air pressure beneath the geomembrane will be greater than the load on the geomembrane during the rising process of the groundwater level to 3.7 m, thereby resulting in air bulging. The rationality of the French drain spacing is further determined in Cases 2 and 3. When the groundwater level rises higher than the French drain bottom and top, the French drain soil is saturated first to form an airtight channel, which then leads to the rapid increase of air pressure beneath the geomembrane, as shown in Figure 9C. However, the air pressure beneath the geomembrane in the center reservoir for French drain spacings of 25 m and 50 m is less than the load over the geomembrane and no air bulging occurs, even when the groundwater Figure 8. Variation of pore-air pressure beneath the geomembrane with the rate of rise in the groundwater level. (A) A 1.0-mthick backfill, no reservoir water, and 15.0 m of the initial groundwater level elevation and 2.0-m increase from the initial groundwater level during reservoir operation; (B and C) a
64
r 1.0-m-thick backfill, dead pool elevation, 15.0 m of the initial groundwater level elevation and 2.0-m increase from the initial groundwater level during reservoir operation. See Figure 4 for the location of the analysis points.
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 53–66
Bulging Geomembranes Used as Liner
is lower than 50 m but that (2) the rising rate of air pressure beneath the geomembrane for the French drain spacing of 25 m is higher than 50 m in Cases 2 and 3 because the large permeability coefficient of the French drain leads to more area of saturated soil and airtight formation for the French drain spacing of 25 m. Therefore, the original design with French drains spaced 50 m apart is reasonable and economical. CONCLUSIONS Extensive and thick, permeable strata occur beneath many reservoir sites. Accordingly, an antiseepage design with full geomembrane can effectively solve the leakage problem in unlined reservoirs. However, new engineering problems arise, including the production of air bulging beneath the geomembrane, and eventually cause geomembrane rupture and new leakage in the reservoir. The groundwater level may be relatively low during reservoir construction because of seasonal variations and dewatering. The pores of the thick unsaturated soil above the groundwater level trap ample air, and the laid geomembrane isolates the pore air from the atmosphere. The rise in the groundwater level and the compression of soil both decrease the pore-air volume and increase the pore pressure, thereby prompting the occurrence of air bulging and geomembrane rupture. The design of the Datun Reservoir in Dezhou City, Shandong Province, is studied to elucidate the problem of air bulge formation beneath the geomembrane. The following conclusions are noted:
Figure 9. Relationships among air pressure beneath the geomembrane, French drain spacing, and the load over the geomembrane when the rise of the groundwater level is greater than 3.7 m with the following conditions: (1) the backfill over the geomembrane is 1.0 m, (2) the pool elevation is 21.00 m, (3) the groundwater level elevation is 15.0 m, and (4) the rise in the groundwater level is greater than 3.7 m. (A) Duration is 10 days; (B) pore-air pressure in the center reservoir; (C) the relationship between the air pressure beneath the geomembrane and load over the geomembrane for French drain spacings of 25 and 50 m.
level rises to 6.26 m (the highest groundwater level is y 5 21.26 m during flood season). Figure 9C also reveals that (1) the air pressure beneath the geomembrane for the French drain spacing of 25 m in Case 1
1. During reservoir construction, high pore pressure is produced in the embankment. This pressure insignificantly influences air bulging beneath the geomembrane. 2. The decrease in pore-air pressure during unloading is less than the increase in the loading at the similar load on geomembrane because the rebound deformation is less than the compaction deformation. The poreair pressure beneath the geomembrane is smaller than the load on the geomembrane because of reservoir operation; however, the lowest water level elevation of 21.00 m is considered the most critical. 3. The rise of the groundwater level significantly affects the air bulging beneath the geomembrane. The critical rises in the groundwater level are 2.5 m on the sides and 1.5 m at the reservoir center area during construction and 1.75 m during reservoir operation. 4. Pore-air pressure is reduced with an increase in the rate of rise in the groundwater level. However, the groundwater level insignificantly influences the pore-air pressure beneath the geomembrane.
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 53–66
65
Cao, Yuan, He, Liu, and Yin
5. The French drain is set to prevent air bulging beneath the geomembrane from occurring. On the basis of the relationship between the top and bottom surfaces of the French drain and the groundwater level, three cases are considered. Case 1: When the groundwater level is lower than or equal to the bottom of the French drain, the French drain is fully effective. Case 2: When the groundwater level rises beyond the top of the French drain, the French drain fails to drain gas completely. Case 3: The French drain is partially submerged and only partly effective. 6. The reasonable spacing of the French drain must be satisfactory to dissipate air pressure beneath the geomembrane fully in Case 1 to ensure that no air bulging occurs, even when the groundwater level rises beyond the French drain bottom and top in Cases 2 and 3. 7. The 75-m French drain spacing is not reasonable because the air pressure beneath the geomembrane will be greater than the load on the geomembrane during the rising process of the groundwater level to 3.7 m, which results in air bulging. However, the air pressure beneath the geomembrane in the center reservoir for the French drain spacings of 25 m and 50 m is less than the load on the geomembrane and no air bulging occurs, even when the groundwater level increases to 6.26 m. The original design with French drains spaced 50 m apart is reasonable and economical.
ACKNOWLEDGMENTS This work was supported by the Fundamental Research Funds for the Central Universities (No. 2010B03414). The authors would like to thank the reviewers for their critical review and suggestions for improving the quality of this manuscript. REFERENCES AITCHISON, G. D., 1961, Relationship of moisture and effective stress functions in unsaturated soils. In Pore Pressure and Suction in Soils Conference: British National Society of International Society Soil Mechanics Foundation Engineering, London, U.K., pp. 47–52. BATHURST, R. J., 2007, Geosynthetics Classification. IGS Leaflets on Geosynthetics Applications: Electronic document, available at http://www.geosyntheticssociety.org BISHOP, A. W.; ALPAN, I.; BLIGHT, G. E.; AND DONALD, I. B., 1960, Factors controlling the strength of partly saturated cohesive soils. In Conference Shear Strength Cohesive Soils: American Society of Civil Engineers, New York, pp. 503–532. BOUAZZAA, A. AND VANGPAISALB, T., 2006, Laboratory investigation of gas leakage rate through a GM/GCL composite liner
66
due to a circular defect in the geomembrane: Geotextiles Geomembranes, Vol. 24, No. 2, pp. 110–115. CAO, X. S., AND YIN, Z. Z., 2009, Simplified computation of twodimensional consolidation of unsaturated soils: Rock and Soil Mechanics, Vol. 30, No. 9, pp. 2575–2580 (in Chinese). CAO, X. S.; YIN, Z. Z.; AND LING, H., 2007, Study of twodimensional consolidation and simplified calculation of unsaturated soil. In Proceedings of the 3rd Asian Conference on Unsaturated Soils, Nanjing, China: Science Press, Beijing, China, pp. 341–346. CAO, X. S.; YUAN, J. P.; HE, G. L.; WANG, B.; LIU, Y. H.; AND YIN, Z. Z., 2015, In situ test and analysis method of air bulging under geomembranes in a shallow-lined reservoir: Geotextiles Geomembranes, Vol. 43, No. 1, pp. 24–34. CHEN, Z. H.; XIE, D. Y.; AND LIU, Z. D., 1993, Consolidation theory of unsaturated soil based on the theory of mixture(I): Applied Mathematics Mechanics, Vol. 14, No. 2, pp. 127–137 (in Chinese). DIXON, R. M. AND LINDEN, D. R., 1972, Soil air pressure and water infiltration under border irrigation: Proceedings Soil Science Society America, No. 36, pp. 948–953. FOOSE, G. J.; BENSON, C. H.; AND EDIL, T. B., 2001, Predicting leakage through composite landfill liners: Journal Geotechnical Geoenvironmental Engineering, Vol. 127, No. 6, pp. 510–520. FREDLUND, D. G., AND RAHARDJO, H., 1993, Soil Mechanics for Unsaturated Soils: John Wiley and Sons, Inc., New York. GONG, A. M. AND SONG, T. W., 1999, The application of total geomembrane liner to solve the leakage problems in Golden Temple Reservoir: Water Conservancy Science Technology Economy, Vol. 3, No. 1, pp. 46–47 (in Chinese). GU, G. C., 1994, Status and development of the geomembrane for the anti-seepage design of earth rock dam in China: Water Resources Hydropower Northeast China, Vol. 124, No. 10, pp. 41–46 (in Chinese). GU, G. C., 2009, The application experience of the geomembrane for the anti-seepage design: Advances Science Technology Water Resources, Vol. 29, No. 6, pp. 34–38 (in Chinese). JIANG, H. B.; SHI, K. B.; AND LI, Y. J., 2010, Leakage analysis of leak proof geomembrane for large area of reservoir basin: Hydro Science Engineering, Vol. 4, pp. 58–61 (in Chinese). JOHNSON, A. I., 1967, Specific Yield Compilation of Specific Yields for Various Materials: U.S. Geological Survey Water-Supply Paper 1662-D, p. 74. LATIFI, H.; PRASAD, S. N.; AND HELWEG, O. J., 1994, Air entrapment and water infiltration in two-layered soil column: Journal Irrigation Drainage Engineering, Vol. 120, pp. 871–891. LI, W. L.; LI, Z. Q.; WEI, X. Y.; AND LI, Y. T., 2013, Air expansion caused by leakage water resulting from geomembrane defects: Chinese Journal Geotechnical Engineering, Vol. 25, No. 6, pp. 1161–1165. SHI, K. B. AND LI, Y. J., 2005, Discussion on application of antiseepage for whole bottom of Shengli Reservoir with geomembrane: Technology Water Resources Hydropower, Vol. 2, No. 11, pp. 143–144. YI, T. P. AND CHEN, P. S., 1999, Design and construction of whole seepage control using geomembrane: Technique Seepage Control, Vol. 3, No. 1, pp. 25–27 (in Chinese). YUAN, J. P.; CAO, X. S.; HE, G. L.; WANG, B.; LIU, Y. H.; AND YIN, Z. Z., 2014, Field test study of mechanism of bulge phenomenon under geomembrane in plain reservoir: Rock Soil Mechanics, Vol. 35, No. 1, pp. 67–73 (in Chinese). ZHANG, K.; LIU, S. H.; WANG, L. J.; SUN, L.; AND FU, Y. C., 2012, Numerical simulation of air field under geomembrane in antiseepage plain reservoir: South North Water Diversion Water Science Technology, Vol. 10, No. 5, pp. 97–101 (in Chinese).
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 53–66
Assessment of the CO2 Storage Potential in the Deep Saline Formation of Offshore Bohai Basin, China GUANBAO LI1 First Institute of Oceanography, State Oceanic Administration, Qingdao 266061, China
CHUANLIN HUO National Marine Environmental Monitoring Center, Dalian 116023, China
TIANYUN SU BAOHUA LIU First Institute of Oceanography, State Oceanic Administration, Qingdao 266061, China
Key Terms: Carbon Capture and Storage (CCS), Potential Assessment, Storage Capacity, Offshore Bohai Basin
storage capacity for hundreds of years of CO2 emissions from the Circum-Bohai-Sea region. Considering the storage capacity and risk, the Bozhong Sag area may offer the most promising early CO2 storage option in the basin.
ABSTRACT Regions of China have generated large CO2 emissions, and investigating the means for reducing them has become a priority. The geological storage of CO2 is one technique gaining interest, and the offshore Bohai Basin is believed to be a promising option to aid emissions mitigation. The Cenozoic sedimentary rocks, which measure more than 10,000 m in thickness and have diverse depositional systems and good reservoirseal assemblages, appear to offer the most favorable circumstance for CO2 storage in the offshore Bohai Basin. The Minghuazhen Formation of the Pliocene– Miocene consists of thick and extensive mudstone, especially its lower member, and is regarded as an excellent regional seal cover. The Miocene sandstone, including the lower member of the Minghuazhen Formation and the Guantao Formation, has high porosity and permeability and is a preferred target storage reservoir. The Dongying Formation and the Shahejie Formation of the Oligocene–Eocene are also storage options. Based on the pore volume method suggested by the U.S. Department of Energy, we calculated the effective storage capacity of the Cenozoic sedimentary rocks in the offshore Bohai Basin. The mean storage capacity is approximately 192 Gt, representing a huge capacity on a par with that of the Pearl River Mouth Basin, and potentially offering 1
Corresponding author email: gbli@fio.org.cn.
INTRODUCTION Tremendous energy consumption driven by fast industrial and economic growth and the coal-dominated energy structure make China one of the largest emitters of greenhouse gases in the world (Chen et al., 2009; Boden et al., 2011). The Circum-Bohai-Sea region, one of the most densely populated and heavily industrialized regions in China, has displayed huge CO2 emissions. The recent statistical data from the Report on Industrial Competitiveness of China (Zhang, 2011) show that three provinces in this region, i.e., Shandong, Hebei, and Liaoning, are ranked first, third, and sixth, respectively, in terms of CO2 emission in China, and the CO2 emission in this region accounts for nearly 30 percent of the total emission in China, if the emissions in Beijing and Tianjin are included. To mitigate CO2 emissions and eliminate their great threat to human health, environmental protection and sustained social development are urgent in this region. As a new technique to reduce the greenhouse gas emissions in significant scale and sustained impact (Zeng et al., 2004; Zhou, 2005; and Sun, 2006), CCS (carbon capture and storage) or CCUS (carbon capture, utilization, and storage, the substitute of CCS in China) is a growing priority in China (Best and Beck, 2011; Dahowski et al., 2012). Since 2005, CCS/CCUS has been integrated into the national
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 67–79
67
Li, Huo, Su, and Liu
medium- to long-term science and technology development plan, and a series of national research programs (e.g., 973 Program and 863 Program) have been launched. Meanwhile, several companies in China (e.g., Shenhua, SINOPEC, Petro China) are increasingly engaged in many CCS/CCUS pilot projects. In a report released by the Ministry of Science and Technology (MOST) of China in 2011, titled “Carbon Capture, Utilization and Storage Technology Development in China,” details on the principles, technologies, pilot activities, and international cooperation opportunities of CCUS development in China can be found (MOST, 2011). Furthermore, milestone goals from 2010 to 2030 were set out in a technology roadmap of CCUS in China that was published in 2013 (Zhang et al., 2013). In terms of CCS, CO2 is captured and injected back into three types of geological media, deep saline aquifers, abandoned oil and gas fields, or unminable coal beds, and stored there safely. Although onshore and offshore areas are both suggested for CO2 storage, most efforts have been predominantly focused on the former. Several preliminary evaluations of the geological storage of CO2 in China have been performed that also focus on offshore storage potential (Chen et al., 2009; Zheng et al., 2010; Jiao et al., 2011; Vincent et al., 2011; Dahowski et al., 2012; Qiao et al., 2012; and Wei et al., 2013). However, as suggested by Dahowski et al. (2012), possible high demands for CO2 storage and potential limits to the onshore capacity in China based on a cost curve analysis and detailed source-sink matching show that near-offshore storage potential could provide value to coastal regions and should be evaluated more closely. A comprehensive analysis by Schrag (2009) also proposed that offshore geological storage has more technical advantages than onshore geological storage and might be a more promising CO2 storage option in the future, with relatively easy pressure management and less impact on the population, land, and groundwater. Moreover, some pilot projects have been completed, not only for the storage evaluation of offshore sites, such as Mexico Bay, USA (NETL, 2008), the Gippsland Basin and Otway Basin of Australia (Cook et al., 2000; Bradshaw and Dance, 2005), and the Pearl River Mouth Basin of China (Zhou et al., 2011), but also on the storage implementation of the Sleipner project in the North Sea. In the industrialized Circum-Bohai-Sea region, the dense population and shortage of land and groundwater might lead to a number of challenges for onshore geological CO2 sequestration. In this case, the offshore Bohai Basin, surrounded by the CircumBohai-Sea region, is an alternative option. In this basin, there are many favorable conditions to implement offshore geological storage, including the
68
huge thickness of sedimentary deposits, the rich data and techniques on the accumulation of oil and gas exploration, and the close distance to CO2 sources. At this early stage of CCS in China, there is still a tradeoff between onshore and offshore storage; therefore, it is important to evaluate both potential options that may be available so that the trade-off can be appropriately weighed. In China, however, only a preliminary evaluation has been performed in the offshore Pearl River Mouth Basin (Zhou et al., 2011). As another important potential site for offshore CO2 storage, the Bohai Basin should be given more attention, and its storage capacity and suitability evaluation will be a prerequisite to make further decisions. That subject is the focus of this study. GEOLOGICAL BACKGROUND AND CO2 STORAGE CONDITIONS IN THE OFFSHORE BOHAI BASIN Geological Settings The Bohai Sea, one of the inland seas of China, lies between 37u079N and 41uN and 117u359E and 121u109E. It is circled by Shandong Peninsula and Liaodong Peninsula and covers a total area of 78,000 km2. Its mean water depth is 18 m. Approximately 95 percent of the sea is shallower than 30 m. The region deeper than 30 m is mainly located at the Bohai Strait, where the maximal water depth is 86 m (Qin et al., 1990). The offshore Bohai Basin is an extension of the Bohai Bay Basin in the sea area, which is a Cenozoic rift basin formed on the North China Craton (Lu et al., 1997; Zhu et al., 2009). The total area of the offshore Bohai Basin exceeds 50,000 km2. The basin is divided into five tectonic units, which are the Bozhong Depression, Liaodong Bay Depression, Jiyang Depression, Huanghua Depression, and Chengning Uplift. The Bozhong Depression is the only one that is completely part of the offshore Bohai Basin; all the others are extensions of onshore units (Figure 1). Each of the depressions and uplifts is composed of several sags and rises, which are the basic tectonic units of the basin (Figure 1). In the Cenozoic, extensive faults caused the base rock to tilt or twist, which resulted in grabens or half grabens, forming the prototype of the Bohai Rift Basin (Zhu et al., 2009). Cenozoic Stratigraphy The Cenozoic stratum of the offshore Bohai Basin is mainly made up of terrigenous clastic rocks. Its maximal thickness is close to 12 km at the Bozhong Sag, in which the Neogene units exceed 5 km at most (Li and Lu, 2002). During the long and complex geological
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 67–79
CO2 Storage of Offshore Bohai Basin
Figure 1. Sketched geological map of the offshore Bohai Basin. The red points represent the petroleum wells (after Li and Lu, 2002). The shaded area in the center of the basin shows the region in which all the parameters for the storage estimates are available (see the text for details). The inlet at the upper-left corner shows the location of the offshore Bohai Basin (in blue), the Pearl River Mouth Basin (in red), and the Circum-Bohai-Sea region (in gray). A NE directional profile shows the stratigraphic units of the basin.
evolution including Paleogene rifting and Neogene postrifting and depression, diverse sedimentary systems were formed, including alluvial fan, fan delta, braided river delta, river delta, submarine fan, sub-lacustrine fan, lake, and extremely shallow water deltas (Figure 2) (Yang and Xu, 2004; Zhu et al., 2009). The uplift in the basin and circumferential highland provide a large amount of terrigenous clastic material. Petroleum Geology The source rocks of the offshore Bohai Basin are concentrated in the Paleogene, including the Shahejie
Formation and the lower Dongying Formation, in which a dark mudstone was well developed. The thickness of the mudstone is over 1,000 m and its volume exceeds half of each formation. The Bohai Basin is characterized by a multicycled, source-reservoir-seal assemblage. Among the four main reservoir rocks, the Paleogene and the Neogene sandstones are the most important for current petroleum exploration in the offshore Bohai Basin. They are, on average, 50–60 percent sand and have favored features of a wide distribution, excessive layers, and good reservoir performance (Cai, 2005).
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 67–79
69
Li, Huo, Su, and Liu
Figure 2. Offshore Bohai Basin stratigraphy (after Yang and Xu, 2004, modified). In addition to the source-reservoir-seal assemblage of petroleum geology, the reservoirs and seals for CO2 storage are suggested.
70
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 67–79
CO2 Storage of Offshore Bohai Basin
The lower Minghuazhen Formation consists of thick shallow lake mudstone and is regarded as the first regional seal. In the sag or the slope, the seals are developed better than those in the heave. In the heave, which had a high content of sandstone, the lower Minghuazhen Formation has superior self-reserved and self-capped conditions (Zhu et al., 2009). In addition, the lacustrine mudstone of the Upper Minghuazhen Formation is also a good seal, especially in the Bozhong region. The first members of the Shahejie Formation and the lower Dongying Formation mudstone are other regional seals for the lower reservoir. By 2006, in total, 511 wells had been drilled, and 41 gas and oil fields had been found in the offshore Bohai Basin. The proved oil and natural gas reserves were 2.03 3 109 m3 and 54.6 3 109 m3, respectively (Zhu et al., 2009). Data from the CNOOC, Ltd., Annual Report (2007–2013) show that approximately 280 new wells were drilled, and the incremental proved oil and natural gas reserves were 1.12 3 109 m3 and 13.6 3 109 m3. CO2 Storage Conditions Good reservoir and seal assemblage are necessary for CO2 storage. Several sets of mudstone were developed in the offshore Bohai Basin in the Cenozoic Era based on the data obtained by petroleum exploration. Of them, the Minghuazhen Formation is believed to be a good candidate for the uppermost regional seal for CO2 storage (Figure 2). The Lower Minghuazhen Formation is mainly made up of shallow lacustrine deposits, in which the mudstone has an accumulative thickness up to 1,000 m and a single-layer maximal thickness close to 200 m in the Bozhong area and is thinned toward the surrounding area in a circular-strip shape. Like the Lower Minghuazhen Formation, the lacustrine mudstone of the Upper Minghuazhen Formation is also a good seal. Its accumulative thickness is thinned from Bozhong toward the surroundings in a circular-strip shape, and its single-layer mudstone is 15–80 m in thickness. Additionally, the first member of the Shahejie Formation and the Lower Dongying Formation can also be used as another set of regional seals. The sandstone stratum of the Lower Minghuazhen Formation and its underlying formation are the potential reservoir for CO2 storage in the offshore Bohai Basin (Figure 2). The sandstone of the Minghuazhen Formation is mainly fluvial facies and developed better in the surrounding areas than in the center of the basin. In the surrounding areas, the sandstone constitutes over 50 percent of the unit and
consists mainly of coarse sandstone and pebbly sandstone with a single-layer thickness up to 50 m. In the center, the sandstone decreases and becomes thinner, and sand-mud interbeds are prevalent. The porosity and the permeability of the Minghuazhen Formation are 22–41 percent and 48–14,220 mD, respectively (Yang and Xu, 2004; Zhu et al., 2009). The sandstone percentage of the Guantao Formation is 32–96 percent and varies in different regions. In the center and northern parts of the basin, the sandstone of the Guantao Formation is over 70 percent and is mainly composed of thick and massive glutinite. In the southwestern part of the basin, it is only 30–70 percent and is mainly composed of pebbly sandstone, medium and fine sandstone, and coarse sandstone, which are interbedded with red and green mudstone and have a porosity of 11–30 percent and a permeability of 30–1,482 mD. The sandstone of the Upper Dongying Formation, which is mainly composed of delta front and floodplain sand, is well developed and widely distributed. Its sand percentage is 20–50 percent on average, and its porosity and permeability are up to 8–33.5 percent and 8,883 mD, respectively. The third member of the Shahejie Formation, which is composed of thick glutinite and thin turbidite, is most widely distributed in this formation, but its physical properties vary in different areas. For example, at a well in the Bozhong Depression, the porosity reaches 10.5–11.8 percent, and the permeability is up to 59.7–358.4 mD, whereas at an adjacent well in the same area, the porosity and the permeability are only 3.78 percent and 0.02 mD, respectively. There are several sets of storage-seal assemblages in the offshore Bohai Basin that may offer potential for secure, long-term CO2 storage. The first is the mudstone of the Minghuazhen Formation, which is used as the seal of the reservoirs of the Minghuazhen Formation, the Guantao Formation, and the Upper Dongying Formation. The second is the mudstone of the Lower Dongying Formation, which is used as the seal of the reservoirs of the Lower Dongying Formation, and the first member of the Shahejie Formation. The third is the self-reserved and selfcapped assemblage of the third member of the Shahejie Formation. METHOD AND DATA SOURCES Two aspects important in evaluating CO2 storage potential (i.e., the capacity calculation and preliminary risk evaluation) are focused on in the work of CO2 storage potential assessment (Bachu, 2003; Ramı´rez et al., 2010). The former means to assess the accommodation of storage space, and the latter stresses the risk and uncertainty assessment of CO2 storage. Here, only
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 67–79
71
Li, Huo, Su, and Liu
the storage capacity of the deep saline aquifer in the offshore Bohai Basin is calculated, despite the fact that there are two other storage media, the abandoned oil and gas fields and the unminable coal beds (IPCC, 2005; Bachu et al., 2007a; Bachu, 2008; and USDOE, 2008), due to insufficient data from the unminable coal beds and the currently active condition of the oil and gas field. The oil and gas production started in the late 1960s and is still in an increasing stage. Most oil and gas fields are far from depletion and are not likely to be ready for CO2 storage anytime soon, as noted by Bachu (2006). Method of Storage Capacity Calculation for Deep Saline Aquifers An understanding of the effective capacity is necessary for basin-scale potential evaluation (Bachu et al., 2007a), which is defined by the Carbon Sequestration Leadership Forum (CSLF) as a subset of the theoretical capacity, including consideration of many geological and engineering limits of reservoirs, such as physical features (porosity, permeability, temperature, and pressure), buried depth, seal stability, and safety (Bachu et al., 2007a). Two methods were proposed by Bachu et al. (2007b) and USDOE (2008), which are believed by Bachu (2008) to be equivalent, and both could be utilized for the storage capacity evaluation on the country scale, region scale, and basin scale. Here, we use the method proposed by USDOE (2008), which better fits the data collected in the offshore Bohai Basin. According to USDOE (2008), the effective storage capacity can be expressed as: MCO2 ~A|h| Q|rCO2 |E,
ð1Þ
where MCO2 is the CO2 effective storage capacity, A is the reservoir area, h is the reservoir thickness, Q is the porosity of the reservoir, rCO2 is the density of CO2 in the reservoir, and E is the effective coefficient of storage. Equation 1 can also be expressed in the integral mode, if the reservoir can be considered as the assemblage of many three-dimensional (3D) units dxdydz with porosity Q(x, y, z), sandstone percentage d(x, y, z), and CO2 density rCO2 (x, y, z). ððð MCO2 ~E|
Qðx,y,zÞ|dðx,y,zÞ
|rCO2 ðx,y,zÞ|dxdydz:
72
ð2Þ
In discrete mode, Eq. 2 can be expressed as X X MCO2 ~E| Dhij |Qij |dij |rij : DAi | i j ð3Þ Supposing that the target reservoir is made up of many cubic units with area of DAi and thickness of Dhij, their center burying depth is dij, the sandstone percentage is dij, the porosity is Qij, the temperature is Tij, and the pressure is Pij. Here, rCO2 is the CO2 density at temperature Tij and pressure Pij. Data Sources The presence of exploratory wells is sparse in the offshore Bohai Basin, only approximately 70 per 10,000 km2. When the evaluation wells are included, there are 145 wells per 10,000 km2. Therefore, the basin is in a low to medium exploration degree. Fortunately, more than 25 3 104 km of 2D seismic data and 4 3 104 km2 of 3D seismic data give a relatively detailed depiction of the stratum layering. In the present paper, the data for reservoir thickness, physical features, and the geothermal field in the offshore Bohai Basin are collected (Table 1), and the following assumptions are made for the calculation: (1) The basin can be divided into many calculation units that are equal in area and thickness, and the capacity of each unit is calculated. Next, Eq. 3 is used to calculate the total storage capacity of the basin. (2) Because the physical parameters of the stratum below the third member of the Shahejie Formation are not well known, only the storage capacities of the Lower Minghuazhen Formation, the Guantao Formation, the Dongying Formation, and the first to third members of the Shahejie Formation are calculated. (3) The minimal depth for submarine storage should ensure CO2 that reaches the supercritical state. The simple way to do so is to set the depth to 800 m or 1,000 m. Herein, the critical depths of all grids are calculated based on the local conditions and applied as the minimum depth for storage. (4) The maximal depth for CO2 storage was suggested to be 3,500 m (Bachu, 2003), but recent results indicate 2,500 m (Chadwick et al., 2008). For the purpose of comparison, this analysis evaluates capacities based on not only these two maximum depths but also the storage capacity of the entire depth of the Cenozoic reservoirs (Figure 3). (5) The reservoir porosity shows a good linear statistical relationship with the reservoir depth
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 67–79
CO2 Storage of Offshore Bohai Basin Table 1. Principles for parameter determination and the main data sources. Parameter DAi Dhij Qij dij rij
E
(6)
(7)
(8)
(9)
Principles for Parameter Determination Area of given grid (1 km 3 1 km) Thickness of given grid (100 m) Statistics relationship of porosity and burial depth Same sandstone percentage was used for the same stratigraphical unit Interpolating based on relationship of CO2 density with temperature/pressure conditions; applying hydrostatic pressure of each depth as pressure; calculating the temperature according to surface temperature distributions and temperature gradients of different depths
E 5 0.02 was used to calculate the mean value of storage capacity; E 5 0.0051 and E 5 0.055 were used to calculate the upper and lower limit of capacity, respectively.
less than 4,000 m in the offshore Bohai Basin. Such a relationship has been applied to determine the porosity of all calculation units at different depths. For the reservoirs deeper than 4,000 m, the porosity is set as 5 percent. Assuming that all the calculation units within a stratigraphic unit have the same sandstone percentage, the product of the sandstone percentage and the calculation unit thickness is used as the actual reservoir thickness for storage. The temperature of each of the calculation units is determined using the isolines of the regional geothermal gradient and the temperature at different depths (1,000 m, 2,000 m, 3,000 m, and 4,000 m). The temperature of the reservoir shallower than 1,000 m is calculated based on the temperature gradient and the temperature at a depth of 1,000 m. The reservoir temperature at depths ranging from 1,000 m to 4,000 m is calculated by linear interpolations based on the temperature at the upper and the lower interfaces, and the reservoir temperature below 4,000 m is calculated by extrapolation based on the temperature at 3,000 m and 4,000 m. The CO2 density is a function of the temperature and pressure conditions (Span and Wagner, 1996). Here, the CO2 density of each calculation unit is calculated by searching reference tables based on the static pressure and the temperature calculated according to point 7. There are many references on the effective storage coefficient E (USDOE, 2008, 2010;
Data
Data Source
– – Porosity of reservoirs in different depths
– – EC (1990)
Contour map of sandstone percentage for each stratigraphical unit (N1m, Ng, Ed, Es) Isopach map of each Q, N2m N1m, Ng, Ed, stratigraphical unit Es1-2, Es3
EC (1990)
Isobath map of Bohai Sea Relationship of CO2 density to temperature and pressure Temperature contour map of different depths (1,000 m, 2,000 m, 3,000 m, 4,000 m) and temperature gradient contour map –
Zhu et al. (2009) EC (1990) Qin et al. (1990) Span and Wagner (1996) Zhu et al. (2009)
USDOE (2010)
IEAGHG, 2009; and Wildgust, 2010). Zhou et al. (2011) applied the effective coefficients E 5 0.01 and E 5 0.04 with P15 and P85 confidence for the calculation of the lower and the upper limits of the storage capacity and E 5 0.026 for the calculation of the mean value of the storage capacity in the Pearl River Mouth Basin, following Wildgust (2010). The most updated results come from the third edition of the Carbon Sequestration Atlas of the US and Canada (USDOE, 2010), in which E was suggested to be 0.0051, 0.02, and 0.055 for clastic lithology of saline aquifers. The probabilities of these values are P10, P50, and P90, respectively. In this paper, USDOE (2010) is followed. (10) Not all the parameters mentioned above are available for the whole study area. For example, the sandstone percentage data of the Guantao Formation cover only the Bozhong Depression, Huanghua Depression, Chengning Uplift, and a small part of both the Jiyang Depression and Liaodong Bay Depression. In this case, the calculation parameter in the void region is extrapolated from that in the covered regions. The shaded region in Figure 1 shows the range of area where all parameters used for the capacity calculations are available. Outside this region, at least one parameter is calculated by extrapolation. As a result, the calculated capacity for this region is believed to have a higher confidence than the other regions.
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 67–79
73
Li, Huo, Su, and Liu
Figure 3. Isopach map of the reservoir for CO2 storage in three cases: (A) the reservoir between the supercritical depth and the lower interface of the third member of the Shahejie Formation, (B) the reservoir between the supercritical depth and 3,500 m, and (C) the reservoir between the supercritical depth and 2,500 m. (D) An E-W directional profile showing the variation of the reservoir for CO2 storage in the three cases mentioned above.
RESULTS AND DISCUSSION Total CO2 Storage Capacity of the Offshore Bohai Basin Table 2 shows the resulting CO2 storage capacity estimates for the offshore Bohai Basin. The total reservoir pore volume between the critical depth and the lower interface of the third member of the Shahejie Formation for each stratigraphic unit from the Lower Minghuazhen Formation to the Shahejie Formation is 1.7 3 104 km3, where a total of 52–564 Gt CO2 can be stored, with the mean value being 205 Gt if the storage effective coefficient is considered. From the critical depth to a depth of 3,500 m, the total pore volume is 1.6 3 104 km3, and the mean effective CO2 storage capacity is 192 Gt, decreasing only approximately 6 percent compared with the
74
reservoir over the lower interface of the third member of the Shahejie Formation, of which the maximal sediment thickness is nearly 10 km. For the reservoirs from the critical depth to a depth of 2,500 m, the total pore volume is 1.3 3 104 km3, and the mean value of effective storage capacity is 156 Gt. Both values decrease by more than 20 percent compared with the reservoir over the lower interface of the third member of the Shahejie Formation. Thus, a serious underestimation of capacity may be caused if 2,500 m is regarded as the lower limit of the storage depth. Li et al. (2009) published an estimated CO2 storage capacity of the saline aquifer in the offshore Bohai Basin at 109 Gt, a value less than but still falling within the range of our results. This estimate was based on a conservative solubility method (Li et al., 2009). The USDOE (2008) method considers only the
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 67–79
CO2 Storage of Offshore Bohai Basin Table 2. Calculated storage capacity of the stratum at different storage depths, with E 5 0.02. Effective Capacity (Gt) Formation
Dc 5 Es3
Dc 5 3,500
Dc 5 2,500
Lower Minghuazhen Guantao Dongying Shahejie Total
54 71 25 55 205
54 70 24 44 192
54 60 19 23 156
Dc 5 depth where CO2 reaches critical state; Es3 5 third member of Shahejie Formation.
effective pore volume and does not aim to identify any specific mechanisms. The similarity of the two results based on different methods might mean to some degree that the solution trap is prominent among the four trapping mechanisms. The results of this analysis are more comparable to the storage capacity of the Pearl River Mouth Basin calculated by Zhou et al. (2011), in that the same method was used. The effective CO2 storage capacity in the Middle to Lower Miocene and Paleogene strata of the Pearl River Mouth Basin is estimated at 118– 473 Gt, with a mean value of 308 Gt. The storage capacity of the strata at depths from 800 m to 2,500 m is 81–324 Gt, with a mean value of 210 Gt (Zhou et al., 2011). For the purpose of comparison, the results have been transformed to a common set of storage efficiency coefficients of 0.0051, 0.02, and 0.055, which have a probability of P10, P50, and P90 percent, respectively. An effective capacity of 60–650 Gt, with a mean value of 237 Gt, has been obtained for the whole Cenozoic reservoir, and an effective capacity of 41–445 Gt, with mean value of 162 Gt, was found for the reservoirs at depths over 2,500 m. Although the area of the Pearl River Mouth Basin exceeds 20 3 104 km2, which is far larger than the offshore Bohai Basin, the storage capacities of the two basins are similar, whether comparing the whole Cenozoic reservoir or those at depths over 2,500 m. The reasons are that, on the one hand, the offshore Bohai Basin has a larger proportion of sag area and a thicker reservoir, and the strata between the critical depth and the depth of 2,500 m are almost completely occupied by the Cenozoic units, which have an average thickness close to 1,400 m (Figure 3C and D). In the Pearl River Mouth Basin, however, the uplift covers approximately 40 percent of the total area, and the average thickness of the whole Cenozoic group is less than 1,000 m, and lower than 600 m at a depth of 800–2,500 m. On the other hand, it may be related to the good physical properties of the reservoirs in the offshore Bohai Basin, where the sandstone percentage averages 43 percent in the Lower Minghuazhen Formation, 74 percent in the Guantao
Formation, 30 percent in the Upper Dongying Formation, and 45 percent in the Shahejie Formation (EC, 1990). Moreover, the porosity of the reservoirs in the offshore Bohai Basin is also high, being 25–35 percent at depths less than 2,000 m, 20–30 percent at depths of 2,000–2,700 m, and up to 10–20 percent at depths of 2,700–3,600 m (EC, 1990). These values are all much higher than those used by Zhou et al. (2011), who used 0.5 and 0.2 for the sandstone percentage and porosity of the Miocene series and 0.37 and 0.1 for the sandstone percentage and porosity of the Paleogene series, respectively. The storage capacity of each stratum is calculated, and the results are shown in Table 2. At the depth from the critical depth to 3,500 m, the capacity of the Lower Minghuazhen Formation, the Guantao Formation, the Dongying Formation, and the Shahejie Formation accounts for 28 percent, 36 percent, 13 percent, and 23 percent of the total, respectively, that is, the capacity of the Miocene series comprises two thirds of the total. Even including the strata at all the depths, the storage capacity of the Miocene series still exceeds 60 percent of the total. Therefore, the Miocene strata could be the major target reservoir for CO2 storage in the offshore Bohai Basin. This result is in accordance with that in the Pearl River Mouth Basin (Zhou et al., 2011). Regional Differences of Storage Capacity Figure 4 shows a contour map of the CO2 effective storage capacity density (CSCD) at the depths from the critical depth to 3,500 m in the Bohai Basin, which represents the CO2 amount that can be stored in a unit area of the basin. The regional difference of CSCD is apparent. The maximal CSCD, exceeding 5.5 3 104 t/km2, lies at the junction of the southern Liaodong Bay Depression and the Bozhong Depression, where the thermal gradient is below 25uC/km, a very low value in the Bohai Basin. The thermal gradient may be one of the factors controlling CSCD. Low thermal gradients can give rise to a stratum temperature lower than the surroundings, which have a high thermal gradient. As a result, the density of the stored CO2 is high, so that the CO2 storage capacity can be enhanced. This is further supported by the occurrence of minimal CSCD in the Huanghekou Sag in the Jiyang Depression and the Shaleitian Rise in the Chengning Uplift, where the thermal gradients are both higher than 40uC/km, and the thickness of the Cenozoic group is less than in the offshore Bohai Basin. As for each of the depressions or uplifts, the Liaodong Bay Depression has the largest CSCD, and the total storage capacity there is also the largest, with an average of 71 Gt. The Bozhong Depression has
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 67–79
75
Li, Huo, Su, and Liu
Figure 4. Contour map of CSCD (CO2 storage capacity density) in the offshore Bohai Basin; values are in 104 t/km2.
a lower CSCD than the Liaodong Bay Depression, but its storage capacity, reaching 65 Gt, is the second highest due to its large area. The CSCD and storage capacities of the Huanghua Depression, the Chengning Uplift, and the Jiyang Depression are all relatively smaller, with an average of 21 Gt, 19 Gt, and 16 Gt, respectively. Uncertainty caused by incomplete data should not be underestimated and might be one reason for the regional differences mentioned above. However, the main reason is the basin features. Summarily, the CSCD distributions of all the units are consistent with their geological features, such as the geothermal gradients and the formation thickness. For example, the thicknesses of the Quaternary and the Pliocene units in the Bozhong Depression are the largest in the offshore Bohai Basin, which may provide larger capping pressure and safety while reducing the stratum thickness used for the CO2 storage within the limited depth scope (critical depth to 3,500 m), as shown in Figure 3D. Analysis of the Storage Potential of the Offshore Bohai Basin and Storage Area Selection The mean storage capacity of the offshore Bohai Basin is 192 Gt, which can hold nearly 1,000 years of
76
CO2 emissions of the Circum-Bohai-Sea region, if the annual emissions within 50 km and/or 100 km from the coast (approximately 158 Mt/yr and 254 Mt/yr, respectively; Li et al., 2006) are taken into account. If taking the annual emissions from the large sources of the five provinces or cities in the Circum-Bohai-Sea region in 2011 (approximately 1 Gt; Zhang, 2011) into account, nearly 200 years of CO2 emissions could be stored in the basin. Even calculated using the minimal value of the storage capacity, 50 years of emissions could be stored. These early results suggest that the offshore Bohai Basin can provide sufficient space for storing the CO2 emissions from sources in the Circum-Bohai-Sea region for a long time. On the premise of the sufficient storage capacity, a series of factors for risk management should also be carefully considered, which include (1) faults, especially the displacement, age, and number of active faults; (2) the magnitude, frequency, and type of natural earthquakes; (3) the density, status, time, and accessibility of drilled wells; and (4) the thickness, composition, and sealing capability of the cap rocks (Ramı´rez et al., 2010). There are many active faults in the offshore Bohai Basin (Figure 5), and many of them have cut upwards to the seafloor. Two huge regionally intersecting fault
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 67–79
CO2 Storage of Offshore Bohai Basin
Figure 5. The suggested potential storage area in the offshore Bohai Basin. The isopach of the seal, the epicenter location of earthquakes larger than ML 4.0 since A.D. 408 (from China Earthquake Data Center), the oil and gas fields (after Zhu et al., 2009), and the active faults are also shown. The suggested early option for CO2 storage is denoted as a shadowed area.
zones, namely the Tanlu Fault Zone and the PenglaiZhangjiakou Fault Zone, show strong activity. Many destructive earthquakes larger than magnitude 7 have occurred in history, and the latest one (with a registered magnitude of 7.4) occurred in 1969 in the sea outside the Yellow River mouth. On the whole, the strongly active tectonics and earthquakes in the offshore Bohai Basin lead to adverse effects on the security of the CO2 storage, even though some areas of the region may be favorable. Among the sags and rises in the offshore Bohai Basin, the Bozhong Sag might be one of the most promising options for an early storage area selection. The Bozhong Sag is the largest sag in terms of area and Cenozoic thickness in the Bohai Basin and possesses a CO2 storage capacity larger than 31 Gt, which could contribute significant storage capacity to nearby sources. Additionally, it has the following favorable conditions: (1) The scarcity of large-scaled active faults and earthquakes within the sag can provide good storage security; (2) the huge thick mudstone in the Minghuazhen Formation can be the regional seal and take advantage of the good physical
features of the Miocene reservoir; and (3) there is less impact on the oil and gas fields located around the sag (Figure 5). However, additional evaluation of the suitability of the Bohai Basin for CO2 storage would be needed before selection of storage targets could begin. CONCLUSIONS Geological storage of CO2 is a promising technique for reducing emissions of CO2 to the atmosphere. In this paper, the CO2 storage capacity of the offshore Bohai Basin of China is calculated based on the published data and methods, and the storage potential is evaluated. The following conclusions have been drawn. (1) The Cenozoic group of the offshore Bohai Basin has good reservoirs and seal conditions and therefore has strong potential for CO2 storage. Both the Lower Minghuazhen Formation and the Guantao Formation of the Miocene series have thick strata with good physical properties
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 67–79
77
Li, Huo, Su, and Liu
and should be the first option for storage. The next option is the Dongying Formation and Shahejie Formation of the Oligocene series. (2) The offshore Bohai Basin has a very large potential CO2 storage capacity. In the reservoirs at depths between the critical depth and 3,500 m, including stratigraphic units from the Lower Minghuazhen Formation to the third member of the Shahejie Formation, a total of 49–530 Gt of CO2, with a mean of 192 Gt, may be stored, which could provide an important CO2 storage resource for the Circum-Bohai-Sea region. (3) Reservoir characteristics vary considerably throughout the offshore Bohai Basin, and the resulting calculated storage capacities vary across different tectonic units. The Liaodong Bay Depression and the Bozhong Depression have a high storage capacity that accounts for two thirds of the whole basin. If initial risk management factors are considered, the Bozhong Sag may be the most promising region and potentially the first choice for storage in the basin due to its large capacity, stable tectonics, and reduced potential to impact known oil and gas resources.
REFERENCES BACHU, S., 2003, Screening and ranking of sedimentary basins for sequestration of CO2 in geological media: Environmental Geology, Vol. 44, No. 3, pp. 277–289. BACHU, S., 2006, Characterization of sites for geological storage of carbon dioxide: Environmental Geosciences, Vol. 13, No. 2, pp. 67–70. BACHU, S., 2008, Comparison between Methodologies Recommended for Estimation of CO2 Storage Capacity in Geological Media—Phase 3: CSLF Task Force on CO2 Storage Capacity Estimation and the USDOE Capacity and Fairways Subgroup of the Regional Carbon Sequestration Partnerships Program, Report: CSLF-T-2008-04, 21 p. BACHU, S.; BONIJOLY, D.; BRADSHAW, J.; BURRUSS, R.; CHRISTENSEN, N. P.; HOLLOWAY, S.; AND MATHIASSEN, O. M., 2007a, Estimation of CO2 Storage Capacity in Geological Media— Phase 2: Final Report CSLF-T-2007-04 from the Task Force for Review and Identification of Standards for CO2 Storage Capacity Estimation, 43 p. BACHU, S.; BONIJOLY, D.; BRADSHAW, J.; BURRUSS, R.; HOLLOWAY, S.; CHRISTENSEN, N. P.; AND MATHIASSEN, O. M., 2007b, CO2 storage capacity estimation: Methodology and gaps: International Journal of Greenhouse Gas Control, Vol. 1, pp. 430–443. BEST, D. AND BECK, B., 2011, Status of CCS Development in China: Energy Procedia, Vol. 4, pp. 6141–6147. BODEN, T. A.; MARLAND, G.; AND ANDRES, R. J., 2011, Global, Regional, and National Fossil-Fuel CO2 Emissions: Carbon Dioxide Information Analysis Center, Oak Ridge National Laboratory, U.S. Department of Energy, Oak Ridge, TN: Electronic document, available at http://dx.doi.org/10.3334/ CDIAC/00001V2011
78
BRADSHAW, J. B. AND DANCE, T., 2005, Mapping geological storage prospectivity of CO2 for the world sedimentary basins and regional source to sink matching. In Wilson, M.; Morris, T.; and Gale, J. (Editors), Proceedings of the 7th International Conference on Greenhouse Gas Control Technologies (GHGT,7): Elsevier, Vancouver, Canada, pp. 583–592. CAI, Q., 2005, Oil & Gas Geology in China Seas: Ocean Press, Beijing, pp. 111–125. (in Chinese with English abstract). CHADWICK, A.; ARTS, R.; BERNSTONE, C.; MAY, F.; THIBEAU, S.; AND ZWEIGEL, P., 2008, Best practice for the storage of CO2 in saline aquifers. In Observations and Guidelines from the SAC Sand CO2 STORE Projects. Keyworth, Nottingham: British Geological Survey Occasional Publication No. 14, 277 p. CHEN, W.; TENG, F.; XU, R.; XIANG, X.; ZENG, R.; DOMPTAIL, K.; ALLIER, D.; AND LE NINDRE, Y.-M., 2009, CCS scenarios optimization by spatial multicriteria analysis: Application to multiple source-sink matching in the Bohai Basin (North China): Energy Procedia, Vol. 1, No. 1, pp. 4167–4174. COOK, P. J.; RIGG, A.; AND BRADSHAW, J., 2000, Putting it back where it came from: Is geological disposal of carbon dioxide an option for Australia?: Appea Journal, Vol. 11, pp. 1–13. DAHOWSKI, R. T.; DAVIDSON, C. L.; LI, X. C.; AND WEI, N., 2012, A $70/tCO2 greenhouse gas mitigation backstop for China’s industrial and electric power sectors: Insights from a comprehensive CCS cost curve: International Journal of Greenhouse Gas Control, Vol. 11, pp. 73–85. EDITORIAL COMMITTEE OF “PETROLEUM GEOLOGY OF OIL & GAS BEARING AREAS ON THE CONTINENTAL SHELF AND ITS NEIGHBORING REGIONS,” (EC), 1990, Petroleum Geology of China (PGC) (Volume 16): Oil & Gas Bearing Areas on the Continental Shelf and Its Neighboring Regions: Petroleum Industry Press, Beijing, pp. 3–260. (in Chinese with English abstract). INTERNATIONAL ENERGY AGENCY GREENHOUSE GAS PROGRAMME (IEAGHG), 2009, Development of Storage Coefficients for Carbon Dioxide Storage in Deep Saline Formations: IEA Greenhouse Gas R&D Programme, Cheltenham, UK, 118 p. INTERGOVERNMENTAL PANEL ON CLIMATE CHANGE (IPCC), 2005, In Metz, B.; Davidson, O.; De Coninck, H.; Loos, M.; and Meyer, L. (Editors), IPCC Special Report on Carbon Dioxide Capture and Storage: Cambridge University Press for the Intergovernmental Panel on Climate Change, Cambridge, UK, Chapters 28–34. JIAO, Z.; SURDAM, R. C.; ZHOU, L.; STAUFFER, P. H.; AND LUO, T., 2011, A feasibility study of geological CO2 sequestration in the Ordos Basin, China: Energy Procedia, Vol. 4, pp. 5982–5989. LI, G. AND LU, M., 2002, Atlas of Oil and Gas Basins in China, 2nd ed.: Petroleum Industry Press, Beijing, pp. 269–283. (in Chinese with English abstract). LI, X.; LIU, Y.; BAI, B.; AND Fang, Z., 2006, Ranking and screening of CO2 saline aquifer storage zones in China: Chinese Journal of Rock Mechanics and Engineering, Vol. 25, No. 5, pp. 963–968. (in Chinese with English abstract). LI, X.; WEI, N.; LIU, Y.; FANG, Z.; DAHOWSKI, R. T.; AND DAVIDSON, C. L., 2009, CO2 point emission and geological storage capacity in China: Energy Procedia, Vol. 1, No. 1, pp. 2793–2800. LU, K.; QI, J.; AND DAI, J., 1997, The Structural Model of Cenozoic Hydrocarbon Basin in Bohai Bay. Geological Publishing House, Beijing, p. 251. (in Chinese with English abstract). MINISTRY OF SCIENCES and TECHNOLOGY OF CHINA (MOST), 2011, Carbon Capture, Utilization and Storage Technology Development in China: Electronic document, available at www. acca21.org.cn/gest/etc/CCUS_en.pdf
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 67–79
CO2 Storage of Offshore Bohai Basin QIAO, X.; LI, G.; LI, M.; AND WANG, Z., 2012, CO2 storage capacity assessment of deep saline aquifers in the Subei Basin, East China: International Journal of Greenhouse Gas Control, Vol. 11, pp. 52–63. QIN, Y.; ZHAO, Y.; CHEN, L.; AND ZHAO, S., 1990, Geology of Bohai Sea: Ocean Press, Beijing, p. 354. (in Chinese with English abstract). RAMI´REZ, A.; HAGEDOORN, S.; KRAMERS, L.; WILDNBORG, T.; AND HENDRIKS, C., 2010, Screening CO2 storage options in the Netherlands: International Journal of Greenhouse Gas Control, Vol. 4, pp. 367–380. SCHRAG, D. P., 2009, Storage of carbon dioxide in offshore sediments: Science, Vol. 325, pp. 1658–1659. DOI:10.1126/ science.1175750 SPAN, P. AND WAGNER, W., 1996, A new equation of state for carbon dioxide covering the fluid region from the triplepoint temperature to 1100K at pressures up to 800 MPa: Journal of Chemical Reference Data, Vol. 25, No. 6, pp. 1509–1596. SUN, S., 2006, Geological problems of CO2 underground storage and its significance on mitigating climate change: China Basic Science, Vol. 3, pp. 17–22. (in Chinese with English abstract). U.S. DEPARTMENT OF ENERGY (USDOE), 2008, Methodology for Development of Geologic Storage Estimates for Carbon Dioxide, Appendix B: Carbon Sequestration Program, USDOE, 18 p, electronic document available at http://www. natl.doe.gov U.S. DEPARTMENT OF ENERGY (USDOE), 2010, Summary of the Methodology for Development of Geologic Storage Estimates for Carbon Dioxide, Appendix B: Carbon Sequestration Program, USDOE, pp. 136–152, electronic document available at http://www.natl.doe.gov VINCENT, C. J.; POULSEN, N. E.; RONGSHU, Z.; SHIFENG, D.; MINGYUAN, L.; AND GUOSHENG, D., 2011, Evaluation of carbon dioxide storage potential for the Bohai Basin, northeast China: International Journal of Greenhouse Gas Control, Vol. 5, pp. 598–603.
WEI, N.; LI, X.; WANG, Y.; DAHOWSKI, R. T.; DAVIDSON, C. L.; AND BROMHAL, G. S., 2013, A preliminary sub-basin scale evaluation framework of site suitability for onshore aquiferbased CO2 storage in China: International Journal of Greenhouse Gas Control, Vol. 12, pp. 231–246. WILDGUST, N., 2010, Storage capacity coefficients (2009–13): IEAGHG Greenhouse Issues, Vol. 96, p., 5. YANG, Y. AND XU, T., 2004, Hydrocarbon habitat of the offshore Bohai Basin, China: Marine and Petroleum Geology, Vol. 21, No. 6, pp. 691–708. ZENG, R.; SUN, S.; CHEN, D.; AND DUAN, Z., 2004, Decrease carbon dioxide emission into the atmosphere: Underground disposal of carbon dioxide: Bulletin of National Natural Science Foundation of China, Vol. 18, No. 4, pp. 196–200. (in Chinese with English abstract). ZHANG, Q., 2011, Annual Reports on Industrial Competitiveness of China (2012) No. 2: Social Sciences Academic Press, Beijing, pp. 41–47. ZHANG, X.; FAN, J.; AND WEI, Y., 2013, Technology roadmap study on carbon capture, utilization and storage in China: Energy Policy, Vol. 59, pp. 536–550. ZHENG, Z.; LARSON, E. D.; LI, Z.; LIU, G.; AND WILLIAMS, R. H., 2010, Near-term mega-scale CO2 capture and storage demonstration opportunities in China: Energy and Environmental Science, Vol. 3, No. 9, pp. 1153–1169. ZHOU, D., 2005, CO2 sequestration: A new issue of geology: Progress in Natural Science, Vol. 15, No. 7, pp. 782–787 (in Chinese). ZHOU, D.; ZHAO, Z.; LIAO, J.; AND SUN, Z., 2011, A preliminary assessment on CO2 storage capacity in the Pearl River Mouth Basin offshore Guangdong, China: International Journal of Greenhouse Gas Control, Vol. 5, pp. 308–317. ZHU, W.; MI, L.; AND GONG, Z., 2009, Petroleum Reservoir Forming and Prospecting on Offshore Bohai Sea: Science Press Beijing, p. 368. (in Chinese).
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 67–79
79
The Integration of Data Review, Remote Sensing and Ground Survey for a Regional-Level Karst Assessment ROBERT KENNETH DENTON JR.1 GeoConcepts-Engineering, Inc., 19955 Highland Vista Drive, Suite 170, Ashburn, VA 20147
ASHLEY HOGAN Geocapital-Engineering LLC, 4545 42nd St. N.W., Washington DC 20016
RONALD DREW THOMAS ECS Mid-Atlantic LLC, 14026 Thunderbolt Place, Chantilly, VA 20151
Key Terms: Karst, Sinkhole, Remote Sensing, Pipeline, Isopod, Habitat, Conservation Plan ABSTRACT Detailed karst terrain assessments require the identification and survey of surface features such as closed depressions, sinkholes, and cave entrances. Typically, surveys are carried out at sites encompassing several hundred acres or less; however, the traditional methods have proven impractical from a time and expense viewpoint for extensive, regional-level surveys in well-developed karst terrain. This subject survey covered a 76-mi (122.3-km) length of a natural gas transmission pipeline. The goal of the study was to assist the United States Fish and Wildlife Service in assessing habitat vulnerability for the Madison Cave Isopod, a federally protected threatened species found only in phreatic ground water of the Great Valley of Virginia and West Virginia. The survey used an integrated approach involving the evaluation of topographic maps, digital elevation models, shaded relief maps, satellite imagery, and historic aerial photographs to identify “concentrations” of karst features. Ground surveys were then undertaken by walking sections of the pipeline right-of-ways that occurred within karst concentrations, and documenting the features using GPS instrumentation. The karst survey encompassed 48,640 acres (19.2 hectares), and was the largest utilityassociated karst assessment ever conducted in the Commonwealth of Virginia. Approximately 216 closed depressions and 28 cave entrances were located, identified, and described based on their geology, physical appearance, and drainage characteristics. The use of an integrated approach significantly reduced the time and cost of the study. The study’s findings and 1
Corresponding author email: rdenton@geoconcepts-eng.com.
recommendations have been used to develop conservationbased avoidance and minimization measures intended to limit the impact to the species’ habitat. INTRODUCTION AND BACKGROUND The cirolanid isopod Antrolana lira Bowman 1964, commonly known as the Madison Cave Isopod (MCI), is a free-swimming crustacean native to the phreatic karst aquifer of the Great Valley of Virginia. First collected in 1958 by Thomas Barr of the University of Kentucky, the holotype and six paratype specimens were subsequently described by Thomas Bowman of the United States National Museum in 1964. The type locality was limited to two pools in Madison Cave (Figure 1) and a small pool in an adjacent cave named Steger’s Fissure, both located in the Grottoes area of Augusta County, Virginia (Bowman, 1964). Studies suggested that A. lira was threatened by human visitation to its only known habitat, and by mercury pollution of the nearby South River (Bolgiano, 1980; Collins, 1982). As a result, the taxon was proposed as a threatened species by the U.S. Fish and Wildlife Service (USFWS) on January 12, 1977 in the Federal Register (42FR 2507-2515). The threatened status of A. lira was finalized in a rule issued by the USFWS on October 4, 1982 (47FR 43699-43701). Subsequent investigations (Orndorff and Hobson, 2007) have extended the range of A. lira beyond its original type locality to a much larger area of the Great Valley (Figure 2). The MCI appears to have a specialized habitat preference for the high conductance, ionically saturated phreatic waters reposing in a specific group of carbonate rocks ranging in age from the Lower Cambrian through the Middle Ordovician, and solely within the Great Valley section of the Valley and Ridge Physiographic Province.
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 81–92
81
Denton, Hogan, and Thomas
Figure 1. The carbonate saturated water of the East Lake in Madison Cave, the type locality of A. lira. Note the calcite “rafts” floating on the surface of the water in the right center part of the photograph. (Plastic tubes are housings for pressure and conductivity sensors.)
It is notable that A. lira has not been found in the western portion of the Great Valley north of the North Fork of the Shenandoah River. Nevertheless, the carbonate rocks in which the preferred habitat is located are contiguous along the strike through this section. Thus, it is difficult to explain the absence of the MCI beyond the North Fork, especially considering that other consequent streams and rivers (i.e. streams that cross the bedding of the local rock), such as the Maury River in Rockbridge County, have not acted as a boundary to the MCI’s range. Nevertheless, the area north of the North Fork is still considered part of the taxon’s “suspected” range. The phreatic aquifer in which A. lira lives is connected to the surface through the epikarst zone. There are few perennial surface streams in karst areas, and much of the surface runoff is diverted underground through closed drainages. Recharge to the
82
Figure 2. Map showing the original type locality of A. lira near Grottoes, Augusta County, Virginia, and the inferred range of the taxon based on subsequent studies, with the three sections of the karst survey AOI, a one mile wide corridor paralleling the existing pipeline ROW.
subsurface can be rapid, specifically where water plunges into sinkholes with open throats, percolates through the bed of losing streams, or flows into cave entrances. In contrast, groundwater recharge can also be slow in places where water percolates down through the surface soils and infiltrates a diffuse network of cracks and fissures located at the soil/ bedrock interface. In many areas where the bedrock is mantled with clay-rich, cohesive soils, the infiltration would be expected to be extremely slow; however, periods of extended drought can open up deep fissures in the soil that can rapidly channel water directly to the underlying bedrock. Similarly, bedrock pinnacles that extend above the soil surface can channel significant quantities of water to the underlying epikarst, particularly if the surrounding soils dry out, shrink, and contract away from the surface of the pinnacles. All of these factors can affect the ionic concentration of the underlying phreatic aquifer and the habitability of the aquifer for the MCI.
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 81–92
Regional Karst Assessment
Figure 3. Idealized geological column of the three segments of the survey area.
Because of the myriad of variables associated with the habitat of A. lira, effective survey methods to determine the taxon’s abundance and suitable habitat have yet to be developed. Thus, the USFWS assumes the presence of the species if a project is located within a potential habitat area. Currently, potential habitat areas include the phreatic aquifers of any of the Cambrian and Ordovician carbonate rocks within the Great Valley. Accordingly, the USFWS recommends implementation of measures to minimize effects on A. lira and its habitat, among which are karst assessments intended to identify all surface karst features within an area of potential impact. GeoConcepts Engineering was contracted in June 2009 by The Conservation Fund on behalf of the USFWS, to provide a karst survey for the NiSource/ Columbia Gas Transmission Pipeline sections that extend through potential habitat areas of A. lira in the Great Valley. The following report summarizes the procedures and findings of that survey with comments on the utility of integrating the data review, remote sensing, and subsequent ground survey phases of the study.
GEOLOGICAL SETTING According to the Geological Map of Virginia (Virginia Division of Mineral Resources, 1993), the project is located entirely within Great Valley Section of the Valley and Ridge Physiographic Province. The Great Valley is a generally downwarped trough (synclinorium) of Paleozoic limestones, shales, and sandstones; it lies between the Blue Ridge Massif on the east and the Allegheny Mountains to the west. The Valley extends between the two mountain uplands from northeast to southwest, parallel with the strike of the bedrock. Specifically, the survey area has been mapped as underlain by a series of carbonate and clastic rocks ranging in age from Lower Cambrian to Middle Ordovician. A geologic column for the study area is presented as Figure 3. Karst Geology As in any region where soluble bedrock is present, a karst landform regime has developed in the Great Valley. All of the bedrock units present throughout
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 81–92
83
Denton, Hogan, and Thomas
the project area are capable of undergoing dissolution with the exception of the clastic rocks of the Middle and Upper Martinsburg Formation. Folding and faulting of the local carbonate rocks has opened up numerous fractures both parallel with the axis of the geologic structures, as well as perpendicular to them. Surface fractures and joints weather differentially, producing a pinnacled or “sawtooth” profile at the bedrock/soil interface (referred to as the epikarst zone). In contrast, rockenclosed fractures can be secondarily enlarged by the action of carbon dioxide charged groundwater, in some cases forming water-filled or air-filled conduits. As the regional terrain is “mature” karst, nearly all the fractures have undergone successive cycles of sediment filling and flushing. In areas such as the survey area, where there is little topographic relief and a relatively minimal groundwater gradient, the great majority of solution fissures are sediment-filled (Ford and Williams, 1989; Moore and Wade, 1989). METHODOLOGY For ease of explanation, the Area of Interest (AOI) for the assessment comprised an area of “covered lands” encompassing K mi (0.804 km) on either side of the NiSource Pipeline in sections that crossed potential or suspected MCI habitat. Within the covered lands was the pipeline Right of Way (ROW), which encompassed 80 ft (24.3 m) on either side of the pipeline. The entire length of the pipeline mapped as passing through known or suspected MCI habitat was divided into three discreet survey sections (Figure 2): 1. Northern Section (20 mi; 32.2 km) passing through Clarke, Warren, and Shenandoah counties; 2. Central Section (17 mi; 17.3 km) passing through Rockingham and Page counties; 3. Southern Section (39 mi; 62.8 km.) passing through Augusta and Rockbridge counties.
Data Review and Remote Sensing Phase (DR-RSP) Because the AOI encompassed approximately 76 sq mi (196 km2), it was beneficial to identify potential karst features remotely and/or by database review, and then confirm their actual presence in the field. This process significantly reduced the amount of time spent on location and on survey tasks. An inventory of known karst features located within the covered lands was reviewed from the following sources:
84
1. Caves of Virginia (Douglas, 1964); The Cave Database of the Virginia Speleological Survey (VSS; proprietary); 2. Maps of selected karst features (sinkholes, caves, and springs) available from the Virginia Division of Mines and Mineral Resources and the United States Geological Survey (USGS); 3. 2-ft and 4-ft contour interval maps for the counties containing AOIs (to determine the presence of surface karst features not included in the previously listed databases based on the presence of closed, descending contours or other suspect karst “fingerprint” features). Contour maps were generated from Triangular Irregular Network (TIN) data for each county available from the Virginia Geographic Information Network (VGIN); 4. Aerial photographs (both recent and historical) available from VGIN; 5. USGS 7.5-minute topographic quadrangles. In addition, readily available geological literature for bedrock and structural characteristics was reviewed (Virginia Division of Mineral Resources, 1993; Orndorff and Goggin, 1994; Gathright et al., 1996; Orndorff et al., 1999; Rader and Gathright, 2001; Campbell et al., 2006; 2007; and Southworth et al., 2009). The survey relied upon the highest resolution mapping available for the particular AOI being examined. The result of the DR-RSP was to identify areas of surface karst “concentrations” or clusters within the AOI that were then subject to closer scrutiny through field reconnaissance. Initial Assessment The initial assessment of karst concentration locations was determined using two studies undertaken by the Virginia Division of Mines and Mineral Resources (Hubbard, 1983; 1988). At the time of our investigation (2009) the Hubbard studies had not yet been rendered as digital files, and the published, paper maps lacked the resolution necessary to identify specific features smaller than several hundred feet in diameter. Nevertheless, the Hubbard maps would provide important clues to areas of significant karst feature concentrations (Hubbard, 2003). Cave location data was also used with the knowledge that many of the cave entrance coordinates were determined by manual approximation using 7.5-minute topographic quadrangles. Thus, reported cave entrance locations were expected to have significant errors of up to several hundred feet or more.
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 81–92
Regional Karst Assessment
Figure 4. (a) Excerpt of the USGS New Market, Virginia 7.5minute topographic quadrangle. Red polygons are sinkholes referenced from the Virginia Department of Mines, Minerals and Energy (DMME) datasets (Hubbard, 1983). (b) DEM dataset showing suspect closed depression (yellow dashed circle). This feature did not appear on the topographic quadrangle shown in Figure 4a and was not included in the reference DMME karst dataset.
Cartographic Data Review Four principal forms of cartographic data were used in the DR-RSP of the survey: 1. 2-ft (0.61 m), and 4-ft (1.22 m) contour maps of the various counties under study; 2. USGS 1/3 Arc-Second National Elevation Dataset (NED) digital elevation models (DEMs), Earth Resources Observation Systems (EROS) Web Map Service, and NED Shaded Relief Map series; 3. USGS topographic maps; 4. Aerial photographs, both recent and historical. The route of the pipeline ROW and covered lands was mapped using Esri’s ArcMapH geographic
Figure 5. (a) One-meter contour map derived from the DEM shown in Figure 4b of a previously unmapped feature in the New Market, Virginia Quadrangle (all contour elevations in meters). (b) Two-foot (0.61-m) contour map of the suspect closed depression shown in Figure 5a. Contours derived from TIN county data for the Commonwealth of Virginia.
information system (GIS) software, and then each section was scrutinized by toggling among each of the four data sources. Initial review consisted of examining topographic maps and comparing them to the areas of significant karst concentrations indicated in the Hubbard studies (Figure 4a). Then, the DEMs were examined to determine if smaller features not visible in the 7.5minute topographic quadrangles could be resolved (Figure 4b). Ultimately, final reliance was placed upon close-interval contour maps (Figure 5a and b), which were carefully examined for areas of closed descending contours. The contour maps were also rendered as hillside shaded relief maps for rapid scanning and evaluation. In this way, nearly the entire length of the covered lands could be scrutinized within a reasonable amount of time.
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 81–92
85
Denton, Hogan, and Thomas
Figure 6. (a) Aerial photograph of the location of suspect feature shown in Figures 4a to 5b. (b) Google Earth imagery of the suspect closed depression shown in Figures 4a to 6a derived from DEMs. This feature was subsequently confirmed by field observation as a sinkhole approximately 100 ft in diameter lying within a larger close depression approximately 300 ft in diameter and encompassing approximately 1.15 acres (0.404 hectares).
Once an area was identified as a closed depression, it was then compared to the other data sources, in particular high-resolution aerial photographs (Figure 6a and b). Many areas which appeared as putative sinkholes could be eliminated as closed depressions related to farm ponds or man-made excavations in this way, the usual signature being a berm on the downgradient side of the depression. Finally, suspect closed sinkholes were examined using the historical capabilities of the Google EarthTM application. This allowed us to compare closed depressions during relative “wet” and “dry” years, based on comparative climate data. Typical wet years used for this analysis were 2010, 2007, 2003 and 1997; dry years were 2005, 2002 and 2000. In many cases, structures that appeared as relatively nondescript vegetated closed depressions during dry years
86
exhibited the presence of standing water, or active resurgence flow, during wet years (Figure 7a and b). In cases where the depression exhibited an active outflow of water, the structure could be classified as an “estavelle”, i.e., a sinkhole that acts as a recharge point during dry seasons and a resurgence point during wet seasons. In some instances there was no outflow during wet seasons, but the depression clearly contained standing water, qualifying it as a “turloch” (i.e., dry lake). When a feature was identified as a sinkhole with reasonable certainty it was then outlined using the GIS software. Attributes were then assigned including county ID (a unique identifier created for this study), USGS quadrangle, bedrock geology, area (in square meters) and the northing and easting of the feature’s centroid. The entire set of closed depressions was then used to determine the specific areas of the pipeline ROW that would be examined closely during the field survey phase of the project. It should be noted that for the purposes of this study we considered any enclosed topographic basin with no sign of external drainage, regardless of its size, as a closed depression. However only features that could be verified as dolines (conical or funnelshaped features formed in the carbonate bedrock) or cover collapse sinkholes (in soil) were designated as “sinkholes”. We referred to any opening within a depression leading into the subsurface that could accept drainage as a “throat” rather than “ponor”. Definitions used herein were based the United States Environmental Protection Agency lexicon of cave and karst terminology (Field, 2002). Field Reconnaissance Phase (FRP) Upon completion of the DR-RSP for each section, we commenced the field reconnaissance phase. Field reconnaissance was based on two principal criteria: 1. Any suspect closed-depression or cave entrance that intercepted or appeared to receive drainage from the ROW was to be examined, regardless of whether there was a significant karst concentration in the area; 2. The ROW was to be examined by direct observation in any area where there was a significant concentration of karst features identified during the DR-RSP, regardless of whether they intercepted the ROW. In addition, suspect closed depressions were examined by direct observation if they occurred near, or could be reasonably accessed from, an intersecting roadway or residence. In all cases, the field staff
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 81–92
Regional Karst Assessment
entrance was located within a sinkhole, then the bounds of the sinkhole were mapped as a polygon. It should be noted that many of the cave entrance locations in our database were originally mapped using manual methods; thus, the coordinates for the entrances had a considerable degree of error. In a number of cases, the cave entrances had been obliterated by development, filled by the owner, or could not be located; however, the location of the now filled entrance was retained in the database. In all cases where a cave entrance was located, the entrance was mapped using GPS, and the cave data attribute table was edited using the corrected entrance locations. INTEGRATION OF THE DATA REVIEWREMOTE SENSING PHASE AND THE FIELD RECONNAISSANCE PHASE
Figures 7. (a) Area of suspected flooding sinkholes (estavelles) south of Forestville, Shenandoah County, Virginia; parapets of the sinkholes are within the dashed lines. Photograph (a) was taken during a relatively dry period in late December 2002; photograph (b) was taken during late April 2007, during a particularly wet spring season (Google Earth, 2011).
contacted the landowners to gain access to the property to inspect the suspect karst structures, particularly those that fell outside of the ROW but within the covered lands. Much to their credit, all the contacted landowners readily gave permission to access their property, allowing the field staff to examine nearly all of the closed depressions and caves located within the covered lands that were identified during the DR-RSP. The actual field survey was performed by walking the ROW, observing and noting any previously identified surface karst features such as sinkholes, subsidence areas, or cave entrances. Any feature noted within the examined area was recorded using a handheld sub-meter accuracy GPS unit. Cave entrances were recorded as points, and sinkholes were recorded as vertex-bounded polygons. If a cave
The DR-RSP of the survey commenced in midAugust 2009, with the supporting field work commencing during September 2009. As each section of the survey area was examined during the DR-RSP, field crews would be dispatched to examine any suspect karst feature concentrations, and these data were used to correct and amend the findings of the DR-RSP observations. The entire field verification of the DR-RSP was completed in less than 2 months, with field work concluding by late October 2009. The final report summarizing the survey results was delivered in December 2009. The integration of the two methods allowed for a rapid turnaround in the delivery of a final report to our client. In the process, several important deficiencies of each method (if they had been performed individually) were revealed: 1. Remote sensing and data review, especially the use of 2-ft contour maps and DEMs, often misidentified man-made features (e.g., farm ponds, silage pits) as natural closed depressions. The integration of field verification with remote sensing allowed for the refinement of the DR-RSP data. 2. Field reconnaissance teams, without the benefit of the DR-RSP generated maps, would have been compelled to walk the entire length of the survey’s extent, including the 1/2-mi (0.80-km) wide covered land zone. We estimate this would have required at least 6 to 8 months of field work. 3. Some of the karst features that were surveyed did not appear on USGS topographic maps; however, they were observed and noted during the DR-RSP. This allowed the field crews to focus their efforts on suspect structures with greater accuracy.
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 81–92
87
Denton, Hogan, and Thomas Table 1. Summary of areal extent of geological formations and karst features. Percentage of Closed Depressions in Areal Extent Geology in Covered Number of Closed Geology per square Number of (mi2/km2) Lands Depressions (N) mile Caves (N)
Geology Martinsburg & Oranda Formations Middle Ordovician Limestones Beekmantown Group Conococheague Formation Elbrook Formation Waynesboro Formation & Tomstown Dolomite Shady Dolomite
13.8/35.74
6.1
4
0.3
0
15.16/39.26 27/69.93 27.02/69.98 24.63/63.79
6.7 11.9 11.9 10.9
88 71 18 14
5.8 2.6 0.7 0.6
13 4 5 6
14.15/36.65 4.7/12.17
6.2 2.1
10 11
0.7 2.3
0 0
4. There were instances where the karst features were actually too large or so obscured by vegetation that, without the DR-RSP-generated maps, they would have been bypassed by the field teams.
SURVEY RESULTS It is difficult to concisely present the findings of a survey of this size; however, patterns regarding the distribution of the various karst features that could be potentially affected by the pipeline can be summarized as follows. The areal distribution of geological formations throughout the survey area is shown in Table 1. As can be seen from these data, the extent of the potentially karst-forming units within the covered lands is distributed unevenly, with the majority (35%) of areal extent being underlain by the Elbrook, Conococheague, and Beekmantown units. Interestingly, the distribution of the closed depressions/ sinkholes did not follow the same trend (Table 1). Of the 227 features identified in the DR-RSP of the project, 216 were natural karst features, while the remaining 11 were man-made depressions, primarily farm ponds. Of the 33 caves documented in the various databases, 27 entrances were observed and located during the study, with the remaining entrances assumed to have been filled or otherwise destroyed. The mapped closed depressions/sinkholes demonstrated a clustering of surface-expressed karst features in the Middle Ordovician Limestones, with nearly twice the number per square mile than in the Beekmantown Group (Figure 8a, b, and c). These two units showed the most surface karst development within the study area. In general, the majority of the sinkholes overall were bowl-shaped, soil-bottomed and vegetated, with gradual slopes leading down from their well-defined edge or “parapet”, but lacked any
88
Caves in Geology per Square Mile 0 0.9 0.1 0.2 0.2 0 0
obvious throat or opening into the subsurface. Sinkholes of this type were especially common in the northern and central sections of the survey (i.e., the “Lower” Great Valley), typically considered a mature karst terrain. In contrast, sinkholes with open throats were more prevalent in the southern section of the survey. A total of eight sinkholes with open and/or debris-clogged (but readily apparent) throats were observed and documented exclusively in the Rockbridge County section of the survey. There was an interesting pattern regarding the number of closed depressions/sinkholes detected in the DR-RSP phase of the project versus the features that had been mapped for the same areas in prior studies. Of the 113 features observed and documented in the northern section there were at least 73 features that had not been documented in any prior study. In contrast, of the 22 features in the central section, and 91 in the southern section, there were 17 and 8 features, respectively, that had not been documented previously. The reason so many features had been previously documented in the southern section may be because the closed depressions/sinkholes in this area (Rockbridge County) tended to be better developed; i.e., deeper and more conical in shape (see Figure 9). It should be noted that one of the reasons for the low number of surface features and caves we documented in the Lower Cambrian Waynesboro Formation and Tomstown/Shady Dolomite was that these units were only present in the current survey’s study area near the eastern edge of the Great Valley along the western base of the Blue Ridge. This area is characterized by two very different types of surface karst features: sinkholes and caves located in widely scattered hills that are the erosional remnants of the original limestone upland; and broad shallow dishshaped alluvium-filled sinkholes in the gently rolling terrain between the hills. The alluvium-filled sinkholes are particularly common in the easternmost portion of the Great Valley, where outwash from the eroding
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 81–92
Regional Karst Assessment
Figure 9. Well-developed conical sinkholes southwest of Lexington in Rockbridge County, Virginia, representative of the typical karst landform in this section of the Shenandoah Valley. The majority of open-throat sinkholes were found in this section of the current survey.
Blue Ridge Mountains provided the gravelly sand overburden that filled previously existing karst depressions. The fact that these sinkholes formed prior to the deposition of the alluvium is supported by the discovery of fossilized pollen dating from the Early Tertiary to Late Cretaceous in sediment
Figure 10. Broad, shallow sinkhole developed in the alluvial sediments in eastern Page County, Virginia near the western pediment of the Blue Ridge Mountains.
Figure 8. (a) Karst feature locations (green centroids) documented during the current study in the northern section of the route overlaid on the regional karst-forming carbonate bedrock. (b) Central section, (c) Southern section. (Geological Abbreviation Key: s 5 Shady Dolomite; s 5 Rome Formation; wb 5
r Waynesboro Formation; wbt 5 Waynesboro & Tomstown Formations; e 5 Elbrook Formation; O co 5 Conococheague Formation; Ob 5 Beekmantown Group; Oeln 5 New Market, Lincolnshire and Edinburg Formations; DSu 5 Silurian-Devonian Carbonates, undivided.)
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 81–92
89
Denton, Hogan, and Thomas
Figure 11. Entrance to Buzz Hole-GeoConcepts Cave located in Rockbridge County, Virginia. Previously unexplored, this cave was explored and partially mapped as part of the current study. It is being monitored during high water periods for the intrusion of the phreatic aquifer as possible habitat of the MCI.
preserved in the sinkholes below the overburden (Pierce, 1965; Tschudy, 1965). Many of the features we mapped in the eastern part of the survey within Page and Augusta counties were far too shallow to be readily apparent using 10- or 20-ft (3.04- or 6.09-m) contour interval topographic maps, aerial photographs, or the available 1/3 arc-second DEMs. However they were detected by examination of 2-ft (0.61-m) and 4-ft (1.22-m) contour interval maps
generated from TIN data. Nevertheless, we assume that only the largest buried subsurface karst features were expressed as surface depressions in the alluvium (Figure 10), and that smaller features in the underlying carbonates were hidden. Thus, the burial of the underlying karst probably accounted for the lower percentage of surface features in the eastern valley. Similar to the data for closed depression density, the highest occurrence of caves (i.e., openings into the subsurface large enough to allow a human to enter) per square mile was noted in the Middle Ordovician Limestones; however, caves in the Beekmantown Group were clearly outnumbered by the caves within the Conococheague and Elbrook Formations. This was an intriguing pattern, especially considering that the Conococheague and Elbrook Formations showed relatively few surface karst features, compared to the overlying Ordovician units. One of the ancillary benefits of our study was to provide the VSS with accurate cave entrance coordinate data for the mapped caves. Many of these cave entrances had been located long before the advent of handheld GPS survey equipment, and the entrance locations were grossly inaccurate. We provided the VSS with accurate locations for more than 20 caves within the survey area, as well as survey information for a “new” cave and the relocation of a “lost” cave. While surveying in the Lexington area of Rockbridge County our survey team noted that the pipeline ran along the upper parapet of a steeply
Figure 12. Survey map of Buzz Sink (GeoConcepts) Cave.
90
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 81–92
Regional Karst Assessment
of a triangular, 6-ft (1.82 m) deep sinkhole, partially filled with leaves (Figure 13). The actual entrance passage to Madden’s Cave was found to be clogged with dirt and plant debris that had washed in from above. It did not appear that anyone had entered the cave for many years. Madden’s Cave is approximately 100 ft (30.4 m) in depth and 275 ft (83.8 m) in length, and has been completely mapped. It is considered “very significant” by the VSS and the Virginia Cave Board because it is the type locality for two unique species: a cave beetle and a pseudoscorpion. There is also a record of the Indiana Bat (Myotis sodalis), a federally listed Rare, Threatened and Endangered Species (RTES) for this cave. Madden’s Cave is not expected to be affected by the pipeline because of its upgradient location and distance from the ROW. Figure 13. Entrance County, Virginia.
to
Madden’s
Cave
in
Shenandoah
ACKNOWLEDGMENTS sloping sinkhole near a newly constructed home. At the southeastern terminus of the depression they discovered a steep-walled sinkhole, with an open throat at its base that appeared large enough to permit human entry (Figure 11). This structure was reported to Bob Thren, chairman of the Rockbridge County section of the VSS, who was aware of its existence, but admitted that it had never been investigated or mapped. The cave was registered with the VSS and dubbed “Buzz Sink (GeoConcepts) Cave” in honor of the landowner’s nickname and the exploratory crew from GeoConcepts. A team from GeoConcepts, the Virginia Department of Conservation and Recreation–Virginia Natural Heritage Program, and the VSS returned to the newly discovered cave on November 3 and again on December 15, 2009, to determine its extent (Figure 12). The cave was primarily vertical and several hundred feet of passage were mapped to a depth of approximately 80 vertical ft (24.3 m). This was particularly intriguing, because caves of this type (i.e., vertical or semi-vertical caves that descend rapidly to the phreatic water table) are where populations of the MCI have most often been found. There is evidence that the lower levels of the cave are completely flooded during seasonal water table high stands. The cave is still under exploration and is planned to be monitored periodically for the presence of the MCI. To the north in Shenandoah County, in the heavily forested bluffs above the North Fork of the Shenandoah River, the survey relocated the entrance to Madden’s Cave. Madden’s had been considered a lost cave because the entrance could not be located when the VSS resurveyed Shenandoah County in the 1980s and 1990s. The entrance was located at the base
The authors would like to thank NiSource Gas Transmission and Storage for allowing the survey data compiled for the Multi-Species Habitat Conservation Plan to be presented in this report. We would also like to thank Wil Orndorff of the Virginia Department of Conservation and Recreation–Natural Heritage Program, for providing invaluable data, guidance and assistance. Mapping of Buzz Sink (GeoConcepts) Cave could not have been completed without the help of Bob Thren, Virginia Speleological Survey–Rockbridge County Chairman and Nathan Farrar, Virginia Speleological Survey–Page County Chairman. We would also would like to thank Cody Sheaffer and Sam Consolvo (GeoConcepts) for assistance in the preparation of various tables and figures. Finally, we want to thank the management of GeoConcepts Engineering for their endless patience and unflagging support of this work. REFERENCES BOLGIANO, R. W., 1980, Mercury Contamination of the South, South Fork Shenandoah, and Shenandoah Rivers: Virginia State Water Control Board Basic Data Bulletin 47. BOWMAN, T. E., 1964, Antrolana lira, a new genus and species of troglobitic cirolanid isopod from Madison Cave, Virginia: International Journal Speleology Vol. 1, pp. 229–236. CAMPBELL, E. V.; EVANS, N. H.; SPENCER, E. W.; AND WILKES, G. P., 2007, Geologic Map of Rockbridge County, Virginia: Department of Mines, Minerals and Energy, Division of Mineral Resources Publication 170, Plate 1. CAMPBELL, E. V.; WILLIAMS, S. T.; DUNCAN, I. J.; HIBBITS, H. A.; FLOYD, J. M.; REIS, J. S.; AND WILKES, G. P., 2006, Interstate 81 Corridor Digital Geologic Compilation: Virginia Department Mines, Minerals, and Energy—Division of Mineral Resources Open File Report 06-01. COLLINS, T. L., 1982, An Ecological Study of the Troglobitic Cirolanid Isopod, Antrolana Lira Bowman, from Madisons Saltpetre Cave And Stegers Fissure, Augusta Co., Virginia:
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 81–92
91
Denton, Hogan, and Thomas Unpublished MS Thesis, Old Dominion University, Norfolk, VA, 41 p. DOUGLAS, H. H., 1964, Caves of Virginia: Virginia Cave Survey, Falls Church, VA, 761 p. FIELD, M., 2002, A Lexicon of Cave and Karst Terminology with Special Reference to Environmental Karst Hydrology: United States Environmental Protection Agency, EPA/600/R-02/003. FORD, D. C. AND WILLIAMS, P. W., 1989, Karst Geomorphology and Hydrology: University Press, Cambridge, U.K. GAITHRIGHT, T. M. II; MCDOWELL, R. C.; ORNDORFF, R. C.; AND RADER, E. K., 1996, Geologic Map of Clarke, Frederick, Page, Shenandoah, and Warren Counties, Virginia: Lord Fairfax Planning District, U.S. Geological Survey Publication 143. GOOGLE EARTH, 2011, “Timberville Sinkholes” 38.679224u N and 78.750552u W. Imagery Dates: December 30, 2002 and April 29, 2007. Accessed: September 26, 2011. HUBBARD, D., Jr., 1983, Selected Karst Features of the Northern Valley and Ridge Province, Virginia: Virginia Division of Mineral Resources Publication 44. HUBBARD, D., Jr., 1988, Selected Karst Features of the Central Valley and Ridge Province, Virginia: Virginia Division of Mineral Resources Publication 83. HUBBARD, D. A., Jr., 2003, Use of regional sinkhole mapping for sinkhole susceptibility maps, In Beck, B. F. (ed.), ASCE Geotechnical Special Publication No. 122, pp. 61–71. MOORE, C. H. AND WADE, W. J., 2013, Carbonate Reservoirs: Porosity and Diagenesis in a Sequence Stratigraphic Framework, 2nd Ed., Elsevier. ORNDORFF, R. C.; EPSTEIN, J. B.; AND MCDOWELL, R. C., 1999, Geologic Map of the Middletown Quadrangle, Frederick, Shenandoah, and Warren Counties, Virginia: United States Geological Survey, Geologic Quadrangle Map GQ-1803. ORNDORFF, R. C. AND GOGGIN, K. E., 1994, Sinkholes and KarstRelated Features of the Shenandoah Valley in the Winchester 309 x 609 Quadrangle, Virginia and West Virginia: U. S.
92
Geological Survey Miscellaneous Field Studies Map, No. 2262. ORNDORFF, W. D. AND HOBSON, C. S., 2007, Status Survey for the Madison Cave Isopod (Antrolana lira) in Virginia, 2005–2007: Natural Heritage Technical Report 07-11. Virginia Department Conservation Recreation, Division Natural Resources, Richmond, VA, 17 pp. PIERCE, K. L., 1965, Geomorphic significance of a Cretaceous deposit in the Great Valley of southern Pennsylvania: U. S. Geological Survey Professional Paper 525-C, pp. C152 –C156. RADER, E. K. AND GATHRIGHT, T. M., 2001, Geologic Map of the Augusta, Page, and Rockingham Counties Portion of the Charlottesville 30 x 60 Minute Quadrangle: Virginia Department Mines, Minerals, and Energy—Division of Mineral Resources Publication 159 SOUTHWORTH, S.; ALEINIKOFF, J. N.; BAILEY, C. M.; BURTON, W. C.; CRIDER, E. A.; HACKLEY, P. C.; MOOT, J. P.; AND TOLLO, R. P., 2009, Geologic Map of the Shenandoah National Park Region, Virginia: U. S. Geological Survey Open File Report 2009–1153. “THE CAVE DATABASE OF THE VIRGINIA SPELEOLOGICAL SURVEY,” Virginia Speleological Survey Proprietary Database, accessed October, 2009. THE NATIONAL MAP (Digital Elevation Models, 7.5-Minute Topographic Quadrangles) Courtesy of the United States Geological Survey. Retrieved from http://viewer.national map.gov/basic/ TSCHUDY, R. H., 1965, An Upper Cretaceous Deposit in the Appalachian Mountains, in Geological Survey Research 1965: U.S. Geological Survey Professional Paper 525-B, pp. B-64–B-68. VIRGINIA DIVISION OF MINERAL RESOURCES, 1993, Geologic Map of Virginia: scale 1:500,000. VIRGINIA GEOGRAPHIC INFORMATION NETWORK, 2009, Triangular Irregular Network Dataset for Virgina Counties. Retrieved from http://vgin.maps.arcgis.com/home/
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 81–92