Issuu on Google+


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

Ernest Orlando Lawrence Berkeley National Laboratory is an equal opportunity employer.


Earth Sicences Division

Annual Report 1998-1999

A Perspective from the Division Director

1

Resource

3

TABLE

Ecology

5

OF

Geophysics and Geomechanics

7

CONTENTS

Geochemistry

9

Departments

Hydrology and Reservoir Dynamics

11

Center

13

Research

for

Environmental

Biotechnology

Programs

Fundamental

and

15

Exploratory

Research

17

Identifying Limits of Film Flow in Unsaturated Fractures Tetsu K. Tokunaga and Jiamin Wan

19

Microbubble Generation, Stability and Transport: A Potential Subsurface Remediation Technique Fred Gadelle, Tetsu K. Tokunaga and Jiamin Wan

21

Particle Motion in Film Flow Srinivas Veerapaneni, Jiamin Wan and Tetsu K. Tokunaga

23

Glass Casts of Rock Fracture Surfaces: A New Tool for Studying Flow and Transport Jiamin Wan, Tetsu K. Tokunaga, Thomas R. Orr, Jim O’Neill and Robert W. Connors

25

A Model for Non-Isothermal Multiphase Multi-Species Reactive Chemical Transport In Porous and Fracture Media Tianfu Xu, Karsten Pruess, Eric Sonnenthal, Nicolas Spycher and George Brimhall

27

The Center for Isotope Geochemistry Donald J. DePaolo

29

Molecular Modeling of Clay Mineral Surface Geochemistry: Hydrated Cesium-Smectites Rebecca Sutton, Garrison Sposito, Sung-Ho Park and Jeffery A. Greathouse

31

Isotopic Effects in Dual-Porosity Fluid-Rock Systems Donald J. DePaolo

33

Helium Isotopes in Long Valley Basalts: Implications for Future Volcanic Activity Allen Dodson, Donald DePaolo and B. Mack Kennedy

35

Isotope Constraints on Fluid Sources: San Andreas Fault System, California E. Pili, B.M. Kennedy, M.S. Conrad and D.L. Shuster

37

The Center for Computational Seismology Thomas V. McEvilly, Ernest L. Majer and Lane R. Johnson

39

Log-Permeability Estimation Using Multiple Geophysical Data Sets Within a Bayesian Framework 41 Susan S. Hubbard, Yoram Rubin and Ernie Majer Modeling the Observed Controlled-Source Waveform Changes at Parkfield, California Valeri A. Korneev, Tom V. McEvilly and Eleni D. Karageorgi i

43


ESD

Table of Contents

Overlap Domain Decomposition Technique for Modeling Wave Propagation Jianli Fan, Kurt T. Nihei, Larry R. Myer and James W. Rector

45

Frictional Effects on the Volumetric Strain of Sandstone Kurt T. Nihei, L.B. Hilbert Jr., Nigel G.W. Cook, S. Nakagawa and Larry R. Myer

47

Stochastic GILD Coupled Modeling and Inversion Ganquan Xie and Jianhua Li

49

Nuclear

Waste

51

Ambient Hydrologic Testing in the Exploratory Studies Facility: Niche Study Robert C. Trautz and Joseph S.Y. Wang

53

Spatial and Temporal Flow Variability in the Paintbrush Nonwelded Tuff Rohit Salve and Curtis M. Oldenburg

55

Field Investigations of Fracture Flow in Welded Tuffs Rohit Salve and Joseph S.Y. Wang

57

A Probe for Measuring Saturation Changes in Rock Rohit Salve, Tetsu K. Tokunaga and Joseph S.Y. Wang

59

The Restricted Interval Guelph Permeameter Barry Freifeld and Curt Oldenburg

61

Laboratory Measurement of Water Imbibition Into Low-Permeability Welded Tuff Qinhong Hu, Peter Persoff and Joseph S.Y. Wang

63

High Resolution Studies of Water Seepage in Unsaturated,Heterogeneous Rock Fractures Tai-Sheng Liou and Karsten Pruess

65

An Active Fracture Model for Unsaturated Flow and Transport Hui-Hai Liu, Gudmundur S. Bodvarsson and Christine Doughty

67

DCPT: A Dual-Continuum Random Walk Particle Tracker for Transport Lehua Pan, Hui-Hai Liu, Mark Cushey and Gudmundur S. Bodvarsson

69

Continual Development of the UZ Model for Yucca Mountain, Nevada Yu-Shu Wu, Jennifer Hinds, C. Frederick Ahlers, Hui-Hai Liu, Lehua Pan, Anne C. Ritcey, Mark Cushey, Charles Haukwa, Eric Sonnenthal and Gudmundur S. Bodvarsson

71

Development of WinGrider: An Interactive Grid Generator for TOUGH2 Lehua Pan, Charles Haukwa, Yu-Shu Wu and Gudmundur S. Bodvarsson

73

T2R3D – A TOUGH2 Code for Tracer Transport in Heterogeneous Media Yu-Shu Wu and Karsten Pruess

75

Laboratory Studies on Heat-Driven Multiphase Flows in Rock Fractures Timothy J. Kneafsey, Karsten Pruess and Jeffery J. Roberts

77

Thermal-Hydrological Modeling of the Drift Scale Test at Yucca Mountain, Nevada Sumit Mukhopadhyay and Yvonne W. Tsang

79

Evolution of CO2 From Heated Rock at Yucca Mountain Mark Conrad and Eric Sonnenthal

81

Prediction and Analysis of Coupled Processes in the Drift Scale Thermal Test Eric Sonnenthal, Nicolas Spycher, John Apps, Mark Conrad and Ardyth Simmons Testing of a Coupled THM Model for Unsaturated Media Against Laboratory Experiments Jonny Rutqvist, Jahan Noorishad and Chin-Fu Tsang ii

83

85


Earth Sciences Division

Annual Report 1998-1999

Coupled Analysis of a THM Field Experiment in an Unsaturated Buffer-Rock System Jonny Rutqvist, Jahan Noorishad and Chin-Fu Tsang Full-Scale Tomographic Seismic Imaging of the Potential Repository Horizon at Yucca Mountain, Nevada Roland Gritto, Thomas M. Daley,Valerie A. Korneev, Mark A. Feighner, Ernest L. Majer and John E. Peterson Investigation of Geologic Water Storage Near Cuzco, Peru Jerry P. Fairley

Energy

Resources

87

89

91

93

Development of Single-Well Seismic Imaging Thomas M. Daley and Ernest L. Majer

95

Vertical Seismic Profiling at the Rye Patch Geothermal Field, Nevada Mark A. Feighner, Thomas M. Daley and Ernest L. Majer

97

Characterization of Fractured Geothermal Reservoirs Using Inverse Modeling Stefan Finsterle, Grimur Bjรถrnsson and Karsten Pruess

99

Comparison of the Two Most Important Salton Trough Geothermal Fields Marcelo J. Lippmann, Alfred H. Truesdell and George A. Frye

101

Surveillance and Supervisory Control of Waterflood Tadeusz W. Patzek and Asoke De

103

Control of Fluid Injection into a Low-Permeability Rock Tadeusz W. Patzek and Dmitriy B. Silin

105

Dynamic Reservoir Characterization Through the Use of Surface Expression Data Don Vasco and Kenzi Karasaki

107

Natural Geochemical Tracers for Injectate Fluids at Dixie Valley, Nevada B. Mack Kennedy, Cathy Janik, Richard Benoit and David L. Shuster

109

TVD Schemes for Phase-Front Propagation in Geothermal Reservoirs Curtis M. Oldenburg and Karsten Pruess

111

Fluctuations in Elastic Waves Due to Random Scattering From Inclusions Valeri Korneev and Lane Johnson

113

Network Modeling of Multiphase Flow Processes in Rock Pingan Hunag and Larry R. Myer

115

Environmental

Remediation

Technology

117

Experimental Studies of Ferrofluids for Subsurface Applications Sharon E. Borglin, George J. Moridis, Curtis M. Oldenburg and Alex Becker

119

Laboratory Monitoring of VOC Biodegradation in Unsaturated Fractured Rock Jil Geller, Simon Davis, Mark Conrad, Hoi-Ying Holman, Jennie Hunter-Cevera and Karsten Pruess

121

Removal of Uranium (VI) from Contaminated Sediments by Surfactants Fred Gadelle, Jiamin Wan and Tetsu K. Tokunaga

123

Multi-Scale Investigations of Liquid Flow in the Vadose Zone of Fractured Basalt Boris Faybishenko, Paul A. Witherspoon, Christine Doughty, T.R. Wood, R.K. Podgorney and Jil T. Geller

125

iii


ESD

Table of Contents

Poland Petroleum Refinery Sludge Lagoon Biopile Demonstration Terry C. Hazen

127

Aerobic Bioremediation of a Municipal Solid Waste Landfill Terry C. Hazen

129

New Software Tool for Visualizing Fracture Data From Oriented Cores Janet S. Jacobsen

131

Enhanced Data Analysis for the Vadose Zone Monitoring System Curtis M. Oldenburg, April L. James and Peter T. Zawislanski

133

Numerical Simulation of Ferrofluid Flow Curtis M. Oldenburg, Sharon E. Borglin and George J. Moridis

135

Advanced Detection of Exposures and Biological Responses to Organic Toxins Hoi-Ying N. Holman, Regine Goth-Goldstein, Michael C. Martin, Marion Russell and Wayne R. McKinney

137

An Algal-Bacterial Treatment System for Drainage Selenium Removal Nigel Quinn, Tryg J. Lundquist, Bailey F. Green, W.J. Oswald, Terrance Leighton, Robert B. Buchanan and Max Zarate

139

Characterization of VOC Biodegradation on Rock Surfaces by SR FTIR Imaging Hoi-Ying N. Holman, Jil Geller, Jennie C. Hunter-Cevera and Karsten Pruess

141

Stimulation of Methyl Tert-butyl Ether Biodegradation Using a Co-Substrate Approach William T. Stringfellow

143

Climate Variability and Carbon Management

145

Regional Climate Simulation for the Western United States Using the RCSM Jinwon Kim and Norman L. Miller

147

The Regional Climate Center Norman L. Miller and Jinwon Kim

149

Climate Change and Wildfire Severity in California Margaret S. Torn, Evan Mills and J. Fried

151

A Simple Physical Model of the Global Hydrologic Cycle: Implications for Paleothermometry Melissa Hendricks, Don DePaolo and Ron Cohen

153

Publications

155

Staff

169

iv


Earth Sciences Division

Annual Report 1998-1999

MEETING THE CHALLENGES OF THE

21ST CENTURY

A PERSPECTIVE FROM THE DIVISION DIRECTOR SALLY M. BENSON

(510) 486-5875 EMAIL: SMBENSON@LBL.GOV

Livermore National Laboratory, to create the new DOE Center for Research in Ocean Carbon Sequestration. • A re s e rvoir simulator for predicting production from methane hydrate reservoirs has been developed. • ESD researchers have played a leading role in defining the national research agenda for sequestration of CO2 in geologic formations. This has also been an exciting year for many of our ongoing research programs. Our core fundamental research program in flow and transport, geochemistry and subsurface imaging continues to push the state of the art and provide the foundation for new insights, approaches and technologies that are relevant to all of our mission-focused programs. Furthering our understanding and quantification of vadose zone flow in fractured and heterogeneous rocks continues to be a central goal of our fundamental research program. This year’s highlights include new insights about the episodic and chaotic nature of vadose zone flow in fracture media, better models of colloid transport and a greater understanding of the range of conditions over which film flow on fracture faces may occur. Our fundamental research program in geochemistry, with primary emphasis on isotopes and reactive transport modeling, has broadened to include experimental and theoretical research in molecular geochemistry. Development of TOUGHREACT, a version of the TOUGH2 family of codes that simulates reactive subsurface transport, represents a major new research capability for our group. Our subsurface imaging program focuses on devel-

A

s we approach the millennium, the importance of the earth sciences in shaping our future grows larger than ever. Not only must we continue to address the issues that were so pressing during the past century—energy and water resources, ground water and surface water pollution and natural hazards— we must face the challenges of the coming century. The potential for global climate change and the need for safe and secure nuclear waste management have emerged at the forefront of the concerns faced by our nation and others around the world. As one of the U.S. Department of Energy’s national laboratories, our mission is to provide a robust scientific foundation to better understand and identify solutions to these pressing issues. This year has been an exciting one.We have launched several new initiatives to prepare ourselves for the future, and we have made great strides in our ongoing research programs. In this report we present examples of both new and continuing research. While it is not a complete accounting, it is representative of the nature and breadth of our research effort. New programs in regional climate modeling, carbon cycling in the oceans and terrestrial biosphere, methane hydrates and carbon sequestration are the cornerstones of a major new research t h rust related to global env i ronmental ch a n ge. Significant progress has already been made in each of the following areas: • The Regional Climate System Model (RCSM) has been used to predict California stream flow in a double CO2 environment. • We have been selected, in part n e rship with Law re n c e 1


ESD

Challenges of the 21st Centruy: A Perspective from the Division Director

oping higher resolution imaging methods for geologic structures, transport properties and fluid composition. This year an award-winning project to construct permeability maps from high resolution cross-borehole seismic tomography was completed, providing the launching point for a new initiative in joint inversion of hydrologic and geophysical imaging data. Scientists working in our Nuclear Waste Program continue to provide leadership in identifying and addressing key research questions related to flow and transport in the unsaturated zone at Yucca Mountain—a site in Nevada that is being investigated as a potential repository for spent nuclear fuel from commercial power plants. As our understanding of the natural flow system improves, we are focusing on the interaction between the waste package and the natural environment. One key question is: How much water will actually seep into the drift and contact the metal canisters that enclose the spent nuclear fuel rods? We have conducted a set of field experiments and theoretical studies which show that only a small fraction of the water percolating through the mountain will seep into the drifts. Some experiments even suggest that there may be a threshold below which no water will seep into the drifts. If this is true, canister corrosion rates and radionuclide fluxes will be lower than curr ently expected. Over the next several years researchers will address these and other issues about the interaction between the engineered and natural systems. Scientists and engineers in our Energy Resources Program work closely with the fossil fuel and geothermal industries to develop more effective methods for locating and producing energy resources. This year we have made a major step toward development of new seismic methods for characterizing fractured reservoirs, a problem of utmost importance to the oil, gas and geothermal industries.At the Conoco Oil Company Test Site in Oklahoma, we completed the first successful field demonstration of high resolution single-well seismic imaging of fluid-filled fractures.These same tests provided field validation of new theoretical models for the seismic response of fractures developed as part of our core fundamental research program. Looking closer to home, researchers in the Environmental Remediation Technology Program continue to lead the soil and groundwater cleanup activities taking place at our own site. Since the inception of Lawrence Berkeley Laboratory in the mid1930s, the site has been used as a major research facility. With that comes the pollution typically associated with light industrial operations. Several plumes of mixed solvents, hydrocarbon fuels, freon and tritium are located across the site. The complexity of the geologic environment here makes locating the source of these plumes, tracking their migration and designing effective

remediation schemes a major challenge. Persistence, a good deal of hydrogeologic sleuthing and a pragmatic approach to remediation have resulted in significant progress towards resolving site cleanup issues. From this and many related experiences we have come to believe that detailed source characterization is the necessary foundation for all successful and cost-effective groundwater remediation activities. Circumventing this time-consuming and often costly step will be paid for many times over. Finally, one of the major highlights of the year was an International Symposium on the Dynamics of Flow in Fractured Rocks held in honor of Dr. Paul Witherspoon’s 80th birthday. Dr. Witherspoon, the founding director of the Earth Sciences Division, continues to play an active role in many of our activities. In honor of his birthday, at which he gave a keynote speech on “Thirty Years of Fracture Flow Research at Berkeley,” more than 250 participants from around the world joined us for an outstanding technical workshop.

Organization of this Report This report is divided into five sections that correspond to the major research programs in the Division: • Fundamental and Exploratory Research • Nuclear Waste • Energy Resources • Environmental Remediation Technology • Climate Variability and Carbon Management These pro grams draw from the each our disciplinary departments: E c o l o gy, Geophysics and Geomechanics, Geochemistry, Hydrogeology and Reservoir Dynamics, and the Center for E nv i ronmental Biotech n o l o gy. S h o rt descriptions of these departments, along with a listing of our personnel are provided as introductory materi a l . A list of publications for the period June 1997 to June 1999 are provided in Appendix A.

Acknowledgments The Earth Sciences Division consists of 199 scientists and engineers. We gratefully acknowledge the support of our many sponsors in the Department of Energy. We also appreciate the support we receive from other federal agencies: the U.S. Bureau of Reclamation, the U.S. Department of Defense and the U.S. Environmental Protection Agency. These activities support and enrich our programs by broadening the range of applications of our research.We also appreciate support from our industrial collaborators and colleagues, both financial and through partnerships that bring ideas, data and experience to the division.

2


Ecology Geophysics and Geomechanics Geochemistry Hydrology and Reservoir Dynamics Center for Environmental Bioremediation

Departments


Earth Sciences Division

Annual Report 1998-1999

This page is intentionally left blank

HTTP://WWW-ESD.LBL.GOV 4


Earth Sciences Division

Annual Report 1998-1999

ECOLOGY DEPARTMENT

TERRY C. HAZEN

CONTACT: (510) 486-6223 TCHAZEN@LBL.GOV

The current expertise of the Ecology Department (ED) is focused on bioremediation, ecosystem engineering, microbial ecology and environmental risk assessment. Department scientists support re s e a rch in the Env i ronmental Remediation Technology Program, Center for Environmental Biotechnology, Energy Resources Program, Nuclear Waste Program and Climate Change and Carbon Management Program.

als and solvents on basalt rocks from DOE waste sites. Indeed, they have shown that naturally occurring bacteria can detoxify Cr6 to Cr3, which is very stable. This promises to be a rapidly developing area for basic research in bioremediation.

Ecosystem

Engineering

The Ecology Department has researched selenium transport in the Grassland Water District for many years. Recent research has focused on better methods for compliance monitoring and management.The U.S. Bureau of Reclamation has sponsored this work in an effort to better manage selenium loading in the San Luis Drain. Microbial studies have shown that selenium in-transit losses occur and that the fate of this selenium is bed sediments. Manipulation of the microbial ecology of the drain may stimulate the bioremediation of selenium from the water,thereby reducing the mass loading of selenium to the San Joaquin River.

Bioremediation Ecology Department staff are internationally known for their research in bioremediation.This work has included both in-situ, ex-situ and end-of-pipe treatments and natural attenu a t i o n (intrinsic bioremediation).The department has published papers on bioremediation of chlorinated solvents, polycyclic aromatic hydrocarbons, poly-chlorinated biphenyls, fuels, BTEX, MTBE, actinides, selenium, uranium and chromium. In addition to Department of Energy and Department of Defense waste sites, ED researchers have also studied municipal landfills, agricultural drainage areas, sewage treatment systems, groundwater treatment systems, pulp mills, oil refineries, oil production areas, explosives-contaminated soil, nuclear waste storage areas, solvent-contaminated sites, canneries, cresote treatment areas and nitrogen fertilizer factories. Researchers have several patents on bioremediation technologies, many of which have been licensed and are being used at sites around the world. A fascinating new area has developed over the last two years using the unique capa t i blities at LBNL’s Advanced Light Source (ALS). Ecology Department researchers have shown that by using infrared analyses they can detect the juxtaposition of bacteria, toxic met-

Microbial

Ecology

The Ecology Department has considerable expertise in monitoring and ch a ra c t e rizing microbes in all types of soil, gro u n dwater, food, freshwater, marine, animal and human env i ro n m e n t s . Research e rs have developed a large number of state-of-the-science techniques to detect and identify microorganisms in the e nv i ronment using nu cleic acid probes, polymerase chain re a ction (PCR), polar lipid fatty acids, fatty acid methyl esters , s i g n ature enzymes, fl u o rescent antibodies and direct fluorochrome staining. E c o l o gy department re s e a rchers have been active ly ch a racterizing microbes from the Chern o byl nu clear site and from 5


ESD

Ecology Department

Funding

L a ke Baikal in Siberi a .The department has also developed expertise using the ALS to visualize individual microbes on mineral surfaces and identify their juxtaposition with minerals and organic compounds in-situ.

Environmental

Risk

Ecology Department research is funded by several DOE progra m s : (1) Office of Science, Office of Biological and Env i ronmental Research ; (2) Office of Environmental Manage m e n t , Offices of Science and Te ch n o l o gy and Environmental Restoration; (3) Office of Fossil Energy, Office of Gas and Petroleum Technologies; and (4) Office of Non-proliferation and National Security, Office of Research and Development. Support is also received from the Department of Defense, U.S. Army Corps of Engineers; Department of the Interior, Bureau of Land Management; and from the State of California, Department of Water Resources and Department of the Interior, Bureau of Reclamation under the CALFED program. Finally, funding is received through LBNL’s Laboratory Directed Research and Develpment Program.

Assessment

Using the ALS and va rious new techniques for enzyme assay s , ED re s e a rchers have developed unique methods for a phy s i ologi c a l ly based protocol to estimate bioava i l ability and health h a z a rds of petroleum products from soils to humans. This re s e a rch is providing practical and realistic tools for evaluating various remedial tech n o l o gies that cost-effe c t i ve ly pro t e c t humans—especially ch i l d re n — f rom exposure to residual petroleum hy d rocarbons in surface soils of petroleum-contaminated sites.

http://www-esd.lbl.gov 6


Earth Sciences Division

Annual Report 1998-1999

GEOPHYSICS AND GEOMECHANICS DEPARTMENT

ERNIE L. MAJER CONTACT: (510) 486-6709 ELMAJER@LBL.GOV

The Geophysics and Geomechanics Department performs a wide variety of work ranging from fundamental to applied research.

and geothermal applications. This both strengthens the applied work and provides feedback into the fundamental studies.

Integrated Approach to Future Work Scientific

Thrusts The future thrusts of the department are to continue to d evelop, test and apply high-resolution geophysical methods for not only ch a ra c t e rizing static pro p e rties of the subsurface, but for estimating the dynamic pro p e rties as well.We plan to accomplish this through an integrated effort of theoretical, l ab o ratory and field pro grams. A specific thrust will be in the joint use of seismic and electrical methods for subsurface imagi n g . We have found that to address complex issues such as site remediation, fl ow and transport in fra c t u re systems, vadose zone transport, CO2 s e q u e stration and reservoir stimulation and definition we must use an integrated approach to geophysics and geomechanics.

The department is organized into four different groups: • Center for Computational and Applied Seismology • Potential Methods • Geosciences Measurement Center • Rock and Soil Physics These groups work closely together to address issues in subsurface imaging. The scientific thrusts have been in joint inversion, wave propagation in complex media (seismic and electromagnetic), coupled-process definition and heterogeneity definition.Much of the work focuses on developing and applying highresolution geophysical methods to derive physical properties affecting flow and transport in heterogeneous media. A prime example is the work funded by the Department of Energy’s Natural and Accelerated Bioremediation Research (NABIR) program for the bacterial injection work. Another primary thrust is using geophysical methods for fracture quantification.This is evident in fundamental to very applied studies for DOE’s Fossil Energy, Env i ronmental Restora t i o n , Nuclear Waste and Geothermal programs. The Geophysics and Geomechanics Department is unique within the national laboratory and academic communities in having equally strong theoretical, modeling, lab, field/data acquisition and processing/interpretation capabilities.The department also works very closely with industrial partners in oil and gas

Funding The work of the Geophysics/Geomechanics Department is derived from a variety of DOE and Work-for-Others sources.The primary funding is received from DOE’s Office of Science (Basic Energy Sciences/Geosciences and Office of Health and Remediation), Fossil Energy, Geothermal Te ch n o l o gy, Environmental Remediation and Nuclear Waste Isolation. Other funding sources include the U.S. E nv i ronmental Protection Agency, U.S. Air Force Office of Scientific Research and the U.S. G e o l o gical Survey’s Eart h q u a ke Hazard Reduction Pro gram. Support has also been received from a variety of oil companies, including Chevron, Conoco,Texaco, Exxon and Shell Oil. 7


Earth Sciences Division

Annual Report 1998-1999

This page is intentionally left blank

HTTP://WWW-ESD.LBL.GOV

8


Earth Sciences Division

Annual Report 1998-1999

GEOCHEMISTRY DEPARTMENT

DONALD J. DEPAOLO

CONTACT: (510) 495-2228 DJDEPAOLO@LBL.GOV

The Earth Sciences Division’s Geochemistry Department combines expertise in chemical and isotopic analysis, molecular env ironmental science and minera l o gy, along with data-gatheri n g m e t h o d o l o gyover the full ra n ge of earth env i ronments to enabl e ch a ra c t e rization of ge o chemical systems from the macroscopic to the molecular.The department comprises four groups with complementary interests and capabilities, as described below.

Aqueous

cycles. Current work includes: state-of-the-art molecular dynamics modeling of the interlayer solvated cations in clays; studies of the solvation environment of contaminant and nutrient molecular units in aqueous solution; determination of the molecular identity of initial iron oxide precipitates on quartz surfaces; and characterization of the "acid-mine-drainage" mineral schwertmannite via simulation, x-ray scattering and x-ray spectroscopy methods. Many of these efforts involve newly developed capabilities utilizing synchrotron x-ray sources. Important new work on the aqueous behavior of humic and fulvic acids, hydroxyl speciation near cations in water, and the nature of organic contaminants on mineral surfaces has been carried out recently at Berkeley Lab’s Advanced Light Source.

Geochemistry

Studies in this group address issues of environmental contaminant sequestration, migration, dissolution and oxidation-reduction via a variety of natural and anthropogenic operants. Recent work has included characterization of the selenium speciation, transport and reaction rates within soil horizons at the Kesterson Reservoir, where national attention has focused on the selenium poisoning of wildlife from buildup of agricultural runoff. Other work has determined inorganic chemical processes that reduce the dangerous selenite species to elemental selenium. Related investigations have examined arsenic transport and redox reactions in soils, and microbial effects on the speciation of selenium in hydrologic systems. The important but overlooked effects of the vadose zone air-water interface on the transport of colloids has been identified and quantified by department scientists. Fundamental studies on the nature of the aqueous solution/mineral interface, and on the structure of near-aqueous solvated ions and colloids are also being performed, with the aim to provide improved modeling capability for contaminant migration and other surface processes, such as weathering, sediment transfer, ion exchange and the biogeochemistry of nutrient

Isotope

Geochemistry

The Isotope Geochemistry group operates the Center for Isotope Geochemistry, which was established in 1988 and includes six important analytical facilities: stable isotope, noble gas and cosmogenic isotope laboratories; a soil carbon laboratory; an analytical chemistry laboratory; and a thermal ionization mass spectrometry laboratory located on the UC Berkeley campus.These facilities provide state-of-the-art characterization of all types of earth materials for research throughout the department and elsewhere in the division. Further, they support the Center's goals of finding new ways to utilize isotopic ratio methods to study earth processes, and applying isotopic and chemical analysis procedures to specific environmental and energy problems of national interest. 9


ESD

Geochemistry Department

Geochemical

Current research programs include: (1) the development of models that use isotopic composition data from element pairs in fluids to constrain the geometry and spacing of fractures in rock matrices; (2) the analysis of rock samples from the San Andreas fault zone to determine the source of fluids that produce the lubrication and hence reduce the friction on this fault; (3) the implementation and analysis of large-scale experiments simulating the effects of nuclear waste heat generation within the nuclear repository in Yucca Mountain; (4) the application of helium and neodymium isotopes to determine magma chamber recharge rates in areas having possible volcanic hazards or the potential for geothermal energy extraction; and (5) the use of general atmospheric circulation modeling in concert with wildfire historical analysis to study the effect of climate on wildfire severity in California.

Transport

A major effort of this group is the simulation and study of coupled mineral-water-gas reactive transport in unsaturated multiphase systems, in particular fractured rock and nonisothermal systems. This allows the prediction of processes accompanying the emplacement of high-level nuclear waste at Yucca Mountain, Nevada, and can be utilized to study several types of natural geochemical systems. The simulation has been implemented by introducing chemical reactivity into the existing multiphase flow code TOUGH2, developed at LBNL, resulting in a general reactive chemical transport code called TOUGHREACT. TOUGHREACT was first developed on a PC, and after testing on other platforms, has been ported to the Cray T3E supercomputer at NERSC, where a parallelized version has been implemented. TOUGHREACT has been applied to several different reactive transport scenarios: (1) supergene copper enrichment processes deriving from the oxidative weathering of pyrite (FeS2) and chalcopyrite (CuFeS2). Here the associated acid generation activates metal ion transfer into the unsaturated zone where ore deposits of chalcocite (CuS) and covellite (Cu2S) can form; (2) prediction of the hydrothermal and chemical processes introduced by a strong heat source in unsaturated fractured rock such as at Yucca Mountain, which also complements work by the isotope geochemistry group in identifying transport paths and mechanisms at that site; (3) refinement of kinetic models of dissolution and precipitation of silica and calcite at low (10-100°C) temperatures.The natural evolution of groundwater chemistry at Yucca Mountain is also being interpreted using extant chemical analyses of pore and ground waters. This work will be coordinated with ongoing TEM studies of devitrified tuff at Los Alamos National Laboratory, New Mexico, that will identify the mineral phases and their relation to rock texture and void geometry.

Atmosphere and Oceans The focus of this group is on the characterization of conditions and chemical components in the oceans and atmosphere, and the development of process models using these inputs combined with other hydrospheric data to explain and predict climatic change. The group operates the Regional Climate Center (RCC), where large-scale simulations using the Regional Climate System Model (RCSM) are used for weather forecasts, climate prediction and basic research.The RCC's numerous ongoing collab o rations incl u d e : streamfl ow simulations with the US Geological Survey; runoff contaminant monitoring and management with the US Bureau of Reclamation; development of landslide hazard prediction models with faculty at UC Berkeley; development of snow cover and snow water equivalent maps for California with UC Santa Barbara; and development of a shared information distribution system with DOE/ACPI (Accelerated Climate Prediction Initiative) collaborators. Researchers in the group are currently working with the National Energy Research and Scientific Computing Center (NERSC) at LBNL to develop a new high-performance version of the RCSM. The U.S. Department of Energy has also recently funded LBNL as co-host (with Lawrence Livermore National Laboratory) for the development of a new center for global climate change.The Center for Research on Ocean Carbon Sequestration will be headed at LBNL by ESD’s Jim Bishop, and will include collaborators from Massachusetts Institute of Te ch n o l o gy, Rutgers University, Scripps Institute of Oceanography, Moss Landing Marine Laboratories and the Pacific International Center for High Technology Research. At Lawrence Livermore the leader will be Ken Caldeira.The goal of the center will be to research the feasibility, effectiveness and environmental acceptability of ocean carbon sequestration. The other new DOE center will concentrate on carbon sequestration in terrestrial ecosystems.

Funding Funding for the Geochemistry Department comes from a va riety of sources, including: the U.S. D e p a rtment of Energy, Office of Science, Office of Basic Energy Sciences, Divisions of Materials Sciences, and Engi n e e ring and Geosciences; DOE Office of Env i ronmental Managment, Office of Science and Te ch n o l o gy ; DOE Office of Energy Efficiency and Renewabl e E n e rgy, Office of Utility Te ch n o l o gi e s , Office of Geotherm a l Te ch n o l o gies; DOE Office of Civilian Radioactive Waste Management; U. S . E nv i ronmental Protection Agency; U.S. Navy; National A e ronautics and Space Administration, Office of Space Science and NASA Earth Enterprise; National Science Fo u n d a t i o n , Office of Polar Pro gra m s ; the University of California Campus-Lab o ratory Collab o ration Hydrology Project; and the Laboratory Directed Research and Deve l o p m e n t Program at LBNL.

http://www-esd.lbl.gov 10


Earth Sicences Division

Annual Report 1998-1999

HYDROLOGY AND RESERVOIR DYNAMICS DEPARTMENT

CHIN-FU TSANG CONTACT: (510) 486-5782 CFTSANG@LBL.GOV

The Hydro geology and Reservoir Dynamics Depart m e n t (HRD) is one of the most active groups in the world doing research in different fields of hydrogeology and reservoir engineering. HRD conducts research for various projects of the Department of Energy in the areas of nuclear waste geological disposal, environmental management, fossil fuel development, geothermal engineering and basic energy sciences. The department is also involved in projects with the U.S. Environmental Protection Agency and U.S. Department of the Interior, and participates in various international collaborative projects with Japan, Sweden, Russia and others. HRD research covers several areas, briefly described below.

Contaminant

modeling studies and pilot studies in the field. Besides these two efforts, advanced well-test methods are also being designed for characterization of the subsurface hydrology, and various analysis and modeling techniques are being developed.

Vadose Zone Hydrology The study of the soil and geologic layer between the ground surface and the water table is a ver y active area of HRD’s research, with applications in both the nuclear waste disposal and environmental remediation programs of DOE. A number of basic studies were made on fast paths in the vadose zone, such as the development of flow channeling in the unsaturated medium and on the chaotic model of flow in such systems. Various field-testing techniques have been developed, including new geophysical methods of measuring soil tension and new designs for studying water percolation in unsaturated fractureporous media in the field.The latter are especially noteworthy, as they are being applied a series of experiments in the underground "exploratory shaft" at Yucca Mountain, Nevada.There, the effects of fractures, a major fault and a porous matrix on the flow through the system and the threshold for seepage into the tunnel are being studied. Another major set of experiments was conducted at Idaho National Engineering and Environmental Laboratory, where a multi-scale study was made to study the characteristics of unsaturated flow through complex structures that strongly vary with scale,ranging from centimeters to tens of meters. Extensive modeling efforts accompany these field activities both for the experimental design and for data evaluation.

Hydrology

Work conducted in this area cove rs va rious theoretical and ex p e rimental studies, including a major evaluation of contamination at LBNL. In this multi-year evaluation, details of the hydrogeologic structure of the site were explored, m e a s u red and modeled, then the contaminant plumes (which are relatively minor) were characterized, m o n i t o red and studied. Finally, a control and remediation stra t e gy was developed in collab o ration with local and federal regulatory agencies.The work invo l ves the multi-disciplinary e ffo rt of geologists, hydrologists, geophysicists and hydrogeochemists, chara c t e rizing many of HRD’s research projects. Another major effort is the development and study of in-situ barriers and their emplacement for control and containment of contaminant plumes.This includes the choice of gelling barrier materials, their characterization in laboratory studies, development and optimization of emplacement methodology through 11


ESD

Hydrology and Reservoir Dynamics Department

Heterogeneous/Fracture

Flow

Systems

For more than two decades, HRD has been active in the field of fracture hy d ro l o gy, being one of the fi rst groups to develop a fra ct u re network model and a channeling model of fl ow through va riable-aperture fra c t u re systems, and to apply the annealing model to fra c t u re hydrology. Our work is characterized by close intera ction between modeling and field data evaluation, with complementary lab o ra t o ry studies.This work continues and has been further generalized to research into strongly heterogeneous media, i n cluding data evaluation of heterogeneous porous-fra c t u re systems and studies with the stochastic continuum and double-perm e ability models. A hierarchical model has also been developed for the study of mu l t i - l evel fracture systems. Furthemore, n ew fi e l d m e a s u rement techniques are being developed to measure strongly va rying perm e abilities in the borehole. General considerations of predictive evaluation have been developed with recommendations for an approach to iterative site characterization and perfo rmance modeling for such heterogeneous systems.

Integrated

Site

into the TOUGH codes. This includes both homogeneous reactions, s u ch as aqueous complexation and re d ox reactions, and hete ro geneous reactions such as ion exchange, adsorption, mineral dissolution and precipitation, and gas dissolution and exsolution.

Coupled Processes in Fractured Rocks The coupling of mechanical stress and temperature effects on permeability of fractured ro cks is important in injection testing, s t i mulation of oil and gas reservoirs and nuclear waste repository performance. HRD’s work involves the development of a coupled thermo-hydro-mechanical (THM) simulator, with fully coupled HM processes and thermal convection and temperature-dependent pro p e rty parameters. Recent improvement of the code to handle processes in unsaturated systems has been motivated by nuclear waste storage in unsaturated media and isolation of waste canisters in bentonite near-field barriers. The developed model has been tested against a number of lab o ra t o rymeasurements and applied to a study of a major THM ex p e riment in the Kamaishi Mines in Japan. The capability of the coupled THM code has also been applied to the understanding of HM effects occurring in pressure injection tests of a borehole intersecting fractures in hard rocks.A number of field tests in Sweden and Iceland are being analyzed to unders t a n d injection-induced fracture opening and propagation.

Characterization

The need for a well-designed and optimal site characterization program at a site is well recognized for many geosciences-related problems of national concern. HRD research emphasizes field measurement using hydrologic, geophysical, chemical and geomechanical methods and integrates the analyses of these different types of data to obtain the best conceptualization of the site. A large-scale example is the characterization of the fracture tuff formation at Yucca Mountain,where DOE is conducting an eightyear heater test over a block of the order of 100 m with temperatures of up to 200°C.To understand the flow of fluids, including evaporation and condensation, in the complex unsaturated fractured porous rocks, measurements of air permeability, water sampling with isotope chemistry analysis and ground penetrating radar, are conducted in the same area. Integrated analysis of the data helps provide a good understanding and characterization of the system. In addition to such major field studies, basic studies related to integrated data analysis techniques are also performed.

Production

Optimization

and

Testing

HRD is also ve ry active in the study of oil and gas reservoirs.This includes optimization and control theory to maximize oil pro d u ction with the hydrofra c t u ring process, using injection wells. Advanced and unconventional well-test methods to determine and ch a ra c t e rizeproduction zones have been developed. Po re netwo rk models are being developed to understand dra i n age and imbibition processes in reservoir rocks during flooding operations. An interesting application being studied is the diatomic fields, which re present potentially billions of barrels of high-quality oil. However, production from the diatomites requires a secondary re c overy process because of their low permeability, even though they have high porosity. R e s e a rch into hy d rofra c t u ring with flooding is being conducted to explore optimal production strategies.

Flow and Transport Modeling HRD has a long history in nu m e rical modeling of flow and t ra n s p o rt in ge o l o gical media. A suite of nu m e rical models using finite diffe rence finite element and integrated finite difference methods has been deve l o p e d . The most well-tested and applied computer code is the TOUGH fa m i ly of simulators, which calculates flow and tra n s p o rt of multi-phase, multi-component fluids in complex fra c t u re - p o rous media. A number of equation-of-state packages we re developed for different fluids appro p riate for env ironmental, nu clear waste disposal, oil and gas and geothermal reservoir applications. Associated with these codes are the iTOUGH codes, which perform the inverse calculations of parameter estimation for s u ch complex systems. Current development i nvo l ves the mplementation of re a c t i ve chemistry

Funding Funding for the Hydrology and Reservoir Dynamics Department comes pri m a ri ly from the U.S. D e p a rtment of Energy, including: O ffice of Science, O ffice of Basic Energy Sciences, Division of Engineering and Geosciences; Office of Biological and Environmental Research; Assistant Secre t a ry for Energy Efficiency and Renewable Energy, Office of Geotherm a l Technologies; O ffice of Civilian Waste Management; and Assistant Secre t a ry for Environmental Management. Other funding is provided through the Lab o ra t o ry Directed Research and Development Program at LBNL.

http://www-esd.lbl.gov 12


Earth Sciences Division

Annual Report 1998-1999

CENTER FOR ENVIRONMENTAL BIOTECHNOLOGY

JENNIE C. HUNTER-CEVERA CONTACT: TERRY C. HAZEN (510) 486-6223 TCHAZEN@LBL.GOV

The Center for Env i ronmental Biotech n o l o gy (CEB) is a virt u a l center for integrated multi-disciplinary research in three main focus areas: biogeochemical transformations, e nv i ronmental diagnostics and env i ronmental health risk assessment based on bioava i l ab i l i t y. The Center’s core staff members represent four major LBNL divisions: Earth Sciences, L i fe Sciences, Environmental Energy Tech n o l o gies and Engi n e e ri n g . In addition, the Center has collaborating scientists from Information and Computing Sciences, A c c e l e rator and Fusion Research, Nuclear Physics, Materials Sciences and the Advanced Light Source. CEB’s responsibilities are to coordinate, integrate and provide multi-disciplinary research teams to address specific environmental problems at Department of Energy sites (mixed wastes) and Department of Defense sites (explosives, polycyclic aromatic hydrocarbons, polychlorinated biphenyl and heavy metals) and within the State of California (mining, forestry, waster quality and agricultural). CEB also manages the U.S. Army Core of E n gi n e e rs' BEST (Biore m e d i a t i o n , Education, Science and Technology) Program and houses the Cal/EPA Bioremediation Validation and Certification Program. In addition, CEB coordinates LBNL’s efforts in non-proliferation detection reagents, antibiotic discovery and enzyme screening.

Biogeochemical

intrinsic endolithic (rock-inhabiting) bacteria to transform VOCs, with respect to the range of critical environmental factors. In addition, stable isotopes are used to monitor the biotransformation processes. Comparative analysis of results from the agar printing-off experiment suggest that at the mesoscale level, there is a significant heterogeneity in the distribution and diversity of cultivable microbial consortia inside the vesicular basalt. It appears that many of the endolithic microorganisms live in pore space with or without direct connection with the surface.

Infrared

Microspectroscopy

Pollution of subsurface geologic zones and the possibility of using the intrinsic endolithic (rock/minera l - i n h abiting) bacteria to either detoxify or immobilize the pollutants have stimulated new interest in the exploration of endolithic bacteria and their longt e rm surv i val in the ge o l o gic env i ronment.We have developed and d e m o n s t rated the applicability of surface-enhanced infra re d absorption-reflection (SEIRA) microspectro s c o py to study— quick ly and with minimum sample pre p a ration—in-situ relationships between the microbial localization and the microscopic physiochemical structures of ge o l o gic materials such as rocks. U n l i ke traditional microscopy approaches, SEIRA micro s p e ctro s c o py can provide researchers simultaneously the biologi c a l , geochemical and physical characteristics of the intact enviro nmental samples. This study has demonstrated that surfa c e enhanced infrared absorption-reflectance (SEIRA) micro s p e ctro s c o py using a metal-ove r l ayer is a promising tool for studying the in-situ localization of bacteria within ge o l o gic materi a l s .

Transformations

Volatile organic compounds (VOCs) are the most common organic contaminants detected in polluted subsurface soil and fractured rock vadose zones. We are examining this microbial biodegradation in fractured rocks by investigating the use of

13


ESD

Center for Environmental Biotechnology

Biodegration of Fuels Another focus area in biogeochemical transformations is our work on the biodegradation of methyl tert-butyl ether, MTBE, an additive in gasoline that is a potential carcinogen. By designing a bioreactor that simulates the actual “process reactor,� we have isolated indigenous microorganisms attached to the carbon material that have the capability of biodegrading.These microorganisms are currently being identified and optimized for use in an ex-situ bioreactor. A pilot plant has been constructed in Southern California for treatability studies with Kinder Morgan Energy Partners, U.S. Filter/Envirex and North Carolina State University.

More re c e n t ly, we have concentrated on using isotopic measurements to detect in-situ biodegradation of more recalcitrant contaminants, such as ch l o rinated solvents and gasoline ox y genates.

Mass Spectometric DNA Diagnostics Biological activity has often been attributed to changes in pollutant pro files found in contaminated soils when abiotic processes actually caused pollutant removal. LBNL engineers are working with CEB molecular biologists to evaluate a monitoring strategy that relies on the combined use of DNA diagnostic procedures and mass spectrometry as the detection scheme. The intent is to track bioremediation by measuring the occurrence of genes in soil samples that are known to code for enzymes capable of degrading specific pollutants.This type of test is commonly performed with PCR and gel electrophoresis, but matrix-assistedlaser-desorption ionization mass spectrometry offers the possibility for automation and high throughput as needed to track the course of bioremediation over large polluted areas. We are currently working with the naf gene from Psuedomonas stuterzi.

Biodegration of Mixed Wastes We are working on the biosorption of actinides in the presence of organics and chelating agents and assessing the fate and transport of actinides.This combination of actinides and organics represents a truer picture of the actual mixtures found at the Hanford Site near Richland, Wash. Results indicate that the fate and transport of actinides is very different in a mixture versus pure actinides and that certain bacteria are capable of absorption actinides at high levels.We have also begun a collaboration with the Seaborg Center for Actinide Chemistry to better understand environmental parameters affecting the interaction of actinides and microorganisms.

Algal-Bacterial

Selenium

Environmental

Assessment

Assessing env i ronmental risk based on bioava i l ab i l i t y, i.e., what is actually being taken up by human, animal and plant cells, is a challenge for CEB. T h e re is much data in the literature on highdose ex p o s u re rates or single-dose, single-compound ex p o s u re rates to animals. CEB has taken the approach of examining the we a t h e red material as it exists in nature and exposing human cell lines to this material over time. In many instances our cell-line models consist of co-culturing two different types of cell lines on thin membranes. We are assessing the effects of ingested we a t hered PAH mixtures, organochlorines and dioxin by simulating the human digestion system. One team is examining the digestion of the material by human enzymes, u p t a ke by intestinal cell lines (KACO2) and conve rsion by metabolic cell lines such as liver, k i dney and breast cell lines. Using a combination of molecular biom a rkers and the synchrotron, we can measure what material is left behind in the intestinal cell lines, what the compounds may be metabolized to and any DNA damage and repair using specific molecular biomarker assays developed by our cell biologi s t s .

Treatment

A selenium and nitrate treatment system has been operated on agricultural land within the Panoche Water District for the past three years. Two side-by-side systems consisting of a shallow, high-rate algal pond, a deep reduction pond, a sedimentation unit and filter bed were constructed close to the tile sump draining a 1,000-acre field. Removal rates vary from 40 to 80% depending on the configuration of the system. Ongoing research is focused on increasing the removal efficiency and throughput of these systems under a variety of environmental conditions. The ultimate objective is to develop a low-cost, easy-to-implement technology to help farmers reduce selenium loads exported to the San Joaquin River and San Francisco Bay-Delta.

Environmental

Risk

Diagnostics Partners and Funding

CEB, in collab o ration with Earth Sciences Division’s Center for Isotope Geochemistry, has been employing stable isotope monitoring to validate bioremedial activities in field sites.The isotopic ratios of soil, gas and gro u n dwater compounds are monitored to distinguish by p roducts of biodegradation of petroleum hy d ro c a rbons (e.g., CO2, CH4) from other potential sources. Furthermore, distinctive shifts in the isotopic compositions of the contaminants and metabolic by p roducts have been used to d i ffe rentiate between specific metabolic pathways used by the microorganisms to degrade the contaminants.

CEB research is funded by the U.S. D e p a rtment of Energy and D e p a rtment of Defense. Industrial collab o ra t o rs are Kinder Morgan and Geokinetics and the Petroleum Env i ronmental Research Forum. Academic part n e rs include the University of Califo rnia at Davis, California Polytechnic Institute, Unive rsity of Utah and the BEST Program partners: Jackson State University, Ana G. Mendez Unive rsity System, Southern Unive rsity of Mississippi, University of Texas at El Paso and University of California at Berke l ey.

http://www-esd.lbl.gov 14


Earth Sciences Division

Annual Report 1998-1999

This page is intentionally left blank

HTTP://WWW-ESD.LBL.GOV

15


Earth Sciences Division

Annual Report 1998-1999

This page is intentionally left blank

HTTP://WWW-ESD.LBL.GOV

16


Earth Sicences Division

Annual Report 1998-1999

FUNDAMENTAL AND

EXPLORATORY RESEARCH PROGRAM SALLY M. BENSON CONTACT: (510) 486-5875 SMBENSON@LBL.GOV

The Fundamental and Exploratory Research Program (FERP) area covers fundamental earth sciences research conducted in support of the Department of Energy’s science mission, which includes research in the natural sciences to provide a basis for new and improved energy technologies and for understanding and mitigating the environmental impacts of energy use. This part of the Earth Sciences Division’s program also includes exploratory research in important new energy and environmental topics conducted under the Laboratory Directed Research and Development (LDRD) program. The scientific insights and breakthroughs achieved in FERP often become the underpinnings for projects that support DOE’s applied research and development program offices. Over the years, the basic earth sciences research program at Berkeley Lab has focused on three broad earth sciences problems, described below. • Fundamental studies of chemical and mass transport in geologic media with special reference to predictive modeling of multiphase, multicomponent, non-isothermal fluid flow in saturated and unsaturated fractured rocks. • The development of new isotopic techniques for understanding the nature of a broad range of global processes— from the relatively short-term effects of natural fluid migration in the crust to longer-term global climate variations. • Fundamental studies in the propagation of waves through geologic media with emphasis on new computational techniques for analyzing seismic/acoustic and broadband electromagnetic signals for high-resolution imaging of near-surface structures, such as possible fracture flow paths, and for inferring the types of fluids present in pores and fractures.

Results from these re s e a rch endeavo rs have had major impacts on applied energy, environmental and radioactive waste management programs. Current research projects are briefly described in the following sections.

Fluid-Chemical Transport Investigations Building on their previous findings that film flow is an important mechanism for fast fluid transport along fractures in the unsaturated zone, researchers can now demonstrate that even in rocks with low matrix permeability (<10-15 m2) but containing fracture apertures larger than 50 microns, fast paths for fluid flow will occur in the vadose zone, and these films will transport colloids. This work verifies that contaminants sorbed onto colloids smaller than the film thickness may be transported effectively from the vadose zone to the ground water. E fforts continue to accurately model subsurface multiphase fluid and heat flow, along with solute tra n s p o rt and chemical reactions. By incorporating reactive chemistry into the fra m ework of the exiting TOUGH2 code, ESD research e rs have been able to model ore-fo rming processes such as supergene copper e n ri chment and to predict the therm a l , hy d ro l o gical and ch e m ical processes that are likely to occur around a thermal source that s i mulates conditions in a high-level nu clear waste repository. Molecular modeling of cesium cation (137Cs+) - smectite clay interlayer systems has confirmed the previous findings from bulk diffusion experiments that clay liners will impede the mobility of radioactive 137Cs+, a fact important to the design of nuclear waste containment facilities. Prior to this study, detailed experi17


ESD

Fundamental and Exploratory Research Program

mental characterization of this system proved difficult due to the high degree of disorder within these clays.

Isotope

the division, resulting in collaborations on a wide variety of fundamental imaging problems, some of which are reported here. Researchers have successfully demonstrated the use of joint ge o p hy s i c a l - hy d ro l o gical data sets for estimating stochastic hydrologic parameters of a test site. Using data collected at the Oyster, Va., bacterial transport test site, they have been able to integrate hydraulic conductivity information from flowmeters and radar cross-hole tomograms to obtain improved images of permeability. Researchers have completed a major study of wave propagation along the San Andreas fault zone as part of the Parkfield Prediction Program. On the basis of more than 6,000 natural earthquakes and 720 source-receiver paths obtained from a controlled-source program, they have developed a detailed elastic model confirming that there are temporal velocity changes occurring in a region suspected to be the nucleation area for past and future magnitude-6 earthquakes. These velocity variations are strongly believed to be related to changing fluid conditions in the shallow section of the fault zone. R e s e a rchers have also developed and tested advanced techniques for modeling elastic and electro m agnetic wave propagation through media heterogeneous in two and three dimensions.In one s t u dy they treated elastic wave propagation as a series of forward s c a t t e ring problems, where the medium is described as a random d i s t ribution of scattere rs of va rious sizes and physical parameters. Analytical results based on simple models compare well with nu m e rical simulations for a wave pro p agating through a medium containing a random distribution of spherical scattere rs . In another study, researchers developed a new coupled integral equation-differential equation approach for the nonlinear inversion of electromagnetic, seismic velocity and hydrologic conductivity data sets. New GILD and SGILD methods provide a high-resolution, robust and stable algorithm suitable for high-performance parallel machines.

Geochemistry

The Center for Isotope Geochemistry (CIG) is a state-of-the-art analytical facility established in 1988 for the measurement of concentrations and isotopic compositions of elements in rocks, minerals and fluids in the earth’s crust, atmosphere and oceans. Fundamental research conducted at this center is directed at finding new ways to use isotopic information to study earth processes such as long-term climate changes and the way mantle-derived or deep crustal fluids move through the crust. In a effort to reconstruct global climate and climate changes during the past 20,000 years, CIG researchers measured the oxygen and hydrogen isotope ratios (δ18O and δD) in Antarctic ice cores from three locations to develop a model that relates isotopic compositions to water available in the ancient atmosphere and past surface temperatures. They have found clear evidence in the ice cores for the temperature transition from the last glacial maximum to the warmer and wetter Holocene, and found evidence that temperatures during the last glacial maximum were substantially lower than previously estimated on the basis of δ18O data and the modern spatial relationships. The presence of He, C, and O isotopes in approximately 250 samples of fault gouge, breccia and host rocks collected along the San Andreas and adjacent faults confirms that a significant fraction of He is of mantle origin and is accompanied by deep crustal water and CO2.These findings support earlier results suggesting that deep crustal and mantle fluids enter and lubricate the fault zone, thus causing the low-friction conditions observed from seismological and deformation data. In their continuing study of a present-day volcanic system, researchers have found that co-variations between He and Nd isotopes in olivines from continental basalts can be used to differentiate between separate magma chambers and to assess the rates for heat and magma recharge into the crust.

Funding Funding for research in the Fundamental and Exploratory Research Program comes from a variety of sources, including primarily the Office of Basic Energ y Sciences, Department of Engineering and Geosciences, and Office of Biological and Environmental Research, both in the Office of Science of the U.S. Department of Energy, and the Laboratory Directed Research and Development Program at LBNL. Additional support comes from DOE’s Office of Environmental Management, Environmental Management Science Program.

Advanced Computation for Earth Imaging The Center for Computational Seismology (CCS) serves as the LBNL and UC Berkeley nucleus for seismic research related to data processing, advanced imaging and visualization. In recent years, a great deal of cross-fertilization between seismologists and other geophysicists and hydrogeologists has developed within

http://www-esd.lbl.gov 18


Earth Sciences Division

Fundamental and Exploratory Research Program

Annual Report 1998-1999

Identifying Limits of Film Flow In Unsaturated Fractures Research

Objectives

be associated with the matric potential at which the local rock matrix is Tetsu K. Tokunaga In an earlier study (Tokunaga and e ffe c t i ve ly saturated (satiated), and and Jiamin Wan Wan,1997), we identified film flow as the upper limit associated with local a mechanism that can contribute to saturation of the fracture aperture. Contact: Tetsu Tokunaga fast flow along unsaturated fractures. The critical matric potential at (510) 486-7176, tktokunaga@lbl.gov The combination of low-permeability which the rock matrix is effectively rock, large fracture aperture and nearsaturated and ab ove which wa t e r zero matric potential was qualitatively identified as necessary for films can begin to emerge on fracture surfaces is approximately development of fast film flow. However, more quantitative deterequal to the air-entry matric potential. Because of hysteresis in minations of the ranges of matrix permeabilities, fracture aperthe potential-saturation relation, and also because of generally tures and matric potentials that permit development of thick, unknown wetting history, the actual critical matric potential will transmissive films along fracture surfaces remained unspecified. typically have a magnitude between about 50% and 100% that of This study is directed at identifying the approximate region of the air-entry value.The lower range of matric potentials,at which stability for thick films, within the parameter space defined by thick films begin to emerge from fracture surfaces,was estimated rock matrix permeability, local fracture aperture and matric through correlations between matrix permeabilities and airpotential. entry matric potentials of a wide range of porous media, including soils, glass bead packs, rocks and ceramics (Figure 1).We seek Approach only rough correlations since the range of permeability spans nearly 10 orders of magnitude. This correlation was shown to Our initial studies on flow in unsaturated fractures identified predict the air-entry matric potential within one order of magnifilm flow as a mechanism capable of permitting fast flow and tude, for 92% of the data (N = 76).The correlation was also fairly transport. That study revealed limitations of earlier conceptual consistent with predictions based on Miller-Miller geometric models for unsaturated flow in fractured rock, introduced the similitude, even though the highly varied sample set does not rigfilm flow hypothesis and provided experimental results suporously conform to prerequisites for geometric similitude. The portive of the hypothesis. The water â&#x20AC;&#x153;filmsâ&#x20AC;? investigated in the exponent in the regression fit is -0.425, whereas Miller-Miller similitude predicts a value of -0.5. previous study as well as the present one develop on rough surUpon effectively saturating the rock matrix, local topographic faces typical of rock fractures, range in average thickness from minima on the fracture surface become progressively waterabout 1 to 50 Âľm and flow in the laminar regime. The initial filled as the matric potential is brought closer to zero. Filling of study showed that film flow becomes important when matric local topographic minima on fracture surfaces progresses from potentials along rock surfaces are high enough (less negative) to finer roughness features to coarser ones, as the matric potential effectively saturate the surficial rock matrix. Here, more specific is brought closer to zero, in accordance with continuously constraints are identified for parameter ranges within which film increasing the radius of curflow can occur. vature characteristic of the The first step in this study air-water interface. Thus, the consisted of identifying a sinave rage film thickness on gle parameter to represent rough fra c t u re surfa c e s each of the primary system i n c reases as the matric components, the ro ck potential appro a ches zero, matrix, fracture and water. In primarily because of surface this simplification, the capillary re l a t i o n s . In this selected para m e t e rs we re progression to thicker averthe rock matrix permeability, f ra c t u re apert u re and age films, the transmissivity matric potential, re s p e cand hydraulic diffusivity also tively. Since the matrix perincrease, as shown in earlier meability and local fracture work. aperture are essentially fixed In typical rock fractures, at any given location along a film thickening has a finite fra c t u re surface, one can limit imposed by the fracture then consider the range of aperture. The upper range of matric potentials over which matric potentials, at which stable thick films exist. The Figure 1. Correlation between permeabilities and air-entry matric potentials for a films give way to locally satlower energy limit will then wide variety of porous media (glass beads, sands, clays, rocks and ceramics). urated fracture apertures, is 19


ESD

Identifying Limits of Film Flow in Unsaturated Fractures

s t ability re gion for thick estimated from parallel plate water films in unsaturated capillary considera t i o n s fractures. This finding is of used in previous models for interest because earlier conpartially saturated fractures. ceptual models did not recThese two limiting matric ognize the existence of potentials, one for ro ck thick water films in unsatumatrix saturation (satiation), rated fractures (i.e., the midand the other for local aperdle region in Figure 2), and ture saturation are plotted as because a significant vo lsurfaces with respect to ume of the fracture flow their dependence on matrix parameter space is associpermeability and apert u re ated with thick films. Note size in Figure 2. The combithat low permeability rock nation of local material (with permeability <10-15 properties for a given segm2), with apertures larger ment of fracture, the local than about 50 Âľm have a sigmatrix permeability and nificant near-zero matric fracture aperture, specifies a potential range over which given vertical line intercepting the base of this parame- Figure 2. Approximate regions of local stability for (1) unsaturated matrix flow, (2) film flow can occur. This combination of permeabiliter space. The energy status film flow over effectively saturated rock matrix, and (3) saturated fracture flow. ties and apertures is quite of water at this location on common, indicating that film flow may be important in many the fracture surface specifies a particular point along this vertical fractured vadose environments. line.Thus, at lower (more negative) matric potentials, characterized by unsaturated matrix flow, systems occupy the region Related Publication below the lower surface shown in Figure 2. At higher (closer to zero) matric potentials, the local rock matrix becomes effectively Tokunaga,T. K., and J.Wan,Water film flow along fracture surfaces saturated, permitting stable thick films along fracture surfaces. of porous rock, Water Resources Research, 33, pp. 1287-95, This condition lies between the two surfaces. At still higher 1997. matric potentials, the local aperture becomes water-saturated, thereby locally eliminating water films. This last state of local Funding aperture saturation occupies the upper region of the parameter space shown in Figure 2. This work has been supported by the Office of Science, Office Significance of Findings of Basic Energy Sciences, Division of Engineering and Geosciences of the U.S. Department of Energy under Contract The result summarized in Figure 2 identifies the approximate No. DE-AC03-76SF00098.

http://www-esd.lbl.gov 20


Earth Sciences Division

Research

Fundamental and Exploratory Research Program

Objectives

Annual Report 1998-1999

Microbubble Generation, Stability and Transport: A Potential Subsurface Remediation Technique

ble suspension was injected into the bottom end of the wa t e r - s a t u ra t e d Bubble suspensions, also known as columns. Microbubbles were genercolloid gas aphrons (CGAs), h ave ated either under atmospheric presbeen used for scouring organics from Fred Gadelle, Tetsu K. Tokunaga sure or under pressure (30 psig). For and Jiamin Wan the micro b u bbles ge n e rated under contaminated aquifers and for deliverambient conditions, the suspension ing oxygen for bioremediation. CGAs Contact: was injected using a syringe pump. have typical bubble size ranging from Fred Gadelle (510) 486-2226, fgadelle@lbl.gov Injection of microbubbles generated 30 to 100 µm.Therefore, bubbles have at 30 psig was controlled with an in-line differential pressure a short lifetime (a few hours) and can only flow through large flowmeter. Microbubble concentration and size distri b u t i o n pores and preferential flow paths.The objective of this research were measured using a Coulter Multisizer II. is to generate small size and long-lasting microbubbles (0.7-15 µm) that present suitable physico-chemical conditions at the gasResults water interface for preferential sorption of contaminants and to provide a steady flow of microbubbles upward through contamMicrobubble Generation and Characteristics: Coalescence inated zones, including those advectively inaccessible zones, for between bubbles and gas diffusion from the bubbles toward in-situ remediation. water is responsible for the lack of stability of gas bubbles in Approach pure aqueous solutions. The presence of surface-active compounds such as surfactants greatly improves the stability of the The microbubble suspension was generated by mixing surbubbles by forming a film at the gas-water interface. Formation factant solutions at high speeds. The generator, built in-house, of stable microbubbles is accomplished by mixing surfactant comprises a 5-cm disk mounted at the end of a shaft, two vertisolutions at high speed. Stable microbubbles were only formed cal baffles and a 4.5-L mixing beaker.The disk, when rotating at with mixtures of water-soluble surfactants and solid non-soluble high speeds, entrains air into the solution; upon hitting the bafsurfactants. The best formulation is a combination of sodium fles, the entrained air breaks into microscopic bubbles. The dodecyl sulfate, SDS, and sorbitan monostearate, Span 60. The microbubble generator was located in a stainless steel chamber number of microbubbles greatly depends on the concentration to generate microbubbles under pressure.This design would also and size of Span 60 particles. It was found best to grind Span 60 allow the use of oxygen or other gases as opposed to air. to a fine powder before mixing with SDS.The optimum SDS conConcentration, size distribution and stability of microbubbles centration is 1 g/L. Figure 1a is a photograph of a SDS/Span 60 generated with several surfactants and surfactant mixtures were microbubble suspension taken seven days after generation.Two measured to determine the optimal “microbubble formulation.” size fractions are shown: <1 µm and 2-5 µm. The microbubble Microbubble transport experiments were conducted in verticoncentration is 1.5 x 109 bubbles/mL and the size ranged from 0.7 to 7 µm, with 75% of the microbubbles <2 µm (Figure 1b). cal sand columns under steady flow conditions.Three grain sizes The specific volume of this sample is 12 x 109 µm3/mL, which were used: 415-500 µm (coarse sand), 150-212 µm (medium corresponds to an air volume of 1.2% (assuming the contribution sand), and 53-106 µm (fine sand). One pore volume of microbub-

Figure 1b. Microbubble size distribution on a number and volume basis. Concentration = 1.5x109 microbubble/mL and 12.2x109 µm3/mL.

Figure 1a. Photograph of a microbubble suspension. Objective = 40x.

21


ESD

Microbubble Generation, Stability and Transport: A Potential Subsurface Remediation Technique

of the surfactant coating to the b re a k t h rough in the medium size of the microbubbles is negsand (size exclusion effect) and ligible). Concentrations as high re c ove ry was ~67%. Flow as 2.5-3.0x109/mL have been through the fine sand was signifachieved with air content of 4 icantly re t a rded and re c ove ry to 7%, as determined by the was <35%, suggesting significant multisizer or by direct gravimetlosses of microbubbles in the ric measure m e n t s . Stability p o rous media. M i c ro b u bble experiments have revealed that losses could be attributed to microbubbles, as well as diluted microbubble breakage (large air microbubble suspensions, 1/10 bubbles were observed in the and 1/100 in deionized water column effluent) and/or irreand salt solution, are stable for versible sorption phenomena. several weeks. Significance The effect of pre s s u re on of Findings concentration and stability of the micro b u bble suspension G e n e ration of long-lasting was also investigated. For example, microbubble concentration Figure 2. Microbubble breakthrough curves through water-saturated and stable microbubble suspensions is accomplished using a (specific volume) incre a s e d sand columns. mixture of surfactants (SDS and from 1.1 (24.3) to 1.6 (37.0) to Span 60). These microbubble suspensions are characterized by 1.8 x 109/mL (38.8 x 109 µm3/mL) when the pressure under which they were ge n e rated increased from 5 to 18 to 30 psig, very large specific volumes and concentrations. Good microbubrespectively. C o nversely, microbubbles are significantly altered ble re c overy is ach i eved in coarse- and medium-size sand when subjected to static pre s s u res slightly ab ove ambient pressure columns, suggesting that microbubbles can effectively be trans(6 to 10 psi): microbubble concentration (specific volume) was ported in such porous media. These results indicate that reduced from 1.6 (45.3) to 1.2 (4.0) to 0.76 x 109/mL (0.79 x 109 microbubble suspensions could be used to deliver in-situ signifiµm3/mL) when the applied pressure increased by 10 and 20 psi, cant amounts of air (or oxygen) for bioremediation of contamirespectively.These numbers also indicate that the larger micro b u bnated soils and groundwaters.Furthermore, the large specific surbles (4 to 15 µm) are most sensitive to pre s s u re increase. face area of the microbubbles, up to 50 cm2/mL, suggests that microbubble suspensions could potentially be used as a sorptive Microbubble Transport in Porous Media: Low-concentraphase to remove contaminants from the subsurface. tion and small-size (<3 µm) microbubbles, generated under atmospheric pressure, were injected with the syringe-pump into Related Publication coarse sand. Figure 2 shows essentially conservative transport with some re t a rd a t i o n . Ap p rox i m a t e ly 100% re c overy wa s Gadelle, F.,T.K.Tokunaga and J.Wan,Transport of microbubbles in obtained. Flow experiments through fine sand were inconclusive porous media, Environ. Sci.Technol., submitted. as the backpressure caused microbubbles to break/dissolve in the syringe-pump. To circumvent this problem and to increase Funding the amount of air delivered, microbubbles were generated under pressure and then directly injected from the pressure chamber This work has been supported by the Assistant Secretary of into medium and fine sand columns. In these experiments, Environmental Manage m e n t , Env i ronmental Manage m e n t microbubble size ranged from 0.7 to 10 µm and concentration Science Program, of the U.S. Department of Energy under (specific volume) was 1.3 x 109 microbubble/mL (ca. 9 x 109 Contract No. DE-AC03-76SF00098. µm3/mL). As shown in Figure 2, microbubbles exhibited early

http://www-esd.lbl.gov 22


Earth Sciences Division

Fundamental and Exploratory Research Program

Annual Report 1998-1999

Particle Motion in Film Flow Research

Objectives

submerged sphere, additional forces, such as surface tension acting along Tra n s p o rt of colloids thro u g h the three-phase contact line between Contact: unsaturated subsurface environments the film surface, particle and air, and Srinivas Veerapaneni has significant environmental implicapressure force also act on the particle (510) 495-2947, sveerapaneni@lbl.gov (Figure 1). The translational and rotations. For instance, movement of contional velocities of the particle can be obtained by balancing all taminants from vadose zone to the groundwater table may be the forces (in the direction parallel to the solid plane) and facilitated by their sorption onto the colloids. Despite its importorques (about z-axis) acting on a particle (Equations 1 and 2) tance, the effect of flowing thin liquid films on colloid transport in partially saturated rock fractures and porous media is poorly (1) understood. In a recent study, Wan and Tokunaga presented a Fdrag + Fgrav + Ffric + Fsurf .ten. + Fpressure = 0 film-straining model to predict the effect of thin water films on colloid tra n s p o rt in part i a l ly saturated porous media. Tdrag + Tfric = 0 (2) Experimental results confirmed model predictions that the movement of colloids is inhibited if the particle sizes are larger The drag forces acting on the particle are corrected for the presthan the film thickness. However, the mechanistic understanding ence of the solid boundary beneath the particle. The effect of of the effect of the characteristics of films, particles and medium free interface above the particle is neglected, limiting the validsurfaces on particle motion is still lacking. The objective of this ity of the model when the free surface influence on particle research is to provide the fundamental understanding of particle motion is significant. Particle roughness, expressed as a percentmovement in thin water films, upon which the predictive modage of particle size, and friction coefficient between particle and els of unsaturated colloid transport will be based.To address this glass plate are used as the adjustable fitting parameters in the issue, a study is undertaken to examine the effect of particle size model.The estimation of surface tension and pressure force for a (dp)/film thickness (ho) ratio on the motion of a sphere in a liquid film flowing down an inclined flat surface. particle partially submerged in a film requires knowledge of capillary rise hc and angle Îą (Figure 1). Although these two paramApproach eters can be estimated by calculating capillary rise profile using the Young-Laplace (Y-L) equation, we believe that the interface Theoretical considerations: The dominant fo rces and profiles for the physical parameters of this study may have sigtorques acting on a particle completely submerged in a flowing nificantly larger radius of curvature than predicted by Y-L equaliquid film and moving in contact with a smooth plane surface tion. Model predictions are therefore limited to particles smaller are fluid drag, friction, lift, buoyancy and gravity. For a partially than film thickness (i.e., dp/hc <1).

Srinivas Veerapaneni, Jiamin Wan and Tetsu K. Tokunaga

Figure 1. Forces and torques acting on a particle in a flowing liquid film.

23


ESD

Experiments: A liquid film is generated on an incl i n e d smooth glass plate with fi l m t h i cknesses ranging from 140 to 700 Âľm under steady-state flow. Motions of spheri c a l hy d rophilic part i cles in the range of 20 to 800 Âľm in diameter were re c o rded by a CCD c a m e ra attached to a long working distance microscope. Film velocity and thickness and part icle size were directly measured. A schematic of the experimental setup is shown in Fi g u re 2.

Particle Motion in Film Flow

Figure 2. Schematic of the experimental setup.

Results

the velocity of the part i cle.This may be attributed to the surface tension acting on the particle along the three-phase contact line. The proximity of the f ree interface to the part i cl e surface is also likely to infl uence the motion of the part i cle. It is interesting to note that t h e re appears to be a critical range of dp/ho ratio when gravity begins to negate the decrease in part i cle ve l o c i t y caused by surface tension. (4) When the particles are larger than the film thickness, the influence of gravity force i n c reases with part i cle size, resulting in increased particle velocities with size.

Results from a typical ex p e riment are shown in Figure 3.The Significance velocities of the part i cles, n o rof Findings malized with maximum fluid velocity at the undisturbed airResults from this study indiwater interface, are plotted as a cate that particle motion in film function of part i cle size, n o rm a lflow is strongly dependent on ized with the film thickness the particle size/film thickness (dp/ho). Model predictions for particles smaller than film thickratio.When particles are smaller ness are also shown in the fi g u re than film thickness, significant as a solid line. Four re gions can t ra n s p o rt of part i cles occurs Figure 3. Normalized particle velocity as a function of particle size. be identified in the fi g u re as diswith flow, aided by fluid drag. cussed below. When particles are comparable (1) When particle diameters are smaller than 50% of the film to film thickness, surface tension may retard particle motion conthickness, particle velocity increases nearly linearly with parsiderably. Motion of particles larger than film thickness may be ticle size, reflecting the motion of the particle in a flow field aided by gravity. characterized by constant fluid shear (linear velocity profile) Related Publications that is prevalent at these depths. (2) When the particle size is in the range of 50-100% of film thickVeerapaneni, S., J. Wan and T. K.Tokunaga, Motion of particles in ness, the increase in the velocity of the particle with its size film flow, Environ. Sci.Technol., submitted. is relatively low, compared to Region I. This is due to the decrease in the fluid shear rate, as the velocity profile in this Funding region is nonlinear. The velocity of the particle peaks when its diameter is close to the film thickness. The measured This work has been supported by the Office of Science, Office velocities of the particles in regions I and II (i.e., dp/ ho 1) agree well with model predictions, as indicated in Figure 3. of Basic Energy Sciences, Division of Engineering and (3) When part i cle size is comparable to or slightly greater than the Geosciences, of the U.S. Department of Energy under Contract film thickness, it is observed that there is a significant drop in No. DE-AC03-76SF00098.

http://www-esd.lbl.gov 24


Earth Sciences Division

Research

Fundamental and Exploratory Research Program

Objectives

L ab o ra t o ry fl ow experiments on transparent fra c t u re replicas perm i t direct visual investigation of fl ow p rocesses and identification of controlling mechanisms of large r - s c a l e phenomena that are critical in many areas. These include flow and transp o rt around repositories for high-level nu clear wastes, contaminant migration and remediation and enhanced p e t roleum re c ove ry. Fra c t u re fl ow visualization ex p e riments have been conducted in parallel rough surfa c e glass plates and fracture casts made with epoxy resin. Rough glass plates do not represent actual natural fractures, but can provide wa t e r - wetting ch a ra c t e ristics similar to many rock surfaces. On the other hand, e p ox y fra c t u re casts can provide re a s o n ably accurate reproductions of natural fract u re aperture fi e l d s , but have hydrophobic surfaces. For some purposes it is desirable to combine the wetting ch a ra c t e ristics of glass with the fra c t u resurface topography reproduction obtained through casting. Glass casts have the combined adva nt ages of closely reproducing natural f ra c t u re surface roughness and of being tre a t able to provide the wide range of we t t abilities found for natural rock surfaces. H e re we present a new method for replicating natural rock surfaces using molten glass.

Glass Casts of Rock Fracture Surfaces: A New Tool for Studying Flow and Transport Jiamin Wan, Tetsu K. Tokunaga, Thomas R. Orr, Jim O’Neill and Robert W. Connors

(510)

Contact: Jiamin Wan 486-6004, jmwan@lbl.gov

Figure 1. Photographs of the surface of a granite fracture (a), and its glass cast (b). Both are 120 x 160 mm.

Annual Report 1998-1999

burning off the wax positives. In step four,glass fracture casts (positives) are made from the investment molds.The final step entails finishing the glass fracture replicas. Five sides (4 edges and back) of each half cast are ground and polished as a pair. The topography of a glass cast surface was compared with that of the o ri ginal rock fracture using laser profilometers and an atomic force microscope (AFM). For coarse-scale measurements, surfaces were pro filed with an LK-081 CCD laser displacement sensor (Keyence Corp., Wo o d cl i ff Lake, N.J.) mounted on a Vi ew Precis 3000 coordinate measuring machine (Vi ew E ngineering Division, G e n e ra l Scanning, Inc., Simi Valley, Calif.). For higher resolution measurements over smaller areas and narrower topographic ranges on rock and glass casts, a UBM laser (UBM, Sunnyvale, Calif.) was used.The UBM laser has a spot size of 1 µm, a z measurement range of 100 µm, and a z resolution of 0.06 µm.This system was also used to obtain surface pro files on the roughened quartz glass samples. For higher resolution topography, small areas of glass casts were scanned with an AFM (Autoprobe M5, Park Scientific Instruments, Sunnyvale, Calif.). We used this AFM in the standard contact scanning mode to obtain info rmation on finer-scale topography and also to obtain values of surface roughness for comparison with laser profilometry results.

Approach Results The glass fracture pairs reported in this paper were cast from two different original rock fractures. The larger (120 x 160 mm) glass fracture cast replicated a Swedish gra n i t e . The smaller (70 x 70 mm) cast replicated a fractured gabbro (Dixie Valley,Nev.). The casting procedure involves five steps. In the first step, silicone rubber molds (negatives) are made of each rock slab. In step two, wax patterns (positives) are made from the silicone rubber molds. In step three, investment molds (negatives) are made by

Figure 2. Comparisons of laser surface profiles on the granite fracture and glass cast. These (a) 100 mm and (b) 10 mm long profiles were measured in 50 µm steps.

25

One set of casts will be discussed with respect to each characterization of the finished products.The simplest characterization is that of visual comp a risons between an ori ginal ro ck fracture surface and its corresponding glass cast. A photograph of one side of the granite fracture (Figure 1a) and its corresponding finished glass fracture cast (Figure 1b) qualitatively shows reproduction of fracture surface texture and roughness. Reflected light was used for photographing the rock,


ESD

Glass Casts of Rock Fracture Surfaces: A New Tool for Studying Flow and Transport

while only transmitted light photograaccurate reproduction of surface wettaphy was suitable for the glass cast bility is a very important factor. Epoxy (because of strong reflections and glare cast surfaces are quite hy d ro p h o b i c . under reflected light). Due to differWater “wets” dry epoxy surfaces with ences in lighting, these photographs do contact angles of about 90˚ (Figure 3a). not provide a good comparison of the In contrast, clean glass cast surfaces prodetailed surface textures. Diffe rences vide excellent wa t e r - we t t ab i l i t y, with between casts and original rock surnear zero contact angle (Figure 3b). faces were quantified through surface Previously mentioned techniques can profile measurements. In Figure 2a, typibe used to alter glass surface wettabilical coarse-scale (100-mm line scans, 50ties. µm in-line steps, at 60-mm lateral line Significance separation) laser profiles of a rock surof Findings face and glass cast are compared. The root mean-squared roughness (rm s r ) A method for casting tra n s p a re n t values of the cast along these two proglass replicas of ro ck fra c t u res wa s files (2.22 and 1.14 mm) are very similar deve l o p e d . The glass casts obtained to that of the rock original (2.25 and using this method provided close repro1.04 mm). However, since the rmsr is duction of major features of natural fracstrongly influenced by larger amplitude ture topography.The glass casts are genfe a t u res encountered at larger diserally more water-wettable than epoxy tances, comparisons over shorter intervals are needed to obtain more direct Figure 3. A comparison of surface wettability between casts, and can also be treated to exhibit an epoxy and a glass fracture cast. (a) Drop of water on information on replication of finer-scale fracture surface of an epoxy cast, where the contact specific desired wettabilities. Thus, for features. Example profiles of the frac- angle is ∼90 degrees. (b) Drop of water on fracture sur- visualization studies of multiphase fluid ture and cast surface over shorter dis- face of a glass cast, where contact angle is near zero. environments in fractures, glass casts are suitable to use for a wide range of natutances (10-mm lengths) are shown in ral fracture surface wettabilities.They can be used to study mechFigure 2b. Note that the rmsr values are substantially lower than anisms controlling multiphase fluid flow and contaminant transin the longer profile scans, and that rock and glass casts still yield port (including solutes, colloids, microorganisms and NAPLs), to similar rmsr values, ranging from 0.63 to 0.84 mm. examine remediation techniques at small scale and to conduct Although fair reproduction of individual surface topography is laboratory studies for enhancing petroleum recovery redistribuachieved in glass casts, mated pairs can still exhibit fairly wide tion seen by the active testing data. ranges in average aperture and saturated transmissivity.The average aperture and transmissivity of the granite fracture pair under Related Publication 5.7 kPa average stress were 122 µm and 6.1 x 10-7 m2 s-1, respectively.Three glass casts of this granite fracture pair were molded. Wan, J., T.K. Tokunaga, T. Orr, J. O’Neill and R.W. Connors, Glass The average apertures of three different glass cast pairs were 75, 173 and 255 µm under the same ave rage normal stre s s . casts of rock fractures: A new tool for studying flow and Transmissivities of these three glass cast pairs were 2.1 x 10-7, 1.0 transport,Water Resour. Res., in review. x 10-6, and 1.9 x 10-6 m2 s-1, respectively. All average apertures Funding and transmissivities were determined with a relative uncertainty of ±5%. This work has been supported by the Office of Science, The artifact of slight bowing of one surface relative to its Office of Basic Energy Sciences, Division of Engineering and opposing surface tends to increase the cast transmissivity, and Geosciences, of the U.S. Department of Energy under Contract this effect will be more problematic with larger area casts. For No. DE-AC03-76SF-00098. The authors thank Dr. Peter Persoff purposes of observing flow and transport under a given cast (LBNL) for providing the Dixie Valley rock fractures and epoxy aperture field, this is not problematic.When transparent fracture casts and for helpful internal review comments. replicas are used in studies of multiphase fluid statics and flow,

http://www-esd.lbl.gov 26


Earth Sciences Division

Research

Fundamental and Exploratory Research Program

Objectives

A Model for Non-Isothermal Multiphase Multi-Species Reactive Chemical Transport In Porous and Fracture Media

Annual Report 1998-1999

ently for each component, whereas the reaction equations are solved on a Coupled modeling of subsurface gri d bl o ck basis using a Newtonmultiphase fluid and heat flow, solute Raphson iteration. An improved equitransport and chemical reactions can librium-kinetics speciation model for Tianfu Xu, Karsten Pruess, be used for the assessment of acid Eric Sonnenthal, Nicolas Spycher simulating water-rock-gas interaction and George Brimhall mine drainage remediation, waste disis used. Quasi-stationary approximaposal sites, hydrothermal convection, tion and an automatic time stepping Contact: contaminant transport and groundwas cheme are implemented in Tianfu Xu ter quality.We have developed a comTO U G H R E AC T. The code was fi rs t (510) 486-7057, tianfu_xu@lbl.gov pre h e n s i ve nu m e rical simulation developed on a PC and then tested on model, TOUGHREACT, which considers nonisothermal multiVAX and UNIX systems. Later, the simulator was ported to the component chemical transport in both liquid and gas phases. A Cray T3E at the National Energy Research Scientific Computing wide range of subsurface thermo-physical-chemical processes is Center (NERSC) at Berkeley Lab, and a parallelized version was considered. The model can be applied to one-, two- or threedeveloped, resulting in significant improvement of computing dimensional porous and fracefficiency. tured media with physical and Results chemical hetero ge n e i t y. The model can accommodate any The model was extensively number of chemical species verified and validated for a wide present in liquid, gas and solid range of subsurface physical and phases. A variety of equilibrium chemical processes. Four applichemical reactions is considcations were carried out using ered, such as aqueous complexTOUGHREACT. Here we report ation, gas dissolution/ex s o l utwo applications. The fi rst is tion, cation exchange and sursupergene copper enrichment, face complexation. Mineral diswhich involves oxidative weathsolution/precipitation can proering of pyrite (FeS2) and chalceed either subject to local copy rite (CuFe S2) and associequilibrium or kinetic condiated acidification, which causes tions. mobilization of metals in the Approach unsaturated zone, with subsequent formation of enriched ore The coupled model is impledeposits chalcocite (CuS) and mented by introducing reactive covellite (Cu2S) in the reducing conditions below the wa t e r chemistry into the framework table (Figure 1). A total of 52 of the existing nonisothermal aqueous species, 10 primary multiphase flow code TOUGH2 minerals and six secondary min(Pruess, 1991), resulting in the erals are considered. The aquegeneral reactive chemical transous complexation and gas dissoport code TOUGHREACT. Our lution are assumed at equilibmodel uses a sequential iterari u m . M i n e ral dissolution and tion approach, which solves the precipitation are subject to t ra n s p o rt and reaction equakinetics. The alteration of pritions separately. Flow and transmary minerals and the developport in ge o l o gic media are based on space discretization by ment of secondary minerals premeans of integral finite differdicted by our model are consisences. An implicit time-weighttent with observations in supering scheme is used for flow, gene copper deposits in the transport and geochemical reac- Figure 1. Simulation of a one-dimensional supergene copper enrichment Atacama Desert, Northern tion. The chemical tra n s p o rt system. Top: schematic representation; bottom: resultsâ&#x20AC;&#x201D;change of min- Chile. equations are solved independ- eral abundance after 100 years (positive indicate precipitation). The second application con27


ESD

A Model for Non-Isothermal Multiphase Multi-Species Reactive Chemical Transport in Porous and Fracture Media

Primary Species: H2O H+ Ca2+ Mg2+ Na+ K+ SiO2(aq) HCO3SO42ClAlO2-

Secondary Species: CO2(aq) CO32OHAl3+ Al(OH)2+ HAlO2 AlOH2+ CaCl+ CaCl2(aq) CaCO3(aq) CaHCO3+ CaSO4(aq) HSiO3-

HCl(aq) KCl(aq) KHSO4(aq) KSO4MgCl+ MgCO3(aq) MgHCO3+ MgSO4(aq) NaCl(aq) NaCO3NaHCO3(aq) NaHSiO3

Minerals and Gas: anhydrite calcite quartz cristobalite amor. Silica microcline albite kaolinite illite sepiolite smectite-Na smectite-K smectite-Ca smectite-Mg

Table 1. Aqueous species, minerals and gas considered in the simulation.

sists of predicting thermal, hydrological and chemical processes induced by emplacement of a strong heat source in unsaturated fractured rocks to simulate a high-level nuclear waste repository. A dual permeability model was used. Aqueous species, minerals and gas considered in the simulation are shown in Table 1. Mineral dissolution and precipitation reactions proceed according to kinetics. Preliminary modeling results (Figure 2) indicate the importance of considering hy d ro chemical interactions between fracture and matrix for this type of system.The simulations are useful to investigate mineral dissolution and precipitation under boiling conditions in fractured unsaturated rock.

Significance

of

Figure 2. Hydro-Thermo-Chemical modeling for drift scale heater test. Top: closeup view of mesh around heater drift; bottom: resultsâ&#x20AC;&#x201D;change of mineral volume.

SPY289M4, Berkeley Lab, MOL.19981130.0132, 1998. Xu,T., K. Pruess and G. Brimhall, An improved equilibrium-kinetics speciation algorithm for redox reactions in variably saturated flow systems, Computers & Geosciences, in press. Xu, T., S.P. White and K. Pruess, Modeling of pyrite oxidation in s a t u rated and unsaturated subsurface fl ow systems, Transport in Porous Media, in press. Xu,T., and K. Pruess, Coupled modeling of non-isothermal multiphase flow, solute transport and reactive chemistry in porous and fractured media: 1. Model development and validation, American Journal of Science, submitted. Xu, T., E.L. Sonnenthal, N. Spycher, K. Pruess and G. Brimhall, Coupled modeling of non-isothermal multiphase flow, solute transport and reactive chemistry in porous and fractured media: 2. Model applications, American Journal of Science, submitted.

Findings

The model is well suited for flow and reactive transport in variably saturated porous and fractured media. Major features of the model include heat driven fluid flow and effects on chemical reactions, and gaseous species transport and interaction with the aqueous phase.The capabilities of the model have been illustrated with a few examples. The full potential is yet to be explored.

Funding Related

Publications This work was supported by the LBNL Laboratory Directed Research and Development Program and by the Director, Office of Civilian Radioactive Waste Manage m e n t , through M e m o randum Purchase Order EA9013MC5X between TRW Environmental Systems, Inc., and the Ernest Orlando Lawrence Berkeley National Laboratory under U.S. Department of Energy Contract No. DE-AC03-76SF00098.

Pruess, K., TOUGH2: A general numerical simulator for multiphase fluid and heat flow, Berkeley Lab report LBL-29400, 1991. Sonnenthal, E., N. Spycher, J.Apps and A. Simmons,Thermo-hydrochemical predictive analysis for the drift-scale heater test, Yucca Mountain Project Level 4 Milestone

http://www-esd.lbl.gov 28


Earth Sciences Division

Fundamental and Exploratory Research Program

Annual Report 1998-1999

The Center for Isotope Geochemistry The Center for Isotope observed and simulated climate feaGeochemistry (CIG) was established tures suggests that the RCMS is capaDonald J. DePaolo in 1988 with three major goals. The ble of long-term regional climate simfirst is to maintain a state-of-the-art ulation. Another project couples genContact: Donald J. DePaolo analytical facility for the measurement eral circulation models, which simu(510) 643-5064, djdepaolo@lbl.gov of the concentrations and isotopic late global climate change scenarios, compositions of elements in rocks, with Califo rnia Department of minerals, fluids and gases in the earth’s crust, oceans and atmosForestry wildfire models to predict the impact of future climate phere.The second is to develop new ways of using isotope ratio change on the occurrence and magnitude of wildfires at a local measurements to study earth processes. This involves improvescale.The average insured cost of wildfires in the United States is ments in analytical techniques, as well as exploration of the natabout $300 million dollars per year. Insurers and climatologists ural isotopic variations of key elements and development of conhave long known that fire danger is intimately linked to climate. ceptual models relating isotopic variations to earth processes. For instance, local and regional droughts linked to the recent El The third objective is to apply well-known isotopic and chemiNiño led to devastating fires in Florida, Indonesia and elsewhere. cal approaches to specific energy and environmental problems, In most cases, climate change driven by a two-fold increase in such as groundwater contamination and remediation, geotheratmospheric carbon dioxide would lead to dramatic increases in mal resource development, nuclear waste isolation and global cliboth the area of land burned by California wildfires and the nummate change. ber of potentially catastrophic fires. The Center’s lab facilities include laboratory and field equipIn many natural systems, fracture permeability exerts a dominant influence on fluid flow. A simple model is under development for oceanographic studies, a stable isotope laboratory, a ment to test the feasibility of using the isotopic compositions of noble gas laboratory, a cosmogenic isotope laboratory, a soil carelement pairs in fluids to constrain fracture-matrix geometry and bon laboratory and an analytical chemistry laboratory.There is a spacing. Fluids acquire heat from matrix blocks by conduction thermal ionization mass spectro m e t ry lab o ra t o ry on the and solubility and ionic diffusivity limits the exchange of chemiUniversity of California’s Berkeley campus. In the coming year cal and isotopic constituents. Therefore, the thermal, chemical the Center will obtain a multiple-collector, magnetic sector mass and isotopic evolution of fluids flowing through fractured rock spectrometer with a plasma ion source, which will greatly depends strongly on fracture geometry and spacing. Reservoir expand its analytical capabilities. ESD researchers have estabmodeling relies on geometric information from rock outcrops lished the Regional Climate Center (RCC) and are associated and core, but the geometry of the reservoir fractures carrying with the Center for Isotope Geochemistry to take advantage of the bulk of the fluid is generally not known.The sensitivity of isothe overlap in research interests. The RCC specializes in a topic ratios to matrix block size (or average fracture spacing) is Regional Climate System Model (RCSM) that downscales global related to the differing solubility and diffusivity of the elements. model information to provide research information, climate predictions and impact assessments on a regional scale.The model The degree to which a matrix block is isolated from a fracture can be applied to weather forecasting, soil water content, river fluid decreases with increasing solubility and diffusivity. flow, hydrology at a waterR e s e a rchers measure d shed scale, climatic trends, helium, carbon and oxygen water re s o u rces, crop isotopes in samples fro m responses and ecological fault zone gouge, breccia and and environmental impacts. host rock of the San Andreas A pressing question in cliFault. Their study confirms matology is what the impact that a significant fraction of of future climate change will the helium in the fault zone be on regional and local fluids has come from the scales. Climate researchers, mantle and is accompanied using a re gional cl i m a t e by deep crustal or metamorhindcast, are evaluating the phic water and carbon dioxeffectiveness of the Regional ide. This supports their earClimate System Model and lier work that suggested fluits various components, such ids are entering the fault as the Mesoscale from the mantle and acting Atmospheric Simulation and to lubricate the fault, which S o i l - P l a n t - S n ow models, to would explain the we l l reproduce western U.S. cliknown dearth of friction on Figure 1. Mono Lake, California. Photo by Roy Kaltschmidt, LBNL. mate. Agreement between the San Andreas Fault. In a 29


ESD

The Center for Isotope Geochemistry

project with researchers from the U.S. Geological Survey and Oxbow Geothermal, it is shown that noble gas concentrations, water isotopes and chloride concentrations in geothermal production streams provide a quantitative measure for tracing the return of injectate to geothermal reservoirs. The rate at which cooler injectate fluids invade a production reservoir is extremely important for establishing injection programs and constraining future reservoir models. Other research has shown that co-variations between helium and neodynium isotopes in continental basalts can be used to differentiate between magma sources or chambers and to assess the present rate of magma chamber recharge with new mantle material.This provides valuable information for assessing volcano hazards and the potential of a region for geothermal energy development. The drift-scale heater test at Yucca Mountain, Neva d a , the proposed nu clear waste repository site, is being conducted to test the effect of heat generated by the stored nu clear waste. Researchers a re monitoring the time evolution of the CO2 carbon isotopic composition in gas released during the heating test.The changing isotopic composition will help quantify factors such as the degree of porewater degassing and identify zones of calcite deposition within the therm a l ly disturbed re gion which bear dire c t ly on fluid movement and changing perm e ability of the system. The Environmental Measurements Laboratory (EML) consoli-

dates the inorganic and organic chemical analytical facilities of the Earth Sciences Division.The EML provides chemical characterization of soil, rock, mineral and fluid samples for many researchers and projects within the division and the UC campus. The EML is equipped with state-of-the-art instrumentation, including ICP-MS with laser ablation capabilities, atomic adsorption spectrometry, GC/MS, HPLC, and facilities for standard wet chemical analysis.

Funding The research of the Center for Isotope Geochemistry has been supported by the Office of Science, Office of Basic Energy Sciences, Materials Sciences and Engineering and Geosciences divisions; the Environmental Science Management Program; the Office of Energy Efficiency and Renewable Energy, Office of Geothermal Technologies; the U.S. Environmental Protection Agency, via Energy Efficiency and Renewable Energy, Office of Building Technologies and State Community Programs of the U.S. Department of Energy. Research has also been supported by the U.S. Navy; the Office of Space Science of the National Aeronautics and Space Administration; the Office of Polar Programs of the National Science Foundation; and the LBNL Laboratory Directed Research and Development Program.

http://www-esd.lbl.gov 30


Earth Sicences Division

Research

Fundamental and Exploratory Research Program

Objectives

Molecular Modeling of Clay Mineral Surface Geochemistry: Hydrated Cesium-Smectites

Annual Report 1998-1999

Center (NERSC) at Berke l ey Lab. Molecular dynamics simulations were obtained with the code MOLDY, w ri tten by Keith Refson and compiled on the Cray T3E at NERSC.

The 137Cs+ cation, a by p roduct of Rebecca Sutton, nuclear fission reactions, has held the Garrison Sposito, Sung-Ho Park status of a significant env i ronmental and Jeffery A. Greathouse contaminant since the late 1940s. An Results u n d e rstanding of hydrated Cs-smectite Contact: interlayer systems is necessary to preRebecca Sutton (510) 643-9951, An iterative MC method was used dict the perm e ability of clay liners at sutton@nature.berkeley.edu to determine the number of water nuclear waste containment facilities to 137Cs+ ra d i o a c t i ve waste. Detailed molecules present in stable Cs-smecex p e rimental characterization of Cs-smectite interlamellar propertite systems with layer spacings of ~12 Å. Previous experimental ties has proved difficult because of the great degree of disorder work indicated that a water content of 0.3-0.7 monolayers, or present in these clays.As a result, the bulk diffusion experiments 0.03-0.07 kg water/kg clay, would suffice. Once we had estabperfo rmed to study these systems are difficult to interpret in term s lished the water content of a stable Cs-smectite system, we were of molecular interactions.N ew info rmation provided by solid-state able to collect thermodynamic information such as layer spacNMR spectro s c o py of 133Cs+ ing, total potential energy, density adsorbed to mineral surfaces adds profiles and atom-atom radial disdetail to the picture of Cs-cl ay tribution functions. molecular interactions slowly fo rmWe then perfo rmed 800 ps MD ing, a picture we can further elucicalculations on the MC-equilidate using molecular modeling b rated clay systems in order to coltechniques. We have used Monte lect tra j e c t o ry info rmation and Carlo (MC) simulations to deterlearn about the motions of cations mine the configuration of water and water in the interlaye r. Comparison of Cs-smectite sysmolecules and Cs+ cations present in stable Cs-smectite structures.The tems with the same water content clays we have examined include but diffe rent cl ay minera l s beidellite, a smectite with negative revealed the dramatic effect of clay ch a rge sites from isomorphic subch a rge sites on Cs+ mobility. The t e t ra h e d rally-charged beidellite stitution present only in the tetrasystem and the octahedra l lyhedral sheet, montmorillonite, a charged hectorite system both smectite with both tetra h e d ral and held Cs+ in a nearly fi xed location octahedral charge sites, and hec(Figure 1).The specific cation locatorite, a smectite with ch a rge sites tions va ried with the type of clay only in the octahedral sheet. ch a rge site, with the near-surface C o m p a rison of these clay systems tetra h e d ral ch a rges drawing the allows us to determine the effect of Cs+ closer to the surrounding clay clay charge site location on their laye rs , while the more distant octathermodynamic properties. h e d ral ch a rge held all of the Cs+ at Subsequent molecular dynamics the midplane of the interlaye r. (MD) ex p e riments on the MC-equilH oweve r, when these two types of ibrated Cs-smectite systems reveal the trajectories of interlayer ions ch a rge sites were combined in and water over time. montmorillonite, we did not observe intermediate behavior in Figure 1. Y-Z coordinates of Cs+ cations collected every 0.1 ps for Approach the location of the cations. 400 ps of MD simulation of Cs-smectite systems with 0.3 monolayers I nstead, the Cs+ within montmori lof water. The z-axis is the same as the crystallographic c-axis of the Monte Carlo calculations were clay mineral, with the midplane of the interlayer region at 0 and the y- lonite was mu ch less tightly bound perfo rmed using the code MONTE, axis is +the same as the b-axis. The six colors represent the six differ- to a particular location, showing a ent Cs in each simulation cell. Cations near the edge of the simulaw ritten by Neal Skipper and Keith tion cell can be seen moving into the neighboring cell, resulting in two much greater ease of movement. A closer examination of the Refson and compiled on the Cray J- clusters of position on either side of the graph shown. The surround90s at the National Energy ing clay surfaces have oxygen center of mass coordinates at ± 2.85 motions exhibited by Cs+ within Å, ± 2.95 Å, and ± 2.93 Å along the z-axis for Cs-beidellite, Cs-montsmectite systems reveals very litResearch Scientific Computing morillonite, and Cs-hectorite, respectively. 31


ESD

Molecular Modeling of Clay Mineral Surface Geochemistry: Hydrated Cesium-Smectites

clusters. Black points represent the surface oxygens of one of the clay mineral surfaces. Below the trajectory data lie visualizations of the clay system, with the simulation cell outlined in black. Green spheres represent Cs+ ions, while blue spheres are water O and red spheres are clay surface O. Small white spheres represent water H and small grey spheres represent clay mineral Si.

Figure 2. X-Y trajectories of Cs+ cations and a selected water molecule over 400 ps of MD simulation. The x-axis is the same as the crystallographic a-axis of the clay mineral and the y-axis is the same as the b-axis. Two systems are shown: Cs-hectorite with 0.3 monolayers of water and Cs-hectorite with 0.7 monolayers of water. In the latter system, a Cs+ cation (in brown) hovers at the edge of the simulation cell, jumping from one side to the next, creating two trajectory

Related

tle true diffusion over the 800 ps time span examined. Instead, the Cs+ hovers near clay charge sites, though it is capable of jump diffusion, especially in the wetter systems (Figure 2). Water is much more mobile than the ions in the interlayer, exhibiting conventional diffusion behavior. Interlayer water molecules tended to diffuse more rapidly in wetter clay systems.These polar molecules were attracted both to the interlayer cations, forming hydration spheres around them, and to the charged clay surface.

Significance

of

Publications

Chang, F.-R.C., N.T. Skipper, K. Refson, J.A. Greathouse and G. Sposito, Interlayer molecular structure and dynamics in Li-, Na-, and K-montmorillonite-water systems, in Kinetics and Mechanisms of Reactions at the Mineral/Water Interface (D. L. Sparks and T. Grundl, eds.), American Chemical Society, Washington, D. C., in press. Sposito, G., N.T. Skipper, R. Sutton, S.-H. Park, A.K. Soper and J.A. Greathouse, Surface geochemistry of the clay minerals, Proc. Nat.Acad. Sci., 96, in press. Sposito, G., S.-H. Park and R. Sutton, Monte Carlo simulation of the total radial distribution function for interlayer water in sodium and potassium montmorillonites, Clays Clay Miner., 47, in press.

Findings

The results from these ex p e riments confirm the findings of bulk d i ffusion experiments, which predict a low mobility of Cs+ in clays due to their retention near clay ch a rge sites. Such agreement indicates that the potential functions used to describe the Cs-smectite system may be re a s o n ably accurate, and provides further evidence supporting the use of clay liners within nuclear waste containment facilities in order to retard the movement of 137Cs+. Cs-smectite molecular simulations can also aid an understanding of basic ge ochemistry, t h rough comparison of these simulations with those of other alkali metal-smectite systems in order to identify trends ex h i bited by alkali metal cations.

Funding This work has been supported by the Office of Science, Office of Basic Energy Sciences, Division of Engineering and Geosciences of the U.S. Department of Energy under Contract No. DE-AC03-76SF00098.

http://www-esd.lbl.gov 32


Earth Sciences Division

Research

Fundamental and Exploratory Research Program

Objectives

Isotopic Effects in Dual-Porosity Fluid-Rock Systems

Annual Report 1998-1999

1/ 2

Ê ˆ D i f mr f ˜ L e,i = Á R (1 f )r K Ë m s r /f ¯

The thermal and chemical behavior Donald J. DePaolo of fractured rock systems depends to Where φ is matrix porosity,ρ is dena substantial degree on the ave rage sity, and Di is the ionic diffivity. For Contact: two elements of differing solubility, spacing between fractures. Fluid movDonald DePaolo (510) 495-2228, djdepaolo@lbl.gov the ratio of the reaction lengths is: ing through a system of fractures i n t e racts therm a l ly and ch e m i c a l ly 1/ 2 L e, i Ê D i K r / f ( j ) ˆ with the matrix blocks between them mainly by heat conduc˜ =Á tion and diffusion of chemical constituents dissolved in the pore L e ,j Ë D jK r / f (i ) ¯ fluid or vapor phase.The spacing of fractures can be estimated For oxygen versus strontium, for example, this ratio is typically in some cases from rock outcrops and cores, but the spacing of about 10 to 100. If the average block size falls between Le,Sr and the fractures actually carrying the bulk of the fluid is not usually Le,O, then the isotopic effects in the fracture fluids provide inforknown directly. mation on the matrix block size. Isotopic ratios of certain pairs of elements dissolved in fluids In cases where the reaction length is very large relative to the can be used in theory to measure the effective matrix block size matrix block size, a steady-state system behaves chemically as if (or average fracture spacing) in some fractured rock systems.The it had a single-porosity, and reaction effects on the fluid are research described here was aimed at producing a simple theory determined mainly by the reaction rate, R, which describes the that relates isotopic ratios in fluids and rocks to fracture spacing, solution-precipitation rate averaged over the minerals in the and at using available data to assess whether the expected rocks.The advective reaction length, which can be measured in effects are present in natural systems. the field, in this case is:

Approach

L e ( single ) =

v f f rf R eff (1 - f )r s K r / f

The sensitivity of isotopic ratios to matrix bl o ck size stems f rom the diffe ring solubilities of the elements, w h i ch can be In general, the effective reaction rate as inferred from the expressed in terms of a ro ck/fluid concentration ratio (Kr/f), effects of water-rock interaction on the fluid moving through the fractures in a dual porosity sysand diffe ring fluid-phase ionic tem can be shown to be related diffusivities. For example, Kr/f for ox y gen is about 0.8, to the actual reaction rate ® in w h e reas Kr/f for Sr and C is typthe matrix (with some simplifyically ~10 to 1000. The degre e ing assumptions) by: to which the cores of matri x 2 L2 2 2 L -1 bl o cks are ch e m i c a l ly isolated R eff = 8 R 2e  (1 + n p 2e ) L b n odd Lb f rom the fra c t u re fluid is less for high solubility elements ( l a rge Kr/f) than for low solubilFor Le < L b, then Re ff R(Le/L b). ity elements. As long as the bl o ck dimension The degree of isolation of is larger than the re a c t i o n matrix bl o cks can be deterlength, diffusion in the matrix mined from the diffusive reacbl o cks re t a rds chemical and isotion length (Le), which depends topic ex ch a n ge between the on the fluid-rock reaction rate ro cks and the fluids by the fa c(=R) and the effective diffusivity tor Le/Lb. for a dissolved element in Results matrix pore fluid. If this reaction length is smaller than the The model was applied to average block dimension (Lb), then the interiors of the blocks published data from mid-ocean are not in equilibrium with the ridge hy d ro t h e rmal systems, f ra c t u re fluid. The re a c t i o n Figure 1. Calculated dual porosity effect for fluids moving in fractures and where Sr and O isotopes have length for element “i” that is communicating with matrix blocks by fluid-phase diffusion. R is the solu- been measured (although they tion-precipitation time constant in the matrix blocks, and R eff is the applicable to the matrix blocks apparent reaction time constant that is sensed by fluid moving in the were not reported on the same is given by: samples). These systems involve fractures under steady-state conditions. 33


ESD

Isotopic Effects in Dual-porosity Fluid-Rock Systems

c i rculation of seawa t e r hy d ro ge n , helium, b o ro n , t h rough the basaltic rocks of carbon, sulfur, lead, urathe oceanic cru s t . Reaction nium, neodymium, thorium is believed to take place and radon. mostly at a temperature of In non-steady systems, about 350°C, so these sysfor example those with tems are similar to some time-varying fl ow, the c o m m e rcial geothermal sysresponse time for re-estabtems. Fluids exit the ve n t s lishment of a new steady with somewhat lower temstate between the fracture peratures. A number of fluids and the matrix blocks assumptions need to made, will also vary element by but several important para melement. This effect could e t e rs are well known for this also be used to estimate the example — the isotopic effective matrix block size compositions of the initial or porosity. It is theoretiand final fluid, the isotopic cally possible to constrain composition of the ro ck s , all of the parameters if isoand the concentrations of tope ratios of three or four the elements in both ro cks elements are measured conand fluids. The reaction rate Figure 2. Calculated relationship between shifts in 87Sr/86Sr of Sr in hydrothermal currently. is less well constrained; it is fluids and shifts of δ18O. The distribution of available data suggests that fracture spacings are 2 to 4 times the value of LSr. Related assumed to be 10-3 yr-1. Publications The model calculations suggest that fracture spacings in the MOR hydrothermal systems Johnson,T.M., and D.J. DePaolo, Interpretation of isotopic data in are in the range of 2 to 8 times the reaction length for Sr.The latgroundwater-rock systems: Model development and applicater is estimated to be about 10 cm, so the fracture spacings are tion to Sr isotopic data from Yucca Mountain, Water estimated to be 20 to 80 cm. The estimated fracture spacing is Resources Res., v. 30, pp.1571-87, 1994. proportional to the estimated reaction rate, so if the reaction rate DePaolo, D.J., and S.R. Getty, Models of isotopic exchange in reacis lower, the fracture spacings could be larger. tive fluid-rock systems: Implications for geochronology in Significance of Findings metamorphic rocks, Geochim. Cosmochim.Acta v. 60, no. 20, pp. 3933-47, 1996. It is normally assumed that fluid-rock systems behave as sinFunding gle-porosity systems, or that dual-porosity effects are the same for all elements. If the model results above are applicable, then This research was supported by the Office of Science, Office instead it may be possible to use isotopes of multiple elements of Basic Energy Sciences, Division of Engineering and to infer important aspects of the structure of geothermal and Geosciences of the U.S. Department of Energy under Contract groundwater systems. Other elements that have variable natural No. DE-AC03-76SF00098. isotopic abundances and could be used in a like manner include

http://www-esd.lbl.gov 34


Earth Sciences Division

Research

Fundamental and Exploratory Research Program

Objectives

Helium Isotopes in Long Valley Basalts: Implications For Future Volcanic Activity

Annual Report 1998-1999

Approach

Various geophysical and geochemiMagmas have higher helium abunAllen Dodson, Donald DePaolo cal observations in the Long Valley, dances and 3He/4He ratios than the and B. Mack Kennedy crust into which they intrude, and California, area suggest that magma depending on the region of mantle may be present at relatively shallow Contact: which has melted, the helium isotopic crustal levels beneath the re gion. Allen Dodson (510) 486-2430, jadodson@lbl.gov ratios of the magma will be different. Evidence includes inflation of a resurAlthough most of the helium in the gent lava dome within the Long Valley magma is lost upon eruption, enough remains trapped in olivine caldera; recent (1989-present) and numerous shallow (6-2 km) crystals to be measured.We have measured helium isotope ratios earthquake swarms beneath Mammoth Mountain, thought to be in olivines from basalts erupted over the past 3.2 million years to related to magma injection; increased emissions of cold CO2 from the flanks and summit of Mammoth Mountain, responsible better understand the relationship between helium isotopic for significant tree kill zones; and a significant increase in magratios currently measured at the Mammoth Mountain Fumarole matic helium and CO2 in gases from a fumarole near the (MMF) and those found in the magma during periods of eruption Mammoth Mountain summit that coincided with the earthquake and recorded in the rocks. swarms. Understanding the relationship of these observations to Samples were collected from more than a dozen olivine-bearone another, and their relationship to the geologic history of ing basalt flows. From approximately 1-2 kg of rock 0.5 to 1.5 Long Valley, is important both for evaluation of volcano hazards grams of the freshest, largest olivine grains were selected for and to understand the potential of the region for geothermal analysis. The olivine grains were crushed under vacuum to energy development (a small geothermal plant already exists in release gases from fluid inclusions, which contain a representative sample of the magma chamber atmosphere during crystal the area).

Figure 1. Helium isotopic compositions for basalts associated with the Long Valley-Mono Lake volcanic complex.

35


ESD

Helium Isotopes in Long Valley Basalts: Implications for Future Volcanic Activity

growth. The chemically purified and cryogenically separated helium was admitted into a mass spectrometer for determination of the isotopic composition.

(SM) basalts, for instance, we re erupted from sourc e vents along the we s t e rn e d ge of the caldera and fl owed eastwa rd into the c a l d e ra moat. R e c e n t drilling into the re s u rgent Results dome found no noticeabl e heat anomalies or evidence A spatial pattern in the for magma at depth, suphelium isotopic composiporting this interpre t a t i o n . tion of the different basalt The fact that the MMF flows is evident (Figure 1). helium ratios are similar to Basalts along the edges of those of the caldera - rim the caldera, ranging in age basalts, and are probably from 3.2 million years to less not the same as those of than 100,000 years, have relthe Mono and Inyo Cra t e rs atively uniform helium isoro cks, has important implitopic ratios of 5.8 to 6.2 cations for volcanic hazard times that of air. However, Figure 2. Correlation between helium and neodynium isotopic compositions. The evaluation. The Mono and the three we s t e rnmost caldera basalts are similar to other basalts found along the eastern Sierra. The I nyo Cra t e rs magmas have basalts, which have lower helium isotopic compositions, may have basalts in the area, which are western u n d e rgone significant re s ievolved from a different magma system. outside the Long Valley dence at depth prior to caldera â&#x20AC;&#x201D;June Lake (JL), Devils Postpile (DP) and Black Point eruption, as demonstrated by their silicic nature. This type of (BP) â&#x20AC;&#x201D; have much lower helium isotope ratios, from 4.6 to 5.2 volcanic system re q u i res significant re ch a rge of basaltic mag m a times that of air. Since the ages of these basalts overlap with to provide heat to the magma ch a m b e r, and there fo re such systems can be quite large and produce large, violent volcanic those of the caldera basalts, it appears that these two groups of eru p t i o n s. basalts are derived from two different magmatic systems, isolated In contrast, the caldera-rim volcanic system appears to be a from each other in the crust. relatively small one in which basaltic magmas frequently rise Fi g u re 2 illustrates that the caldera basalts are isotopically through the crust and erupt as small basaltic lava flows. While similar to others found in the eastern Sierra; they are pri m a ri ly this may provide some comfort, the fact that the helium isotope melts of mantle lithosphere with a small contribution from the ratios in the MMF fluids are similar to those exhibited by the sysconvecting asthenospheric mantle. Their helium and n e o dymium isotopic ratios lie along a mixing line between tem during eruptive periods suggests that the potential for such these two types of mantle. The we s t e rn basalts, on the other an eruption is significant. hand, have lower helium isotopic ratios than either the MMF Related Publications t o d ay or other eastern Sierra basalts. Their neodymium isotopic ratios are similar to those of Mono and Inyo Craters rhyo l i t e s , Dodson, A., D.J. DePaolo and B.M. Kennedy, Helium isotopes in suggesting that they may be related to these ro cks. lithospheric mantle: Evidence from tertiary basalts of the U n fo rt u n a t e ly, the mineralogy of the rhyolites is not suitable for Western U.S., Geochimica et Cosmochim-ica Acta 62, pp. helium analy s i s . 3775-87, 1998. During the recent period of eart h q u a ke swarms beneath Dodson, A., D.J. DePaolo and B. Mack Kennedy, Helium Isotopes Mammoth Mountain, the helium isotopic composition in gasses at Long Valley, California: Indicators of magmatic source f rom the Mammoth Mountain fumarole increased significantly, regions and plumbing systems, poster presented at the to about six to seven times the air ratio.The similarity between Geological Society of America Annual Meeting, Toronto, this ratio and those of the caldera - rim basalts indicates that the Ontario, 1998. m agmatic system sampled at the MMF is probably the one re s p o n s i ble for generation of these small-scale basalt fl ow s .

Funding Significance

of

Findings This project has been supported by the Office of Science, O ffice of Basic Energy Sciences, Division of E n gi n e e ring and Geosciences of the U.S. D e p a rtment of E n e rgy under Contract No. D E -AC03-76SF00098.

Our data suggests that magma may be located to the west of the re s u rgent dome, rather than beneath the dome itself. The North Moat (NM) and South Moat

http://www-esd.lbl.gov 36


Earth Sciences Division

Research

Fundamental and Exploratory Research Program

Objectives

Isotope Constraints on Fluid Sources: San Andreas Fault System, California

Annual Report 1998-1999

along the San Andreas and adjacent faults (the San Gabriel Fault, a deeper Fluids are suspected to play a major equivalent of the SAF, and the Santa E. Pili, B.M. Kennedy, role in earthquake mechanics, espeYnez Fault, a former strand of the SAF) M.S. Conrad and D.L. Shuster cially in the case of the weak San from South San Francisco to East Los Andreas Fault (SAF). Models deve lAngeles. The samples consisted of Contact: oped to explain the weakness of the gouge, fault breccia, slickenslide, cataB. Mack Kennedy (510) 486-6451, bmkennedy@lbl.gov fault include either low-friction fault clasite material from defo rmation zone materials or super-hy d ro s t a t i c zones, vein fillings, and their undefluid pressures within the fault zone. Models invoking high fluid formed host rocks. The samples are being analyzed for helium pressures are similar but rely on different fluid sources. During isotopic composition in fluid inclusions and the isotopic comthe earthquake cycle, fault-zone fluid pressure increases to near position of carbon and oxygen in the bulk samples. lithostatic values and induces rupture. Dilation accompanies rupResults ture, locally lowering the fault zone fluid pressures, and the cycle begins again. Potential fluid sources include meteoric, crustal, The helium isotopic composition of noble gases in fluid incluand mantle.A recent study of groundwaters associated with the sions from the various fault zone samples are in the range ~0.1 – SAF found elevated 3He/4He ratios, providing evidence for a geopressured mantle fluid source (Kennedy et al., 1997). Because 2.5 Ra (Ra is the 3He/4He ratio in air).This indicates that past fluids percolating through the SAF system contain mantle helium mantle fluids must pass through the lower plastic crust, they contributions of ~1 to ~32%, similar to that measured in presententer the base of the fault zone in the brittle upper crust at or day groundwaters associated with the fault (Kennedy et al., near lithostatic pressures. In transit, the 3He/4He ratios are diluted with radiogenic 4He that is produced locally in the crust, 1997). This confirms the involvement of mantle fluids and generating a vertical gradient in the fault-zone helium isotopic shows, from structural relationships observed in the field and in composition that depends on the vertical rate of fluid flow. thin sections, that these fluids are directly associated with the Calculated flow rates vary from ~1 – 10 mm yr-1, sufficient to process of faulting. maintain near lithostatic fluid pressures at the shallower depths Calcite is the dominant vein material and repeatedly occurs as of the seismogenic zone of the fault. an accessory mineral in deformation zones.The C- and O-isotope One can assume that the mantle 3He is associated with other compositions of carbonates from veins, deformation zones and more abundant mantle volatiles, certainly CO2 and perhaps their hosts are summarized in Figure 2. By comparing relative water. However, using the mantle CO2/3He ratio, the CO2 flux depletions in the 13C and 18O isotopic compositions of the defor(~3 × 10-4 kg km-2 sec-1) inferred from the helium isotopic data mation zone or vein carbonates with host rocks at various scales, is inadequate, by at least an fluid infiltration can be idenorder of magnitude, to retified. establish fa u l t - weakening At each sampling site, sevfluid pre s s u res on a time eral trends in the C- and O-isoscale relevant to earthquake tope depletion are observe d : cycles (Figure 1). This proj(1) The deformation zone ect is an isotopic study and vein material are designed to (a) compare the most depleted co deformation zones and vein p a red to their host fillings with their hosts and rocks. the fluids associated with (2) Veins that cut through these materials to confirm deformation zones are the presence of mantle even more isotopically helium in San Andreas fault depleted. zone fluids, and (b) identify (3) With increasing disand characterize the various tance from what can be potential fluid sources. structurally defined as the core of fault zones, Approach host rocks are less is t o p i c a l ly affected and Approximately 250 samthe density of veins and 1. Computed times required to generate lithostatic fluid pressures from ples from more than 20 Figure d e fo rmation zones hydrostatic values by CO2 flux at a 10 km depth. Values used for rock porosity (φ) localities we re collected and compressibility (Cr) are indicated. decreases. 37


ESD

Isotope Constraints on Fluid Sources: San Andreas Fault System, California

Figure 2. Carbon- and oxygen-isotope compositions of carbonates from gouges, cataclasites, fault veins, slickensides (altogether called deformation zones), veins and host rocks from the San Andreas, Santa Ynez, and San Gabriel faults.

Some or all of the CO2 may be of deep crustal origin.This supports the model invoking a deep source of fluids at or near lithostatic pressure weakening the fault zone (Rice, 1992).

(4) Host-rock carbonates display progressive evolution towards greater depletion in the order limestone, marble, gneiss, granite and basalt, mirroring the increase in metamorphic grade or deep crustal origin of the host-rocks. We infer the following from these trends and the isotopic compositions: (1) the fault zones have been infi l t rated by fluids of deeper origin during deformation; and (2) the fluids are dominated by crustal or connate water Âą CO2. Meteoric water does not appear to represent a significant contribution and the CO2 is inferred to have a metamorphic or mantle origin.

Significance

of

Related

Publications

Kennedy, B.M.,Y.K. Kharaka,W. Evans,A. Ellwood, D.J. DePaolo, J. Thordsen, G. Ambats and R.H. Mariner, Mantle fluids in the San Andreas fault system, Science, 278, pp. 1278-81, 1997. Rice, J.R., Fault stress states, pore pressure distributions, and the weakness of the San Andreas fault, in Fault Mechanics and Transport Properties of Rocks, Academic Press, San Diego, pp. 475-503, 1992.

Findings Funding

The infiltrated deformation zones, veins and host rocks show that fault zones in the San Andreas system maintain a higher permeability than that of adjacent regions.The noble gas and stable isotope compositions both provide evidence that mantle-derived fluids are involved in faulting and that the mantle helium is accompanied by deep crustal or metamorphic water Âą CO2.

This project has been supported by the Office of Science, Office of Basic Energy Sciences, Division of Engineering and Geosciences of the U.S. Department of Energy under Contract No. DE-AC03-76SF00098.

http://www-esd.lbl.gov 38


Earth Sciences Division

Fundamental and Exploratory Research Program

Annual Report 1998-1999

The Center for Computational Seismology Research

Objectives

for 1999. The list of doctoral thesis topics developed with some CCS resource support over the years is a good measure of results and major accomplishments. On average, two or three Ph.Ds per year are produced (six in 1998) spanning a wide spectrum of research interests.

Thomas V. McEvilly, The Center for Computational Ernest L. Majer Seismology (CCS) serves as the core and Lane R. Johnson data processing, computation and visualization facility for seismology-related Contact: Thomas McEvilly research at LBNL. Pursuing an objective (510) 486-7347, tvmcevilly@lbl.gov of providing modern tools for seismological research, the Center is designed and operated to provide a Significance of Findings focused env i ronment for research in modern computational seism o l o gy by scientists whose efforts at any time may be distributed Findings for a facility and scientific environment such as that among diverse research projects. A large number of varied, s e p ap rovided by CCS must be defined in the context of the multidisrately funded research projects from many diffe rent sponsors rely ciplined research base that is supported there, rather than projectupon this resource for intellectual exchange as well as computas p e c i fic accomplishments (those appear in other sections of this tional needs. U s e rs include LBNL scientists, along with collab o ra ting UC professors, postdoctoral fellows, visiting scientists, graduate report). S i g n i ficance lies in the enhanced productivity and innoand undergraduate students. Doctoral theses and journal publ i c avation that are produced by the ri ch mix of intellectual pursuits tions reveal a spectrum of effo rt from the most fundamental theothat come together in CCS. It is fair to attribute a large part of the retical studies to field applications at all scales. s c i e n t i fic reputation in seismology at LBNL to the CCS enviro nment.

Approach Related CCS provides a specially equipped and staffed computational facility to support and advance a wide-ra n ging program of seism o l o gical re s e a rch. B eyond computers , work stations, seismic processing pack ages and visualization capabilities, it is a physical fa c i lity in which scientists pursuing individual re s e a rch interact with other scientists and technical support staff in a multidisciplinary intellectual environment. The data management and processing techniques at CCS are integrated with the data acquisition instrumentation in LBNL's Geophysical Measurements Facility (GMF) within which field projects are designed and managed. A wide range of research projects relies upon CCS resources for development and application of methods for characterization, process definition and process monitoring in the rock-fluidthermochemical subsurface environment. Consequently, CCS research reaches from the most fundamental investigations to those driven by the most applied technologies. CCS supports research in the general areas of wave propagation, geophysical inverse methods, earthquake and explosion source theory, seismic imaging, borehole geophysics, four-dimensional process monitoring and visualization technology.

Publications

Recent Ph.D. theses: Fan, J., O verlap domain decomposition technique for modeling wave propagation, 1998. Hubbard, S., Stochastic characterization of hydrogeological properties using geophysical data,1998. Kaelin, B., Seismic imaging of the shallow subsurface with high frequency seismic measurements,1998. Nakagawa, S., Acoustic resonance characteristics of rock and concrete containing fractures, 1998. Parker, P., Genetic algorithms and their use in geophysical problems, 1999. Seifert, P., Effects of pore fluids in the subsurface on ultrasonic wave propagation, 1998. Yi,W., Numerical investigation of wave propagation in fractured rock, 1998. Selected 1999 journal publications: Kaelin, B., and L.R. Johnson, Using seismic crosswell surveys to determine the aperture of partially water saturated fractures, Geophysics, 64, pp. 13-23, 1999. Keers, H., L. Johnson and D. Vasco, Crosswell imaging using asymptotic waveforms, Geophysics, submitted. Nadeau, R.M., and T.V. McEvilly, Fault slip rates at depth from re c u rrence intervals of repeating micro e a rthquakes, Science, submitted. Gritto, R., V.A. Korneev and L.R. Johnson, Nonlinear 3-dimensional inversion of low frequency scattered elastic waves, Pure & Appl. Geoph. in press. Hubbard, S., Y. Rubin and E. Majer, Spatial correlation structure estimation using ge o p hysical and hy d ro ge o l o gical data, Water Resources Research, in press.

Results Results from the diverse seismological program at CCS are best demonstrated in the CCS research output. Major accomplishments flow largely from the breadth of research support provided by CCS and the cross-fertilization between applications and fundamental studies.The list of publications, produced with CCS support to varying degrees, displays the range of research accomplishments. For example, at the time this report is being prepared (March, 1999), there are more than 20 articles either published, in press or submitted to major peer-reviewed journals 39


ESD

The Center for Computational Seismology

Figure 1. CCS resources, March, 1999.

Rubin,Y., K. Grote and S. Hubbard, Precision moisture content Rector, J.W., Q. Dong and T.W. Patzek, Passive characterization of estimation using radar data; Applications to transportation hy d ro f ra c t u res using signal from hy d raulic pumps, J. studies, Geophysical Research Letters, submitted. Petroleum Sci. and Eng., in press. Korneev,V.A., and L.R. Johnson, Fluctuations in elastic waves due Herman, G.C., P. Milligan, R.Huggins and J.W. Rector, Imaging shalto random scattering from inclusions, Journ. Acoust. Soc. of low objects with scaattered guided waves, Geophysics, in Am., submitted. press. Fan, J.K., K.T. Nihei, L.R. Myer and J.W. Rector, Overlap domain Byun, J., and J.W. Recto,Wide angle effects in crosswell reflection decomposition technique for modeling wave propagation, imaging, J. Seismic Explor., in press. Geophysics, submitted. Byun, J., and J.W. Recto, Post-Map migration of crosswell seismic Nihei, K.T., M. Schoenberg, W.Yi and L.R. Myer, Fracture channel data, Geophysics, submitted. waves, J. Geophys. Res., 104, in press. Funding Nihei, K.T., S. Nakagawa and D.L. Hopkins, Defect detection in bonded structures using the reverberant wavefield, This work has been supported by the Office of Rev. of Prog. in Quant. Nondestr. Eval., 18, in press. Science, Office of Basic Energy Sciences, Division of Mars, J., J.W. Rector and S. Lazaratos, Filter formulation and Engineering and Geosciences, of the Department of wavefield separation of cro s swell seismic data, Energy under Contract No. DE-AC03-76SF00098. Geophysical Prospecting, in press. http://www-esd.lbl.gov 40


Fundamental and Exploratory Research Program

Earth Sciences Division

Research

Objectives

Annual Report 1998-1999

Log-Permeability Estimation Using Multiple Geophysical Data Sets Within a Bayesian Framework

where y is a realization of the log-perS u b s u r face investigations often meability and 〈y〉 and σy are the mean require characterization of hydraulic and standard deviation of Y that comparameters. Conventional sampling or pletely describe the distribution. Susan S. Hubbard, Yoram Rubin and Ernie Majer b o rehole techniques for estimating A log-permeability pdf can be these parameters are costly, time-conobtained from borehole permeability Contact: suming and invasive. The difficulty of measurements using a variety of estiSusan Hubbard collecting re p re s e n t a t i ve and suffimation or inversion techniques; the (510) 486-5266, sshubbard@ccs.lbl.gov cient hy d raulic pro p e rty measureresult is considered to be a prior pdf, ments using conventional sampling techniques, the large spatial and is annotated by fY(x)(y). Log-permeability estimates, obtained from geophysical tomographic data together with petrophysical variability of hydraulic properties in natural geologic systems relationships, can be used to update this prior pdf in a Bayesian over a wide range of scales and the dependence on the meassense.The updated log-permeability pdf is referred to as a posteurement support scale render measurement of hydraulic proprior pdf, and is denoted by fY(x)(y).The posterior pdf of Y at a sinerties difficult.The purpose of this study is to explore the use of gle location in space, x0, given additional geophysical informajoint geophysical-hydrogeological data for estimation of hydrotion at that location, gx0, is given by: logical parameters. As the ability to detect changes in physical and hydrological properties using geophysical data varies with f'Y(x0 ) (y) = fY(x0 ) y g(x 0 ) = fY(x0 ) (y yˆ(g(x0 ))) (2) the method selected, it is becoming more common for site characterizations to use multiple geophysical methods. In this study, where y(g(x0)) is the prediction of permeability based on geoˆ we focus on the improvement to the log-permeability estimate ˆ physical data, henceforth abbreviated as y(x0).This implies that offered by systematic incorporation of multiple, co-located tomothe conditional distribution of log-permeability given geophysigraphic data sets with limited hydrological data. Our estimation cal measurements is just a function of the predicted log-permeis perfo rmed in a stochastic framewo rk using a Baye s i a n ability based on those ge o p hysical methods. Using Baye’s approach. Here we present the methodology followed by a case Theorem (Ang and Tang, 1975), the posterior pdf f'Y(x0 ) ( y) can study using co-located hydrological and geophysical data from be written as: the NABIR (Natural and Accelerated Bioremediation Research Program) Oyster Bacterial Transport Site at Oyster,Virginia.

(

-•

Mathematical Statement of the Bayesian Approach: The goal of the estimation procedure is to estimate Y(x) over the entire synthetic aquifer, where Y(x)=ln(k(x)) is log-permeability and x is a vector of coordinates along a vertical plane.The data assumed available for this procedure include a limited number of permeability measurements, as well as complete sets of colocated geophysical data [g(x)], such as seismic velocity, seismic attenuation, electrical resistivity, dielectric constant or radar attenuation measurements obtained from high-resolution seismic, electrical and radar tomography data. Log-permeability is treated as a spatial random function; the available measurements are assumed to be realizations of this random function and the distribution of log-permeability at each location within the aquifer is described by its probability distribution function (pdf). The geophysical data are used to produce log-permeability estimates using a linear relationship between coincident log-permeability and geophysical measurements; thus the geophysical data are assumed to be fixed for this study.All attributes are modeled as second-order stationary random fields whose pdfs are assumed to be normally distributed.The log-permeability pdf is expressed as: 2˘ È 1 Êy - y ˆ ˜ ˙ exp Í - Á 2s Y ÍÎ 2 Ë s Y ¯ ˙ ˚

( (

) )]

f Yˆ(x ) yˆ(x 0 ) Y(x0 ) = y f Y(x ) (y) 0 0 f'Y( x 0 ) (y) = +• Ú f ˆY(x0 ) ˆy(x 0 ) Y(x 0 ) = y fY(x0 )(y)dy

Approach

f Y(x)(y) =

)

[

(3)

Equation (3) shows that the posterior pdf, f'y(x0), can be expressed as the product of two pdfs (the numerator) divided by a normalizing constant. An iterative Bayesian methodology is used for log-permeability estimation. A prior pdf at each location within the aquifer, fy(x)(y), is defined initially using solely hydrological measurements. This prior is then updated using information from one tomographic data set and a petrophysical relationship. For example, if estimates of log-permeability, obtained from one tomographic data set (g1(x)), are available at a particular location x0, the “first posterior” obtained using this conditional information can be expressed as:

(

)

f'Y(xo ) (y) = fY(x0 ) y yˆ (g1(x 0 ))

(4)

Equation (4) can in turn be updated using information available from another co-located tomographic data set (g2):

(

1

)

f'Y(xo ) (y) = f' Y(x0 ) y ˆy (g2 (x 0 ))

(1) 41

(5)


ESD

Log-Permeability Estimation Using Multiple Geophysical Data Sets Within a Bayesian Framework

Equation (5) yields a log-perm e ab i ldata.This field was in turn updated ity field, w h i ch has been updated using Equation (5) with hydraulic using two geophysical data sets, it is conductivity ratio info rmation re fe rred to as the “second posteri o r.” obtained from radar attenu a t i o n The second posterior field can in tomographic data.This second posturn be updated by info rmation terior field, obtained using flowmef rom another co-located data set, ter, dielectric constant, and attenu aand so on. Using the assumption of tion data, is shown in Fi g u re 1.The Gaussianity, B aye’s pro c e d u re (Eq. 3) flowmeter data are superi m p o s e d can be implemented using a simple on top of the estimated hydraulic a n a lytical ex p re s s i o n . Numerical conductivity pro fi l e . A n a lysis of these estimates revealed that mean simulations show that including error and va riance associated with mu l t i p l e , co-located ge o p hy s i c a l the estimates, calculated from the data in the log-perm e ability estimaestimated fields and values from tion pro c e d u re using the iterative flowmeter data that were not used B ayesian technique presented in in the estimation procedure, we re Equations (4) and (5) decreases the Figure 1. Estimated permeability using hydrological and georeduced when ge o p hysical data e rror, va riance and entro py associ- physical data. were included in the estimation pro c e d u re at the Oyster Site. ated with the log-perm e ability estimate (Hubbard, 1998). Case Study: The Oyster bacterial transport site is located on Significance of Findings the southern Delmarva Peninsula, situated on the eastern coast of the United States between the Chesapeake Bay and the The results presented here suggest that updating prior hy d roAtlantic Ocean. A field-scale bacterial transport study within a logical measurement estimates with successive geophysical info rsandy Pleistocene aquifer is being undertaken at the Oyster Site mation improves the log-perm e ability estimation. Numerical studby a multidisciplinary NABIR research team funded by the ies revealed that the offe red improvement is more substantial far Department of Energy Subsurface Science Program.The purpose f rom the wellbore,w h e re the prior estimates are more likely to be of the study was to evaluate the relative importance of physical, diffuse. This approach offe rs a practical and efficient method of chemical and hydrological heterogeneities in controlling bacterial transport.This in-situ bacterial transport investigation was the i n c o r p o rating multiple, co-located tomographic data sets to first of its kind, and extensive characterization using hydrologii m p rove the hy d raulic parameter estimation and retain the spatial structure. cal, geological, geochemical and geophysical methods is required to build an accurate numerical flow model necessary to predict Related Publications bacterial transport. H e re we integrate hy d raulic conductivity info rmation ava i l abl e Copty, N.,Y. Rubin and G. Mavko, Geophysical-hydrological idenf rom electro m agnetic flowmeters and radar tomography data tification of field permeabilities through Bayesian updating, using the Bayesian approach outlined ab ove. For this study, a sinWater Resour. Res., 29(8), pp. 2813-25, 1993. gle tomographic radar profile, collected using a PulseEKKO 100 Hubbard, S.S., Stochastic characterization of hydrogeological system with 200 MHz central frequency borehole antennas, was properties using geophysical data, Ph.D. dissertation, UC analyzed.This profile begins at the bacterial injection well where Berkeley, 1998. fl owmeter data we re collected, and extends 4.7 m down- gradient Hubbard, S.S., Y. Rubin and E. Majer, Spatial correlation structure in the expected injectate flowpath, w h i ch trave rses the locations estimation using geophysical and hy d ro geological data, of three other flowmeters. The flowmeter data are re p o rted in Water Resour. Res., accepted. hy d raulic conductivity ratios, or Ki/Kave, w h e re i re fe rs to the measurement interval and ave refers to the ave rage value over the Funding e n t i re measured interva l . The flowmeter and tomographic data were used to estimate a spatial cova riance stru c t u re following a This study has been supported by National Science method developed by Hubbard et al. (accepted).This correlation Foundation Grant EAR 9628306 and by the Office of Science, stru c t u re was then used in an ordinary kri ging routine, condiOffice of Biological and Env i ronmental Research, Subsurface tional to the flowmeter data, to obtain prior hy d raulic conductivScience Pro gram of the U. S . D e p a rtment of Energy ity estimates along the expected injectate flow centerline. under Contract No. DE-AC03-76SF00098. All computaThe prior field was subsequently updated using the tions we re carried out at the LBNL Center fo r Bayesian methodology (Eq. 4) with hydraulic conductivity Computational Seismology. ratio information obtained from radar dielectric constant http://www-esd.lbl.gov 42


Earth Sciences Division

Research

Fundamental and Exploratory Research Program

Objectives

Annual Report 1998-1999

Modeling the Observed Controlled-Source Waveform Changes At Parkfield, California

velocity models for the area all combine to provide well-determined conA unique data set for the study of straints in modeling the observations. wave pro p agation in the San Andreas, In this study we consider only data Valeri A. Korneev, California, fault zone has been recorded at stations VCA and JCN Tom V. McEvilly acquired in the Pa rk field Prediction from vibrator site VP2.At VP2 we have and Eleni D. Karageorgi E x p e riment (PPE) underway in central the routine Vibroseis monitoring data California. Data have been collected from the repeated point source as Contact: with a 10-station borehole network in well as a cross array of sources with Valeri Korneev (510) 486-7214, vakorneev@lbl.gov a search for evidence of changes asso17 VPs on each leg. We confined our ciated with the nu cleation process of modeling exercise to the VP2 data for the anticipated magnitude-6 eart h q u a ke at Parkfield. More than VCA and JCN for several reasons. Both source-receiver paths are 6,000 eart h q u a kes have been re c o rded since 1987 in the mag n iin the anomalous region and reveal substantial travel-time variatude range -1<M<5. In addition, s e i s m o grams for 720 sourcetions.The two paths are approximately co-linear and orthogonal receiver paths have been obtained for repeated illumination of to the San Andreas fault, permitting the use of a 2-D formulation the network using a large shear-wave vibration (using the in simulating wave propagation. The paths sample segments of Vibroseis machine), f rom June 1987 until November 1996, when similar length on the two sides of the fault zone. Finally, the data profile from the closely-spaced source array at VP2 defines the the program ended.That investigation reported significant travelspatial coherency of the wavefield that is helpful in phase identime ch a n ges in the coda of S for paths crossing the fault zone tification and interpretation of the recorded wavefield. The southeast from the epicenter of the 1966 M6 eart h q u a ke. Pro gressively decreasing travel times in the anomalous re gi o n velocity model used in numerical simulation incorporates the reached 50 msec or more by the end of the study. Changes in freknown properties of the region, where tomographic 3-D velocquency content and polarization we re also found and those ity models have already been determined. A major factor cone ffe c t s ,too, could be localized to the zone of common nu cleation trolling the character of wave propagation at short range from a and rupture onset for the previous M6 eart h q u a kes, and, possibly, surface source is the severity of the shallow vertical velocity grathe re gion of slip initiation for the great eart h q u a ke of 1857.The dient. We found a velocity gradient model by matching the temporal pattern in these va riations appears to be synchronous observed and computed direct arrivals in the early part of seiswith changes in defo rmation and seismicity measured independmograms. For the NE side of the fault, the direct arrivals at JCN ently. Because similar va riations are not seen in the wave fo rm s could be matched with a velocity profile reduced to 0.76 of that re c o rded from micro e a rt h q u a kes in the same part of the fault, in for VCA, and to 0.5 for the narrow fault zone, modeled as a verprevious studies we concluded that changing fluid conditions in tical layer with a thickness of 200 m, bounded by interfaces F1 the uppermost section of the fault zone in response to deeper, and F2. Computations were performed using a 2-D elastic finitetectonic stress perturbations are the likely cause of the temporal difference formulation with a staggered grid.The model was digvariations. Exploring that possibility further for plausible velocity itized on a 2200 x 500 grid with 5 m spacing, which yields a p e rturbations in the shallow fault zone, in this study we model model space of 11 km horizontal and 2.5 km vertical extent, as the observed wave form changes nu m e ri c a l ly. depicted in Figure 1.

Approach

Results

At Parkfield, the San Andreas fault zone is a striking near-vertiA snapshot of elastic field development is shown in Fi g u re 1. cal low-velocity zone which very clearly acts as a waveguide for Two fe a t u res dominate the process: e n e rgy trapping near the surseismic energy from earthquakes on the fault and from surface face by the shallow gradient and wavefield scattering from the sources.Velocity models there show a high Vp/Vs ratio along the fault zone. Most of the energy is confined to the upper part of the section in multiple reflections at the free surface, producing a fault near the surface and at depth within the fault zone, and a complex train of surfa c e pronounced vertical velocity guided waves made up of gradient in the upper 2 km many arriving phases. of the section.The geometry Synthetic seismogra m s of the Vibroseis source and ( h o rizontal component) are re c e i ver netwo rk , the s h own in Fi g u re 2. Initial approximate two-dimensiondirect P and S waves arri ve at ality of the fault zone in the VCA at around 1 and 2 secre gion of the travel-time onds, re s p e c t i ve ly. At JCN anomaly and the existence of detailed P- and S-wave Figure 1. A snapshot of wavefield propagation from VP2 source at 2 seconds. t h ey are seen at 2.2 and 4.4 s. 43


ESD

Modeling the Observed Controlled-Source Waveform Changes at Parkfield

Because the re c e i ve rs are time shifts at both stat i o n s . The located at depth, both up- and match is quite good in character, dow n - going energy are seen,as magnitude and timing. The first unstable wavelet at VCA correwell as horizontally propagatsponds well to the PF2PP reflection ing turning-point wave s . from the fault zone. At JCN the patThe signature of the fault tern of steady increase in the travelzone and shallow gradient on time shift due to pro gre s s i ve the wavefield is dramatic. In involvement of slower S-waves is the interval between the firstFigure 2. Synthetic traces for variation modeling. quite clear. arriving P and S waves at VCA are the surfa c e - ge n e ra t e d Significance multiples and conve rsions. of Findings The latter are especially strong for P-waves, e.g., PS, Previous studies clearly detected PPS, etc. S t rong re flections real changes in travel-times in the are also produced by the eight-year controlled-source monitorfault-zone boundaries, F1 and ing program at Parkfield, and localF2. PF1P, the first F1 reflection at VCA, is small and masked ized the anomalous changes to a Figure 3. Comparizon of the observed and modeled variations. by the large direct S-wave just region southeastward from Middle b e fo re it. F2 re flections at Mountain, roughly in the presumed VCA have passed twice through the fault zone, and these late nu cleation zone for the past and future M6 Park field earthquakes. phases, such as PF2PP (3.5 s) and SF2S (4.7 s), a re quite strong, Their best hypothesis for the phenomenon called for changing arriving well after the direct waves have passed.At JCN the interfluid conditions in the shallow section ab ove the fault zone. Our nal fault-zone reflections produce sequences of strong, distinct study supports that hypothesis and offe rs a more quantitative model for the actual wave pro p agation involved.The final link in arri vals following the direct P- and S-waves.The times in the synthe puzzle lies in the responsible mechanism for the velocity thetic seismograms where large travel-time ch a n ges we re observed in the monitoring project at VCA and JCN contain sigchange in the fault zone. We are inclined to accept the idea of a n i ficant energy that has been scattered from the fault zone.This deeper tectonic deformation that somehow changes the fluid env iresult suggests a re a dy explanation for the cause of the observed ronment in the shallow fault zone.The striking importance of the p ro gressively decreasing travel-times. For the path VP2-VCA the shallow ve rtical velocity gradient cannot be overstated. It is clear changes were seen at arri val times after 3.5 s, i.e., for our model from this study that surface sources employed in highly heterogeseismograms, after the direct waves have passed and the faultneous environments such as the San Andreas fault zone can be zone reflected waves are arri v i n g . On the other hand, the travelexpected to generate an overwhelming near-surface wave field that time changes for the VP2-JCN fa u l t - c rossing path begin with the must be dealt with in looking for deeper images. If the individual phases can be identified, however, t h ey may provide an important arri val of the direct P-wave and occur through the entire seismotool for studying near-surface details of the fault structure. gram. We take these results to be strong evidence that the observed va riations are most likely caused by ch a n ges within the Related Publication fault zone itself. To test the fault-zone hypothesis we modeled t ravel-time va riations that would be produced by a small velocity Ko rn e ev,V.A.,T.V. McEvilly and E.D. Karage o rgi, S e i s m o l o gical studchange at the fault. To compare with seismograms for the refe ries at Park field VIII: Modeling the observed controlled-source ence model described ab ove, we computed new seismograms at wave changes at Parkfield, Bull. Seism. Soc. of Am., submitted. VCA and JCN for a velocity increase of 6% localized in the narrow fault zone.These seismograms are shown in Fi g u re 2 along with Funding their diffe rences from the re fe rence traces. As expected, the changes at VCA appear only after the fault-zone F2 reflections This work has been supported by the U.S. G e o l o gical Survey re a ch the station, while at JCN the travel-time advance begins through NEHRP award 1434-95-G-2540 and through the Office of with the initial P-wave and increases throughout the seismogram. Science, Office of Basic Energy Sciences, of the U.S. Department of The magnitude of the calculated travel-time va riations match the Energy under Contract No. DE-AC03-76SF00098. Data observed data quite closely. In Fi g u re 3 we make a direct processing was done LBNL's Center for Computational comparison with the Vi b roseis data, w h e re the syntheticSeismology. derived va riations are plotted with the observed travelhttp://www-esd.lbl.gov 44


Fundamental and Exploratory Research Program

Earth Sciences Division

Research

Annual Report 1998-1999

Overlap Domain Decomposition Technique for Modeling Wave Propagation

Objectives

compute p(xm,tn+1) in Ω 1, the values of two previous time steps are needed A computationally efficient overlap from the grid points m-2 through Jianli Fan, Kurt T. Nihei, domain decomposition (ODD) techm+2. Assume that the pre s s u res Larry R. Myer and nique based on Huygens’ Principle p(xi,tn) and p(xi,tn-1) i=1,2,..,M at time James W. Rector step n-1 and n are known values. If Ω1 has been developed for modeling is overlapped to cover the grid points wave pro p agation. The ODD techContact: m+1 and m+2, then p(xm,tn+1) can be nique divides a large domain into sevKurt Nihei (510) 486-5349, ktnihei@lbl.gov solved within Ω1. eral smaller overlapping subdomains The grid points m+1 and m+2 that and allows the exchange of waves originally only belonged to Ω2 now also belong to Ω1.Thus, these between subdomains through the overlapping regions without points form the overlap region. The pressure p(xm,tn+1) can be the need for any internal interface connecting conditions. computed from Ω1 just as in the conventional finite difference Calculations are performed independently in each subdomain, method. Similarly, for the grid point m+1 in Ω 2, the addition of and the wavefield in the whole domain is then obtained from the two grid points m-1 and m to Ω 2 allows the calculation of local solutions in the subdomains. The ODD technique is not p(xm+1,tn+1) at the grid point m+1 in Ω2 using the values of p restricted to a specific numerical method for solving the wave from grid points m-1 and m+3. The total overlap region now equation and different methods can be used in each subdomain. spans the grid points from m-1 through m+2. As can be seen This flexibility is particularly advantageous because it enables from this analysis, the overlap region only requires four grid very accurate but computationally expensive methods to be used in selected subdomains to achieve a desired level of accuracy and computing speed. In addition, calculations can be “turned-off” in subdomains that do not have appreciable wave activity, resulting in savings in computing time and memory use.

Approach A detailed description of the overlap domain decomposition (ODD) technique is gi ven in Fan (1998). H e re , we present a specific application of the ODD technique to the finite difference method. To apply the ODD technique to the finite difference (FD) method, we must first determine the appropriate size for the overlap areas. Starting with the 1-D wave equation for a medium with a constant density ρ 1 ∂ 2 p( x, t ) c

2

∂t

2

=

∂ 2 p(x, t ) ∂x

2

(1) ,

where p is the pressure and c is the velocity.The explicit finite difference scheme for fourth-order space and second-order time (2) p( x m , t n +1 ) = 2p(x m , t n ) − p( x m , t n−1 ) +

c 2 Δ t 2  − p( x m +2 , t n ) + 16 p( x m +1, t n )  2  , t ) + 16 p( x , t ) − p( x , t ) 12Δ x  − 30p( x m n m −1 n m −2 n 

accuracy is To apply the ODD technique to Equation (2), we divide the domain Ω into two subdomains Ω 1 (dark shaded area) and Ω 2 (light shaded area) shown in Figure 1, where grid points 1, ..,m2,m-1, m belong to Ω1 and grid points m+1,m+2,..,M belong to Ω2. Grid point m is the last point in Ω1 where the spatial derivative must be computed. Point m+1 is the first point in Ω 2. The amount of overlap required between these two subdomains can be determined as follows.According to Equation (2), in order to

Figure 1. Schematic of the overlap domain decomposition scheme for the 1-D finite difference method with 4 th order space and 2nd order time differencing.

45


ESD

Overlap Domain Decomposition Technique for Modeling Wave Propagation

points (Figure 1) to achieve the same results as the conventional finite difference method applied to the full domain Ω.Additional grid points in the overlap region do not affect the results.

computational wave pro p agation problem to be divided into many smaller p ro bl e m s . The technique can incorporate diffe rent nu m e rical methods for solving the wave equation (e.g., finite difference, psuedo-spectral, staggered grid methods) in each subdomain to a ch i eve a desired level of accuracy, computing speed and memory usage. In addition, the ODD technique has the ability to “ t u rn off” subdomains that contain little or no wave activity. This fe a t u re can be used to save computation requirements in problems with localized wave fields. Finally, the ODD framewo rk is ideal for parallel computing because of the modular nature of the algorithm. Future work will explore the application of the ODD technique to 3D wave pro p agation problems using parallel computers .

Results

Several advantages of the ODD technique can be illustrated with the foll owing example of two - d i m e n s i o n a l elastic wave pro p agation in a low velocity laye r. The computational domain is divided into nine subdomains with overlap areas that are s h a red between neighboring subdomains.These overlap areas allow transfer of the wave from one subdomain to the next, as described in the previous section. During the course of the calculation, the wavefield may not have Related Publications spread over the entire domain, particularly at early time steps. The calculaFan, J., Overlap domain decomposition tions in these inactive subdomains are technique for modeling wave propaganot necessary and can be “turned off.” tion, Berkeley Lab report LBNL-42881, Figure 2 shows snapshots of the wave1998. field for the nine-subdomain model. At 2. Snapshots of the wavefield at two times comFan, J., K.T. Nihei, L.R. Myer and J.W. the earlier time step, the wavefield is Figure puted with the ODD/finite difference method. Rector, O verlap domain decomposipresent only in subdomain Ω 6. Thus, computation of the wavefield need only be performed in this tion technique for modeling wave propagation, Geophysics, subdomain. Similarly, at the later time step, computations in all submitted. the subdomains except Ω1, Ω4 and Ω7 can be turned off.This simFunding ple procedure can save a significant amount of computing time.

Significance

of

Findings

This work has been supported by the Office of Science, Office of Basic Energy Sciences under U.S. Department of Energy Contract No. DE-AC03-76SF00098.

The ODD technique is a flex i ble fra m ework that allows a large

http://www-esd.lbl.gov 46


Earth Sciences Division

Research

Fundamental and Exploratory Research Program

Objectives

Annual Report 1998-1999

Frictional Effects On the Volumetric Strain Of Sandstone

tigation of the effects of friction on the nonlinear, hy s t e retic deformation When subjected to compre s s i ve of sandstones. Kurt T. Nihei, L.B. Hilbert Jr., stresses, a granular rock is reduced in To examine the contribution of Nigel G. W. Cook, S. Nakagawa volume by the deformation of the friction to the volumetric strain of and Larry R. Myer compliant grain contacts and solid granular rock, we use the face-cengrains and by compaction due to frictered cubic packing (fcc) model of Contact: tional slip between grains. While the spheres of identical size and properKurt Nihei (510) 486-5349, ktnihei@lbl.gov first two processes have been studied ties (Figure 1). This model is one of extensively in the context of linear the densest regular arrays of elastic elasticity, fewer studies have examined the role of nonlinear spheres with a porosity of 25.95% and a coordination number of grain contact deformation and frictional integranular slip on the 12. Both the effects of Hertzian grain contact deformation and volumetric strain of granular rocks. An understanding of these Mindlin-type intergranular slip are included in the model. The processes and their contributions to the deformation of rock is primary advantage of using an idealized model of rock is that important in many areas of the geosciences, including reservoir closed-form stress-strain relations can be derived for loading, production and stimulation. The objective of this work is to unloading and reloading paths. investigate the basic character and effects of grain contact nonResults linearity and frictional slip on the volumetric strain of sandstone using a Hertz-Mindlin sphere pack model and stress-strain measAnalysis of the fcc packing stress-strain relations for uniaxial urements on Berea sandstone. Here, we summarize several of the strain consolidation along the z-axis revealed that the elastic findings from the analysis of the sphere pack model. modulus that describes P-wave velocities are path-independent Approach functions of the axial strain There are a number of closed-fo rm analytic solutions for the stress-strain behaviors of regular packings of identical spheres (e.g., simple cubic, fa c e - c e n t e red cubic, hex agonal close-packe d ) that were developed by Mindlin and others . These models have the disadva n t age of lacking grain packing disorder, grain angularity and intergranular cementation that are prevalent in most sandstones. H oweve r, because these models do include the contri b utions of grain contact nonlinearity and frictional slip to the vo l um e t ric strain, they represent a logical starting point for the inve s-

c 33 = =

1/ 2  µ (4 − 3ν )  ε  (1 − ν )( 2 − ν )  zz ,

(1)

where µ and ν are the shear modulus and Poisson’s ratio of the grains, respectively. By integrating this modulus, it is possible to construct a new stress-strain curve that is identical for loading and unloading paths: nonlin

σ zz

ε zz

locked

= ∫ c 33 0

dε zz =

 2µ (4 − 3 ν )  3/ 2 ε  3 (1 − ν )(2 − ν)  zz

.

(2)

Figure 2. Comparison of the volumetric strain resulting from nonlinear grain contact deformation and frictional slip compaction for the fcc sphere packing.

Figure 1. Face-centered cubic packing of identical spheres.

47


ESD

Frictional Effects on the Volumetric Strain of Sandstone

Significance

Simple inversion of this stress-strain relation yields an expression for the volumetric strain resulting from nonlinear deformation of the grain contacts that is path-independent (i.e., it depends only on the final stresses and not on the historical sequence of stresses to which packing was subjected): nonlin

ε zz

=

 3 (1 − ν)( 2 − ν )   2µ (4 − 3 ν) 

2/3

.

Using this expression, the total volumetric strain for uniaxial strain consolidation along the z-axis can be decomposed into contributions due to (path-independent) nonlinear grain contact deformation and (path-dependent) frictional slip compaction

ε

total

nonlin

(4)

friction

Findings

The results of this study highlight the importance of considering frictional effects when attempting to evaluate the volumetric strain of sandstone. For the specific conditions of this study (i.e., uniaxial strain loading, unloading and reloading paths), frictional slip between grains was shown to result in a path-dependent volumetric strain. However, for these conditions, the nonfrictional part of the volumetric strain can be extracted by integration of the path-independent, small-strain elastic moduli obtained from seismic velocity measurements made at different static loads.Additional work is needed to determine both the magnitude and character of the path-dependence of the volumetric strain and the anisotropic elastic moduli of sandstones, and to determine if the analysis presented here can be extended to general three-dimensional loading.

(3)

2/ 3

σ zz

of

.

Related It is of interest to determine the relative magnitudes of the nonlinear and frictional contributions to the total volumetric strain. Figure 2 shows the volumetric strain of the fcc packing for a range of friction coefficients.The frictional slip contribution (for the loading path) approaches zero for large friction coefficients and a maximum (54% of the nonlinear contribution) as the friction coefficient approaches zero. While the nonlinear contribution dominates, the frictional component can be significant when the friction coefficient is low.

Publication

Nihei, K.T., L.B. Hilbert, N.G.W. Cook, S. Nakagaw and L.R. Myer, Frictional effects on the volumetric strain of sandstone, Int. J. Rock Mech., to appear, 1999.

Funding This work has been supported by the Office of Science, Office of Basic Energy Sciences under U.S. Department of Energy Contract No. DE-AC03-76SF00098.

http://www-esd.lbl.gov 48


Earth Sciences Division

Fundamental and Exploratory Research Program

Annual Report 1998-1999

Stochastic GILD Coupled Modeling and Inversion Research

Objectives

the acoustic velocity, electric conductivity and flow conductivity, etc. The Ganquan Xie and Jianhua Li Seismic, electromagnetic and advantages of our SGILD method are hydrological modeling and inversion that: (1) we use the global integral Contact: Ganquan Xie are important to the field of oil exploequation and local differential equa(510) 486-7134, g_xie@lbl.gov ration, environmental sciences, CO2 tions to compensate each other to sequestration and atmospheric sciimprove the ill-posed property and ences. Many imaging approaches in geophysical research areas obtain high resolution; (2) replacing the artificial approximate are based on deterministic and traditional nonlinear inversion. boundary condition, we use an exact integral equation on the Because the data are often incomplete and contaminated by ranboundary to greatly reduce numerical noise; (3) we decompose dom noise, the deterministic approach does not provide accuthe full matrix into several small sparse matrices and one small rate or stable solutions. During the iteration of the traditional full matrix, we solve these small matrices, in parallel, and greatly nonlinear inversion, inexact reflections of the artificial boundary reduce computational time and storage; (4) we can use the condition enter the inversion domain; these degrade resolution SGILD method for large-scale oil exploration, marine exploand even cause divergence.The discretization of the volume interation, and atmosphere prediction.The basic frame of our SGILD gral equation is ill posed and its full matrix needs very large storcoupled inversion is the GILD modeling and inversion method. age and lengthy computational time. The conjugate gradient (CG) method requires a long computation time and often termiNew Magnetic integral equations in the time domain: nates in local minima, thus producing a wrong result. Our objecBecause most hydrological data is measured in the time domain, tive is to develop a new parallel stochastic Global Integral and we develop the joint inversion in the time domain.We derive a new integral equation Local Differential (SGILD) electromagnetic (EM), seismic and hydrological modeling and coupled inversion. This method can σb  σ  overcome the shortcomings of traditional deterministic nonlin1 −ε t 1 − εb t  ∗t − e H(r, t ) = H b ( r, t ) − ℵ ∫  e (1) ear inversion. We have derived new integral equations for mag εb Vs  ε netic, acoustic and flow field in the frequency domain (Xie and   Li, 1995 and 1997) and time domain (Xie et al., 1999) We couple M ∇ × H ∗ t ∇ × G b dr ', the integral equation on the boundary and differential equation in the interior of the domain to develop SGILD modeling. We and differential equation couple the Jacobian volume integral equation on a small selected sub-domain and variation differential equation on the remaining σ   (2) large sub-domain to develop SGILD inversion for updating the 1 −ε t ∂H ∗ t ∇×H +µ =- µMδ' (t )δ (r',r ), ∇× e electric conductivity/permittivity, seismic velocity and hydrologε ∂t   ical conductivity. We use a new hybrid direct-iteration algorithm and a regularizing method associated with the constitutive law for the magnetic field in the time (Xie et al., 1999) to solve the equation. where ℵ is the unit that makes the integral quantity unit of the M Approach right-hand side of the equation a magnetic unit, G b (r ', r ) is the M magnetic Green tensor, E b (r',r ) is the electric tensor, and *t is Stochastic GILD modeling and inversion: The outline of a convolution. the SGILD method is as follows: (1) we decompose the domain Similarly, we derive the integral equation for seismic and into the two subdomains; (2) we use a new stochastic moment hydraulic inversion in the time domain. The new integral equaintegral equation on the boundary and moment differential equation (1) will be important progress in electromagnetic theory tion in the interior of the domain to calculate the moments of and application. the wave field in the modeling; (3) we use new Jacobian volume We use Bayes theory, the new integral equations in the moment integral equations in the small selected subdomain and selected subdomain and differential equations in the remainder variation differential equations in the remainder large subdosubdomain to develop the SGILD geophysical and hydrological main for the joint inversion. Supposing that the parameters and modeling and coupled inversion in the time domain. data are random variables, we derive the new integral and differential equation system for the statistical moments of the mean, GILD regularizing: We find that the ill-posed pro p e rty of covariance, and standard deviation. We use our parallel SGILD the inve rse pro blem is re l a t i ve to the constitutive law.We use a algorithm to solve the moment integral and differential equan ew constitutive term to construct the new GILD re g u l a rizing tions and obtain the high resolution imaging of the mean impedoperator. ance, covariance standard deviations and confidence interval for

(

49

)


ESD

Stochastic GILD Coupled Modeling and Inversion

Figure 1. Coupled SGILD imaging using single hole data.

Results

inversion is a high-resolution, robust, stable and high-performance parallel imaging algorithm.There are obvious improvements of resolution of the moments from the field data. We use these moments to estimate the uncertainty and construct confi d e n c e intervals and consistent conditions. The GILD and SGILD method discussed here is capable of ove rcoming the limitations of the c o nventional inversion and can accurately handle high-resolution t h rough the joint nonlinear inversion of coupled hy d ro l o gical and geophysical data.The GILD method can become a powerful imaging tool for DOE env i ronmental remediation, geophysical exploration, CO2 s e q u e s t ration and atmosphere sciences.

We use the SGILD algorithm for hydrological conductivity, e l e ctric conductivity and seismic velocity imagi n g .An axis symmetric SGILD conductivity code is tested pri m a ri ly using synthetic data. Mean conductivity imaging and standard deviations are presented. In Fi g u re1, 18 injection point sources are located in a single hole.The 72 point re c e i ve rs are located in the same hole. We use the synthetic model to create head response synthetic data with Gaussian noise in one and two months; the maximum standard deviation of data is 5%. The imaging of the mean conductivity is obtained (Figure 1c). The total maximum standard deviation (TSTD) of the conductivity is 11.8%. The local standard deviation (LSTD) of the conductivity of the target at the left side corner is 6%. The local standard deviation of conductivity on the right side is 18.6%; that is because the left side is closed to the area of the data site in the single hole.The 2-D mesh is 148 x 128, 16 x 18 CPU minutes in MPP and 68 iterations are used. The optimized mean regularizing parameter is 0.687456E10-6. A coupled acoustic, electromagnetic and flow head inversion is used to obtain conductivity imaging as shown in Figure 1b. It is obvious that the conductivity imaging from the coupled inversion has higher resolution than the single inversion imaging shown in Figure 1c. SGILD resistivity imaging from practical field data in the geothermal exploration site is obtained.The maximum standard deviation of the resistivity is 18.8%; the local standard deviation of resistivity near the borehole area is 11%.

Significance

of

Related

Publications

Xie, G., and J. Li, 3D extrapolating electromagnetic imaging, Proceedings of Pro gress in Electro m agnetic Research Symposium. p. 576, 1997. Xie, G., J. Li, G. Moridis and E.L.Majer, 3D boundary-domain finite element method for electromagnetic modeling, Proceedings of the SEG 68th Annual Int. Meeting and Exposition, pp. 417420, 1998. Xie, G., J, Li, C.C. Lin and E.L. Majer, Stochastic coupled SGILD geophysical and hydraulic modeling and inversion, SEG 69th Annual Meeting, 1999.

Funding This work has been supported by the Office of Science, Office of Basic Energy Sciences, Division of Engineering and Geosciences of the U.S. Department of Energy under Contract No. DE-AC03-76SF00098.

Findings

The preliminary tests show that the SGILD modeling and

http://www-esd.lbl.gov 50


Earth Sciences Division

Annual Report 1998-1999

NUCLEAR WASTE PROGRAM

GUDMUNDUR S. BODVARSSON CONTACT (510) 486-4789 GSBODVARSSON@LBL.GOV

The role of the Nuclear Waste Program is to assist the U.S. Department of Energy, the United States and other countries in solving the problem of safe disposal of nuclear waste through high-quality scientific analyses and technology development.The primary work of the program involves investigating the feasibility and potential of the Yucca Mountain site in Nevada for permanent storage of high-level nuclear waste. The Nuclear Waste Program also does collaborative work on nuclear waste disposal issues with such countries as Japan,Switzerland and Sweden.The program has established the Center for International Radioactive Waste Studies, the main purpose of which is to foster relationships with other countries in order to promote the exchange of ideas and research results. The Yucca Mountain Site is located about 120 km northwest of Las Vegas in a semi-arid region.The potential repository will be located about 350 m below the surface within a thick unsaturated zone (UZ).The subsurface rocks at Yucca Mountain consist primarily of fractured volcanic tuffs that vary in the degree of welding.To date, a total of 60 deep surface boreholes have been drilled in the area. In addition, an 8-km-long underground tunnel, the Exploratory Studies Facility (ESF), was completed at Yucca Mountain in 1996. Berkeley Lab’s work at Yucca Mountain consists of solving many ground-breaking problems related to multiphase, nonisothermal flow and transport through the UZ. Some of the key questions Berkeley Lab scientists are addressing include: • How much water percolates through the UZ to the repository at Yucca Mountain? • What fraction of the water flows in fractures and how much flows through the matrix blocks?

• How much of this water will seep into the waste canister emplacement drifts? • How will the water containing radionuclides migrate from the repository to the water table? • How will coupled TH (thermo-hydrological), THC (thermohydrologic-chemical) and THM (thermo-hydrologic-mechanical) processes affect flow and transport? The Nuclear Waste Program is organized into three main groups — Ambient Testing, Thermal Testing and Modeling — to address these questions, with support from geophysical studies.

Ambient Testing Group The Ambient Testing group investigates how water fl ow s t h rough the mountain and how mu ch of this water will seep into the emplacement drifts. The group performs va rious tests within the ESF, including the fracture/matrix interaction test, the drift-to-drift test, the Paint Brush Unit test (PTn test), and nich e testing.The fra c t u re / m a t rix interaction test is a relative ly smallscale (seve ral meters) test that focuses on the components of water fl ow in fra c t u res and matrix bl o cks and the interaction between the two continu a . The dri f t - t o - d rift test addresses the same issues, but on a mu ch larger spatial scale (10-20 m). The test in the Paint Brush Unit, which is an unwelded tuff unit, a d d resses issues of episodic fl ow, effects of faults and large - s c a l e features and lateral continuity of fl ow and transport. This unit, being ab ove the potential repository, is key to dispersing fract u re fl ow from the ab ove fra c t u red units and buffe ring the tra nsient behavior of episodic fl ow. The niche studies address perhaps the most crucial pro blem of Yucca Mountain; i.e., d e t e r51


ESD

Nuclear Waste Program

Modeling

mining the fraction of water that will fl ow into the emplacement drifts. The niche studies are carried out by introducing water into boreholes ab ove the opening and measuring what f raction actually seeps into the opening. Results to date suggest that there is a seepage threshold below which no water will seep into the drifts.

Group

Berkeley Lab has the primary responsibility for the development of the UZ flow and transport model.This is a comprehensive 3-D, dual permeability numerical model that represents the entire unsaturated zone at and near Yucca Mountain. The model is intended to integrate in a single computational framework all of the relevant geological, hydrological, geochemical and other observations that have been made at the surface, in boreholes and in tunnels at Yucca Mountain.The model is calibrated against pneumatic moisture tension, matrix potential, temperature, geochemical, perched water and other data from the UZ.The model is then used to predict all of these variables in new boreholes and new drifts to be drilled. The degree of agreement between the model predictions and the subsequent observations gives a measure of the reliability of the model and guidance to what additional data need to be collected and incorporated. A very important sub-model of the UZ model is the seepage model, which is on a tens-of-meters scale, versus the UZ model’s hundreds-of-thousands-of-meters scale. The seepage model, in a similar fashion as the UZ model, predicts the results of the niche tests, and is modified subsequently to match the actual observations. Another sub-model of the UZ model is the Coupled Process THC model, which is calibrated using the heater test data and used to estimate the chemistry of water and gas entering the drifts. All three models—the UZ model, the seepage model and the THC model—are key process models for Total System Performance Assessment (TSPA); performance of the potential repository is only as reliable as these key models.

Thermal Testing Group The Thermal Testing group wo rks in collab o ration with other national laboratories to evaluate the effects of heat on thermodynamic conditions, fluid fl ow and transport, and permanent p ro p e rty changes at and near the emplacement dri f t s .The Yucca Mountain Project has completed the fi rst in-situ heater test, called the Single Heater Test.The project is now invo l ved with a large-scale heater test in a drift over 80 m long. This second heater test is intended to re s e m ble the actual in-place conditions when the high-level ra d i o nu clide waste is placed in the emplacement dri f t s . B e rke l ey Lab’s role in the tests is to ch a ra cterize the heater test ro ck bl o ck (area) prior to testing; monitor potential changes in fra c t u re and matrix saturations through air injections, t racer testing, and ground penetrating radar measurements; and perform pre d i c t i ve therm o - hy d ro l o gical and thermohy d rologic-chemical calculations. The initial ch a ra c t e rizations of the heater test areas are perfo rmed with air injection tests that yield the 3-D permeability stru c t u re of the fra c t u re netwo rk . C o n t i nued air-injection testing during the heating and cooling phases of the heater test yielded changes that can be attributed to ch a n ges in fra c t u re satu rations or mechanical effects. C ross-hole radar tomogra p hy has also yielded ve ry promising results re g a rding ch a n ge in global saturations of the system due to heating. Detailed 3-D thermo-hydro l o gic and thermo-hydrologic-chemical calculations we re used to predict the behavior of the tests. These will be re fined to better fit observations as the test pro gresses. Fi n a l ly, B e rke l ey Lab scientists are invo l ved with m e a s u rements of the isotopic compositions of condensate water that have seeped into the boreholes.

Funding The Nuclear Waste Program’s Yucca Mountain Project research is supported by the Office of Civilian Radioactive Waste M a n agement, U.S. D e p a rtment of Energy, through Memorandum Purchase Order EA9013MC5X between TRW Env i ronmental Safety Systems Inc., and the Ernest Orlando Lawrence Berkeley National Laboratory under DOE Contract No. DE-AC03-76SF00098.

http://www-esd.lbl.gov 52


Earth Sciences Division

Research

Objectives

Nuclear Waste Program Ambient Hydrologic Testing In the Exploratory Studies Facility: Niche Study

Annual Report 1998-1999

drilled within the limits of the proposed niche and were subsequently The potential exists for a capillary removed when the niche was conRobert C. Trautz barrier to form in unsaturated porous structed and before the seepage tests and Joseph S.Y. Wang media at the boundary between a we re perfo rm e d . All the boreholes fine-grained unit, such as a highly fracwere installed parallel to the axis of Contact: tured rock, and an underground cavthe niche. Robert Trautz (510) 486-7954, rctrautz@lbl.gov ity. Conditions such as these exist at More than 200 air-injection tests the proposed radioactive-waste reposwe re conducted in the boreholes itory located at Yucca Mountain, Nevada. If licensed and conprior to excavation to determine the distribution of air permestructed, the monitored geologic repository will include 117 km ability within the rock mass and to select test intervals for subof mine openings or waste emplacement drifts housing containsequent liquid-release tests. Next, a series of liquid-release tests ers of radioactive waste.The repository would be constructed in were conducted by pumping water containing colored or fluothe unsaturated zone at least 200 meters (m) below the land surrescent dyes at a constant rate into a select number of packedface and at least 200-250 m above the regional water table within off zones previously tested with air. A finite amount of dyethe fine-grained Topopah Spring Tuff (Tpt), a densely welded, spiked water was introduced into each test interval, with essenintensely fractured, ash-flow tuff. tially no pressure buildup, to document the relatively undisIt is important to determine whether a capillary barrier will turbed flow path traveled by the wetting front. exist above a waste emplacement drift because such a barrier The niche was excavated dry using a mechanical excavator to can have a direct impact on waste isolation and repository perobserve and photograph the distribution of fractures and dye formance. Water introduced at the land surface during a natural within the rock. Dye was observed along individual fractures as precipitation event, assuming it migrates to the repository level, well as along intersecting fractures to depths ranging from 0 to may be excluded from seeping into the drift and instead may be 2.6 m below the release points.Two primary types of flow paths diverted laterally around the opening if a capillary barrier exists. were observed during the mining operation, including: (1) flow In contrast, if a capillary barrier does not form, then water may through individual or small groups of high-angle fractures, and (2) flow through several interconnected low and high-angle fracdrip into the opening and come in contact with the waste packtures creating a fracture network. age, hastening canister corrosion and failure, and potentially leadAir-injection tests were repeated in the upper boreholes after ing to radionuclide migration to the accessible environment. the niche was excavated to determine the post-construction disAn extensive series of liquid-release tests was performed tribution of air permeabilities. The geometric mean air permeabove a specially constructed drift, called a niche, located in the ability for all the boreholes tested increased dramatically by Exploratory Studies Facility at Yucca Mountain to determine nearly two orders of magnitude after the niche was excavated, whether a capillary barrier exists. In addition,the tests were used probably due to mining-induced damage to the formation. to quantify the seepage threshold flux, defined as the liquidForty short-duration seepage tests were conducted after the release flux at and below which water will no longer seep into niche was excavated by pumping water at a constant rate into the opening. the boreholes located above the Approach niche.After water appeared at the niche ceiling and dripped into a Seven 10-m-long bore h o l e s specially constructed capture syswere drilled at the niche to gain tem located within the nich e access to the rock for air injec(Figure 2), it was collected and tion and liquid-release tests prior weighed. The mass of water capto mining the niche.Three of the t u red ra n ged from 0 to 568.6 borings (shown in Fi g u re 1) grams per test. The seepage perwe re installed approx i m a t e ly centage, defined as the mass of one meter apart, 0.65 m above water that dripped into the capthe crown of the niche in the ture system divided by the mass same horizontal plane. The of water released, ranged from 0% remaining boreholes we re for very low permeability zones

Figure 1. Test configuration above the niche representing a waste emplacement drift.

53


ESD

Ambient Hydrologic Testing in the Exploratory Studies Facility: Niche Study

to 56.2% for pre d o m i n a t e ly gravity driven fl ow through highly saturated fractures.

this investigation and future studies will be used in part to design and construct the monitored geologic repository.

Results The seepage threshold data were interpreted using an analytical solution describing the ex clusion of water from a b u ried cylindrical cavity analogous to a niche. The resulting analysis produced estimates of the quantity 2α-1 defined as the sorptive length, a useful parameter that characterizes the capillary pro p e rties of the unsaturated medium. Values of 2α-1 equal to 19.6 and 981 Pascals were measured and believed to be re p re s e n t a t i ve of fl ow through two types of in-situ fractures,including individual or small groups of high-angle fract u res and fra c t u re networks, respectively. The sorptive number α was used along with wetting front arri val times to estimate fracture-water characteri stic curves (volumetric wa t e r content, θ, versus water potential, Ψ) for several test zones. These curves indicate that the residual θ may be on the order of 0.1% and that the saturated θ (i.e., e ffective porosity) may be as high as 2.4% for the seepage flow paths tested.

Significance of Findings

Related Publications

Figure 2. Arrival of the wetting front containing pyranine at the niche ceiling during a test.

Figure 3. Philip’s et al. (1989) seepage exclusion model versus the experimental results.

Trautz, R.C., J.S.Y. Wang and P.J. Cook,Chapter 2:Compilation of seepage results from the preand post-liquid-release tests, in D rift seepage test and niche moisture study, Phase 1 re p o rt on flux threshold determ i n ation, a i r - p e rm e ability distri b ution, and water potential measurement, J.S.Y. Wang (ed.), Milestone Report SPC315M4, LBNL, June 1998. Trautz, R.C., J.S.Y. Wang, P.J. Cook, R. Salve, A.L. James, M.Q. Hu and S. Finsterle, Chapter 2: Flow characterization and drift seepage evaluation in the niches, in Progress report on fracture flow, drift seepage and matrix imbibition test in the ex p l o ra t o ry studies facilities, J.S.Y. Wang (ed.), LBNL, Sept. 1998. Birkholzer, J.T., G. Li, C.F.Tsang, Y.W. Tsang, R.C. Trautz and J.S.Y. Wang, Drift scale modeling: Studies of seepage into a drift, Milestone Report SP3CKLM4, LBNL, Sept. 1998.

Funding

This work has been supported by the Director, Office of Civilian Radioactive Waste Management, through Memorandum P u rchase Order EA9013MC5X between TRW Env i ro n m e n t a l Safety Systems, Inc., and Ernest Orlando Lawrence Berkeley National Laboratory for the Yucca Mountain Site Characterization Project under U.S. Department of Energy Contract No. DE-AC0376SF00098.

Direct field observation, indirect field evidence and analytical and numerical models were used to confirm that a capillary barrier exists above an underground cavity. In addition, the seepage tests demonstrated that a seepage threshold flux does exist below which water will not seep into the opening.The results of

http://www-esd.lbl.gov 54


Nuclear Waste Program

Earth Sciences Division

Research

Objectives

Spatial and Temporal Flow Variability in the Paintbrush Nonwelded Tuff

Annual Report 1998-1999

focused flow to the PTn from the ove r lying fra c t u red Ti va Canyo n (TCw) unit during transient high infiltration events.

An understanding of pro c e s s e s Rohit Salve and influencing percolation flux, and the Curtis M. Oldenburg partitioning of fl ow between the Approach matrix and fractures and/or faults, is Contact: Rohit Salve important for the development of an (510) 486-6416, r_salve@lbl.gov Field experiments we re conducted a p p ro p riate conceptual model of in the north face of Alcove 4 within flow for Yucca Mountain, Nevada. Of the Exploratory Studies Facility (ESF). particular importance is the role of In the north face of Alcove 4, the PTn the Paintbrush Nonwelded Tuff (PTn), is cut by a normal fault with a small offset (~0.25 m) and a high through which downward percolation must pass prior to reachangle fracture (Figure 1). In-situ flow experiments involved the ing the potential repository horizon (i.e., the fractured Topopah release of different volumes and rates of tracer-tagged water in Spring Welded (TSw) unit). A fundamental question is whether intervals 0.3 m in length along horizontal boreholes intersecting the PTn is effective in dampening pulses of localized infiltration, the fault and matrix. Twelve 0.075-m diameter boreholes were providing a generally uniform percolation flux to the TSw, or drilled perpendicular to the plane of the north face of Alcove 4 whether flow is localized through preferential flow paths in the for liquid release and monitoring. Changes in percolation rates PTn. and saturation we re monitored using a nest of boreholes We have carried out field studies of flow through the variably installed with psychrometers and electrical resistivity probes altered, nonwelded PTn. The overriding objective of this effort (ERPs).A slot 4 m wide, 4 m deep and 0.3 m high was excavated was to gain insight into flow dynamics along a fault and within below the test bed to collect any injected fluid percolating the altered matrix of the PTn. Specifically, the objectives were to downward. measure flow velocities, seepage rates and wetting front migraThe nature of matrix flow in the PTn was investigated using tion within a fault and in the matrix of the PTn.The tests were boreholes 5, 6, 7 and 8 (Figure 1), located in a section of the test designed to mimic conditions analogous to the introduction of

Figure 1. Geological sketch and borehole/slot layout for the north face of Alcove 4.

55


ESD

Spatial and Temporal Flow Variability in the Paintbrush Nonwelded Tuff

bed that was determined to be wetting front can be inferred free of fra c t u res and faults. from the decrease in resistivity Water was continu o u s ly at various distances from the injected under constant head collar (as shown in Figure 2). conditions into a zone located Initially, wetting was observed between 2.44 and 2.74 m from along the projected fault, but the collar in borehole 5, for a with subsequent tests the wetperiod of 29 day s , while ting width incre a s e d . At the changes in saturation and water fault (i.e., at a distance of 1.4 m potential were monitored along from the collar), the ERPs the lengths of boreholes responded to individual release 6, 7 and 8. events, showing a wetting trend Two boreholes (11 and 12) during fluid injection and a drywe re positioned to intercept ing trend between and after the the fault at a distance of ~1.4 m Figure 2. Changes in electrical resistance in borehole 11 following injec- release tests. Further from the from the alcove face and used tion of water along the fault in borehole 12. fault, the probes detected gradto determine specific flow charual increases in moisture for the acteristics of the fault. Here, water was injected under a constant duration of the tests.The travel time of the wetting front within head condition in borehole 12 at a distance of 1.4 m from the the fault from the injection zone in borehole 12 to borehole 11, collar (i.e., along the fault) during seven separate release events. a distance of 1.0 m, was initially 3.3 hours and gradually During and after the release events, changes in saturation and increased to 10 hours. water potential were continuously monitored in borehole 11.

Significance

of

Findings

Results These series of experiments have demonstrated for the first time that active flow tests can be conducted within a fault in the vadose zone of Yucca Mountain. Techniques developed during the design of these experiments have permitted investigations of specific hydrologic properties of a fault located within the PTn. Results to date suggest that the variably altered PTn matrix at Alcove 4 dampens flow pulses. Tests in the fault suggest that as increasing amounts of water are introduced into the fault, the rate at which water can be released to the fault continuously decreases.Travel times within the fault were found to be dependent both on the release rates (i.e., for fast release rates the travel times were faster), and wetting history (i.e., in successive injections travel was faster, but with delays of 4-6 days between injections, travel times were significantly slower).

Liquid-release tests suggest that a large variability exists in the hydrologic response between the fault and matrix in the Alcove 4 test bed.Two components of the test that demonstrate this variability are the measured rates of liquid release and the capacities of the matrix and fault to transmit water. Flow dynamics in the matrix: When water was released into the matrix in borehole 5, the rates at which the matrix absorbed water rapidly decreased from >50 ml/min to <1 ml/min in less than five hours. During this early stage of the test ~1.7 liters of water had been released to the formation. During the next eight days the liquid release rate to the formation continued to decrease, asymptotically approaching 0.14 ml/min after ~4.0 liters of water had been released.This rate then persisted for the remainder of the test. The wetting front in the matrix was detected 0.5 m below the point of release 10 days after the test was started. Here changes in saturation were detected by a single ERP located 2.9 m from the collar in borehole 6. Flow dynamics in the fault: Liquid-release rates in the fault gradually fell from ~200 ml/min to 50 ml/min over a period of 41 hours, during which 200 liters of water percolated into the fault. A wetting front was observed along borehole 11 over a distance of 1.5 m following the seven release events in borehole 12.This

Funding This work has been supported by the Director, Office of Civilian Radioactive Waste Management, through Memorandum P u rchase Order EA9013MC5X between TRW Env i ro n m e n t a l Safety Systems, Inc., and the Ernest Orlando Lawrence Berkeley National Laboratory for the Yucca Mountain Site Characterization Project under U.S. Department of Energy Contract No. DE-AC0376SF00098.

http://www-esd.lbl.gov 56


Earth Sciences Division

Research

Objectives

Nuclear Waste Program

Annual Report 1998-1999

Field Investigations Of Fracture Flow In Welded Tuffs

holes were drilled perpendicular to the alcove wall. Of these, one (boreThe view of fractures as primary hole A) was used for fluid injection conduits for flow in the unsaturated Rohit Salve and Joseph S.Y. Wang while the other three (boreholes B, C zone of Yucca Mountain, Nevada, has and D) were monitored for changes in Contact: Rohit Salve replaced the belief of the last decade, moisture conditions. A slot excavated (510) 486-6416, r_salve@lbl.gov w h i ch held that flow pri m a ri ly b e l ow the test bed was used to occurred through the ro ck matrix observe the location and amount of under ambient conditions. This shift in paradigm has resulted water seeping from the formation above.There were three disfrom indirect evidence of fracture flow as suggested by a relatinct components to the fl ow inve s t i g a t i o n : (1) controlled tively young perched water table below the Topopah Spring release of water into isolated zones, (2) borehole monitoring for Welded (TSw) unit, mineral deposits on fractures, and the detecchanges in saturation and water potentials, and (3) monitoring of tion of bomb-pulse 36Cl in isolated locations within the mounseepage from the slot ceiling. tain. However, direct evidence of fracture flow remains elusive, Experiments were conducted to determine flow dynamics largely because of technical difficulties encountered in locating resulting from liquid release in two zones of different permeand measuring fracture flow in this underground site. abilities within a single borehole (borehole A).These zones were This paper presents the results of a field investigation of fracidentified from air-permeability measurements to lie 0.75-1.05 m ture flow in the Exploratory Studies Facility (ESF) at Yucca (low permeability) and 2.30-2.60 m (high permeability) from the Mountain.The objective of this effort was to estimate hydraulic borehole collar. In both the high permeability zone (HPZ) and parameters such as formation intake rates, flow velocities, seeplow permeability zone (LPZ), a series of tests were conducted to age rates and fracture porosities using techniques developed for determine the temporal changes in the rate at which the formain-situ testing of flow in fractured rock. Field tests were designed tion could take in water. In the HPZ, a second series of tests was to mimic conditions analogous to releases of contaminated fluid conducted during which the injection rates were changed and from the rupturing of containers or from transient high infiltrathe seepage rate into the slot was monitored. During the entire tion events leading to focused flow (i.e., point source releases of duration of the experiments, saturation and water potential 5â&#x20AC;&#x201C;20 liters of fluid). changes along the monitoring boreholes were continuously measured, and water that seeped into the slot was periodically Approach sampled and analyzed for tracer concentrations.

Results

Field experiments were conducted on the right rib of Alcove 6 in the Main Drift of the ESF (Figure 1a).The test bed was situated in the middle nonlithophysal portion of the Topopah Spring Tuff.Two distinct features defined the layout of the field experiment: horizontal boreholes and a slot (Figure 1b). Four bore-

The rate at which water moved into the LPZ steadily decreased with time, asymptotically approaching a steady rate (~0.35 ml/min). In the HPZ, the rates varied significantly during

Figure 1. Schematic (not to scale) of location (A) and layout (B) of test bed within the ESF at Yucca Mountain.

57


ESD

Field Investigations of Fracture Flow in Welded Tuffs

Figure 2. Seepage response observed in the slot in Alcove 6 in the ESF.

and between tests. In the two monitoring boreholes located below the injection borehole (C & D), changes in saturation were detected both by the electrical resistivity probes (ERPs) and psychrometers in response to liquid releases in both the LPZ and HPZ. For releases in the LPZ, large changes were detected between 0.9 and 1.9 m from the collar; for releases in the HPZ, changes were observed between 1.7 and 3.4 m from the alcove face. Seepage into the slot was observed directly below the injection zone during all eight tests in the HPZ. In the first test, after 0.41 liters of water had been introduced to the formation, water was first seen on the slot ceiling after five minutes. In the second and third tests, water appeared in the slot within three minutes after 0.17 and 0.14 liters, respectively, had been injected. In the final test in this batch, after 1.50 liters of water had been injected at a rate of 5 ml/min, water appeared in the slot after five hours. In the second batch of experiments, travel time for the first drop of water during the first injection (0.14 liters at a rate of ~69 ml/min) was three minutes. In the two subsequent tests, after 0.26 and 0.20 liters of water were injected at a rate of 38 and 29 ml/min, the arrival time of the wetting front was seven minutes. In the final test in this batch, , after 0.90 liters had been injected into the formation at a rate of 14 ml/min, water appeared in the slot after just over one hour. The amount of water recovered in the slot continued to increase as each test progressed. However, significant variability existed in the percentage of water recovered (Figure 2a) and the seepage rate (Figure 2b) during and between tests.This variability was a function of both the amount of water injected and the rate at which water was released into the formation. Early in each test, the amount of water recovered sharply increased until approximately 10 liters of water had been injected, after which the percentage of injected water recovered

approached a relatively constant value. The amount of injected water recovered was consistently higher for all injected volumes at release rates of 28 and 38 ml/min. In all of the tests during which there was seepage, 0.5-1.3 liters of water entered the slot after water supply to the formation was switched off. The constant head test had a â&#x20AC;&#x153;steppedâ&#x20AC;? nature to the post-injection recovery; here, during the first 15 minutes, the 0.8 liters of collected water appeared in four bursts, each containing 0.1-0.3 liters of water.

Significance

of

Findings

From these experiments, we were able to establish the effectiveness of new techniques developed for in-situ characterization of certain fundamental flow parameters (e.g., travel times, percolation and seepage rates). Results from liquid-release tests in the LPZ and HPZ suggest the presence of both open- and closeended fractures and provide estimates of the volume of water occupied by both fast flow paths and close-ended fractures.This data, when coupled with data from the seepage response to different liquid injection rates, has led to certain hypotheses, which, when tested, should provide new directions for conceptualization of flow within these unsaturated, fractured systems.

Funding This work has been supported by the Director, Office of Civilian Radioactive Waste Management, through Memorandum P u rchase Order EA9013MC5X between TRW Environmental Safety Systems, Inc., and the Ernest Orlando Lawrence Berkeley National Lab o ra t o ry for the Yucca Mountain Site Characterization Project under U. S . Department of Energy Contract No. DE-AC03-76SF00098.

http://www-esd.lbl.gov 58


Earth Sciences Division

Nuclear Waste Program

Annual Report 1998-1999

A Probe for Measuring Saturation Changes in Rock Research

Objectives

comprised two electrical leads sandwiched between pieces of filter paper Rohit Salve, Tetsu K. Tokunaga In-situ measurements of hydrologic (Figure 1a). To maintain a consistent and Joseph S. Y. Wang p a ra m e t e rs at Yucca Mountain, geometry between probes, all electriContact: N evada, fo rm an integral part of cal leads were cut to the same length Rohit Salve efforts directed towards characterizaand meshed through a nylon fabric (510) 486-6416, r_salve@lbl.gov tion of flow and transport in the such that the distance between wires unsaturated zone of this potential site for high-level nuclear was the same for all sensors. Meshed nylon fabric was also used waste disposal. An important feature in this environment is the to provide a protective outer sleeve for the probes.The size of fractured, consolidated rock formation, which contains fractures the probes was determined by the type of application and of varying densities, orientations, spatial extents and apertures. ranged from squares with 0.02-m sides used for laboratory colBecause of this, unique problems arise with regard to instrument umn experiments to 0.075 m for borehole monitoring. placement and sampling techniques. While theories, methods Borehole Sensor Trays: To locate ERPs and psychrometers and experimental results developed from unconsolidated media in multiple locations along borehole walls, t rays we re designed to have initially formed the basis for the characterization of unsatup e rmit multiple arrays of sensors to be easily installed and rated consolidated media, re m oved in the boreholes, n ew techniques are now independent of bolehole orirequired (in addition to those entation. These trays we re established) to monitor fracfab ricated from PVC pipes. tured rocks. The outer diameter of each The primary objective of tray was selected to match this effort was to develop a the diameter of the borehole reliable tool for monitoring being instrumented (Figure the migration of a wetting 1b). Psychrometers we re f ront in boreholes located located in small diameter holes (~ 3 mm ID) dri l l e d within fra c t u red ro ck env ithrough the trays. ERPs were ronments. In this paper we attached to the outer surface p resent the design of of the trays with strips of Electrical Resistivity Probes Velcro.This assembly perm i t(ERPs) and Borehole Sensor ted ex t e n s i ve contact Trays (BSTs). We also include between the ERPs and the some results from ex p e riborehole wall while allowing ments conducted to evaluate the psychrometers to contact the ERPs in the laboratory the borehole wall through a and in boreholes within the small cav i t y. On each tray, Exploratory Studies Fa c i l i t y p s y ch ro m e t e rs we re (ESF) at Yucca Mountain. installed at typical distances Approach of 0.25-0.5 m along the borehole while ERPs we re Electrical Resistivity located at 0.25-m interva l s . Probes: The basic concept Two BSTs we re located defining the working of ERPs along each section of boreis that increasing water satuhole to permit sensors to be ration results in decreasing located on opposite sides.To e l e c t rical resistance. Probes i m p rove the contact were developed using a piece b e t ween probes and the of filter paper across which borehole wall, an adjustable changes in electrical re s i s twedge system was used. ance could be measured. The Results paper permits a flexible sensing surface that could be Figure 1. Design of electrical resitivity probes and borehole sensor trays. Water imbibition in a molded to the contours of basalt column: To test ERP the boreholes. These probes 59


ESD

A Probe for Measuring Saturation Changes in Rock

function in the laboratory, six Significance of Findings ERPs (each a square with 0.02 m sides) were attached to the We were able to establish walls of a basalt column (0.20 that ERPs are effective for labom height, 0.06 m diameter) ratory scale experiments lookafter the column had been ing at water movement in the placed in a pool of water 0.01 rock matrix, in which both the m deep.As the column imbibed test bed and the sampling resowater, resistance changes along lution can be relatively small. the surface were continuously From field investigations, the measured. response of the ERPs compared Te m p o ral changes in re s i s twell with the performance of ance measured across the six psychrometers, suggesting that ERPs show the migration of a these probes (with their curwetting front along the basalt rent design) can be effectively column (Fi g u re 2a).The initially used as a qualitative tool to low resistance measured in the detect the arri val (or departwo lower ERPs reflects the ture) of wetting fronts in rock close proximity of the water matrix, fractures and faults. table.The diffe rence in the temUnlike psychrometers, these poral response observed in the probes are relatively inexpent wo higher zones suggests that sive, easy to maintain and have despite the wetting front re a cha low failure rate. This makes ing the higher probes,a gradient them part i c u l a r ly useful fo r in saturation developed with extensive down-hole monitorthe bottom of the column being ing applications in unsaturated wetter than the higher zones. fra c t u red ro ck env i ro n m e n t s Wetting front detection such as that found at Yucca following tunneling activiMountain. ties: At the ESF, plumes of water When installed on BSTs resulting from tunneling activialong with psych ro m e t e rs, ties (cutter head cooling, dust these ERPs provide an effective control, etc.) in the vicinity of a Figure 2. ERP responses observed in (a) laboratory and (b) field compilation of sensors, which new tunnel were tracked with experiments. provide data on relative wetERPs and psych ro m e t e rs. To ness and water potential changes at high spatial resolution.With monitor the plumes, a 30-m-long borehole (0.10 m ID) was constructed sloping downwards at an angle of 30° under the proERPs, the arrival (and departure) of a wetting front can be measposed path of the tunnel. Twenty-seven psychrometers and 54 ured at time intervals of less that a minute, while the spatial electrical resistivity probes located on nine BSTs were installed extent can be measured to a resolution of less than 0.10 m. in this borehole. Changes in water potential and electrical resistBackfill is not needed, increasing the accuracy of measurements ance were monitored along the entire length of the borehole and the relative ease of installing and removing sensors.The BSTs using psychrometers and ERPs located on BSTs as tunneling have the potential for further improving our abilities to measure activities progressed through the formation above. flow parameters in the unsaturated zone of fractured rocks. The ERPs responded in a pattern similar to that of the psyFunding chrometers located adjacent to the probes. For example, at a depth of 9.4 m where the water potential increased steadily from This work has been supported by the Director, Office of –400 m to –70 m, the corresponding ERP measurements followed Civilian Radioactive Waste Management, through Memorandum a similar pattern (Fi g u re 2b). Large fluctuations in water potenP u rchase Order EA9013MC5X between TRW Env i ro n m e n t a l tials in relatively short periods of time (~200 m in four days) we re Safety Systems, Inc., and the Ernest Orlando Lawrence detected by both types of pro b e s .The slower, m o re gradual Berkeley National Laboratory for the Yucca Mountain re c overy observed by psychrometers deeper in the fo rSite Characterization Project under U.S. Department of mation was also well tra cked by the ERPs (e.g., at 21.3 m; Energy Contract No. DE-AC03-76SF00098. Fi g u re 2c). http://www-esd.lbl.gov 60


Earth Sciences Division

Nuclear Waste Program

Annual Report 1998-1999

The Restricted Interval Guelph Permeameter Barry Freifeld and Curt Oldenburg

Research

Objectives (510)

is closed and the test is underway.The volume of water introduced into the formation is read off a scale on the Mariotte Siphon reservoir and recorded as a function of time until a steady-state flow rate is reached. By performing two injection tests at the same location using different applied heads, a method has been developed that allows for the determination of both the field-saturated hydraulic conductivity and the sorptive number.A closed-form expression for flow from a saturated sphere using dimensionless spatial coordinates, neglecting the influence of gravity, has been derived by A.W.Warrick, and is used here:

Contact: Barry Freifeld 486-4381, bmfreifeld@lbl.gov

Contaminant transport in arid climates is often controlled by thick vadose zones.The in-situ determination of hydraulic conductivity and other relevant transport parameters in these unsaturated systems has been hampered by the lack of practical methods for reliably making measurements and interpreting the results. The Restricted Interval Guelph (RIG) permeameter has been developed to fill this void by making measurements in saturated/unsaturated systems.Whereas the original Guelph permeameter is known to produce unreliable results for heterogeneous systems, the RIG permeameter was developed specifically to overcome this limitation.A simple analytical method for reducing the data has been developed based on saturated/unsaturated flow theory.The analytical data analysis approach is compared with a numerical simulation that more accurately reproduces actual test geometry. Finally, sensitivity analysis is used to verify the applicability of the RIG methodology for determining transport parameters for a field application in a rhyolitic tuff.

K fs =

αq  1 1   − . 8π Ψ  r r0 

(1)

The ab ove equation is based upon a solution of Richards equation where water is the mobile liquid phase, and air is immobile. ψ( m ) is the head applied during testing and q (m3/s) is the measured s t e a dy-state vo l u m e t ric fl ow rate. The field saturated hydraulic Approach conductivity, Kfs(m/s), and the sorptive number, α( m-1), are the u n k n owns to be determined through the simultaneous applicaThe RIG permeameter is similar to the original Guelph pertion of Equation 1 for two injection tests conducted with different meameter, in that it uses a Mariotte Siphon reservoir to produce heads at the same location.The radius at which the pressure head is applied, t a ken as an equivaa constant head injection test. lent spherical radius, is r, and r0 However, unlike the Guelph is the radius of the saturated permeameter, the re gion re gion (ψ = 0). The equiva l e n t tested is isolated using either spherical radius corresponds to a single pneumatic packer or a sphere with the same surface a straddle packer (Figure 1). a rea as is wetted during the This results in a fixed interval actual field test.The radial coorto be tested, which will not dinates are nondimensionalized change even when changing by α/ 2 . the applied head.The applied Forward simulations of RIG head, H in Figure 1, is equal to permeametry were carried out the diffe rence in elevation using the general multiphase between the air reference line and multicomponent poro u s in the Mariotte Siphon resermedia tra n s p o rt simu l a t o r voir and the elevation of the TOUGH2. Similar to the analytiborehole zone being tested. cal approach, the single-phase A test is conducted by first equation of state module that isolating the region of interest governs flow in the numerical with the pneumatic packer(s). model solves Rich a rds equaIt should be noted that the tion. A radial mesh with 871 borehole can be oriented in grid blocks was constructed to any direction.The air exit line more re a l i s t i c a l ly re p ro d u c e is opened as water is introduced through the water line. the actual test geometry. An Figure 1. Schematic of Restricted Intervals Guelph permeameter. When only water flows out N.O.=normally open and N.C.=normally closed. inverse model based upon the from the air exit line, the line ITOUGH2 code was used to 61


ESD

Borehole

G1 G1 G2 G2

The Restricted Interval Guelph Permeameter

Interval Length (cm)

6.75 6.75 8.65 8.65

Head (cm)

33.0 66.0 25.0 42.5

Steady State Q (ml/sec)

Analytical Kfs (m/s)

0.0488 0.0632 0.0037 0.0041

Analytical α (m-1)

Numerical Kfs (m/s)

Numerical α (m-1)

1.6×10-7

1.3

2.0×10-7

2.5

7.5×10-9

0.73

1.0×10-8

1.6

Table 1. Results of RIG permeametry in an unsaturated rhyolitic tuff showing a comparison between a simple analytical model and a numerical model which more accurately depicts testing geometry.

number, α. Both initial liquid saturation and formation porosity were shown to play a secondary role in determining the amount of water introduced into the formation. The capillary pressure parameters were shown to have minimal effect on test results.

verify the sensitivity of the volumetric flow rate to the parameters to be estimated,Kfs and α, and validate the RIG methodology.

Results

Significance

RIG permeametry was used to estimate the field-saturated hydraulic conductivity and sorptive number for an unsaturated rhyolitic tuff at Yucca Mountain, Nevada. Measurements were carried out at the ends of two 2.54-cm-diameter, 1-m-deep, horizontal boreholes. The first borehole, which we refer to as G1, was drilled into a high porosity (φ=0.44) rhyolitic tuff, containing abundant coarse pumice and lithic fragments. Borehole G2 was drilled 0.7 m higher in elevation at the argillically altered upper portion of the same stratigraphic unit, which had a significantly reduced porosity of 0.25. Table 1 lists relevant test conditions for each injection as well as the steady-state flow rates recorded. S i multaneous application of Equation 1 was used for each of the two applied heads and measured steady-state flux rates to solve for the hydraulic conductivity, Kfs, and the sorptive number, α. Although the actual injection geometry is cylindrical, the test interval is small enough that we use an equivalent spherical radius based on the surface area of the cylinder.The G1 borehole was determined to have a field-saturated hy d raulic conductivity two ord e rs of magnitude less than the hy d raulic conductivity calculated for the G2 borehole, which is located in the argi l l i c a l ly altered layer. The numerical model provided similar results.Table 1 also shows that the value for α is in good agreement between the analytical and numerical model. ITOUGH2 simulations were used to determine the sensitivity of the volume of water released to porosity, capillary pressure parameters, sorptive number, absolute permeability and the initial liquid saturation. It was determined that the most sensitive parameter was absolute permeability, k, followed by sorptive

of

Findings

The RIG permeameter has shown itself to be a versatile instrument for vadose zone characterization. Its simplicity and low cost make it an attractive method for measuring in-situ field-saturated hydraulic conductivity and estimating the sorptive number. Good agreement between the analytic and numerical results indicates that the analytical spherical flow model used is a valid approach for analyzing the data collected. Furthermore, the sensitivity analysis shows that RIG permeametry as carried out here is an effective method for estimating k and α for the nonwelded rhyolitic tuff that was tested.

Related

Publication

Freifeld, B. M., and C. M., Oldenburg, The Restricted Interval Guelph Permeameter:Application and analysis, Berkeley Lab report LBNL-42135, 1999.

Funding This work has been supported by the Director, Office of Civilian Radioactive Waste Management, through Memorandum P u rchase Order EA9013MC5X between TRW Env i ro n m e n t a l Safety Systems, Inc., and Ernest Orlando Lawrence Berkeley National Laboratory for the Yucca Mountain Site Characterization Project under U.S. Department of Energy Contract No. DE-AC0376SF00098.

http://www-esd.lbl.gov 62


Earth Sciences Division

Research

Objectives

Nuclear Waste Program Laboratory Measurement Of Water Imbibition Into Low-Permeability Welded Tuff

Annual Report 1998-1999

Results

The partitioning of water between An example of cumulative imbibifractures and matrix is determined by tion versus time for a tuff sample is Qinhong Hu, Peter Persoff the relative rates of imbibition and shown in Figure 1.The measured data and Joseph S.Y. Wang fracture flow. This partition is very without corrections are compared important for the repository design with (1) evaporation correction, and Contact: Qinhong Hu and performance assessment for the (2) both evaporation and imbibition (510) 495-2338, q_hu@lbl.gov proposed high-level waste repository corrections. The data agree with the at Yucca Mountain, Nevada. In this work, laboratory imbibition final weighing of the core sample, which was independently experiments were developed to accurately measure water imbimeasured. Over a total of 109 imbibition tests, the average difbition into welded tuff of low permeability. For water flow ference between the corrected mass for water imbibition and through fractured rocks with low matrix permeability, the fast the balance-indicated mass is 20.7 % (with a standard deviation flow component is through the fracture network. Some water of 2.3%). Thus it is important to account for buoyant force flowing along the fractures will imbibe into the matrix blocks decrease in imbibition measurements. because of strong capillarity from the extremely small-sized Plots of cumulative imbibition versus the square-root of time pores of the matrix. Imbibition of water into a porous medium are shown in Figure 2, with the sorptivity determined from the typically proceeds in proportion to the square root of time, with slope of the linear segment. Linear segments are evident only for the proportionality constant denoted as the sorptivity. Sorptivity the “early” times, which differ between the two cores: a relatively is jointly controlled by the perm e ability and capillary pressure, uniform Core-A and a fractured Core-1. Deviation from linearity and is a useful parameter for characterizing transient imbibition in Figure 2 can arise from several sources, including (1) the leading edge of the wetting front reaching the opposite boundary, processes,e s p e c i a l lywhen the influence of gravity is insignificant. (2) heterogeneities in hydraulic properties along the length of Approach the sample, and (3) gravitational effects. For these tuff samples under initially dry conditions and short core length, gravitational Rock cores (5.08 cm in diameter and 2.0 cm in length), oveneffects can be neglected. dried at 60oC, were used for the imbibition experiment. Cores The measured porosity for is 0.160 for Core-1 and 0.0821 for were cut and machined from a sample block, which was colCore-A; the difference is attributed to the presence of the fraclected within the middle non-lithophysal zone of the Topopah ture and the altered band regions around the fracture in Core-1. Spring welded unit at Yucca Mountain, where the potential During imbibition tests for Core-1, it was observed that water repository for waste canister emplacement will be located. was imbibed preferentially into the band regions, which preTo measure imbibition rates into cores, the rock core was sumably possess a higher permeability. The wetting fro n t placed in a holder that was connected to a bottom-weighing reached the core top in less than 20 minutes (which correelectronic balance. The rock core was suspended inside a consponds to a water saturation of 35% for the whole core); this corstant humidity chamber, with the core bottom immersed into responds very well to the time when the cumulative imbibition water in a reservoir at about a 1-mm depth. A PC and LabVIEW starts to show curvature (Figure 2B). Wetting front propagation program were used to acquire and store the balance readings, at a for Core-A imbibition was more or less uniform, travelling less user-specified logging interval, over time. than 0.3 cm from the core bottom in 30 minutes. For Core-A, the The water level in the reserlinear segment ends after 29 voir fell because of imbibition hours, which is likely related to and evaporation, and the core the arrival of the wetting front to submergence depth decreased. the top of the core. At this time, This reduced the buoyant force the water saturation of Core-A is on the sample. As a result, the about 80%. sample appeared (from the balThe corrected increase in mass ance reading) to gain more water of water includes not only water than it actually did. A mathematimbibed but also water transical formulation for the buoyant ported by vapor-phase diffusion force change was developed and into pore space and condensaemployed to correct the experition in pores. The core sample, mental results to obtain “true” treated by coating the core walls water mass imbibed into the and covering the core top (but core sample. a small hole Figure 1. Buoyant force change correction during water imbibition into leaving for air to escape), was hung rock core. 63


ESD

Laboratory Measurement of Water Imbibition Into Low-Permeability Welded Tuff

above the water reservoir inside well with independently measthe humidity-controlled chamured total uptake. High contrasts ber; core weight gain by conin sorptivities between fra cdensation was compared to the tured and unfractured cores are n o n - t reated case. The re s u l t s related to high contrasts in perindicate that the coating and meabilities between fra c t u res coverage can control/minimize and the tuff matrix. Capillary capillary condensation into the condensation can provide a sigcore. For Core-A, the condensanificant contribution to the tion component could be as water uptake into cores during large as 18%. imbibition ex p e riments. Rock The measured sorptivity difcores were treated by coating fe rence for the core sample with and cove rage to reduce this d i ffe rent cove rages is found to effect. The method developed be consistent with the contri b uresults in an accurate measuretion from capillary condensament of sorptivity, as exhibited tion. The ave rage sorptivity is from the consistent and reproFigure 2. Relationship of cumulative imbibition and water saturation vs. 3 3 . 9 5Ă&#x2014;10-5 (m s-0.5) for the fra c- the square-root of time. (A) Core-A. (B) Core-1. ducible results of multiple meastured Core-1, and 3.58Ă&#x2014;10-6 (m surements. The method is 0.5) for the non-fra c t u red Core-A, with the sorptivity ratio (S expected to be very useful for measuring imbibition rates of very C o re th / S ) equal to 9.49. Since perm e ability is correlated to the 4 low permeability materials and/or materials with initial partial 1 Core-A p ower of sorptivity, the perm e ability for Core-1 is about 8,100 saturation. times greater than that for Core-A. Hence, the perm e ability for the Related Publication f ra c t u re surface zones that have undergone perm e ab i l i t y - e n h a n cing alterations is as large as 8,100 times that of the underlying Hu, M.Q., J.S.Wang, P.Persoff and R.C.Trautz,Water imbibition and ro ck matri x . tracer penetration into welded tuff, in Proceedings, Fall Significance of Findings meeting of the American Geophysical Union, EOS Trans. AGU, 79(45), p. F383, 1998. A well-controlled method has been developed for measuring Funding imbibition rates and determining sorptivities for low permeabilThis work has been supported by the Director,Office of Civilian ity media.Automatically recorded balance readings were used to R a d i o a c t i veWaste Management, through Memorandum Purchase quantify the imbibition of water into rock cores, even when the Order EA9013MC5X between TRW Env i ronmental Safety Systems, total mass of imbibed water was small and the duration of the Inc., and Ernest Orlando Lawrence Berkeley National Laboratory experiment was long.A formulation was developed to correct for for the Yucca Mountain Site Characterization Project under U.S. the apparent balance increase because of the buoyancy force Department of Energy Contract No. DE-AC03-76SF00098. decrease. Weight readings corrected for this effect agree very

http://www-esd.lbl.gov 64


Earth Sciences Division

Research

Objectives

Nuclear Waste Program High Resolution Studies Of Water Seepage in Unsaturated, Heterogeneous Rock Fractures

Annual Report 1998-1999

but we found that this is not adequate to capture the correlation structure Field evidence suggests that in near asperity contacts. Therefore, we semi-arid environments such as Yucca have developed enhancements to the Mountain, Nevada, water is able to Metropolis algorithm that can better Tai-Sheng Liou and Karsten Pruess migrate downward rapidly along prefrepresent the neighborhood of aspererential paths through fracture netity contacts. Contact: wo rks in part i a l ly saturated ro ck s , Flow simulations were carried out Tai-Sheng Liou without being imbibed into the rock with the general-purpose flow simula(510) 486-7083, tsliou@lbl.gov matrix. This raises concerns for the tor TOUGH2, using a special equationdisposal of nuclear waste because nuclides may transport downof-state flow module EOS9 which solves Richardsâ&#x20AC;&#x2122; equation; that stream with water.To help understand mechanisms of fast prefis, it is assumed that the gas phase acts as a passive bystander erential flow in unsaturated fractures, we have developed a highwith negligible pressure change during variably saturated water resolution numerical model for simulating fluid flow and transseepage. port in unsaturated fractured rocks. Such a numerical model Results re q u i res a detailed description of fra c t u re hetero ge n e i t y. Therefore, the objectives of our research are to develop a statisFi g u re 1 shows two realizations of permeability fields tical simulator that is able to capture realistic heterogeneities of annealed to the same variogram natural fractures and to provide a with the standard and modified better understanding of water seepMetropolis algorithms. It is obvious age behavior in unsaturated fracthat the permeability field tured rocks by simulating flow in annealed with the modified large ensembles of synthetic hetMetropolis algorithm shows more erogenous fractures. consistent correlation near asperi t y Approach contacts than that annealed with the standard Metropolis algorithm. In this research, we focus on In the subsequent flow simulasmall-apert u re fra c t u res in hard tions, water is injected uniformly over the entire top boundary at a rocks, such as tuffs, granites and constant rate of 10-3 kg/s.To model basalts. Fra c t u res are conceptualthe effect of normal stress on ized as two-dimensional heterogewater seepage, we generated fracneous porous media that can be tures with increasing fractions of characterized by a heterogeneous asperity contacts. One realization permeability field. Elements that are of water seepage at steady state is essential for a synthetic fracture to shown in Figure 2. Also shown in replicate natural fractures include Fi g u re 2 are the corresponding the presence of asperity contacts, a bre a k t h rough curves (BTCs) for gradual ch a n ge from asperities the seeps in Figure 2, as well as for towards larger aperture, small-scale other seeps with different fractions wall roughness, and a finite-size spaof asperity contacts. tial correlation length. We have d eveloped a statistical simu l a t o r Typical phenomena seen in our based on the algorithm of simulated nu m e rical seepage ex p e riments annealing for ge n e rating re a l i z ainclude flow bypassing, ponding, tions of natural fractures. Simulated fingering and flow exclusion. For annealing (SA) is an algorithm that each fraction of asperity contacts, operates in analogy to the physical we generated 30 realizations of process of annealing, moving a syssynthetic fractures and calculated tem towards a state of global minitheir effe c t i ve permeability. mum energy through successive Plotting the arithmetic mean of perturbations. The cl a s s i c Figure 1. Anisotropic permeability fields annealed with standard these effe c t i ve permeabilities and modified Metropolis algorithms. Permeability is the product of Metropolis algorithm is commonly a permeability modifier ( Îś) and a reference permeability (10-9m2). shows that effective permeability used as a perturbation mechanism, Asperity contacts are simply modeled as grid blocks with Îś=0. decreases nonlinearly with i n c re a s65


ESD

High Resolution Studies of Water Seepage in Unsaturated, Heterogeneous Rock Fractures

Figure 2. Liquid seeps at steady state in fractures subject to increasing normal stress. Water is injected uniformly at a constant rate of 10-3 kg/s over the entire top boundary. BTCs are also shown.

ing fraction of asperity contacts. Furthermore, BTCs in Fi g u re 2 show faster bre a k t h rough in fractures with increasing fraction of a s p e rity contacts. Thus, we have the remarkable result that flow can be faster in fractures with smaller effective perm e ab i l i t y. Dispersion of seepage in fractures depends strongly on asperity contacts. For fractures with a small fraction of asperity contacts, their effect on seepage is small, and all realizations have similar breakthrough behavior, with weak flow dispersion. For a large fraction of asperity contacts, only a limited number of fingers can break through and the dispersion again is weak. For intermediate fractions of asperity contacts, however, many fingers can be formed. Some of them are fast and some are slow, resulting in stronger dispersion. Simulation results not shown here indicate that seepage depends strongly on the spatial correlation of asperity contacts, but less on the spatial correlation of the permeability field. Especially strong effects on BTCs can result from asperity contacts with long-range spatial correlations, which can generate tortuous flow paths.

Significance

of

Combining these two findings suggests that water seepage in unsaturated fractures may be faster than in saturated ones, and seepage may become faster as normal stress increases. On the other hand, the asymptotic decrease of effective permeability with normal stress implies the existence of residual flow in fractures subject to large normal stress. Ponding at asperity contacts may have significant effects on water seepage and solute transport. It will in general delay downward seepage and cause the breakthrough time to increase. Dispersion may either increase or decrease due to ponding, depending on the geometric connection between ponded and flowing regions.

Related

Publications

Liou, T. S., K. Pruess and Y. Rubin, Numerical simulation experiments on water seepage in unsaturated, heterogeneous fractures, in Proceedings of the TOUGH Workshop â&#x20AC;&#x2122;98, pp. 198â&#x20AC;&#x201C;204, Berkeley Lab report LBNL-41995, 1998. Pruess, K., On water seepage and fast preferential flow in heterogeneous, unsaturated rock fractures, J. Contam. Hydr.,Vol. 30, pp. 333-362, 1998.

Findings

Our statistical simulations are capable of reproducing smallscale heterogeneities of natural rock fractures. Flow simulations carried out in synthetic fractures show seepage behavior in general agreement with field observations, such as fast preferential flow, flow fingering, bypassing and ponding. In addition, numerical simulations show a nonlinear decrease of effective permeability with the fraction of asperity contacts (normal stress).

Funding This work has been funded by the Office of Science, Office of Health and Env i ronmental Sciences, Biological and Environmental Research Program, of the U.S. Department of Energy under Contract No. DE-AC03-76SF00098.

http://www-esd.lbl.gov 66


Earth Sciences Division

Research

Objectives

Nuclear Waste Program An Active Fracture Model For Unsaturated Flow and Transport

Annual Report 1998-1999

distribution model. The latter assumes that liquid water occupies fra c t u res with small apert u res fi rst and then fra c t u res with re l a t i ve ly large apert u res as water potential increases. In contra s t , the active fra ct u re model presumes grav i t y - d o m inated, non-equilibrium, p re fe re n t i a l liquid water fl ow in fractures. The water distribution is not contro l l e d

Fracture/matrix (F/M) interaction is a key factor affecting flow and Hui-Hai Liu, t ra n s p o rt in unsaturated fra c t u re d Gudmundur S. Bodvarsson and Christine Doughty ro ck s . In classic continuum approaches, it is assumed that flow Contact: occurs through all the connected fracHui-Hai Liu tures and is uniformly distributed over (510) 486-6452, hhliu@lbl.gov the entire fracture area, which generally gives a relatively large F/M interaction. However, fractures by the capillari t y. seem to have limited interaction with the surrounding matrix at The fraction of active fractures in a connected fracture netYucca Mountain, Nevada, as suggested by geochemical nonequiwork, fa, should be determined by flow and transport conditions and fractured rock properties. An expression for fa must satisfy librium between the perched water (resulting mainly from fracthe following conditions: (1) all connected fractures are active if ture flow) and pore water in the rock matrix. the system is fully liquid saturated; (2) all fractures are inactive if Because of the importance of the F/M interaction and related the system is at residual saturation; and (3) fa should be related issues, there is a critical need to develop new approaches to to water flux in fractures. Based on these considerations, we accurately consider the interaction reduction inferred from field express fa as a power function of effective water saturation in data at the Yucca Mountain site. Motivated by this consideration, connected fractures, Se, we have developed an active fracture model based on the hypothesis that not all connected fractures actively conduct γ water in unsaturated fractured rocks. f a = Se

Approach

where γ is a positive constant depending on properties of the corresponding fracture network and can be considered as a measure of the “activity” of connected fractures. A generalized version of van Genuchten capillary pressure and relative permeability relations for the fracture continuum were developed based on the above equation for fa and the following considerations: (1) van Genuchten relations are assumed to be valid for active fractures; (2) inactive fractures are filtered out from the fracture continuum in describing flow and transport in fractures; and (3) the apparent F/M interface are a reduction factor (R) results not only from the actual interface area reduction, but also from the difference between active fracture spacing and fracture spacing, determined from the geometry of the corresponding fracture network. Note that the derived relation for the F/M interfa c e area reduction factor is used as an additional constitutive relation. Figure 1 shows the factor as a function of the effective water saturation.

Physically, the F/M interaction reduction can result from three mechanisms. First, because of gravitational instability and a heterogeneous aperture distribution, water flow through an unsaturated fracture is likely to be channelized, which can considerably reduce the F/M interface area. Second, fracture coating can reduce imbibition into the matrix, significantly in some cases. Third, because of the combination of the large nonlinearity i nvo l ved in an unsatura t e d system and heterogeneities of fracture structure at different scales, s i g n i ficant fingering flow can occur even in a wellconnected fracture network. In other words, only a portion of fractures in a connected network contribute to water flow, while other fractures are simply by-passed. For largescale flow and transport, the third mechanism is the most important. The portion of the connected fra c t u res that active ly conduct water are defined as active fractures. It is important to note diffe rences between the active fra c t u re model and the conve n t i o n a l , capillary equilib- Figure 1. The fracture-matrix interface area reduction factor (R ) curves for γ = 0, 0.5 and 0.9. rium-based, f ra c t u re wa t e r 67

Results An inve rse modeling approach was used to determine the “activities” of connected fracture net-


ESD

An Active Fracture Model for Unsaturated Flow and Transport

Figure 2. Comparison between simulated matrix water saturation profile corresponding to Borehole SD-7 and measured data.

works, characterized by the Îł factor, for the Topopah Spring welded (TSw) unit of Yucca Mountain by matching currently available matrix water saturation and potential data from several deep boreholes (e.g., Figure 2).The ITOUGH2 code was used for the inversion. The estimated Îł values are about 0.43, indicating that about 18-27% of connected fractures are active in the TSw unit under ambient conditions.This relatively high percentage is consistent with various field observations, including the relatively uniform matrix saturation and in-situ water potential values measured for most of the units. Furthermore, calcite coatings, signatures of water flow history in fractures, were found in about 10% of the fractures within the welded units. Tracer transport simulations were also performed to check the consistency of the model with geochemical data. The calculated 50% tracer concentration breakthrough times of fracture continua at the bottom of the TSw unit are between 3,500 and 5,000 years, and are generally in agreement with the perched water age data.

Significance

of

anism in an unsaturated fracture-matrix system. This makes it necessary to formulate the fracture water flow differently from water flow in unsaturated porous media.This study introduced a new model to incorporate fingering flow at a fracture network scale into the dual continuum approach. The simulation results based on the new model are generally in agreement with the field observations at the Yucca Mountain site.

Related

Publications

Liu, H.H., C. Doughty and G.S. Bodvarsson, An active fracture model for unsaturated flow and transport in fractured rocks, Water Resour. Res., 34(10), pp. 2633-2646, 1998.

Funding This wo rk has been supported by the Director, Office of Civilian Radioactive Waste Manage m e n t , t h rough Memora n d u m P u rchase Order EA9013MC5X between TRW Env i ronmental S a fety Systems, Inc., and Ernest Orlando Law rence Berke l ey National Laboratory for the Yucca Mountain Site Characterization Project under U.S. Department of Energy C o n t ract No. DE-AC03-76SF00098.

Findings

Fingering flow in a fracture network is a common flow mech-

http://www-esd.lbl.gov 68


Earth Sciences Division

Research

Objectives

Accurate and efficient simulation of chemical transport processes in the unsaturated zone of Yucca Mountain, Nevada, is important to evaluate the perfo rmance of the potential repository.The scale of the unsaturated zone model domain for Yucca Mountain (50 km2 area with a 600-m depth to the water table) requires a large gridblock approach to efficiently analyze complex flow and transport processes. The conventional schemes based on finite element or finite difference methods perform well for dispersion-dominated transport, but are subject to considerable numerical dilution/dispersion for advection-dominated transport, especially when a large gridblock size is used. Numerical dispersion is an artificial, grid-dependent chemical spreading, especially for otherwise steep concentration fronts. One effective scheme to deal with numerical dispersion is the random walk part i cle method (RWPM). While significant progress has been made in developing RWPM algorithms and codes for single-continuum systems, a random walk particle tracker, which can handle chemical transport in dual-continua (fractured porous media) associated with irregular grid systems, is still absent (to our know l e d ge) in the publ i c domain. This is largely due to the lacking of rigorous schemes to deal with particle transfer between the continua, and efficient schemes to track particles in irregular grid systems. The main objectives of this study are to (1) develop approaches to extend RWPM from a single-continuum to a dual-continuum system; (2) develop an efficient algorithm for tracking particles in 3-D irregular grids; and (3) integrate these appro a ches into an efficient and user-friendly software, DCPT, for simulating chemical transport in fractured porous media.

Nuclear Waste Program DCPT: A Dual-Continuum Random Walk Particle Tracker for Transport Lehua Pan, Hui-Hai Liu, Mark Cushey and Gudmundur S. Bodvarsson

(510)

Contact: Lehua Pan 486-2360, lpan@lbl.gov

Figure 1. Comparisons between DCPT and analytical solutions: (a) 1-D dual continuum with linear sorption; (b) 2-D convection and dispersion transport, and (c) particle trail in a concentric velocity field with an irregular grid.

69

Annual Report 1998-1999

Approach In RWPM, chemical tra n s p o rt is re p resented by the movement of a large number of part i cles. In addition to advection determined by the corresponding velocity field, dispersion/diffusion is simulated by random walks of part i cles. In DCPT, RWPM is dire c t ly adopted to simulate part i cle movement in each cont i nu u m . U n l i ke a single continu u m , h oweve r, a dual-continuum system is associated with two very different velocity fields at a gi ven â&#x20AC;&#x153;phy s ical point.â&#x20AC;? The mass tra n s fer b e t ween the fra c t u re and matrix continua is one of the critical p rocesses that control the movement of chemicals. The mass tra n sfer process between the fra c t u re and the matrix is simulated by the part i cle tra n s fer probabilities (PTP) w h i ch determine if a part i cle will l e ave the current continuum at the next time level. Determination of PTP is one of the key and unique elements in DCPT. The challenge is to convert the net mass flow between the fracture and the matrix (in the Eulerian point of view) into the particle transfer probabilities (the Lagrangian point of view). An analytical solution of the mass conservation equation of the particles in the Lagrangian point of view is found, based on which a new scheme to calculate the PTP is developed and incorporated into DCPT (Pan et al., 1999). In most of the currently available RWPM codes, regular grids are gene ra l ly used. For many subsurfa c e contaminant pro blems, such as those in the unsaturated zone of Yucca Mountain, the related subsurface media can be highly heterogeneous. To capture these heterogeneities, irregular grid systems are re q u i re d . An efficient scheme to track particle locations in irregular grids was developed for DCPT (Pan et al., 1999). The main idea of this scheme is to establish a secondary structure of the original 3-D irregu-


ESD

DCPT: A Dual-Continum Random Walk Particle Tracker for Transport

lar grid so that only a very small site-scale unsaturated zone (UZ) portion of the whole grid has to be model for Yucca Mountain. The searched each time. In this way, comparison is fa i r ly go o d , and decaying of the computational effiDCPT gives a steeper concentraciency due to increases in the nu mtion front, resulting from its ability ber of grid blocks can be avoided. to deal with numerical dispersion. DCPT can be used to simulate All the simulations of DCPT reactive transport with linear sorpmentioned above were performed tion. For a single-continuum system, on a Pentium II 300 MHz PC and the retardation factor method has took a few seconds up to about 10 been widely used in RWPM. minutes of CPU time, depending However, this approach can not be the particle numbers used. directly used for a dual-continuum Significance system, because of the coupling of of Findings transport processes in the two continua with different retardation facBased on the newly developed tors. In the light of the imaginary methodologies to extend RWPM porosity method proposed by Liu from a single continuum system to and Bodvarsson (1999), a physically a dual continuum system, a new based conditional pro b ability software program, DCPT, has been approach (Pan et al., 1999) was developed. Comparisons between incorporated into DCPT to describe DCPT simulation results with anathe effects of the linear sorption process on either the transfer prob- Figure 2. Comparisons between DCPT and T2R3D for simulating lytical solutions and re s u l t s radionuclide transport in the unsaturated zone of Yucca Mountain: ability between two continua or the (a) technetium (without sorption); and (b) neptunium (with sorption). obtained from T2R3D show that DCPT could be used to simulate a d ve c t i o n / d i s p e rsion processes in chemical transport processes associated with linear sorption in each continuum. fractured continua without numerical dispersion. DCPT was developed using FORTRAN 90 and following the principle of the objective-oriented programming technologies.

Related

Publications

Results Liu, H.H., and G.S. Bodvarsson, A scheme to determine particle transfer in random walk algorithms for dual continua, Water Resour. Res., in review, Berkeley Lab report LBNL-42886, 1999. Pan, L., H.H. Liu,M. Cushey and G.S.Bodvarsson,DCPT:A new random walk particle tracker for dual-continua, Berkeley Lab report LBNL-42958, 1999.

Figure 1 shows comparisons between simulation results of DCPT and several typical analytical solutions in 1-D and 2-D domains.The comparisons were designed to verify DCPTâ&#x20AC;&#x2122;s capability to (1) calculate particle transfer between the two continua and consider the sorption (Figure 1a); (2) incorporate dispersion and diffusion in multi-dimensional domains (Figure 1b); and (3) t ra ck part i cle movement in an irregular grid (Fi g u re 1c). Particularly, Figure 1c depicts a trail of a particle in a concentric velocity field (radial velocity is zero everywhere) simulated by DCPT, which is a circle theoretically. The satisfactory matches indicate that the methodologies and algorithms developed are valid and successful. Figure 2 shows the comparison between simulation results obtained from DCPT (2000 particles) and T2R3D (Wu et al., this report) for two 1-D problems under unsaturated conditions.The corresponding 1-D column is extracted from the 3-D grid of the

Funding This work has been supported by the Director, Office of Civilian Radioactive Waste Management through Memorandum P u rchase Order EA9013MC5X between TRW Env i ro n m e n t a l S a fety Systems and the Ernest Orlando Law rence Berkeley National Laboratory for the Yucca Mountain Site Characterization Project under U.S. Department of Energy Contract No. DE-AC0376SF00098.

http://www-esd.lbl.gov 70


Earth Sciences Division

Research

Objectives

Nuclear Waste Program Continual Development Of the UZ Model for Yucca Mountain, Nevada

Annual Report 1998-1999

The general approach of the UZ model development consists of the The site characterization and modfollowing major steps: Yu-Shu Wu, Jennifer Hinds, eling studies continued in 1998 for (1) Development of a conceptual investigating the feasibility of using C. Frederick Ahlers, Hui-Hai Liu, hydrogeological model for moisLehua Pan, Anne C. Ritcey, the unsaturated zone (UZ) at Yucca t u re , gas and heat fl ow, and Mark Cushey, Charles Haukwa, Mountain, Neva d a , as a permanent t ra c e r / chemical tra n s p o rt Eric Sonnenthal and storage facility for the geologic disprocesses in the UZ system; Gudmundur S. Bodvarsson posal of high-level nuclear waste. As (2) R e p resentation of the hy d roContact: part of these efforts, the 3-D site-scale ge o l o gical system using a 3-D Yu-Shu Wu UZ flow and transport model develdiscrete computer model; (510) 486-7291, yswu@lbl.gov oped at LBNL has been improved sig(3) I n t e gration and incorpora t i o n nificantly during the fiscal year. The primary goals of this model of all available field and laboratory observation data into the are to simulate and investigate the existing natural state of the model; mountain’s hydrogeological system, and to predict possible (4) Model calibration and sensitivity studies, including both future system responses in support of the Yucca Mountain inverse and forward modeling efforts to estimate model Project. In 1998, modeling studies and development efforts input parameters and conditions; focused on supporting the total system performance assessment(5) Model applications, including predictive studies. viability assessment (TSPA-VA) and license application (LA) activResults ities. The UZ model and simulation results of flow fields were used directly in the current TSPA-VA efforts by the Yucca A comprehensive summary of the development and results of Mountain Project. the 1998 UZ model was reported by Bodvasson et al., 1998.The Approach major achievements and emphases of the year have been to incorporate all new site data into the model, update the UZ flow In developing the UZ model, the and transport conceptual models, UZ hydrogeological system of the and focus on specific features and site is re p resented using a 3-D processes that may be important numerical model grid; model condito repository performance.A sumtions are described based on the mary of these achievements folmost updated ge o l o gical fra m elows: work model and observed data.The • New data have been incorporated from the new boreholes, model domain encompasses a the ESF and East-West Cross 5,000-m by 9,000-m area, as shown Drift. in Figure 1, with a 500-700 m depth • Matrix and fracture properties to the water table. Model paramewere updated with a new set t e rs are estimated using a comof calibrated hy d rologica bined, inverse and direct modeling propeRties. a p p ro a ch , dual-permeability con• Submodels of the Ghost Dance cept for fracture-matrix interactions Fault and the Solitario Canyon and several infiltration rates. During Fault were developed to evalui nve rse modeling, the state va riables are calibrated using observed ate flow within and near faults. matrix liquid satura t i o n , wa t e r • The 3-D UZ model was potential, pneumatic and temperaupdated with a new hydrolot u re data from bore h o l e s . The gical pro p e rty set, i m p roved p a rameter sets estimated fro m ge o l o gical and minera l o gi these simultaneous 1-D inversions models, and a refined model are further modified, when necesgrid, as shown in Figure 1.The sary, to account for spatial heteroUZ model utilizes a dual-pe geneities, such as perched-water meability concept with a dua occurrences, geochemical data and p o rosity re p resentation fo Figure 1. A plan view of the 1998 site-scale model domain, grid and groundwater travel times through incorporated major faults, and locations of boreholes used for the faults. Simulations were commodel calibration. the 3-D modeling studies. pleted that predict ambient 71


ESD

Continual Development of the UZ Model for Yucca Mountain, Nevada

Figure 2. Three-dimensional perched water bodies near the base of the TSw.

conditions in new wells SD-6 and WT-24 and the East-West Cross Drift.The UZ Model successfully predicted the location of perched water in borehole WT-24. • A 2-D cross-sectional, drift-by-drift model was used to evaluate the impact of coarse-grid averaging. • Conceptual models for perched water and flow below the repository were evaluated with a detailed model of the Calico Hills nonwelded unit. Large effects on lateral flow are found from the permeabilities of zeolites, the degree of fracturing within zeolites, and the extent of perched-water bodies. One of the model-predicted perched water bodies is shown in Figure 2. • A detailed model of the Paintbrush bedded tuff has been developed and showed that lateral f low may be limited to less than 100 m.

Significance

of

the unsaturated zone and of the re l evant processes occurring at Yucca Mountain. In addition, the model provides input to va ri o u s other models, such as the ambient and thermal drift-scale model, the mountain-scale thermohy d ro l o gical model and the UZ tra n sport model used for LA.

Related

Publication

Bodvarsson, G.S., E. Sonnenthal and Y.S.Wu, eds., Unsaturated zone flow and transport modeling of Yucca Mountain, N evada-Fiscal year 1998 report,Yucca Mountain Site Characterization Project Milestone, Revision 00D, LBNL, 1998.

Funding This work has been supported by the Director,Office of Civilian R a d i o a c t i veWaste Management, through Memorandum Purchase Order EA9013MC5X between TRW Env i ronmental Safety Systems, Inc., and Ernest Orlando Lawrence Berkeley National Laboratory t h rough U.S. D e p a rtment of Energy Contract No. DE-ACO376SF00098.

Findings

The UZ model provides estimates of important parameters and p rocesses for the TSPA-VA and LA. Updated modeling effo rts re nder more realistic representations of the 3-D nature of fl ow within

http://www-esd.lbl.gov 72


Earth Sciences Division 1998-1999

Nuclear Waste Program Development of WinGrider: An Interactive Grid Generator For TOUGH2

Annual Report

zation functionality. A d vanced database techniques were incorporated to handle a large number of grid blocks Lehua Pan, Charles Haukwa, Designing a proper numerical grid and connections efficiently. Yu-Shu Wu and is one of the major efforts in modeling Development of the new grid sysGudmundur S. Bodvarsson flow and transport processes in comtem starts with assignment of nodes Contact: plex, h e t e ro geneous systems using in map view for each object [e.g., the Lehua Pan the TOUGH family of codes.The probdomain (base nodes), repository or (510) 486-2360, lpan@lbl.gov lem becomes even more diffi c u l t faults] with specified orientation and when using an irregular grid, such as in the case of modeling density. Based on the information of these nodes, a primary 2-D studies of unsaturated zone (UZ) flow and transport for Yucca grid is generated with Voronoi tessellation techniques. The 2-D Mountain, Nevada.The development of an efficient meshmaker grid is then impro ved systematically and/or interactively by was motivated by the requirements of the TOUGH codes for simdeleting physically incorrect or unnecessary connections. A few ulating the subsurface processes of high-level nuclear waste isoiterations of these steps, including adding and deleting some lation in partially saturated geological media of the Yucca nodes, are taken to create a final 2-D grid (a column scheme) that Mountain site. A mountain-scale model often involves various serves as the basis for generating the third dimension of the grid. geological and engineering objects in the complex hydrogeologAll 3-D cells and vertical connections between adjacent cells are ical settings. Designing and generating a suitable irregular grid generated column by column to ensure that each vertical confor such a system has been a tedious and error-prone process, nection connects only adjacent cells and that each cell has at especially when the number of gridblocks and connections least one vertical connection. Lateral connections are then genbecomes very large. Inspecting the grid or extracting some suberated segment by segment with each segment joining two grids or other specific information or use from the grid has also neighboring columns. This ensures that only cells in two adjabeen a time-consuming task. Therefore, user-friendly, efficient cent columns have lateral connections and that no connections grid-generating software is a critical part of successful applicabetween two adjacent columns are missing. tion of the TOUGH codes to solve real world problems. Our Three-dimensional cells are created for each hydrogeological objective is to develop user-friendly and integrated software for unit in each column and then are laterally connected within the generating grids that best represent the geological and engisame layer except where the layer is interrupted by faults or neering features within 3-D model domains for given computing repository drifts. An approach using the three parallel, faultresources. related columns is used to ensure that the grid represents faults correctly. In particular, this scheme preserves three important Approach roles of faults: (1) separator between geological layers that may serve as a structural barrier to lateral flow across it; (2) continuWinGrider was primarily written in the Visual Basic language ous zone that may serve as a fast path for flow along the fault following the principles of object-oriented programming to prodepending on its hydraulic properties; and (3) dipping feature vide graphical user interfaces, interactive operation and visualiwith angles of inclination varying spatially.

Research

Objectives

Figure 1. A snapshot of the WinGrider grid generator.

73


ESD

Development of WinGrider: An Interactive Grid Generator for TOUGH2

Results

such as inclined faults with offset, layering structure, local refinements and repository drifts can be represented in the grid. This grid-maker has been used as a primary tool for designing model grids in modeling efforts for the Yucca Mountain project.

WinGrider is Windows-based software for designing, generating and visualizing numerical grids used in reservoir simulation and groundwater modeling using the TO U G H 2 code (Pan et al., 1999). Fi g u re 1 shows an example of the graphical user interfaces provided by WinGrider.WinGrider has been used for various numerical simulations in the Yucca Mountain project.The generated grids were verified independently in terms of accuracy in representing the geological and engineering systems as well as correctness of connections. Figures 2a and 2b show a map view of the UZ98 grid and a 2D cross-section grid for grid spacing analysis, respectively.

Significance Findings

Related

Publications

Pan, L., J. Hinds, C. Haukwa, Y.S. Wu and G.S. Bodvarsson, WinGrider: A user manual (version 1.0), Berkeley Lab report LBNL-42957, 1999. Pruess, K., TOUGH2: A general purpose numerical simulator for multiphase fluid and heat flow, Berkeley Lab report LBL-29400, 1991.

Funding

of

This work has been supported by the Dire c t o r, Office of Civilian Figure 2. Examples of the grid: (a) map view of UZ98 grid; (b) R a d i o a c t i ve Waste Management, E-W cross-section grid for grid side analysis. We have developed a user-friendly, t h rough Memorandum Purchase efficient grid-generating software that Order EA9013MC5X between TRW should provide an important tool for successful applications of Environmental Safety Systems, Inc., and Ernest Orlando Lawrence the TOUGH family of codes to solve real-world problems. B e rkeley National Lab o ra t o ry for the Yucca Mountain Site WinGrider can be used to generate complex, irregular 2-D or 3-D Characterization Project under U. S . Department of Energy grids with high quality and efficiency. Many important features, Contract No. DE-AC03-76SF00098.

http://www-esd.lbl.gov 74


Earth Sciences Division

Research

Objectives

Nuclear Waste Program T2R3D – A TOUGH2 Code For Tracer Transport In Heterogeneous Media

Annual Report 1998-1999

Approach

Even with the continual progress T2R3D is built on the framework of Yu-Shu Wu made in both computational algothe TOUGH2 code (Pruess, 1991).The and Karsten Pruess rithms and computer hardwa re , it basic mass and thermal energy balremains a challenge to simulate transance equations for three components Contact: Yu-Shu Wu port of a tracer or radionuclide in hetof water, air and a radionuclide/tracer, (510) 486-7291, yswu@lbl.gov erogeneous fractured porous media and heat solved by T2R3D are similar with a numerical method. It becomes even more difficult when in form to those for the standard TOUGH2 EOS3 module. Also, dealing with tracer transport in a multiphase and non-isothermal the integral finite difference method and a first-order, backward flow system using a general 3-D, irregular grid. One of the prifinite difference scheme are used for spatial and temporal dismary problems in solving advection-diffusion type transport c re t i z a t i o n , re s p e c t i ve ly. The tracer tra n s p o rt mech a n i s m s include molecular diffusion and hydrodynamic dispersion in the equations in a complex geological medium is determining how liquid or gaseous phase, in addition to advection terms. Firstto approximate the diffusion/dispersion tensor in order to estiorder decay is taken into account and adsorption of a tracer on mate the dispersive terms of mass transport accurately. Most rock matrix and/or fractures is described by an equilibrium numerical modeling approaches in the literature use schemes isotherm with a constant distribution coefficient. that are based on regular grids and the partial dispersion tensor. The model formulation considers adve c t i o n / d i s p e rsion tra n sVery few studies have addressed transport using an irregular 3-D p o rt processes of a liquid or gas tracer and incorporates a full disgrid for large, complex geological systems. p e rsion tensor, based on a 3-D velocity field on a 3-D, regular or T2R3D, a TOUGH2 module, has been developed to model i rregular, integral finite-difference grid in a heterogeneous ge oradionuclide or tracer transport in heterogeneous, fractured logical system. In addition to advection terms for the tracer tra n sporous media (Wu et al., 1996).The model formulation incorpo→ p o rt , the dispers i ve and diff u s i ve mass flux, FD, is described by rates a full dispersion tensor with a 3-D, irregular, integral finite difference. It takes account of the physical processes of tracer or (β=liquid or gas) (1) F D = - r b D ∑—X b radionuclide transport in a non-isothermal, multi-phase, multidimensional flow environment in the subsurface.

Figure 1. Model grid and simulation results of T2R3D for radionuclide transport through the unsaturated zone of Yucca Mountain, Nevada.

75


ESD

T2R3D – A TOUGH2 Code for Tracer Transport in Heterogeneous Media

Results

where D is the combined diffusion-dispersion tensor accounting for both molecular diffusion and hydrodynamic dispersion; ρβ is fluid density; Xβ is mass fraction of the tracer in phase β.We have incorporated a general dispersion model for 3-D tracer transport into the T2R3D code. D =aT

v b dij + (a L - a T )

v b vb +φSfβSτd btd mijdij mδ vb

As a new member of the TOUGH2 family of codes,T2R3D (Wu et al., 1996) provides a capability for modeling liquid or gas tracer or radionuclide transport in multiphase and non-isothermal flow systems. In particular, T2R3D can be used to simulate tracer transport in complex, heterogeneous fractured rock using a general, irregular 3-D grid. In addition to incorporation of a full dispersion tensor in evaluating dispersive tracer transport, the code takes into account linear adsorption and first-order decay effects. The model formulation and numerical scheme make it easy to include any other transport mechanisms, such as nonadsorption, multi-decay chains, or thermal/mechanical effects. Figure 1 shows an example application, in which T2R3D is used to assess radionuclide transport in the unsaturated zone of Yucca Mountain along a 2-D cross section with an irregular finite-difference grid.

(2)

(β=liquid or gas) where αT and αL are the transverse and longitudinal dispersivities, respectively; vβ is the Darcy velocity vector of phase β through fractures or matrix; τ is the tortuosity of the medium; dm is the molecular diffusion coefficient in phase β; and δ ij is the Kronecker delta function. (δij=1 for i=j, and δij=0 for ij). One of the key issues in implementing the general 3-D dispersion tensor of (2) is how to interpolate velocity fields for determining the dispersion tensor. The averaging or weighting scheme used to evaluate a velocity vector at the interfaces between element blocks is called “projected area weighting method” (Wu and Pruess, 1998). In this method, a velocity component, vn,i, of the velocity vector element n is determined by the vectorial summation of the flow components of all local connection vectors in the same direction, weighted by the projected area in that direction: Â ( A n m n i )( v n mn i )

v n, i = m

(i=x,y,z)

Significance

 ( A nm ni ) m

Related

where m is the total number of connections between element n and all its neighboring elements m, vnm is the flux along connection nm in the local coordinate system, and ni are the directional cosines of connections.The velocity vector v at the interface of element n and m is then evaluated by harmonic weighting to preserve total transit time for solute transport travelling between the two blocks. The mass fraction gradient of the tracer/radionuclide is evaluated at the interface between gridblocks n and m as

(k )

Funding

(4)

(k )

This work has been supported by the Assistant Secretary for Energy Efficiency and Renewable Energy, Geothermal Division, and by the Dire c t o r, Office of Civilian Radioactive Waste Manage m e n t , t h rough Memorandum Purchase Ord e r EA9013MC5X between TRW Environmental Safety Systems, Inc., and the Ernest Orlando Lawrence Berkeley National Laboratory under U.S. Department of Energy Contract No. DE-AC03-SF7600098.

(k )

Xm - Xn

(5)

D m + Dn

Publications

Wu,Y.S., C.F.Ahlers, P. Fraser,A. Simmons and K. Pruess, Software qualification of selected TOUGH2 modules, Berkeley Lab report LBL-39490, UC-800, 1996. Wu,Y. S., and K. Pruess,A 3-D hydrodynamic dispersion model for modeling tracer transport in geothermal re s e rvoirs, Proceedings of the 23rd Workshop, Geothermal Reservoir Engineering, Stanford University, CA, pp. 139-146, 1998.

with DX n m =

Findings

The T2R3D code has found a wide range of applications in field characterization studies of the unsaturated zone transport of environmental isotopic tracers and radionuclides at the Yucca Mountain site, a potential underground repository for high-level radionuclide wastes. The special capability of modeling tracer transport processes through heterogeneous fra c t u red ro ck s under multi-phase and non-isothermal conditions with full consideration of hydrodynamic dispersion will make T2R3D a very useful tool in modeling studies of tracer transport in oil, gas and geothermal reservoirs.

(3)

(k ) Ê (k ) (k ) (k ) ˆ —Xn m = n x DX n m,n y DX n m, nz DX n m Ë ¯

of

The net mass flux of diffusion and dispersion of a tracer/radionuclide along the connection of element n and m is determined by equation (1).

http://www-esd.lbl.gov 76


Nuclear Waste Program

Earth Sciences Division

Research

Objectives

Laboratory Studies on Heat-Driven Multiphase Flows In Rock Fractures

Annual Report 1998-1999

and (3) temperature in a heated fracture. Direct visualization and video Water and vapor flow through nonrecording were used to observe flow Timothy J. Kneafsey, isothermal, fra c t u re d , p o rous ro ck phenomena in the fracture; an infraKarsten Pruess and environments are important in a numred camera and thermocouples were Jeffery J. Roberts* ber of fields, including the geologic used to monitor temperature, and xdisposal of high-level nuclear waste ray attenuation was used to monitor *Lawrence Livermore National Laboratory imbibition.The fracture was made by and geothermal re s e rvoir deve l o pContact: placing a saw-cut slab-shaped block of ment. Here we present the results of a Tim Kneafsey (510) 486-4414, tjkneafsey@lbl.gov Topopah Spring tuff (15 x 23 x 2.6 number of visualization experiments cm) with a porosity of 10% next to a that were performed to gain a better glass plate. The rock and the glass u n d e rstanding of liquid and va p o r plate were placed in a frame and held apart by gold shims, proflow in non-isothermal fractures, and imbibition into the fracture viding a 12.5-micron nominal aperture. Heaters were installed wall. both in the rock slab and on the assembly such that the fracture Approach was heated on the bottom, with a thermal gradient from about 140째C near the rock bottom to about 80째C at the top.Water conSeveral types of experiments have been performed to investitaining potassium iodide to enhance x-ray attenuation contrast gate the behavior of liquids in boiling fractures, including introwas introduced into the aperture. Solid potassium iodide was ducing liquid into natural and saw-cut fractures in Topopah precipitated upon boiling, identifying regions of active boiling. Spring tuff from Yucca Mountain, Nevada. In each experiment, The experiments extended over 24 and 72 hours with flow rates heat was applied to create a temperature gradient and boiling of 2.5 ml/min initially that were reduced to 0.55 ml/min in the zone.Two types of experiments are described here. first experiment and 0.3 ml/min in the second. In the first experiment, we introduced water at three different Results flow rates into the top of a 22-cm natural fracture oriented vertically. Heat was applied at the bottom, creating a boiling zone In the first experiment, dyes stained the rock as the water near the center of the rock with temperatures of about 120oC at the rock bottom and 80째C at the top.Water was introduced at a boiled off. At the lowest flow rate with red dye, water did not point in the center of the fracture top at 7, 14 and 28 mL/hr, and penetrate significantly below the boiling isotherm (Figure 1). a different non-volatile dye was added for each flow rate. The Water evaporation occurred in two bands: near the horizontal dyed water flowed through the heated fracture, and where it boiling isotherm and near the injection point. Yellow dye was boiled, the non-volatile dye was left on the rock. introduced with water at the medium injection rate. Some evapTwo experiments of the second type of were performed in an oration occurred in the same locations where the water introassembled rock/glass fracture to simultaneously observe (1) liqduced at the lowest flow rate evaporated, however f low also uid flow through a fracture, (2) imbibition into the porous rock, extended downward beyond the centrally located nominal boil-

Figure 1. Fracture faces following infusion of water with dyes. The boiling isotherm for each case is shown.

77


ESD

Laboratory Studies on Heat-Driven Multiphase Flows in Rock Fractures

ing isotherm.A large quantity (darker spots in Figure 2c) of yellow dye was deposited occurs in other locations, in a narrow horizontal band indicating salt build-up in the well below the rock center. pore structure of the rock. Additionally, a finger several The bulk sorptivity of the centimeters wide re a ched sample was obtained by plotthe fracture bottom where ting water stored in the samthe temperature was mainple versus the square root of tained at 120째C. Blue dye time. For the entire sample stains in the fracture from over the range of temperawater introduced at the hightures the measured sorptivity est flow rate was visible in compares well with other valmost cases beyond the Figure 2. (a) Visual, (b) temperature and (c) x-ray attenuation images of water ues measured under conextent of the yellow dye, and flow through a non-isothermal fracture. The single arrow in (a) indicates con- trolled conditions. densing water; the double arrow indicates solid potassium iodide also seen in an apparently wider finger, or (c). In (b), the heated zone is at the bottom with red representing 140째C, cyan several fingers, extended to 100째C, and purple <80째C. Darker regions in (c) show greater x-ray attenuation, S i g n i f i c a n c e indicating higher imbibition or precipitation. of Findings the bottom of the fracture. In the second type of experiments investigating fracture flow We have identified and observed several important phenomand imbibition, the time scale of fracture flow was on the order ena which occur in boiling fractures.Water flow into boiling-hot of minutes and the time scale of imbibition was hours. Upon regions was hindered by heat and the accumulation of mineral introduction into the fracture, water flowed in fingers and films precipitates. Higher flow rates allowed deeper penetration into downward toward the heated region, with broader fingers occurthe boiling region, and allowed washout of precipitated minerals. ring in the first experiment with the higher flow rate. As water Imbibition was not uniform across the sample; more permeable penetrated the above-boiling region, boiling occurred and vapor regions and hotter regions showed greater x-ray attenuation, indibegan to condense in cooler regions (Figure 2a). Intermittent cating increased imbibition into the rock and salt deposition. unstable boiling events called rapid evaporation events (REEs) REEs occurred frequently in the boiling region, affecting liquid frequently occurred in both experiments when liquid water and vapor flow in their vicinity. Flow paths were variable superheated and boiled rapidly. These REEs caused pressure through the heated region, particularly at the higher flow rates, pulses that affected flow and often triggered REEs in other parts with flow rate influencing flow paths. of the sample. Solid potassium iodide accumulated in the boiling region at temperatures exceeding the boiling point and hindered Related Publications liquid flow, but the salt buildup was intermittently washed out by flow through these regions. In the first experiment, partial clogKneafsey, T. J., and K. Pruess, Laboratory experiments on heatging of the aperture by salt caused a buildup of heat, which was driven two-phase flows in natural and artificial rock fracreduced during washouts. tures,Water Resources Research, 34, pp. 3349-3367, 1998a. Temperature monitoring showed cooling upon introduction Kneafsey, T. J., and K. Pruess, Preferential flow paths and heat of water. Water pathways were visible as cooler areas extending pipes: Continued laboratory experiments on heat-driven through the boiling region in the thermal images. At the higher flow rate, water flowed through the heated region without flow in natural and artificial rock fractures and scaling relabecoming permanently confined to particular pathways. Water tionships, Milestone SP3CK1M4, Berkeley Lab report LBNLflowed first through one side, cooling it, then the middle, then 42262, 1998b. the other side, intermittently oscillating from side to side on a Funding time scale of hours.At the lower flow rate (Figure 2b), water generally flowed around the hottest regions, but intermittent boiling This work has been supported by the Director, Office of events of water accumulating in the apparatus bottom caused Civilian Radioactive Waste Management, through Memorandum water to flow through these regions. P u rchase Order EA9013MC5X between TRW Env i ro n m e n t a l Water imbibition into the rock increased over time, but was Safety Systems, Inc., and Ernest Orlando Lawrence Berkeley spatially non-uniform. X-ray attenuation in the boiling region was National Laboratory for the Yucca Mountain Site Characterization much higher than elsewhere, indicating both salt-crystal build-up Project under U.S. Department of Energy Contract No. and imbibition (Figure 2c). In some locations, video D E - AC03-76SF00098 and by Lawrence Livermore images and radiographs show salt crystal buildup in the National Laboratory under contract W-7405-ENG-48. same locations, but even greater x-ray attenu a t i o n http://www-esd.lbl.gov 78


Earth Sciences Division

Research

Objectives

Nuclear Waste Program Thermal-Hydrological Modeling of the Drift Scale Test at Yucca Mountain, Nevada

Annual Report 1998-1999

along the heated drift wall to reach as high as 200째C.The heat output of the The Drift Scale Test (DST) is the canister and wing heaters can be second of the two in-situ heater tests adjusted over the full range to achieve being carried out in the underground these targets. Sumit Mukhopadhyay and Yvonne W. Tsang Exploratory Studies Facilities at Yucca Therm a l - hy d ro l o gical simu l a t i o n s Mountain, Nevada (the Single Heater of the DST are being performed with Contact: test, SHT, was the first). The primary the TOUGH2 simulator, an integrated Sumit Mukhopadhyay (510) 495-2440, smukhopadhyay@lbl.gov objective of the DST,as it was with the finite difference simulation program SHT, is to develop a better underfor nonisothermal flow of multicomstanding of the effect and influence of thermal-loading on the ponent, multiphase fluid in porous and fractured media. We use coupled thermal, mechanical, hy d ro l o gical and ch e m i c a l the EOS4 module, which supports the thermodynamics of non(TMHC) responses of the surrounding rock mass. isothermal two-phase flow of components air and water, with vapor-pressure-lowering capabilities. Approach The configurations, parameters, and initial and boundary conditions of the numerical model are designed to resemble the The primary difference between the DST, located in the midactual test conditions of the DST as closely as possible.The DST dle nonlithophysal unit of the Topopah Spring welded tuff of model domain encompasses three different stratigraphic layers Yucca Mountain, and the SHT is that of scale.Whereas the heatof Yucca Mountain; i.e., the upper lithophysal, the middle noning element of the SHT was only 5 m long and placed in a 9.6lithophysal, and the lower lithophysal units of the Topopah cm diameter borehole, the heated drift of the DST is 47.5 m long Spring fractured welded tuff.The heated drift itself is in the midand 5 m in diameter. Heating is provided by nine canister dle nonlithophysal unit. The material properties within each heaters,which are placed on the floor of the heated drift. In addilayer are assumed homogeneous. As far as feasible, the input tion, there are 25 wing heaters, placed perpendicular to the lonparameters of thermal and hydrological properties for the DST gitudinal axis of the heated drift, on each side of it. The wing numerical model have been derived from laboratory and field heaters are used to emulate the effects of adjacent heat-generatpre-test characterization data of the DST block. When site-specific measurements are not available,properties are derived from ing waste-storage drifts. Each wing heater is about 9.5 m long mountain-scale calibration to measured data from numerous sur(consisting of an inner and outer section, each about 4.5 m face-based boreholes. long), and the spacing between each wing heater is about 1.83 The results presented below are based on the hydrology propm. The canister heaters and wing heaters together have a heat erty sets, which are calibrated to an infiltration rate of 0.36 output of approximately 185 KW.The TMHC responses from the mm/yr. While different conDST are measured by ceptual models have been a p p rox i m a t e ly 3,500 senutilized for simulating the sors, which are placed in 147 TMHC responses from the bo reho l e s . Numerous va riDST, here we will present abl e s , i n cluding heater only the thre e - d i m e n s i o n a l power, temperature, thermal dual-permeability model, ex p a n s i o n , moisture and which assumes two separate m e chanical displacements are measured by these senc o n t i nua for the fra c t u re s sors, which are connected to and the matrix.The details of an automatic data collection the model can be found in system. Birkholzer and Tsang (1997, The heating phase in the 1998). DST started on Dec. 3, 1997. Results The planned duration for the heating and cooling phases The key pro c e s s e s is four years each. The heati nvo l ved in the therm a l ing phase has been planned hy d ro l o gical response of the to elevate temperatures of a unsaturated fractured tuff to substantial ro ck vo l u m e heat are as follows. As fo rm a( m o re than 10,000 cubic meters) above 100째C, while Figure 1. Fracture liquid saturation in a vertical plane (about 10 m from the west tion tempera t u re approaches 100oC around the heater, a l l owing the temperatures side of the heated drift) after 12 months of heating. 79


ESD

Thermal-Hydrological Modeling of the Drift Scale Test at Yucca Mountain, Nevada

matrix pore water boils and a distance of 1.67 m from the vaporizes. Most of the vapor ge nheated drift wall, explaining erated moves into the fractures, the first dip in temperature. where it becomes highly mobile Since the wing heater is about and is dri ven by the gas pre s s u re 4.5 m long, one observes the gradient away from the heat upward trend in temperature source. Upon further heating, the after the first dip. The inner zone near the heating plane and outer parts of the wing develops a “dry-out” a re a . heaters are separated by a gap H owever, away from the heating of 0.67 m, resulting in the secplane, when the vapor encounond dip. Notice also the small ters cooler rock, it condenses, t e m p e ra t u re plateau at the and the local fracture liquid satunominal boiling point, at ration builds up. about 12 m from the heated Part of the condensate then drift wa l l , indicating that a m ay imbibe into the matrix, small two-phase zone exists at where it is subject to a very the tip of the wing heaters. strong capillary gra d i e n t Figure 2. Measured and simulated temperatures in borholes 139 and 143 Figure 2 confirms that our simulation results are able to towards the heat source, giving after 12 months of heating. mimic the thermal and hydrorise to a reflux of liquid to the logical processes taking place in the DST quite closely. dry-out areas. Some fraction of the condensate in the fractures may also flow back towards the boiling zone; however, as capilSignificance of Findings lary forces are relatively weak in the fractures, a substantial amount of liquid may drain by gravity.As a result,liquid saturation The agreement between measured data and simulation results builds up below the heated drift. Figure 1 shows the simulated indicates that the thermal-hydrological responses of the DST are liquid saturation in the fractures after one year of heating in a verwell represented by the coupled thermal-hydrological numerical tical plane about 10 m from the west end of the heated drift.The model.While heat conduction accounts for most of the temperasignatures of all the physical processes described above can be ture rise, thermal-hydrological coupling contributes to the heat observed in this figure. transfer by convection, resulting in better agreement between Figure 1 also shows that the drying region is about 4 m thick simulations and measured data. The thermal-hydrological couafter 12 months of heating. Beyond the drying region is the condensation zone, where the liquid saturation increases from prepling also accounts for the moisture redistribution seen by the heat values.The redistribution of the moisture content described active testing data. by the numerical model is consistent with data (not shown here) Related Publications obtained from active testing by neutron logging, electrical resistivity tomography, cross-hole radar tomography and air-permeBirkholzer, J. T., and Y. W. Tsang, Pretest analysis of the thermalability tests. hydrological conditions of the ESF Drift Scale Test, Berkeley Temperature data have been continuously collected in about Lab report LBNL-41044, 1997. 40 boreholes, which allows the display of data either as a snapBirkholzer, J.T., and Tsang, Y.W., Interpretive analysis of the shot or as time evolution at a particular spatial location. Figure 2 thermo-hydro l o gical processes of the Drift Scale Test, shows a snapshot of temperature after 12 months of heating. Chapter 2 of Yucca Mountain Drift Scale Test Progress Temperature data from boreholes 139 and 143 are displayed as a Report, Berkeley Lab report LBNL-42538, 1998. function of radial distance from their collars at the heated drift wall. The simulated matrix temperatures are also plotted along Funding with measured data. Observe that the temperature at the heated drift wall is around 150°C.After a small dip in temperature thereThis work has been supported by the Director, Office of after, it rises to about 165-170°C over a distance of about 4.5 m. Civilian Radioactive Waste Management, through Memorandum Then the temperature dips slightly again, before going back up P u rchase Order EA9013MC5X between TRW Env i ro n m e n t a l over another 4.5 m or so. Beyond this point, the temperature Safety Systems, Inc., and Ernest Orlando Law re n c e decreases as one moves farther away.The interpretation Berkeley National Laboratory for the Yucca Mountain of such a temperature profile is as follows. Boreholes Site Characterization Project under U.S. Department of 139 and 143 are located parallel to the wing heaters on Energy Contract No. DE-AC03-76SF00098. either side of the heated drift.The wing heater starts at http://www-esd.lbl.gov 80


Nuclear Waste Program

Earth Sciences Division

Research

Objectives

Annual Report 1998-1999

Evolution of CO2 From Heated Rock At Yucca Mountain

the drift and wing heaters that extend 10 m into the rock on either side of Yucca Mountain, Nevada, is the site the drift. The entrance to the heater of a proposed high-level nu clear Mark Conrad and Eric Sonnenthal drift is isolated from the observation waste repository. To test the effects drift (OD) by a thermally insulated Contact: that heat generated by the nuclear bulkhead. Gas samples are collected Mark Conrad (510) 486-6141, msconrad@lbl.gov waste may have on the host rock for from a series of 12 hydrological monthe repository, a drift-scale heater test (DST) is being carried out itoring boreholes drilled into the rock around the drift. High-temin the underground Explora t o ry Studies Facility at Yucca perature, inflatable packers placed at approximately 10-m spacMountain.The DST began in December 1997 and is designed to ing divide the boreholes into three or four intervals. One-halfcontinue for a total of eight years. During the initial heating liter gas samples were collected for analysis from the packed-off phase, which will last for the first two years of the project, the intervals after purging approximately three times the volume of rock at the drift wall is being heated to 200°C. Over the next two the interval.The CO2 concentrations in the samples were measured using an infrared analyzer (Li-Cor).The CO2 was then sepayears the temperature of the wall rock will be maintained at rated out of the samples and the stable carbon isotope ratio ana200°C. For the last four years the heaters will be turned off and lyzed using the VG Prism Series II isotope ratio mass spectromethe system will be allowed to cool. As part of the monitoring ter at Berkeley Lab’s Center for Isotope Geochemistry. Where efforts for this project, the concentrations and stable carbon isopossible, aliquots of the CO2 collected for δ 13C analyses were tope ratios (δ13C values) of CO2 in gas samples from the rock around the heated drift are being measured. CO2 and associated saved for 14C analyses. To date, six of those samples have been dissolved inorganic carbon compounds (DIC) in the porewa t e rs analyzed at the Center for Accelerator Mass Spectrometry at in the rock have a strong influence on the chemistry of the Lawrence Livermore National Laboratory. p o rewa t e rs, which in turn controls mineral reactions in the sysResults tem.

Approach

The concentration of CO2 in the rock has increased significantly since heating began.The CO2 concentration of a gas sample collected from a borehole drilled into the unheated side of the access/observation drift (AOD) was 919 ppm. Fi g u re 1 is a plot of the concentrations and δ13C values of CO2 in samples collected f rom boreholes 74 through 78 during December 1998 (slightly over one year after heating began).The CO2 concentrations in several of the hotter intervals are in excess of 20,000 ppm. By contrast, the concentration of CO2 in the AOD was 375 ppm and in

The host rock for the thermal test is the Topopah Springs welded tuff unit. The rock is a highly fractured unit containing approximately 10% matrix porosity, which is about 90% filled with water. The thermal test alcove is located about 250 m below the surface. The heater drift for the DST is 47.5 m long and 5 m across. Heat is provided by a series of canister heaters in the center of

Figure 1. The δ13C values of CO2 in samples collected from boreholes 74-78 at the DST during December 1998. Given in parentheses are the concentrations of CO 2 (in ppm) and the temperatures (in °C) for the sampling intervals.

81


ESD

Evolution of CO2 From Heated Rock at Yucca Mountain

Figure 2. Plot of the δ13C values of CO 2 from borehole 78, interval 3, of the DST versus time. The values in red are the 14C concentrations (in % of modern atmospheric CO 2) in the CO2. Given in parentheses are the temperature (in °C) and the CO2 concentration (in ppm).

where steam generated from boiling of the porewaters is condensing or draining (e.g., interval 3 of borehole 78), hydrolysis can occur, and will produce increased concentrations of CO2. The δ13C values of calcite in the rock are high (between -6‰ and +10‰) and will also produce high-δ13C CO2 when it is dissolved. This makes it difficult to distinguish from CO2 derived from porewater DIC. However, the 14C content of CO2 derived from calcite will be lower than that derived from porewater DIC.The 14C content of CO2 (which will be close to equilibrium with the porewater DIC) at this depth is about 50% of modern, whereas the 14C content of the calcite in the fractures is essentially 0% of modern. T h e re fo re , the trend towa rds lower 14C contents observed for CO2 from borehole 78, interval 3, indicates a shift from degassing of porewater DIC to hydrolysis of calcite. The isotopic compositions of the CO2 in the rock present a powerful tool for quantifying factors such as the degree of degassing of porewater DIC and for identifying areas and rates of calcite dissolution.These factors can have significant impacts on variables such as the pH of the porewaters and mineral equilibria in the rock. In turn, mineral precipitation and dissolution may influence the permeability of the rock and the movement of water within the system, which are important issues for ensuring safe, long-term storage of nuclear waste.

the heater drift was 403 ppm (approximately equal to the CO2 concentration in the outside air used to ventilate the tunnel). The δ13C values of the CO2 are also increasing as the rock is heated. The δ13C value of the CO2 in the gas sample collected from the borehole into unheated rock opposite the heater drift was -13.1‰.The δ13C values of the borehole samples on Figure 1 range from -12.3‰ to -1.5‰.The 14C contents of a limited number of the CO2 samples were also measured.The data for samples from borehole 78, interval 3, are plotted in Figure 2.The general increase in CO2 concentrations and δ 13C values with temperature is clear in this interval.The 14C values of the CO2 decreased from 40.0% of the concentration of modern atmospheric CO2 in February to 10.5% of modern in October.

Significance

of

Findings

The shifts observed in the δ13C values and 14C contents of the CO2 in the heated rock suggest that the increased concentrations of CO2 are derived from two sources. One source is the porewater DIC. As the temperature rises, the partitioning of inorganic carbon between the gas and liquid phases shifts, with a higher proportion going into the gas phase.When the rock reaches the boiling point of water, the CO2 concentrations will increase even more quickly until the rock dries out and all the inorganic carbon is converted to CO 2 or precipitated in carbonate minerals. At 25°C and neutral pH, the δ 13C value of DIC will be 8‰ higher than co-existing CO2.Therefore, as the temperature rises and the DIC is converted to CO2, the δ13C value of the gas-phase CO2 will increase. The other source of CO2 is dissolution of secondary calcite deposited in the fractures in the rock. As the temperature increases, calcite will tend to precipitate rather than dissolve. However, in areas of the system

Funding This work has been supported by the Director, Office of Civilian Radioactive Waste Management, through Memorandum P u rchase Order EA9013MC5X between TRW Env i ro n m e n t a l Safety Systems, Inc., and Ernest Orlando Law re n c e Berkeley National Laboratory for the Yucca Mountain Site Characterization Project under U.S. Department of Energy Contract No. DE-AC03-76SF00098.

http://www-esd.lbl.gov 82


Earth Sciences Division

Research

Objectives

Nuclear Waste Program

Annual Report 1998-1999

Prediction and Analysis Of Coupled Processes In the Drift Scale Thermal Test

the rate of equilibration via diffusion leads to disequilibrium betwe e n The Drift Scale Test (DST) is the wa t e rs in fra c t u res and matrix. second underground thermal test that Because the system is unsaturated and is being carried out in the Exploratory Eric Sonnenthal, Nicolas Spycher, undergoes boiling, the transport of John Apps, Mark Conrad and Studies Facility at Yucca Mountain, gaseous species is an important conArdyth Simmons sideration. The model must also capNevada. The purpose of the test is to ture the differences in initial mineralevaluate the coupled thermal, hydroContact: ogy in fractures and matrix and their logical, chemical and mechanical Eric Sonnenthal evolution. processes that take place in unsatu(510) 486-5866, elsonnenthal@lbl.gov To handle these separate yet interrated fractured tuff over a range of acting processes in fractures and matrix, we have adopted the temperatures from ambient (approximately 25째C) to nearly dual permeability method. In this method, each grid block is bro200째C.The DST was begun on Dec. 3, 1997, with a planned fourken into a matrix and fracture continuum, characterized by its year period of heating, followed by four years of cooling. own pressure, temperature, liquid saturation, water and gas Our objectives were to make predictions of the coupled therchemistry, and mineralogy. mal, hydrological, and chemical (THC) processes, followed by Coupled THC simulations of the DST were carried out with model re finement and comparison to measured data. We the TOUGHREACT code (Xu et al., 1998), which was enhanced expected that some water (formed by condensation of steam in to handle regions of complete dryout, and chemistry under boilfractures) would be collected as was the case for the completed ing conditions (Sonnenthal et al., 1998; Spycher et al., 1998).The Single Heater Test. Throughout 1998, samples of water and gas geochemical module in TOUGHREACT solves simultaneously a were collected from boreholes that enabled us to further refine set of chemical mass-action and mass-balance equations to comour models and understanding of the complex rate-limited pute the extent of reaction and mass transfer between aqueous chemical interactions between rock, water and gas. species, minerals and gases at each grid block. Mineral-water Approach reactions take place under kinetic and/or equilibrium conditions, whereas gas-water and aqueous species interactions are Here we describe our conceptual model for coupled THC assumed to be at equilibrium. Equations for transport and chemprocesses at the DST and their treatment by numerical modelical reactions are carried out sequentially once the equations for ing. A conceptual model for reaction-transport processes in the heat, water and vapor flow are solved. fractured welded tuffs at the DST must account for the different The two-dimensional dual-permeability grid, the thermal and rates of transport in very permeable fractures, compared to a hydrological properties, and pressure-temperature-liquid saturamuch less permeable rock matrix. Transport rates greater than tion boundary conditions were developed by Birkholzer and

Figure 2. Cross-section of DST showing pH and temperature contours in fractures after one year of heating. Hydrology boreholes, packer intervals and the observation drift (OD) are overlain.

Figure 1. Cross-section of DST showing PCO2 and temperature contours in fractures after one year of heating. Hydrology boreholes, packer intervals and the observation drift (OD) are overlain.

83


ESD

Prediction and Analysis of Coupled Processes in the Drift Scale Thermal Test

Tsang (1997) for their predic- Minerals Gaseous Species Significance Primary Aqueous of Findings tions of the thermohydrological Species CO2 behavior of the DST. This pro- Cristobalite-a SiO2 (aq) H2O Previously, there has been litCa2+ vided an excellent starting point Quartz Na+ tle work on modeling reactionon which to build the model for Amorphous silica Cltransport processes in unsatuthe mineralogy, water and gas Calcite HCO3rated systems under boiling conchemistry of the DST.The miner- Gypsum SO42ditions. The DST presented an alogy of the rocks at the DST was H+ unprecedented opportunity to deri ved from studies done at test the conceptual models and L aw rence Live rmore and Los Table 1. Mineral phases, aqueous and gaseous species used in codes on a well-constrained sysAlamos national laboratories.The the DST model simulation. tem over time and spatial scales initial water and gas chemistry greater than simple lab experiments.The agreement of CO2 conwere based on data from the U.S. Geological Survey. centrations between model results and field measurements gives Results confidence that essential aspects of the coupling between thermal, hydrological, and chemical processes have been captured. The predictive modeling for the DST indicated that the pH of The methodology developed for the modeling of the DST can waters was controlled primarily by CO2 degassing during boilthen be applied with more certainty to predictions of the long ing, redissolution in condensate waters in fractures, and interacterm behavior of a potential nuclear waste repository. tion with calcite (Sonnenthal et al., 1998). However, to capture Related Publications the full chemical character of condensate waters it is necessary to consider several primary and secondary minerals (feldspars, Bi rkholzer, J.T., and Y.W.Tsang, Pretest analysis of the therm a l - hy d roclays, zeolites, silica phases, calcite and gypsum). This is apparlogical conditions of the ESF Drift Scale Test,Yucca Mountain ently the result of the extremely small effective reaction rates for Project Level 4 Milestone Report SP9322M4, LBNL, 1997. the aluminosilicates, such as feldspars,that have a measurable, yet Conrad, M., Isotope analyses of samples from the Drift Scale Test minimal, effect on the water chemistry over the time scale of hy d ro l o gy holes, in Second Quarter TDIF Submission days to weeks. (Hydrological, Radar and Microseismic), Chapter 3. Yucca A simulation illustrating the outgassing, transport, and redissoMountain Project Level 4 Milestone Report SP2790M4, LBNL, lution of CO2 and the precipitation and dissolution of calcite, silica phases and gypsum is presented here. Minerals, aqueous and 1998. gaseous species are given in Table 1. Sonnenthal, E., N. Spycher, J. Apps and A . Simmons,Thermo-hydroSimulation results for PCO2 and temperature in fractures after chemical predictive analysis for the Drift-Scale Heater Test, one year of heating are contoured in Figure 1. Regions of Version 1.1, Yucca Mountain Project Level 4 Milestone decreased PCO2 are evident near the drift wall, and along the SPY289M4, LBNL, 1998. wing heaters where dryout has occurred, or where water has S py cher, N., E.L. Sonnenthal and J.Apps, Interpretive analysis of the mostly boiled away. A large halo of increased PCO2 starts at thermo-hydrological-chemical aspects of the sSingle heater about the 90째C isotherm and extends well into the region of test, in Tsang et al., Yucca Mountain single heater test final ambient temperature (below 30째C). The water in fractures, report,Yucca Mountain Site Chara c t e rization Project, Chapter which started with a pH of about 8.3, has dropped to less than 7 4, Berke l ey Lab re p o rt LBNL-42537, 1999. in the condensate areas (Figure 2).The boiling zone, within the Xu, T., F. Gerard, K. P ruess and G. Brimhall, Introducing reactive wide interval between 90째C and 100째C, is characterized by solute transport to TOUGH2:Application to supergene copper enrichment, Proceedings of the TOUGH Workshop, 1998. higher pH waters and reduced PCO2. Numerous measurements of CO2 concentrations in gas colFunding lected from hydrology boreholes around the DST (Conrad, 1998) have shown a large region of increased partial pressures of CO2 (over 50,000 ppmv CO2), compared to the atmospheric concenThis work was supported by the Director, Office of Civilian tration (around 375 ppmv CO2). In terms of the magnitude and Radioactive Waste Management, through Memorandum Purchase distribution of PCO2 around the DST, these data compare remarkOrder EA9013MC5X between TRW Environmental Safe t y ably well to the model results. Some local differences Systems, Inc., and the Ernest Orlando Law re n c e are thought to be related to heterogeneity in the fracBerkeley National Laboratory under U.S. Department of ture system. Energy Contract No. DE-AC03-76SF00098.

http://www-esd.lbl.gov 84


Earth Sciences Division

Research

Objectives

Nuclear Waste Program Testing of a Coupled THM Model for Unsaturated Media Against Laboratory Experiments

Annual Report 1998-1999

independently developed thermohydrological two-phase flow computer ROCMAS is a three-dimensional code TOUGH2. Further code-to-code finite-element computer code develverifications and validations against oped at LBNL for analysis of coupled Jonny Rutqvist, Jahan Noorishad laboratory and field experiments are and Chin-Fu Tsang therm o hy d ro m e chanical (THM) provided by participation in the interprocesses in fractured rock. Recent national cooperative project DECOContact: interests in coupled THM processes VALEX (Development of COupled Jonny Rutqvist associated with nuclear waste reposimodels and their VAlidation against (510) 486-5432, jrutqvist@lbl.gov tories in geological media and, in parEXperiments in nuclear waste isolaticular, the issue of resaturation of a clay buffer around a waste tion). Within this project, the fundamental responses of a bencanister has encouraged major developments of the ROCMAS tonite clay material were investigated in a number of laboratory code in the past three years. The main objective is to develop a tests conducted by the Japan Nuclear Cycle Development tool for analysis of THM processes in practical field scale, includInstitute (JNC).The following tests were included: ing unsaturated rock masses and also detailed behavior of near• Suction test field fractures and an unsaturated clay buffer.This development • Infiltration test is accompanied with validation of the re l evant processes • Thermal gradient test through modeling of laboratory and field experiments. • Swelling pressure test First, a number of suction tests were conducted to determine Approach the relationship between suction pressure and saturation; that is, the water retention curve. Second, several infiltration experiThe ROCMAS code has been extended from analysis of fully ment were conducted on compacted bentonite specimens of 20saturated media to partially saturated media.This development is by-20 mm to determine the relative permeability, kr, and the effective molecular diffusion coefficient, Dv. The samples were based on Phillip and de Vries’ work on moisture and heat transinitially dry and water was supplied through a metal filter at the port in hysteretic, inhomogeneous porous media. In the new bottom. After various infiltration periods, the specimens were development, ROCMAS considers both liquid-water flow and sliced into 2-mm sections for measurement of the water content. vapor flow in air-filled pores due to molecular diffusion; both are The third type of experiment was the thermal gradient test for coupled with temperature and mechanical deformation.The liqdetermination of the thermal vapor diffusion. These tests were uid flow is driven by the pressure gradient and depends on the conducted on compacted bentonite samples of 68% saturation, relative permeability, which is a function of saturation.The vapor 50 mm in diameter and 100 mm in height (Figure 1). Applying flow is driven by the vapor density gradient and depends on an an elevated temperature at the lower boundary of the samples effective molecular diffusion coefficient, Dv. Airflow and convection of vapor with bulk airflow are not considered.Thus, this creates a thermal gradient. Temperature was monitored at variapproach may be limited to relative ly low-tempera t u reor low- perous distances along the sample and water content was measured m e ability systems where the steam convection can be neglected. from the weight loss during a subsequent oven drying. The The new algorithms of ROCMAS are verified against existing fo u rth test was conducted to determine the relationship between total stress and saturation due to swelling of the benanalytical solutions and by code-to-code comparison with the

Figure 1. Schematic view of thermal gradient test.

85


ESD

Testing of a Coupled THM Model for Unsaturated Media Against Laboratory Experiments

Figure 2. Experimental and modeling results of the thermal gradient test after 400 hours of heating.

tonite. The above laboratory tests were modeled with ROCMAS for determination of material properties and for validation of the newly implemented algorithms.

from wet toward dry areas.The experiments were ma t ched using a thermal diffusion enhancement factor, fTv = 1.0 â&#x20AC;&#x201C; 1.7. The swelling pressure test could be well matched with a Bishop type of effective stress law. Howeve r, the ex p e riment was conducted by increasing the saturation from 68% toward full saturation.The results may therefore not be re l evant for ve ry low saturation where the Bishops effective stress law pro b ably is not va l i d .

Results The suction tests resulted in water retention curves that cannot be exactly represented by any standard function. Therefore it is used exactly as is, and tabulated into the ROCMAS code. From the infiltration test, the relative perm e ab i l i t y, kr, and the effective molecular diffusion coeffi c i e n t , Dv, were determined directly from the test results by an analytical method. Thereafter, the actual ex p e riment was modeled with ROCMAS for validation of the code. The results of the modeling show that the liquid water flow and the relative permeability are important at saturations above 30%. Relative permeability at full saturation is kr = 1.0 and decreases to kr = 0.01 at about 30% saturation. Below 30% saturation, kr is so low that the vapor flow due to molecular diffusion becomes dominating, and the effective molecular diffusion coefficient is determined to be Dv = 2.0 Ă&#x2014; 10-6 m2/s. With the properties for isothermal diffusion known from the infiltration test, the properties for thermal diffusion can be backcalculated from the thermal gradient test. Figure 2 presents the results of modeling and experimentation for one of the tests after 400 hours of heating.The moisture transport in this experiment is dominated by the vapor flow due the thermal gradient.As the temperature increases at the hot end of the sample, liquid water vaporizes, increasing the density of vapor. In response to the increased vapor density, the vapor is transported towards cooler regions, where it is again condensed into liquid water.Thus, the bentonite becomes drier at the hot end and wetter at the cooler end. The vapor flow due to the thermal gradient is opposed by a liquid flow that is driven by the liquid pressure gradient, which wants to transport water back

Significance

of

Findings

The agreement between the data and simulated results indicates that the THM responses of the experiments are well represented by the new algorithms simulating coupled THM processes in unsaturated media. Furthermore, the results indicate that the approach of Phillip and de Vries is appropriate for the bentonite material and for the circumstances of these experiments.Thus, convection of vapor with gas flow seems to be minor in comparison to the molecular diffusion.

Related

Publications

Noorishad, J., and C.-F.Tsang, ROCMAS simulator:A thermohydromechanical computer code, in Coupled thermo-hy d ro mechanical processes of fractured media (Stephansson, Jing and Tsang, eds), pp. 551-558, Elsevier, 1996. Rutqvist, J., J. Noorishad and C.-F. Tsang, Coupled analysis of a thermohydromechanical experiment in rock at Kamaishi Mine,Swedish Nuclear Power Inspectorate technical report, submitted.

Funding This work has been supported by a grant from the Swedish Nuclear Power Inspectorate.

http://www-esd.lbl.gov 86


Earth Sciences Division

Research

Objectives

Nuclear Waste Program Coupled Analysis Of a THM Field Experiment In an Unsaturated Buffer-Rock System

Annual Report 1998-1999

for analysis of coupled THM processes in fractured and unsaturated rocks The Kamaishi Mine Heater Test is a (see article, this report). The threemajor field test case in the internadimensional model of the test envitional coopera t i ve project DECO- Jonny Rutqvist, Jahan Noorishad ronment consists of 11,158 elements and Chin-Fu Tsang VALEX (DEvelopment of COupled and includes 13 materials, with the models and their VAlidation against nearby drifts ex p l i c i t ly defined Contact: EXperiments in nuclear waste isola(Figure 1).The rock mass is simulated (510) 486-5432, jrutqvist@lbl.gov tion).The heater experiment was conas an equivalent continuum, but in the ducted in the fractured hard rock and clay buffer system of a near field a few important shear fractures are defined discreetly. potential nuclear waste repository design. A major objective of The in-situ permeability of these fractures could be determined by model calibration using measurements of water inflow the test is to observe near-field coupled thermohydromechanical through the wall of the open test pit.The property values of the (THM) phenomena in-situ and to build confidence in coupled buffer and intact rock matrix, including relative permeability, mathematical models of these processes. molecular diffusion coefficient, thermal conductivity, thermal Approach expansion coefficient, heat capacity and Youngâ&#x20AC;&#x2122;s modulus, were determined from laboratory tests of small samples. The experiment was conducted in a 5-by-7-m alcove excaResults vated off an existing drift located at a depth of about 300 m. In 1995, a test pit 1.7 m in diameter and 5 m in depth was drilled Figure 2a shows a comparison of the predicted and measured in the floor of the alcove. In 1996, a heater was installed into the temperatures at three key points at mid-elevation of the heater. test pit and surrounded by a buffer of bentonite-clay which had The agreement of the temperature is very good in these points an initial water content of 16% (66% saturation). The temperaas well as at other monitoring points in both the bentonite and ture of the heater was set to 100°C for eight and one-half months surrounding rock. After turning on the heater, the temperature followed by a six-month cooling period. The bentonite and the rises in the bentonite and approaches an apparent steady state rock surrounding it were extensively instrumented for monitorafter about one month, with a temperature of about 55°C at the ing of system response. Data included temperature, moisture bentonite-rock interface. The elevated temperature is a major content, fluid pressure, stress, strain and displacements. The driving force for the moisture flow in the bentonite and gives experiment was completed in the beginning of 1998 and thererise to thermal stresses. after the monitoring sensors were calibrated. There are two main processes controlling the water flow in The heater test is modeled with the computer code ROCMAS the unsaturated bentonite. First, there is vapor flow from the (Noorishad and Tsang, 1996), which is a finite-element program

Figure 1. Finite-element model of the Kamaishi mine heater test.

87


ESD

Coupled Analysis of a THM Field Experiment in an Unsaturated Buffer-Rock System

inner hot region towards cooler water content increases. On the regions of the bentonite driven other hand, the bentonite by the temperature gradient. shrinks near the heater where Second, there is a liquid flow the water content is reduced from the fully saturated rockand tensile fra c t u ring may bentonite interface towards the occur. Both the modeling and inner dryer regions of the benthe field measurements show tonite, which is driven by the that the swelling pressure in pressure gradient. The overall this case is low (less than 0.5 agreement of the time history MPa). However, there are conof water content in the bensiderable uncertainties in the tonite is satisfactory with all measurements of the swelling measured general trends cappressure as well as modeling of tured in the modeling (Figure the mechanical behavior of the 2b). Near the heater (BW5) the bentonite, especially when the water content gra d u a l ly degree of saturation is low. decreases to a few perc e n t Significance (10% saturation) at the end of of Findings the heating phase and thereafter increases slow ly duri n g The good agreement the cooling phase. In the midbetween the predictions and section between the heater and measured results regarding temthe rock (BW4) the water conpera t u reand water content inditent increases during the first cates that the therm a l - hydro l o gifew months and there a f t e r cal responses of the Kamaishi decreases. This tempora ry heater test are well represented i n c rease of water content is by the coupled nu m e rical due to a condensation zone model. Uncertainties remain in that is moving outwards during the hydromechanical behavior the heating phase. At BW3, of the bentonite and the infl uwhich is located in the bentonite about 1 cm from the ben- Figure 2. Comparison of model prediction (symbols) with measurements ence of the rock-bentonite inter(lines) for time history of temperature and water content in the bentonite face on the wetting process. tonite-rock interface, the water buffer. content increases to the maxiRelated Publications mum of about 25% (full saturation) within 10 days.Both the modeling and the field experiments indicate that there is no influNoorishad, J., and C.-F.Tsang, ROCMAS simulator: A thermohy d roence of high permeability fractures on the wetting of the benm e chanical computer code, in Coupled therm o - hydrotonite. This indicates that the rock permeability was sufficient mechanical processes of fractured media (Stephansson, Jing over the entire rock-bentonite interface to supply an unlimited and Tsang, eds), pp. 551-558, Elsevier, 1996. amount of water for wetting of the bentonite. However, the Rutqvist, J., J. N o o rishad and C.-F.Tsang, Coupled analysis of a therinflow from the rock into the bentonite was slightly over-premohydro m e chanical ex p e riment in rock at Kamaishi Mine, dicted at locations above and below the heater.This may be due to the fact that the permeability of the rock matrix is overestiSwedish Nuclear Power Inspectorate technical report, submitmated or that there is a sealing effect at the bentonite-rock interted. face that is not captured in the modeling. Funding There are two main processes controlling the changes in m e chanical stress during this ex p e riment. The immediate This work has been supported by a grant from the Swedish response is that the thermal expansion of the rock and bentonite Nuclear Power Inspectorate (SKI). The work was also partially gives rise to thermal compressive stresses in both the rock and supported by the National Energy Research Scientific the bentonite.Thereafter, as the water content changes, Computational Center (NERSC) through U. S . the bentonite either swells or shrinks. It swells and creDepartment of Energy Contract No. D E - AC O 3 ates increased compressive stresses (swelling pressure) 76SF00098. in regions near the rock-bentonite interface where the http://www-esd.lbl.gov 88


Earth Sciences Division

Research

Objectives

Nuclear Waste Program Full-Scale Tomographic Seismic Imaging of the Potential Repository Horizon At Yucca Mountain, Nevada

Annual Report 1998-1999

along the vertical z coordinate (i.e., horizontal layering).The first step is to The primary role of geophysical apply a 2-D velocity model, as derived studies at Yucca Mountain, Nevada, from VSP studies at UZ-16 (Majer et has been the measurement and imagal., 1996) and project the velocities ing of physical ro ck pro p e rties. Roland Gritto, Thomas M. Daley, for the lithologic units onto a 2-D Valerie A. Korneev, Properties such as density, conductivcross-section of the 3-D site scale geoity, bulk and shear moduli are used to Mark A. Feighner, Ernest L. Majer logic model.This original 2-D velocity and John E. Peterson estimate other ge o l o gic properties model has to be adjusted to match the such as stratigraphy, structure, saturaarrival times and waveform between Contact: tion, fracturing and permeability. the elastic waveform simulations and Roland Gritto Although boreholes and tunnels in the recorded field data.The modeling (510) 486-7118, rgritto@lbl.gov the Exploratory Study Facility (ESF) efforts produce evidence that first will allow direct examination of physical properties, there is a arrivals at far offsets are reflected waves from lower structures need to detect and characterize subsurface features away from rather than direct waves. Thus, to eliminate the larger offset and between these access points. Boreholes and tunnels to date travel times in the tomographic imaging inversions, only angles have given a very small window into the entire repository volof less than about 45 degrees are used. ume. In addition, the lateral variability and heterogeneity in the Results Topopah formation make it difficult to extrapolate between observation points. It is necessary to know the location of sigIn order to estimate the variations in seismic wave properties, nificant faults and fracture zones as well as variations in lithology the seismic traces, when aligned along their first arrival travel and rock type to help design and predict the performance of the times and corrected for geometrical spreading, reveal significant potential repository. The current experiment was undertaken to lateral variation in the amplitudes across the tunnel length. In broadly detect and characterize subsurface faults, fracture netorder to quantify this variation, the root mean square (rms) valworks and lithological features within the potential repository, ues of the amplitudes associated with the first arrival are deterthe middle-nonlithopysal zone (Tptpmn) and its vicinity. mined. To support this comparison, the fracture density, deterApproach mined along the tunnel walls, is presented alongside the amplitude values in Figure 1. Both curves are normalized (the fracture To achieve these goals a surface-to-tunnel seismic imaging surdensity to 1, the rms amplitudes to 2) to separate them in the vey was designed. A total of 180 vibroseis (vibrating) sources plot. It is evident that the correlation is good, particularly in the with a spacing of 30 m were sections with increased fracture deployed along Yucca Mountain intensity where the rms ampliRidge over a total length of 5 tudes rise ab ove their backkm, while 224 two-component ground value. geophone sensors were grouted To prove the hypothesis that in the tunnel with a separation this result is caused by site of 15 m between 2680 and 5970 amplification due to a decrease m, producing more than 5 km2 in velocity and density within of interpretable images. In order the fracture zone, a noise investo investigate and image the tigation is perfo rmed. If the repository area, the first arrival increase in seismic amplitudes times and associated amplitudes is due to site amplification, the of the seismic waves are interacnoise should be amplified in the t i ve ly determined. To better same way. Therefore, the rms understand the wave propagaamplitude of the noise is detertion in this survey, a 2-D elastic mined and the values displayed program with parameter adjustin Figure 1 (upper curve). It is ment allowing a simulation of 3evident that the noise is not D ge o m e t ry is used. Such a amplified in the same way as the reduction is justified for the seismic wave s , and there fo re , cases of 3-D pro blems with local effects can be ruled out as cylindrical symmetry where Figure 1. First arrival RMS amplitude values as a function of location in an explanation for the correlathe tunnel. Lower curve: fracture density as measured inside the tunnel. medium properties are laterally Middle curve: RMS amplitude of first arrivals. Lower curve: RMS ampli- tion between the seismic amplihomogeneous and change only tude of noise window. tudes and the fracture density. 89


ESD

Full-Scale Tomographic Seismic Imaging of the Potential Repository Horizon at Yucca Mountain, Nevada

The second hypothesis is that of a guidSignificance of Findings ing effect of the fractures, which keep the propagating waves concentrated within the Surface-to-tunnel tomographic imagi n g fracture zone, producing a pattern of conseems to be suitable to produce estimates s t ru c t i ve interference, and thus preventing of large-scale ve l o c i t y, a t t e nuation and amplitude losses due to spherical spreading fra c t u re density distributions at Yucca of the waves.Two-dimensional elastic modMountain. It is found that certain ge o m eeling of wave pro p agation through a fra ctries of fra c t u re distribution may lead to t u re zone revealed an increase in amplitude constru c t i ve interfe rence and consevalues of the waves comparable to the data quently to an increase in amplitudes of in Fi g u re 1. It remains to be seen, howeve r, the pro p agating wave s . The repository whether this effect is more likely for subhorizon appears to be hetero geneous h o rizontal or ve rtical fractures.Therefore, it re g a rding the degree of fracturing, with is like ly that the intensely fractured zone is the more intensely fra c t u red areas in the extending into the planed repository horisouthern part , while the nort h e rn end of zon, as it affects the amplitudes of the waves the survey area seems to be less fracd u ring their pro p agation from source to t u re d . Furtherm o re the alignment of fracreceiver. t u res may vary from the south towa rd s Numerical simulation indicates that the the nort h . The East-West drift appears to elastic waves pro p agate downwards at a steep angle towards the Tptpmn zone, after be located in an area of relative ly unfracwhich they are re f racted and propagate t u red rock. along this interface. To accommodate this Related wave propagation it is necessary to project Publications the source locations from the surface onto the top of the Tptpmn layer at depth, before Kaelin, B., Seismic imaging of the shallow 2-D tomographic imaging is possible. The subsurface with high frequency seisl ow velocity values and heterogeneity in the Figure 2. Tomogram of fracture density based on velocity estimates. The tomogram represents a mic measure m e n t s , Berke l ey Lab shallow subsurface produce strong va ri a- broad characterization of the variability of fracture report LBNL-42058, Ph.D thesis, UC tions in the travel times and amplitudes density in the mapped zone. Fracture density is Berkeley, California, 1998. b e t ween neighboring source locations. shown as fractures per cubic meter. Majer, E.L., M.A. Feighner, L. Johnson,T. Daley, E. Karageorgi, K.H. These va riations need to be taken into account by static corre cLee, K. Williams and T. McEvilly, Synthesis of borehole and tions, before the new source locations at depth can be used to surface geophysical studies at Yucca Mountain, Nevada, and i m age the repository horizon.The imaging of the P-wave velocity vicinity, Surface Geophysics, 1. Milestone Report OB05M, and attenuation distribution estimates is based on algebraic reconBerkeley Lab, 1996. s t ruction (Peterson, 1986). H owever, the most indicative image of the structure of the repository horizon is a map of fracture density Peterson, J.E., The application of algebraic reconstruction techdistribution. In an attempt to quantify the fracture properties, we niques to geophysical problems, Berkeley Lab report LBNLuse a relationship between seismic velocity (P- and S-wave), fra c21498, Ph.D. thesis, UC Berkeley, California, 1986. t u re density, matrix pro p e rties and fluid saturation deri ved by Funding Kaelin (1998). Fi g u re 2 shows the fracture distribution across the repository horizon. It is evident that the southern part of the inve sThis work has been supported by the Director,Office of Civilian tigated area is dominated by fracturing (west of the intensely fra cR a d i o a c t i veWaste Management, through Memorandum Purchase t u red zone encountered in the tunnel), possibly affected by the Order EA9013MC5X between TRW Env i ronmental Safety Systems, Ghost Dance West fault. In contrast, the nort h e rn part of the area Inc., and Ernest Orlando Lawrence Berkeley National Laboratory shows less fracturing, p a rt i c u l a r ly in the vicinity of the East-West for the Yucca Mountain Site Characterization Project under U.S. drift.The presence of the Sundance fault seems to be manifested Department of Energy Contract No. DE-AC 0 3 -76SF00098. by an interruption of the unfractured area at about 23,3300 m N o rthing (northerly direction).

http://www-esd.lbl.gov 90


Earth Sciences Division

Research

Objectives

Nuclear Waste Program Investigation Of Geologic Water Storage Near Cuzco, Peru

Annual Report 1998-1999

whether the stone walls act to intercept and store groundwater in-situ, a Inca rulers were held accountable field examination of the site for eviJerry P. Fairley by their subjects for provisions during dence of increased groundwater levtimes of adverse conditions. For examels behind the stone retaining walls Contact: ple, a local ruler in Lambayeque, Peru, was conducted. Following the field Jerry Fairley (510) 486-4161, jpfairley@lbl.gov was put to death by his subjects dursurvey, a numerical model was coning an especially severe drought.As a structed using the TOUGH2 flow simresult, the Inca developed sophisticated systems for the storage ulator code to illustrate the concept of geologic water storage at and distribution of goods such as textiles, maize, beans and other Tambomachay and gain insight into the effectiveness of this storagricultural products. In much of southern Peru, precipitation age system. occurs primarily during the monsoon season, and the ability to Results store and retrieve water is necessary to ensure consistent supply. Inca water storage systems consisting of reservoirs, cisterns and Inspection of the site yielded abundant evidence that the aqueducts are known from many sites.The focus of the present stone walls act to intercept and store groundwater.The site visit research is to demonstrate that the stonework at the site of occurred during the monsoon season. At that time, the soils Tambomachay, about 5 km north of Cuzco, is not merely ornabehind the walls were obviously at field capacity, and in areas mental, as has been assumed by previous authors, but rather where foot or livestock trails were incised below the general forms the basis of a sophisticated geologic water storage system land surface, standing water was in evidence. Soils to either side not as yet described in the literature. of the stone walls were damp from recent rains, but appeared Approach well drained. This observation suggests that recharge entering the aquifer from precipitation mounds up behind the stone conThe site of Tambomachay is dominated by a series of masonry struction, resulting in high saturations behind the walls.To either walls, ranging from 1 to 3 m in height (Figure 1).The upper two side of the stone walls, recharge is free to discharge through the stone courses are of high-quality masonry, with finely fitted stone coarse alluvial sediments to the Rio Cachimayo.Further evidence joints and smoothly finished faces, while the lower two courses of increased water table elevations behind the masonry comes are of slightly lower-quality fieldstone construction with mud from inspection of the walls themselves. Although the joints between the individual stone blocks are too tightly fitted to mortar. The tiers are separated by terraces of flagstone and allow significant seepage, precipitate encrustation (presumably pounded earth. Aside from such ubiquitous features of Inca CaCO3) and moss growth near the top of the highest tier indiarchitecture as trapezoidal niches and vertically tapered walls, cate that the water table regularly rises to the height of the stone the principal features of the site are the fountains, which origiretaining walls. nate from a small opening in the main stone course. From this Because the results of the field survey appeared to confirm discharge point, outflow is led through a series of channels to the hypothesis that the stone walls intercept discharge from feed first a single and then a double fountain. On either side of ground water, a model of the site was devised to illustrate the the masonry walls, the land surface slopes down from a limegeneral concept of geologic water storage and to give some stone ridge that marks the rear of the site to a small stream (the Rio Cachimayo) that flows in the bounds on the efficiency of the va l l ey bottom in front of the system. To this end, the system stone construction. was modeled in two cases. Case I The site re c e i ves approx irepresented the hydrologic sysmately 950 mm of rainfall per tem before the construction of year, almost all of which falls durthe stone retaining walls. It was ing the monsoon season. During modeled in two dimensions as a and shortly after the monsoon square bounded by three noseason, the toe of the slope is the flow boundaries and one consite of numerous springs that disstant-head boundary, which repcharge ground water from the resented the Rio Cach i m ayo . hill slope to the Rio Cachimayo. Case II represented the post-conAlthough these springs dry up struction system and was modshortly after the rainy season, the eled in two dimensions as a fountains continue to run yearsquare surrounded by four noround. Figure 1: The site of Tambomachay, about 5 km north of Cuzco in the flow boundaries, with a small In order to determine southern Peruvian highlands. constant head discharge are a 91


ESD

Investigation of Geologic Water Storage Near Cuzco, Peru

representing the discharge point solution of instituting a control for the springs from the stone on the discharge boundary walls. allowed the gro u n dwa t e r The results of modeling for resources to be stored in-situ, and these two scenarios can be seen used at a rate more suitable for in Figure 2, which represents a human needs. cross section through the aquifer It is presently unknown how dire c t ly behind the discharge p revalent ge o l o gic storage of point. In the figure, distance from water was in pre-Columbian Latin the discharge point (x-axis) and America. Although the example potentiometric head (y-axis) have presented in this paper is the only been normalized on the aquifer one observed by the author to ch a racteristic length and the date, the archaeological litera t u re maximum potentiometric head, contains seve ral re fe rences to respective ly, while the time is â&#x20AC;&#x153;elaborated springs,â&#x20AC;? which may shown as the hydrologic equivarefer to Tambomachay-style water lent of the Fourier number (the storage systems.The prevalence of square of the characteristic length this type of storage method is Figure 2. Comparison of pre-construction (Case I) and post-constructimes time, divided by the aquifer tion (Case II) normalized head levels at Tambomachay for several nor- important, because cultural deve ldiffusivity). By comparing the malized times. opment in South America wa s curves for the two cases, it can be heavily influenced by the ava i l ab i lseen that the Case II scenario (post-construction) contains sigity of water. G e o l o gic water storage would allow colonization of nificantly greater quantities of water in storage at all times than otherwise marginal areas and could therefore comprise an important control on settlement patterns not only in South America, but the Case I scenario. By assuming reasonable values for specific possibly across cultures. For example, a correspondence has been storage, hydraulic conductivity, characteristic length, etc., it can noted between ge o l o gic water storage at Tambomachay and the be shown that potentiometric head values in the pre-construcseeps at the feet of shallow escarpments that provided water for tion aquifer would reach a value of 1% of the original value Mesa Verde in the Four Corners re gion of the United States (V. approximately 19 days after cessation of the monsoon rains. Scarborough,personal communication). As at Tambomachay, many Using the same parameter values for the Case II aquifer yields a ge o l o gic water storage systems may still be operational, impacting value of about two years to decrease heads to 1% of their origiland use patterns from pre-Columbian through modern times. nal values. In addition, although both cases exhibit exponential decay of discharge rate, dimensionless discharge for the Case I Related Publication aquifer decreases by more than two orders of magnitude by a dimensionless time of two units. Over the same period of time, Fairley, J.P., G e o l o gic water storage in Pre-Columbian Peru, the Case II aquifer dimensionless discharge decreases by only Berkeley Lab report LBNL-40581,1999. about 10%. The effect of this discharge leveling is a nearly constant flow rate over the entire year for the post-construction Funding aquifer.

Significance

of

Findings

This work has been supported by a grant from the Stahl Foundation and by the Director, Office of Civilian Radioactive Waste Manage m e n t , through Memorandum Purchase Order EA9013MC5X between TRW Environmental Safety Systems, Inc., and Ernest Orlando Lawrence Berkeley National Laboratory under U. S . Department of Energy Contract No. DE-AC0376SF00098.

A comparison of the pre- and post-construction system models clearly indicates the dilemma of the ancient hydrologists: the natural system contained adequate resources for one or more years of human use, but under pre-construction conditions, this water was exhausted in a minimal amount of time. The elegant

http://www-esd.lbl.gov 92


Earth Sciences Division

Annual Report 1998-1999

ENERGY RESOURCES PROGRAM

LARRY R. MYER CONTACT: (510) 486-6456 LRMYER@LBL.GOV

The Energy Resources Department (ER) Is responsible for two major program areas: Oil and Gas Exploration and Development, and Geothermal Energy Development.

• Development of magnetotellurics for marine applications; • Development of mid-frequency electromagnetic techniques with a focus on imaging through casing; • Improved inversion methods for reservoir characterization, with a focus on combining production and geophysical data; • Application of X-ray C7 and NMR imaging to study multiphase flow processes; • Pore-to-laboratory-scale study of physical pro p e rties and processes, with a focus on controlling phase mobility, predicting multiphase flow properties and drilling efficiency; • Development of reservoir process-control methods; • Development of new methods to mitigate environmental effects of petroleum refining and use; • Enhancement of refining processes using biological technologies. Since 1994, the major part of the Oil and Gas Exploration and Development program has been funded through the Natural Gas and Oil Technology Partnership Program. Begun in 1989, the partnership was expanded in 1994 and again in 1995 to include all nine Department of Energy multi-program laboratories, and has grown over the years to become an important part of the DOE Oil and Gas Technologies program for the national laboratories. Partnership goals are to develop and transfer to the domestic oil industry the new technologies needed to produce more oil and gas from the nation’s aging, mature domestic oil fields, while safeguarding the environment. Partnership technology areas are: • Oil and gas recovery technology

Oil and Gas Exploration and Development Multidisciplinary research is being conducted in reservoir characterization and monitoring, optimization of reservoir performance and environmental protection. Using basic research studies as a source of innovative concepts, Energy Resources Department researchers seek to transform these concepts into tangible products of use to the industry within a time-frame consistent with today’s rapid growth in technology. Reservoir characterization and monitoring involve development of new seismic and electromagnetic techniques focused at the inter-well scale. Optimization of reservoir performance involves application of reservoir engineering and geomechanics principles to enhance production.The next major step in research will focus on methods to optimize performance through integration of monitored geophysical data, production data and reservoir simulation. Principal research activities include: • Development of single-well seismic technology, including instrumentation, acquisition and processing; • Applications of seismic methods for characterization of fractured reservoirs; • Use of converted waves in cross-well applications;

93


ESD

Energy Resources Program

• • • •

Diagnostics and imaging technology Drilling, completion and stimulation Environmental technologies Downstream technologies International and national concern about the variable climactic effects of greenhouse gases produced by burning of fossil fuels is increasing, while it is also recognized that these fuels will remain a significant energy source well into the next century. In response to these concerns, the Energy Resources Program is initiating research in methane hydrates and geologic sequestration of CO2. These studies are focused on development of technologies which will minimize the impact of fossil fuel usage on the environment. Projects are typically multi-year, are reviewed and reprioritized annually by industry panels and are collaborations between national laboratories and industry

Geothermal

Energy

and development. In recent ye a rs, DOE’s geothermal pro gram has become more industry-dri ve n , and a significant part of the Berke l ey Lab effort has been directed towa rd industry assistance, e s p e c i a l ly in the area of technology tra n s fer and in understanding the nature and dynamics of The Geysers ge o t h e rmal field in northern California, which has begun to show the effects of ove rexploitation. At present, the main research activities of the pro gram include: • Development and enhancement of computer codes for modeling heat and mass transfer in porous and fractured rocks; • Laboratory investigations of the hydraulic and thermal properties of fractured and intact reservoir rocks; • Isotopic and noble gas studies to characterize geothermal fluids and to identify their sources, their potential migration paths and the connectivity within the geothermal reservoir; • Documentation of the behavior of geothermal fluids under commercial production and injection operations (e.g., field case studies), with specific emphasis on The Geysers field.

Development

The main objective of ER’s geothermal energy deve l o p m e n t p ro gram is to reduce uncertainties associated with fi n d i n g , characterizing and evaluating geothermal resources. The ultimate purpose is to lower the cost of geothermal energy for electrical generation or direct uses (e.g., agri c u l t u ral and industrial applications, aquaculture, b a l n e o l o gy, etc.).With these goals in mind, existing tools and methodologies are upgraded and new techniques and instrumentations are developed for use in the areas of ge o l o gy, geophysics, geochemistry and reservoir engi n e e ring. The pro gram encompasses theore t i c a l , laboratory and fi e l d studies, with an emphasis on a multidisciplinary appro a ch to solving the pro blems at hand. Coopera t i ve wo rk with industry, u n i ve rsities and gove rnment agencies draws from Berke l ey Lab’s 25 years of experience in the area of geothermal re s e a rch

Funding The Oil and Gas Exploration and Development program receives funding from the Fossil Energy, Natural Gas and Oil Technology Partnership Program and Office of Science of the U.S. Department of Energy. Support is also provided by the Gas R e s e a rch Institute (GRI) and direct industry contri b u t i o n s . Industrial collaboration is an important component of the DOE Fossil Energy and GRI programs. The Geothermal Energy Development program receives support from the Office of Energy Efficiency and Renewable Energy, Office of Geothermal Technologies, of the U.S. Department of Energy.

http://www-esd.lbl.gov 94


Earth Sciences Division

Energy Resources Program

Annual Report 1998-1999

Development of Single-Well Seismic Imaging Research

Objectives

Crosswell seismic-imaging applications have yielded a variety of tools and approaches over the past several years. It has become clear that it is possible to leverage this technology to address the next frontier in borehole seismology—that of single-well seismic imaging (SWSI).A multi-participant research program in SWSI is being led by Berkeley Lab and includes Idaho National Engineering and E nv i ronmental Lab o ra t o ry, Sandia National Laboratories and Stanford University. The continuing objective of this project is to identify and provide solutions to fundamental issues surrounding single-well seismic imaging in order to evaluate and develop the technology in a timely and cost effective fashion.

Thomas M. Daley and Ernest L. Majer

(510)

Contact: Thomas Daley 486-7316, tmdaley@lbl.gov

from the Uni-well project in the United Kingdom. The project is designed around four major tasks re p resenting the tech n o l o gi e s d e s c ribed ab ove. Task 1 – Instrumentation Design and Development The objective of this task will be to design and develop prototype instrumentation to augment current technology. The project team draws on past industrial ex p e rience, together with computer and laboratory experiments and analysis, to design, fabricate and field test hardware concepts for enhancing the signal-to-noise ratio for different singlewell source/receiver combinations. Task 2 – Modeling of Wave Propagation in Complex Media Algorithms for performing computational modeling of the seismic wavefield expected in a single-well recording environment will enable us to investigate a large variety of factors that may influence the success of this imaging technique. Efforts include (1) development of elastic or viscoelastic wave propagation algorithms to generate realistic synthetic seismogra m s , and (2) investigation of the influence of geologic fa c t o rs (e.g., curved stra t a , overhung salt flanks, diffuse reflectors) on the quality/utility of acquired data and on various data acquisition issues, such as sourcereceiver offset.

Approach

The current work consists of four interdependent activities which comprise facets of the techn o l o gy required for the ultimate, successful development of singlewell seismic imaging (and required for improvements in cro s swe l l imagi n g ) . T h ey are (1) Hardwa re : sources/re c e i ve rs , telemetry/re c o rd i n g , borehole noise effe c t s , deployment; (2) Modeling: synthetic seismograms, Figure 1. Schematic of Single-Well Acquisition system, including a 3500-m fiber-optic cable for data transmission and borehole para m e t ric studies, i nversion, o p t i- A/D converter for data acquisition. The system is shown with five mal designs for hardware/surveys; three-component sensors. Also available is a sensor string of 16 (3) Field Testing: quality data sets, hydrophones. eva l u a t i o n / validation at well ch a ra c t e rized sites; 4) Data Task 3 – Field Testing and Data Acquisition Processing and Interpretation: algorithms, 3-D imagi n g , noise In order to evaluate and develop a useful methodology, it must reduction, visualization. be tested in env i ronments representative of the sites of eve n t u a l We are working cl o s e ly with the industry’s 12-company Salt application.The on-going objective of this task is to validate and I m aging Consortium, which is focused on the use of single-well test current ava i l able methodologies as well as test new instrus u rveys to image salt dome flanks and the leve raging of resources mentation concepts to identify optimal data collection systems,

95


ESD

Development of Single-Well Seismic Imaging

well re flection survey was acquired over 1000 ft of a zone i n cluding salt and sediments near the base of the salt dome. Task 4 – Development and Application of Software to Process Single-Well Data Travel-time analysis of the axial vibrator seismic data at the Bayou Choctaw test site has been initiated.The goal of this modeling/analysis is to identify events that are potential reflected arrivals from the nearby salt dome flank. Several events in the vertical and horizontal component recorded data have nearly linear moveout at large source-receiver ranges, and can be modeled as mode-converted reflected arrivals.

modeling needs and data-processing schemes. Initially, DOE-sponsored effo rts concentrated on interfacing a dive rse set of sources and receive rs in order to allow field testing. Task 4 – Development and Application of Software to P rocess Single-Well Data The processing and interpretation of single-well imaging data is far from routine.The objective of this task is to evaluate and aid in the development of processing techniques that will be needed for single-well seismic imagi n g , i n cluding analysis of va rious borehole seismic sources such as the orbital vibrator.

Results

Significance of Findings

Our efforts at LBNL have focused on tasks 1, 3 and 4.

The successful acquisition of single-well imaging data with a powerful bore h o l e source (the AC orbital vibrator) and a modern mu l t i channel borehole acquisition system in a oil production environment leads the way to a new technique in exploration geophysics. There are immediate applications fo r i m aging the flanks of salt domes. Another important application will be imaging around horizontal boreholes to find sub-horizontal horizons. The technology developments of this program will also improve the state-of-theart in cro s s - well seismic imaging.

Task 1 – Instrumentation Design and Development The borehole hardwa re originally available was limited by data tra n s m i s s i o n rates possible with standard wirelines, so a fiber optic design was chosen to increase commu n i c a t i o n and bandwidth. rates Successfully addressed were the design, fabrication and testing of fiber optic connectors, including transmission Figure 2. SWSI data set for orbital vibrator source and three-component sensor. In-line and cross-line source orientations (left and right) are shown for horizontal, through a seismic sourc e horizontal and vertical sensors (top to bottom) for source depths 4380 to 3380 ft (Conoco AC orbital vibrator) (left to right). All data had a source/receiver offset of 185 ft. Direct S-wave arrivals, in a borehole environment. tube-wave arrivals and scattered wave arrivals are indicated. Additional processing will enhance the scattered and reflected waves while muting the direct In addition, Sea Con Inc. com- P- and S-waves and tube-waves. pleted the adaptation of the AC orbital vibrator such that it can now be used with the comRelated Publications mercially available P/GSI fiber optic wireline (20,000 ft) in addition to the LBNL fiber-optic wireline (10,000 ft).Also acquired by Majer, E.L., J.E. Peterson,T.M. Daley, B. Kaelin, J. Queen, P. D’Onfro Berkeley Lab was a OYO fiber-optic system, which will accomand W. Rizer, Fracture detection using cross-well and single modate the Exxon multilevel receiver system, the Conoco fivewell surveys, Geophysics, v62, n2, 1997. level geophone string or the Conoco 15-level hydrophone string. Daley,T.M., Single well seismic imaging tests: Nov. 1997 at Bayou This provides great flexibility in data acquisition capability. A Choctaw Site, Berkeley Lab report LBNL-42672, 1998. schematic of the current system is shown in Figure 1. Daley,T.M., Single well seismic imaging in a deep borehole using a piezoelectric orbital vibrator, Berkeley Lab report LBNLTask 3 – Field Testing and Data Acquisition 42673, 1997. The second salt dome field test was carried out by Berke l ey Lab in November 1998 using the AC orbital vibrator in conFunding junction with the CONOCO fi ve - l evel 3C wall lock re c e i ve r string in well Wi l b e rt #28 inside of the salt dome. The well This work has been supported by the Assistant exited the salt, giving ground truth on the location Secretary for Fossil Energy, National Petroleum of the edge. T h ree offsets were acquired (167, 184 Office of the U.S. Department of Energy under and 204 ft), giving a maximum of 15-fo l d . An ex a mContract No. DE-AC03-76SF00098. ple of this data is shown in Fi g u re 2. A full singlehttp://www-esd.lbl.gov 96


Earth Sciences Division

Energy Resources Program

Annual Report 1998-1999

Vertical Seismic Profiling At the Rye Patch Geothermal Field, Nevada Research

Objectives

geophone used 14-Hz ve rtical and horizontal geophones. Six data chanThe Rye Patch Reservoir area in nels were recorded: the three geonorthern Nevada has had preliminary phones, the source pilot, the vibrator ge o t h e rmal re s o u rce ex p l o ra t i o n . reference and the vibrator baseplate Contact: Initial exploration in the late 1980s accelero-meter.The record length was Thomas Daley and early 1990s resulted in only one 12,288 samples at a 1-millisecond (510) 486-7316, tmdaley@lbl.gov successful well (44-28). Other wells (ms) sample rate, giving a 2.3-s correwere either too cold or had no flow. In 1997 TransPacific lated record length.A 10-Hz low-cut filter was used; no high-cut Geothermal Inc. (TGI) proposed a 3-D seismic survey to deterfilter was used except for the anti-alias filter. mine the geologic structure of the potentially fault-controlled Data was acquired from a 600-ft offset location northwest of reservoir.This would be possibly the first application of the 3-D the well, spanning the depth range in the well from 1000 to 4200 seismic method to a geothermal field and therefore of interest to ft at 40-ft intervals.Four sweeps were summed in the field to prothe entire geothermal community. The 3-D seismic method has duce one record.One to three sets of four were recorded at each proven an integral part of modern oil and gas exploration efforts; depth. Well depths were measured from the ground level of however, the heterogeneous and hydrothermally altered nature 4,418 ft. Because of borehole fluid pressures, a lubricator and of geothermal reservoirs makes all seismic imaging more diffipacker were used to place the borehole geophone in the well. cult. It was not known if the methods used in the petroleum The geophone had a temperature monitor, which showed a maxindustry could be directly transferred to the geothermal indusimum temperature of 259°F. try. The VSP dataset was processed to obtain accurate seismic Before conducting a fullvelocity as a function of scale 3-D survey, DOE condepth and to image any tracted LBNL to investigate reflections in the data. The the viability of seismic imagfirst step in ref lection proing in the Rye Patch area. cessing was to balance trace LBNL obtained a vertical amplitudes with an autoseismic profile (VSP) in nonmatic gain control (AGC) of producing well 46-28 to test 200 ms, followed by a frethe seismic reflectivity in the quency-wave n umber (F-K) area and to obtain velocity filter to remove the downgoi n fo rmation for designing ing energy.After this process, and potential processing of there still was some coherthe proposed 3-D seismic ent tube wave noise that was survey. This initial borehole removed using a median dip filter.This was followed by a seismic study would provide 200-ms AGC. After these proa go/no-go decision for the cessing steps, two pro m ilarger, costlier 3-D study. nent, c o h e rent re flectors Approach were present.

Mark A. Feighner, Thomas M. Daley and Ernest L. Majer

Results A vertical seismic profile (VSP) was recorded in Rye Using the velocity model Patch by LBNL between Dec. Figure 1. The VSP data are shown here with a geologic cross-section through the 11 and Dec. 13, 1997. The VSP well (46-28) and the producing well (44-28). A strong reflector can be seen obtained from the fi rst arrival times, the VSP reflecVSP in well 46-28 used from the clastic unit, which is the main production unit in this geothermal field. tion data are mapped to depth. The results are superimposed LBNLâ&#x20AC;&#x2122;s vibroseis source and a single-level, high temperature, hy d raulic wa l l - l o cking, three-component seismometer. The upon a cross-section traversing the VSP well (46-28) and the prosource was a P-wave vibrator. The source sweep was 10 to 80 ducing well (44-28); see Figure 1.The upper reflector correlates Hz, 10 seconds (s) long, with a 0.2 s cosine taper.The borehole with the sandstone/siltstone (upper member) of the Natchez

97


ESD

Vertical Seismic Profiling at the Rye Patch Geothermal Field, Nevada

Pass Formation. This is the main permeable clastic unit which produces the thermal fluids at the Rye Patch wells. It spans the elevation range of 1300-1500 ft above sea level and is strongly coherent to about 180 ft northwest of the well. At this point, there is a loss of the reflection, possibly due to changes in rock properties or the presence of a fault. The deeper reflector appears within the lower member of the Natchez Pass Formation, and may occur at a limestone/siltstone interface. This reflector is about 400 ft below sea level and is coherent over 285 ft northwest of the well.This spans the entire CDP transform range, indicating that this reflector is more continuous and may continue beyond the mapped extent.The depth of this reflector is not certain because it occurs below the last sensor and there is no velocity control from direct arrivals.

Also, seismic attenuation at this borehole was not extreme enough to limit a 3-D survey. Based on this information, a â&#x20AC;&#x153;goâ&#x20AC;? decision was made, and a three-square-mile 3-D seismic survey was collected in August 1998. The processing of this dataset is ongoing, and will be reported at a later date.

Related

Publication

Feighner, M. A.,T.M. Daley and E.L. Majer, Results of vertical seismic profiling at well 46-28, Rye Patch Geothermal Field, Pershing County, Nevada, Berkeley Lab report LBNL-41800, 1998.

Funding Significance

of

Findings This work has been supported by the Assistant Secretary for Energy Efficiency and Renewable Energy, Geothermal Division, of the U.S. Department of Energy under Contract No. DE-AC0376SF00098.

The VSP data collected at well 46-28 did produce a coherent reflection from the permeable clastic unit, which is the main production unit in this geothermal field at a depth of about 3,000 ft.

http://www-esd.lbl.gov 98


Earth Sciences Division

Energy Resources Program

Annual Report 1998-1999

Characterization of Fractured Geothermal Reservoirs Using Inverse Modeling Research

Objectives

in high-salinity geothermal reservoirs, precipitation or dissolution of salt Numerical modeling is an essential may occur, changing fracture porosity tool for the design and optimization and thus the overall permeability of of production and injection operathe reservoir. Changes in sodium chlo*Orkustofnun, Reykjavik, Iceland tions at geothermal reservoirs. The ride concentrations may there fo re Contact: response to production and reinjeccontain information about fluid flow Stefan Finsterle tion is governed by the coupling in the fracture network, indicating (510) 486-5205, safinsterle@lbl.gov between fluid flow in fractures and potential connections between injecheat transfer from adjacent matrix blocks. Furthermore, flow of tion and production we l l s , which may eve n t u a l ly lead to water, steam, gas and heat are strongly affected by the geometric unwanted thermal interference. and hydrological characteristics of the fracture network. The next step is to develop an optimization routine to autoThe reliability of model predictions depends on the accuracy matically improve the match between the observed data and the with which these coupled processes are accounted fo r. model calculation. A number of nonlinear minimization algoMoreover, the salient features of the geothermal reservoir must rithms have been incorporated into the iTOUGH2 inverse modbe captured, including the development of an appropriate coneling code (Finsterle, 1999). By performing synthetic data inverceptual model and the determination of thermal and hydrologic sions, the contribution of each potential observation to the estiparame t e rs . Inve rse modelingâ&#x20AC;&#x201D;automatic calibration of the mation of relevant input parameters can be determined. For numerical model against field dataâ&#x20AC;&#x201D;is a means to obtain modelexample, temperature data obtained in production and observarelated parameters that can be considered optimal for the given tion wells contain aggregate information about hydrologic and conceptual model. However, the large number of parameters thermal properties of the reservoir, which govern the conducneeded to fully describe coupled tive heat exchange from the matrix nonisothermal multiphase flow in blocks to the flowing fluids in the fra c t u re d - p o rous media often fractures. Steam production and flowing enthalpy data as well as leads to an ill-posed inverse probtracer data are likely to contain lem, w h i ch is predisposed to information about effective fracyielding nonunique and unstable t u re pro p e rties on the re l evant solutions. field scale. The general objective of this An error analysis is performed research is to assess the usefulafter calibration to assess the ness of inverse modeling for the uncertainty of the estimated calibration of geothermal reserp a ra m e t e rs and to reveal voir models. More specifically, we unwanted parameter correlations. try to identify the data that conUsing inverse modeling, the layout tain relevant information regardof the monitoring system can be ing fluid and heat flow in the optimized to reduce estimation reservoir. Effective fracture properties will be estimated, potenuncertainty. The appro a ch was tially improving the accuracy of tested using data from a synthetic long-term model predictions. geothermal reservoir, and by simulFigure 1. Identification of best-estimate parameter set by minimiz- taneously matching field data from ing objective function, which includes four datasets. The green line a well completion test and subseApproach shows the inversion search path. quent production. The most important element in geothermal inverse modeling Results is a sophisticated multiphase flow simulator, capable of capturing the coupling between fluid flow and heat transfer. As an The iTOUGH2 code was used to simulate production from a example, extraction of hot fluids and reinjection of cold water hypothetical geothermal reservoir with high salinity and CO2 as leads to vaporization and condensation effects near production the non-condensible gas. Due to precipitation of salt near the and injection wells, respectively. Further-more, as a result of presproduction well and the depletion of fluid reserves in the resersure and temperature variations during production and injection

Stefan Finsterle, Grimur BjĂśrnsson* and Karsten Pruess

99


ESD

Characterization of Fractured Geothermal Reservoirs Using Inverse Modeling

voir, the production of steam Significance of Findings declines rapidly and almost ceases within a relatively short period of The iTOUGH2 analyses show time. After five years of exploitathat a joint inversion of all available tion, condensate is reinjected a few data greatly improves the identifiahundred meters from the producbility of key hydrologic and thertion well. Injection of cold water mal properties affecting geotherleads to a reduction of steam satumal field perfo rmance. Adding ration in the immediate vicinity of tracer concentration data to the the injection wells. Evaporation of analysis considerably reduces the injectate, howeve r, i n c reases the correlation among some of the reservoir pressure, driving steam parameters, allowing for a more towa rds the production we l l , independent and more stable estienhancing both the rate and mation of reservoir properties. enthalpy of the produced fluid. Inverse modeling is a powerful Time series of simulated temperatures, steam production rates, flow- Figure 2. History matching of enthalpy data from the Krafla, Iceland, tool for the design of field tests, for the optimization of re i n j e c t i o n ing enthalpies and NaCl concentra- high-temperature geothermal well KJ-31. operations and for the evaluation of prediction uncertainties tions in the production well are considered to be the data availfrom geothermal reservoir simulations. Combining the characterable for model calibra t i o n .The parameters studied include fracization efforts from a variety of disciplines and making use of all t u re spacing, f ra c t u re absolute permeability, po ro sity, initial available data obtained during testing or production provides the re s e rvoir tempera t u re , heat conductivity and an exponent that basis for a better understanding of nonisothermal multiphase describes the ch a n ge in perm e ability as a function of poro s i t y flow processes, for the development of an appropriate concepch a n ge due to salt dissolution and precipitation. Fi g u re 1 shows tual model and for the estimation of the properties required to contours of the objective function—an integral measure of misperform reliable model predictions. fit between all ava i l able data and the corresponding model output—in the parameter space spanned by log-perm e ability and Related Publications intrinsic fra c t u re poro s i t y. A unique parameter set was accura t e ly identified by the minimization algo rithm within a few Finsterle, S., iTOUGH2 user’s guide, Berkeley Lab report LBNLiterations, as indicated by the green search path. 40040, 1999. In the field study, pressure data from multistep cold water Fi n s t e r l e , S., K. Pruess, G. Björnsson and A. Battistelli, injections into the high-temperature well KJ-31 at the Krafla Characterization of fractured geothermal reservoirs using geothermal field, Iceland, were matched simultaneously with inverse modeling, in Proceedings, Dynamics of Fluids in enthalpy data observed during the initial discharge period after Fractured Rocks, February 10–12, Berkeley lab report LBNLwell completion (Finsterle et al., 1999). Figure 2 shows the 42718, pp. 152–154, 1999. enthalpy data (red symbols), along with the simulation results obtained with an initial parameter set from a conventional well Funding test analysis (dashed line), and the best match obtained using iTOUGH2 (solid line).The estimated fracture properties and iniThis work has been supported by the Assistant Secretary for tial steam saturation also honor the pressure data observed durEnergy Efficiency and Renewable Energy, Office of Geothermal ing the multistep completion test (not shown). While the presTechnologies, of the U.S. Department of Energy, under Contract sure data alone can easily be matched using either a single- or No. DE-AC03-76SF00098, with in-kind contributions from the dual-porosity model, the inclusion of fractures is essential to adeNational Energy Authority, Reykjavik, Iceland. quately account for the boiling of large quantities of cold drilling water, as evidenced by the enthalpy data.

http://www-esd.lbl.gov 100


Earth Sciences Division

Research

Objectives

Energy Resources Program Comparison of the Two Most Important Salton Trough Geothermal Fields

Annual Report 1998-1999

Approach

The Salton Trough is host to the Data published in reports, confermost prolific geothermal power proence art i cles and journal papers were Marcelo J. Lippmann, duction area in North America. It assembled and analyzed to determ i n e Alfred H. Truesdell includes the Imperial Valley of souththe main characteristics of the Cerro and George A. Frye ern California (U.S.) and the Mexicali P rieto and Salton Sea ge o t h e rmal Valley of Baja California (Mex i c o ) , fi e l d s . Additional info rmation wa s Contact: obtained from the Califo rnia Division where a number of high-temperature Marcelo Lippmann (510) 486-5035, mjlippmann@lbl.gov of Oil, Gas and Geothermal Resources. (above 180°C) geothermal areas are under commercial production. At Results present, the installed electrical generation capacity in the area exceeds 1,000 MW (i.e., in Mexico, 620 MW at Cerro Prieto As described by Lippmann et al., 1999, the comparison (CPGF); and in California, 248 MW at Salton Sea (SSGF), 105 MW between CPGF and SSGF shows that they are alike in certain at East Mesa and 84 MW at Heber). aspects and quite different in others. Both have similar geologic The existence of these geothermal fields is due to the region’s frameworks and maximum measured temperatures (350-370°C). particular geologic environment. The Salton Trough is a broad They differ slightly in lithology (more continental, especially structural basin, characterized by high heat flow, tectonic deforlacustrine sediments at the SSGF) and ultimate energy capacities mation and seismicity, as well as volcanism resulting from tec(more than 800 MW at CPGF, more than 1000 MW at SSGF). tonic activity that created a series of pull-apart basins and transHowever, there are large differences in the chemistry of their form faults linking the East Pacific Rise to the San Andreas fault geothermal fluids (Figure 2) and in the amount of evaporites and system (Figure 1). sulfides in their sedimentary columns, both being higher at SSGF. As is typical of geologic systems, all geothermal fields have Different conceptual models apply to the two fields. unique characteristics; however they also have underlying simiThe contrast in fluid chemistries is explained by the positions larities. The CPGF and SSGF were compared because of their of the fields with respect to the crest of the Colorado River delta. large size and high temperature. The purpose was to establish South of the crest, in an area in good communication with the the similarities and differences between these two Salton Trough sea (i.e., the Gulf of California to the south), salts do not accufields from the earth sciences and developmental points of view. mulate in the deltaic sediments because surface and groundwaters tend to flow toward the gulf.On the other hand, at the SSGF, located in the closed Salton depression north of the delta crest,

Figure 2. Chemical compositions of aquifer fluids from Cerro Prieto well M-5 and Salton Sea SSSDP well State 2-14 compared to those of normal seawater and Reykjanes, Iceland, well 8 (geothermal fluid compositions calculated for reservoir conditions).

Figure 1. Generalized map showing the location of the Cerro Prieto (CPGF) and Salton Sea (SSGF) geothermal fields within the Salton Trough, Gulf of California, and the East Pacific Rise tectonic regime.

101


ESD

Comparison of the Two Most Important Salton Trough Geothermal Fields

Figure 3. Schematic model of the Salton Trough when the pull-apart basins “moved” into the areas of the Salton Sea (SSGF) and Cerro Prieto (CPGF) geothermal fields.

salts have accumulated because waters can only leave the basin by evaporation. In both areas concentrated brines infiltrated and accumulated in deep sediments of the Salton Trough. At SSGF the brines had higher salinities and were of continental origin, while at CPGF they were less concentrated and mainly marine. When the pullapart basins “moved” into these two areas and the areas were heated by igneous intrusions,the brines were mobilized, forming diapirs (domes) which did not reach the ground surface (Figure 3). At CPGF mixing of the ascending brines with less saline groundwaters was significant, while at SSGF, because of the higher density contrast between the brines and the local groundwaters, mixing was minor. The need to solve the problems of handling the higher salinity (and corrosivity) of the SSGF fluids, as well as differences in U.S. public policy and economic considerations were the main factors that delayed the development of SSGF compared to that of CPGF in Mexico.

Significance

of

the largest developed liquid-dominated geothermal systems in Mexico and the United States; additional power plants are planned in both fields. A good conceptualization of the geohydrologic systems will help optimize the expansion of the wellfields and reduce the impact of large-scale fluid production and injection on the geothermal reservoirs. Further expansion of geothermal electricity production in the Salton Trough will depend on economic factors, although public policy in favor of renewable energy—e.g., to meet Kyoto global warming go a l s — m ay foster development. Therefo re in the future, the comparison will be extended to other fields in the region.

Related

Publication

Lippmann, M., A. Truesdell and G. Frye, The Cerro Prieto and Salton Sea ge o t h e rmal fi e l d s — A re they re a l ly alike?, Proceedings, 24nd Wo rkshop on Geothermal Reservo i r Engineering, Stanford, CA, Jan. 25-27, 1999 (in press).

Findings Funding

From the exploration and development points of view it is important to recognize differences and similarities between geothermal areas and understand the underlying causes. The presently installed electrical generation capacities at CPGF and SSGF are 620 and 248 MW, respectively, making them

This work has been supported by the Assistance Secretary for Energy Efficiency and Renewable Energy, Office of Geothermal Technologies of the U.S. Department of Energy under Contract No. DE-AC03-76SF00098.

http://www-esd.lbl.gov 102


Earth Sciences Division

Energy Resources Program

Annual Report 1998-1999

Surveillance and Supervisory Control of Waterflood Research

Objectives

layers, poor displacement efficiency and reservoir damage due to cataThe objective of this project is to strophic hy d ro f ra c t u re ex t e n s i o n s . Contact: offer easy-to-use graphical software to Excessive wa t e r flooding, under an Tadeusz Patzek monitor waterflood performance and assumption that maximizing wa t e r (510) 643-5834, Patzek@patzek.berkeley.edu show ways of increasing oil producinjection will maximize oil production while limiting reservoir damage tion, leads to well-failure and causes from water injection.The software should be data-driven, but it reservoir damage. Over time, a constant injection rate at constant should have several process-dependent plug-in models of water pressure causes, albeit inadvertently, hydrofracture extensions injectivity, oil and gas productivity, waterflood response, spatial and reservoir damage. In order to operate the injection wells sucrelationships in injection and production data and their crosscessfully and prevent catastrophic hydrofracture extensions, correlations.We also should be able to perform simple inversions reservoir damage and well failures, we need to measure the of field data. Finally, the software should be intuitive, simply dynamics of hydrofracture growth in response to fluid injection. point-and-click and robust. We have used two different approaches for the characterization We have focused on waterflood in low-permeability rock with of injection hydrofractures. One is based on an inversion of the close well spacing, where fluid flow is linear and transient and injection data, while the other is based on periodic hydraulic all wells are hydrofractured. The injection fractures may grow, impedance testing of the injection wells. catastrophically at times, and production hydrofractures may Hydraulic impedance testing involves monitoring the transhrink. For each injector, the injection rate and pressure can be sient response of the injection well-hydrofracture system to a inverted into the hydrofracture area as a function of time. Our short-duration pressure pulse generated at the wellhead. The approach is object-oriented,making it relatively easy to switch to fracture size is then estimated from the pressure response. other types of waterflood or altogether different fluid injection Results processes.

Tadeusz W. Patzek and Asoke De

Approach

In this project, we have had access to complete production and injection data from several large waterflood projects in the South- and Middle-Belridge Diatomite and Lost Hills.These fields are located in the San Joaquin Valley, Calif. The field data were used to test the various features of Analyzer. When we choose a particular field for analysis, Analyzer displays a map of the field showing all its producers and injectors. A user can click on any given well and examine its production or injection history. In addition, Analyzer allows comparisons of the performances of several producers or injectors through an interactive plot option. A n a lyzer can also quantify i n t e ractions between a single injector and its neighboring producers. By cross correlating the injection and the production data, A n a lyzer identifies injector-producer linkage as water bre a kthrough at a producer caused by a gi ven injector. In addition, A n a lyzer lets us examine the domain of influence of injectors and producers. The influence diagram helps to identify improper

Supervisory control of waterflood requires continuous surveillance of the water injection–oil production relationships. Accordingly, we have developed an interactive software tool called Analyzer using the Matlab 10 programming environment.This software has been developed for waterfloods in lowpermeability reservoirs where all wells are hydrofractured. Analyzer lets us: • Examine the producer and injector well histories, develop maps of field productivity and injectivity, and quantify injector-producer interactions. • Model cumulative oil, water and gas production of individual wells, as well as the field as a whole. • Estimate the reservoir properties based on the primary production data. • Analyze the oil, water and gas responses to waterflood. Waterflood failures are caused by direct injector-producer coupling, water breakthrough in some

Figure 1. Relationship between the water injection and oil production in a waterflood.

103


ESD

Surveillance and Supervisory Control of Waterflood

well placements, and inefficiencies, if Significance of Findings any, of waterflood and oil production. One of the important issues is to We have developed Analyzer, an identify the influence of waterflood on interactive software tool for complete oil production. Analyzer provides varisurveillance of waterfloods. Our softous options for examining the injecware allows one to identify deficiencies t i o n - p roduction relationship. One of in water injection policies, imminent them is a three-dimensional animation injector-producer linkage and wa t e r of the field-wide water injection and oil b re a k t h rough at producers. Analyzer production profiles.We illustrate this in also provides an estimate of the Figure 1, which shows a snapshot in hydrofracture size. In addition, we have time of the field-wide water injectivity shown that an independent measure of and oil productivity for a diatomite the dynamics of hydrofracture growth waterflood. The peaks represent injeccan be obtained through periodic tion highs, whereas the shading reprehydraulic impedance testing at the sents oil production. Regions of low oil injection wells. It is expected that the production are marked in blue, while surveillance software, along with the high production regions are marked in hydraulic impedance testing methodolred. Note that maximum oil productivogy, will provide adequate inputs for a ity is in regions of low - t o - m o d e ra t e supervisory control system that strives water injection, while high injection is Figure 2. Estimation of injection hydrofracture growth associated with low oil productivity.This from the cumulative injection volume and injection pres- to optimize the performance of waterfloods. suggests that the injection policy for the sure data. field is sub-optimal and needs to be Related Publications improved. Successful operation of water injectors requires knowledge of Patzek, T.W., and A. De, Lossy transmission line model of the dynamics of injection hydrofracture growth. Analyzer estihydrofractured well dynamics,paper SPE 46195 presented at mates this growth through an inversion of injection data,as SPE 1998 Western Regional Meeting, Bakersfield, Calif., May shown in Figure 2. 10-13, 1998. Our research has shown that an injection well and its associated fractures can be modeled as a lossy transmission line netFunding work. The model parameters for the well are determined completely by the well geometry and the fluid properties. The fracThis work has been supported in part by the Assistant ture characteristics are determined either by matching the measSecretary for Fossil Energy under the Advanced Computational ured hydraulic impedance test response of a well with the simuTechnology Initiative of the U.S. Department of Energy under lated response or by model-based inversion of the transient Contract No. DE-AC03-76FS00098. Partial support has been proresponse data. An independent measure of the dynamics of vided by CalResources and Chevron Petroleum Technology hydrofracture growth is, therefore, obtainable through periodic Company, Inc., as gifts to U.C. Oil Consortium. hydraulic impedance testing of injection hydrofractures.

http://www-esd.lbl.gov 104


Earth Sciences Division

Energy Resources Program

Annual Report 1998-1999

Control of Fluid Injection Into a Low-Permeability Rock Research

Objectives

have been plugged into the controller design. Tadeusz W. Patzek About one-third of the wo r l d ’s We have also developed the capaand Dmitriy B. Silin crude oil and a comparable fraction of bility of inverting the injection presoil reserves in the United States are sure and rate into controller input. Contact: Tadeusz Patzek l o cked in low - p e rmeability ro ck s . The controller requires three basic (510) 643-5834, Low-permeability fractured oil reserinput para m e t e rs: the history of Patzek@patzek.berkeley.edu voirs, such as the Austin Chalk, the cumulative injection, the history of West Texas carbonates or the California diatomites hold tens of injection pressure and the history of hydrofracture growth. The first two parameters are normally collected and stored anyway, billions of barrels of oil. For example, in the diatomites, 10 years whereas the technique for the estimation of the last one is more of primary recovery yield only 2.5-6% of estimated original oil in complicated.We use inverse modeling to estimate the last paramplace. Primary production leads to reservoir compaction and eter. Such an approach may require some experimental calibrawell damage, and must be stopped. tion, but after that the fracture size will be provided as the conThe low-permeability diatomite reservoirs present a trementroller input at no additional cost. dous target for incremental recovery by water and steam injection. Steam can displace oil without contacting it directly, but it Results is very expensive, and its low density makes stable control of injection a very difficult task.Thus, we have focused our efforts Pressure propagation model—The Gordeyev-Entov self-simion the analysis, modeling and control of water injection into lowlar solution is an exact two-dimensional solution of the boundpermeability fractured rocks. ary-value problem for pressure propagation from a growing The purpose of this project is to design and implement a hy d ro f ra c t u re . Using Duhamel’s “smart” controller of water, steam or principle we have enhanced this CO2 injection into a low-permeability fra c t u red rock. Analysis of the reasons solution to include time-dependent of poor past perfo rmance of such coninjection pressure. trollers in critical situations led us to a As an example, we have modeled new design. Our controller is based on a portion of Section 33 in the South optimal control principles; it “underBelridge Diatomite Field. As we can stands” the dynamics of fluid injection observe from Figures 1 and 2, the and remains stable even during catapressure propagates almost perpenstrophic hydrofra c t u re extension. dicularly to the fracture face. The pressure changes very little at a disApproach tance of 150 ft from the fracture face and remains almost unperturbed C u rrent waterflood projects maineven after 10 ye a rs of injection tain a field-wide injection rate at a con(Figure 3). stant leve l . Thus, the growth of injecAccounting for the layered struction hydrofractures is inev i t able. Ve ry ture of the rock, we have compared low perm e ability rocks re q u i re spethe pressure propagation rates in difc i fic techniques for modeling the ferent layers (see Figure 4). One can propagation of pre s s u reaway from the observe that although the pressure hydrofracture, as well as boundary propagation in some layers is very conditions on the hydrofracture that slow, in others there is a real danger d i ffer from those used currently. of establishing links between the We propose to apply the Gordeevinjector and the neighboring proEntov self-similar solution for characducers (compare layers I and K.) t e rization of pre s s u re propagation. Since this solution is exact, it helps in Fracture size estimation—We comprehending the most important have developed an inversion procefeatures of the process. Moreover, it dure, which allows us to obtain one can be used to build stable numerical of the controller input-parameters schemes of high precision in spite of without additional measure m e n t s and therefore at no extra cost. singularities at the tips of the fracFigure 2. Pressure distribution after 730 days of injection. Comparison of the results of ture. The results of this modeling 105


ESD

Control of Fluid Injection into a Low-Permeability Rock

inversion and the results of other fracture estimation techniques will allow us to scale and tune up the model and the controller. For this purpose we have considered the dimensionless relative fracture size instead of absolute dimensions. In order to obtain the estimate, we solve a Volterra integral equation generated from a Carter-like mass-balance model. Figure 5 illustrates application of this procedure to real data.The top plot shows cumulative injection versus time.The second plot shows the injection pressure history versus time.The bottom plot shows the relative effective hydrofracture size estimate. Since the estimated parameter itself incorporates the fracture size along with the changes in the formation permeability, we call it “effective fracture area.” Figure 3. Formation pressure versus time.

Significance

of

Findings

Our analysis of pressure propagation in low-permeability rocks has provided insights crucial to the design and implementation of a “smart”controller of water, steam or CO2. Our inve rsion technique for estimating the effe c t i ve hydrofracture size allows obtaining an additional controller input parameter at no extra cost. Practical implementation of the results described here—combined with our Waterflood Analyzer software—will lead first to the creation of a cradle-to-grave automated or semi-automated system of control and management of waterflood projects.

Related

Publications

Patzek,T.W., and D.B. Silin, Control of fluid injection into a lowpermeability ro ck : 1. Hydro f ra c t u re growth, SPE/DOE Improved Oil Recovery Symposium,Tulsa, Okla., April 1922, 1998. Patzek,T.W., and D.B. Silin, Control of fluid injection hydrofract u re grow t h ” in Proceedings of the 9th International Scientific and Technical Conference “New Methods and Technologies in Petroleum Geology, Drilling and Reservoir Engineering,” Vol. II, pp. 311-317, Krakow, Poland, July 2-3, 1998.

Figure 4. Pressure in layers after 730 days of injection at X = 0.0.

Funding This work has been funded by the Assistant Secretary for Fossil Energy, Office of Gas and Petroleum Technology, of the U. S . Department of Energy under Contract No. DE-ACO376FS00098.

Figure 5. Estimation of the fracture size.

http://www-esd.lbl.gov 106


Earth Sciences Division

Research

Objectives

Energy Resources Program Dynamic Reservoir Characterization Through the Use of Surface Expression Data

Annual Report 1998-1999

to reliably estimate the distribution of reservoir volume changes to infer the The objective of this project is to flow geometry. Monitoring of the surdevelop a methodology to characterface expressions at or near the surDon Vasco and Kenzi Karasaki ize a reservoir though the use of surface, however, costs very little comface expression data such as the dispared to drilling a corresponding set Contact: placements and surface tilts induced of boreholes and installing pressure Don Vasco (510) 486-5206, dwvasco@lbl.gov by pumping or injection.An inversion sensors in them. This remote-sensing algorithm that jointly inverts such surapproach also provides independent face data and the pressure data is being developed.The results of data of the re s e rvoir dynamics that can be used to the inversion will be the map of the pressure changes in the confirm/refute the reservoir model based on the borehole data reservoir.This can be a cost-effective remote-sensing technology alone. for obtaining information regarding the reservoir flow geometry, A series of fluid injections were conducted in a geothermal which is a critical piece of information for successful reservoir reservoir at Hijiori, Japan, as part of the JAPEX-LBNL collaboradevelopment/management. tive research project. Ten high-precision tiltmeters (Pinnacle Technologies, Inc.) were installed in shallow boreholes near and Approach around injection well HDR-1, located at the southern edge of the Hijiori caldera. The granodiorite bedrock was encountered at a When fluid is produced from or injected into a reservoir, it depth of 1,450 m in HDR-1. The injection interval was from a causes volume changes in the reservoir, which in turn induce disdepth of 2,151 m (bottom of casing) to 2,205 m (bottom of the placements on the ground surface. If the reservoir is horizontal well). A series of four injections were conducted in November and homogeneous, the induced displacements will be distrib1998, during which the surface tilts were monitored. uted concentrically around the production/injection borehole. Results In the case of the surface tilts, tilt vectors will be radially convergent to or divergent from the borehole. If a preferential flow The inversion results of the four injections indicate volume path such as a fault zone exists in the reservoir, which is often change to the east or to the southeast. This agrees with indethe case in geothermal reservoirs, the flow will mostly occur pendent acoustic emission and pump test results.We discuss the along the fault. The distribution of the volume change and subresults from the stage-4 injection sequent surface displacements will conducted on Nov. 20, 1998, in be skewed. An inversion algorithm m o re detail. The duration wa s can then be used to estimate the approximately five hours . In total, distribution of the volume changes a p p rox i m a t e ly 110,000 liters of in the reservoir. For example, a verwater were injected during the six tical fault zone would produce a h o u rs of stages 3 and 4.There was a linear trough in the inverted image. clear tilt signal coinciding with the The larger the volume change at a initiation of the injection.The comparticular location, the more fluid plete set of tilt ve c t o rs , c o rre s p o nhas likely moved into or out of the ding to the first hour of injection, is location. The distribution of volshown in Figure 1a. There is a fair ume change is tightly coupled with bit of scatter in the tilt directions that of the reservoir flow properand magnitudes. H oweve r, the two ties: permeability and compressibils o u t h e a s t e rnmost tiltmeters display ity. A joint inversion of surface displacements and reservoir pressure consistent and strong tilting to the should satisfy both constraints. north. The peak signal, associated S u r face ex p ressions of such with tiltmeter 1, exceeds 0.1 microreservoir dynamics can be moniradians and lies well ab ove the tored by using high precision tiltinstrument noise. The inve rted vo lmeters, GPS (global positioning sysume increase in the granodiorite, tem), laser level gages and eve n based on the one-layer model with a 20 x 20 gri d , is offset to the east by SAR (synthetic apert u re ra d a r ) , Figure 1. (a) Comparison of observed and predicted tilt data overdepending on the magnitude of the laid on the Hijiori site map; (b) Calculated volume change map in m o re than 0.7 km from injection displacements. A large number of the shallow layer (0.5-1.0 km); (c) Calculated volume change map well HDR-1.The pattern of volume i n c rease is elongated in an eastmeasurement points are necessary in the deeper layer (1.0-2.0 km). 107


ESD

Dynamic Reservoir Characterization Through the Use of Surface Expression Data

west dire c t i o n .The location of the peak volume change coincides with the southern edge of the volume change associated with the s t age-3 injection eve n t . Also, the location of the volume change during stage-4 injection appears to be an eastwa rd extension of the volume change in stage 2. The east-west orientation agrees with acoustic emission information, which suggests east-west fl ow from HDR-1. In addition, p revious analysis of 1995 circ u l ation tests in HDR-1 and surrounding wells indicated an extension of fractures and water loss to the east. Because the stage-4 injection involved a much larger volume than the other injections and the data appeared to have a higher signal-to-noise ratio, we constructed a more detailed model of subsurface volume change. In particular, we attempted to fit a two-layer model with a deeper layer (1.0-2.0 km) and a shallower layer (0.5-1.0 km). Each layer consisted of a 15 x 15 grid of cells, in which each cell could undergo a distinct volume change. We conducted an inversion of the stage 4 data (Figure 1a) using this model parameterization.The result is shown in Figure 1b and 1c. In the deeper layer (1.0-2.0 km) we still observe the largest volume change to the southeast of HDR-1 as in our one-layer inversion. However, there is a secondary area of volume change to the west of HDR-1, elongated in the north-south direction. In the uppermost layer (0.5-1.0 km) the deep southeast body extends upward in depth. In addition, there is an arm of volume change extending from the western body to the east. Interestingly, this extension in part coincides with a river, which crosses the caldera in a roughly east-west direction. The river is thought to follow a zone of weakness, hypothesized to be a fault, within the caldera. It is an intriguing possibility that some fluid has migrated along this zone of weakness. However, more (and higher quality) data is required to better constrain the volume change. Note that

the two-layer model has predicted surface tilt which is very close to the observed data (Figure 1a) both in direction and magnitude.We conclude that some degree of shallow (0.5-1.0 km) volume change is needed to match the high displacement gradients observed in the data.

Significance

of

Findings

High-precision tiltmeters can be used to monitor the surface expression of reservoir dynamics.An inversion algorithm can be used to infer the reservoir flow geometry.The technique can be used to better manage/explore operating reservoirs. It can also be used to track the injected mass of CO2 during the planned deep geologic sequestration effort.

Related

Publications

Vasco, D.W., L. R. Johnson and N. E. Goldstein, Using surface displacement and strain observations to constrain deformation at depth, with an application to Long Valley caldera, CA, J. of Geophys. Res., 93, pp. 3232-3242, 1988. Vasco, D. W., K. Karasaki and L. Myer, Inversion of surface tilt caused by fluid migration, J. of Geotech. and Geoenv. Eng., 124, pp. 29-37, 1998. Vasco, D. W., K. Karasaki and C. Doughty, Using surface deformation to image reservoir dynamics, Geophysics (in press).

Funding This wo rk has been supported by JAPEX Geosciences Institute, Japan, through Contract No. BG98-071(00).

http://www-esd.lbl.gov 108


Earth Sciences Division

Energy Resources Program

Annual Report 1998-1999

Natural Geochemical Tracers For Injectate Fluids At Dixie Valley, Nevada Research

Objectives

B. Mack Kennedy, Cathy Janik* Richard Benoit+ and D.L. Shuster

Injection of spent production fluids Results * U.S. Geological Survey back into geothermal reservo i rs from +Oxbow Geothermal Almost immediately after injection which they are produced is widely Contact: commenced and continuing over the recognized as the single-most imporB. Mack Kennedy last 10 years the ch l o ride concentratant factor in maintaining reservo i r (510) 486-6451, bmkennedy@lbl.gov tion of the production fluids has pre s s u re and extending the pro d u cincreased, suggesting significant re t u rn of injectate. H oweve r, at tive lives of geothermal fields. Injectate is always colder than Dixie Va l l ey eve ry part of the reservoir that was ch e m i c a l ly samreservoir fluids, so eve n t u a l ly returning injectate will cool indipled prior to the onset of injection had a different ch l o ride convidual production wells and entire reservoirs.To predict the onset tent. P re - flash ch l o ride contents ranged from 250 to 400 ppm and of cooling, it is necessary to develop re l i able techniques to deters h owed an inve rse ch l o ri d e / e n t h a l py relationship: cooler wa t e rs mine the volume of injectate co-produced with the reservo i r h ave progressively higher ch l o ride contents.T h e re fo re, increasing fluid and the rate at which the injectate return increases. ch l o ride concentrations may not reflect injectate returns. A ny The injected fluids are the brines residual to steam production. infl ow of cooler indigenous water would also increase chloride Therefo re , the injectate is enri ched in ch l o ride and heavy isoconcentrations. topes of water (18O and D) and depleted in low-solubility nonc o n d e n s able gases relative to the production fluids.These natural During the flash process at Dixie Valley, the light water isotopes (16O and H) are pre fe rentially partitioned into the steam tra c e rshave been utilized at seve ral geothermal fields to monitor fraction, l e aving a residual brine enri ched in the heavy isotopes. injectate returns. H oweve r, their re l i ability as a quantitative measThe dominant trend in the isotopic history of the produced fluid ure of injectate re t u rn relies on seve ral assumptions, most notably b e t ween 1986 and 1998 is a pro gressive enri chment of about 2‰ that there is a single homogenized geothermal fluid into which in δ18O and 8‰ in δD following the start of injection.This is also the injectate is mixe d . Since most ge o t h e rmal systems are comconsistent with co-production of injectate and supports the interp rised of fluids from seve ral sources, this assumption is ra re ly p retation of the chloride concentration trends. realized. By comparing different natural tracers, the objective of The noble gases are part i c u l a r ly well suited for tracing injecthis study was to evaluate their quantitative reliability. tate because their solubilities are low, mass-dependent and are Approach k n own as a function of tempera t u re up to the critical point of water. At Dixie Valley, the concenInjection into the Dixie Va l l ey, tration of noble gases in the injecNevada, ge o t h e rmal re s e rvoir tate will be about 1-10% of that in began in September 1988, ab o u t the pre-flashed production fluid three months after the field comand the 1997-98 production fluids are signifi c a n t ly depleted in noble menced production. Since the gases, consistent with the addition onset of production the chemistry of gas-poor injectate. of the production and injectate fl uThe volume fractions of co-proids have been thoroughly docuduced injectate fluid (Vinj/Vtotal), mented, with analyses of quart e r ly calculated from all three independb rine samples (Benoit, 1992) and ent tra c e rs using the 1998 data set, intermittent water isotope analyses. In the past two years, analyses a re in ve ry good agre e m e n t , of low solubility noble gases have despite the assumptions and been obtained to gi ve a complete potential uncertainties.This implies evaluation of the ava i l able natural that ch l o ride concentrations at chemical tra c e rs in a geothermal Dixie Valley can be used with confield that has been well docufidence to assess the volume fra cmented. tion of injectate fluid in the proFigure 1. The volume fraction of injectate fluid calculated from the 36Ar abundances, assuming that the well data represent mixtures of 20°C air-saturated water and spent brine.

109


ESD

Natural Geochemical Tracers for Injectate Fluids at Dixie Valley

duction stre a m s . This is ve ry Significance of Fi n d i n g s important because chloride data exists for the entire 1988-98 proPerhaps the most challenging duction/injection period (Fi g u re aspect of geothermal reservoir 1), w h i ch can be used to model e n gi n e e ring is to be able to prethe change in the Vinj/Vtotal ratios (Fi g u re 2) with continued prodict thermal breakthroughâ&#x20AC;&#x201D;the duction and injection. onset of cooling in production The Section 33 re s e rvo i r wells resulting from cool fluids (dash-dot linein Fi g u re 2) has returning from injection wells. b e h aved in a ve ry different manAddressing this ch a l l e n ge ner than the Section 7 reservo i r requires, among other things, the (dotted line). The very rapid ri s e ability to quantitative ly assess and in the volume of injectate fluid model injectate returns. At Dixie Valley, good agreement between c o - p roduced with Section 33 the calculated volume fraction of reservoir fluids suggests a ve ry high degree of connectivity Figure 2, The yearly average volume ratios of co-produced injectate injectate in the production stream fluid calculated from the yearly average chloride data for the production b e t ween the injectors and the wells and injectate. Open symbols: southern Section 7 production wells. (Vinj/Vtotal), determined using three independent natural tracers Section 33 producing we l l s . Closed symbols: northern Section 33 wells. (Cl, noble gases and water isoH owever, in late 1992, the vo ltopes), suggests that chloride changes can be used as a viable ume of co-produced injectate began to decline.This may reflect a p roxy for monitoring injectate returns. C o n t i nued monitoring of change in the hydraulic connectivity induced by water tabl e the production and injectate fluid chloride, water isotopes, and drawdown isolating the reservoir f rom the pathway supplying the injectate. Alternatively, the early rise in chloride concentration in noble gas concentrations could lead to a better understanding of Section 33 resulted from the invasion of any of the other known the relative flux of the indigenous ge o t h e rmal fluid into the re s e rhigher chloride indigenous fluids. voir and provide additional constraints for future modeling efforts. The volume of co-produced injectate in Section 7 wells has R e la ted P u b li c a ti o n s i n c reased linearly with time at a rate of about 5-7% per year. It is also noteworthy that the ra n ge in chloride concentrations Kennedy, B.M., C. Janik, D. Benoit and D.L. Shuster, Natural ge ob e t ween Section 7 wells has diminished over the same period of chemical tracers for injectate fluids at Dixie Valley, in time, resulting in a rather homogeneous fluid which is about 61% P roceedings of the Twenty-Fourth Workshop on Geotherm a l injectate.The true volume fraction of injectate may be as high as Reservoir Engineering, S t a n fo rd Unive rsity, S t a n fo rd , 100%, because after 1994 the chloride contents in the Section 7 California, in press. producers have exceeded the initial injectate concentrations Benoit, R., A case history of injection through 1991 at Dixie Valley, (Fi g u re 1). Neva d a , Geothermal Resources Council, Trans., 16, pp. 611Assuming that the injectate fluid gra d u a l ly replaced the ori gi620, 1992. nal fluid by pushing the latter out of the fra c t u res and into the p roduction wells, an upper limit on the fra c t u re volume for the Funding system can be estimated from the total volume of fluid injected at Dixie Valley. Once the production fluids reached a chloride conThis work has been supported by the Assistant Secretary for tent equal to the ori ginal injectate concentration, the total volume Energy Efficiency and Renewable Energy, O ffice of Geotherm a l injected would approximate the local fra c t u re volume. This Technologies, of the U.S. D e p a rtment of Energy under Contract re q u i red about six ye a rs and took 206 x 109 lbs of injectate, c o rresponding to a fracture volume of about 0.12 km3. If it is furt h e r No. DE-AC03-76SF00098. assumed that the porosity provided by the fracture network is about 1%, this implies a reservoir volume of about 12 km3.

http://www-esd.lbl.gov 110


Earth Sciences Division

Research

Objectives

Energy Resources Program TVD Schemes for Phase-Front Propagation In Geothermal Reservoirs

Annual Report 1998-1999

uid saturation ahead of the reinjected fluid. In the problem, cold water (T Geothermal reservoir engineering 30Ë&#x161;C) is injected into a 200-m-long Curtis M. Oldenburg requires accurate numerical solution one-dimensional single-continuum and Karsten Pruess of advective-diffusive transport equaflow domain. The system is initially tions.The standard weighting scheme nearly single-phase liquid at the satuContact: Curtis Oldenburg used in numerical simu l a t o rs to rated vapor pressure (P0 = 85.93 bar) (510) 486-7419, cmoldenburg@lbl.gov at T0 = 300Ë&#x161;C. A schematic diagram of a p p roximate saturation or re l a t i ve the system and initial and boundary permeability at gridblock interfaces is conditions are shown in the upper part of Figure 1. The flow upstream weighting, a scheme that is very stable but also known begins by injection of cold water on the left at a rate of 0.4 kg/s to produce numerical dispersion. Numerical dispersion degrades and by production of mass at the same rate from the other side. the accuracy of strongly advective flow problems by artificially The production at the right-hand side lowers the pressure and smoothing sharp fronts. Artificial smoothing in reservoir engiinduces boiling. neering can lead to errors in prediction of thermal breakthrough and tracer arrival times, as well as nonphysical dilution. PhaseResults front propagation is typically self-sharpening due to the effects of relative permeability and is therefore less prone to numerical Results for the LTVD differencing scheme with 100 gridblocks dispersion. However, in phase-front propagation problems with are shown in the lower part of Figure 1, through profiles of liqboiling, phase saturation is strongly coupled to temperature. In uid saturation (dashed lines) and temperature (solid lines). such cases, numerical dispersion of the thermal energy equation Results for upstream weighting were very similar, except the can lead to inaccurate modeling of phase-front propagation. phase front was approximately Numerical dispersion can be 10 m farther advanced to the diminished by grid refinement, right-hand side relative to the but this can greatly increase LTVD result. The advancement execution times and computer of the upstre a m - we i g h t e d memory requirements. phase front relative to the LTVD Another approach for reducphase front occurs because ing numerical dispersion is to upstream weighting produces use higher-order differencing greater smearing of the temperschemes. The objective of this ature front, so that saturation research is to develop and test temperature at prevailing presmethods of efficiently decreassures is reached at a somewhat ing nu m e rical dispersion fo r larger distance from the injecimproved geothermal reservoir tion point.The phase transition simulation. We have impleto two-phase conditions theremented total variation diminfore also occurs farther from ishing (TVD) higher-order difthe injection point. ferencing schemes in TOUGH2. The differences between the Here we demonstrate applicaupstream and LTVD schemes tion of the Leonard TVD decrease with increased spatial (LTVD) scheme to a geotherresolution.We show in Figure 2 mal flow pro blem invo l v i n g a summary of the results of p h a s e - f ront pro p agation and boiling. phase-front location at a time of sixmonths as a function of the Approach number of gridblocks. Note in Figure 2 that the two schemes We compare upstre a m are converging slowly but that weighting to the LTVD scheme the LTVD scheme is closer to for a phase-front propagation the grid-conve rged result at Figure 1. Upper: Boundary and initial conditions for the one-dimensional pro blem where boiling pro- injection and production problem. Lower: Liquid saturation and tempera- much coarser resolution. In this gressively diminishes the liq- ture for the geothermal injection and production problem with LTVD. problem, the improved approxi-

111


ESD

TVD Schemes for Phase-Front Propagation in Geothermal Reservoirs

Figure 2. Phase front location vs. grid resolution for upstream weighting and LTVD schemes at t = six months.

mations in the thermal energy equation using LTVD give rise to the more accurate phase-front locations.

Significance

of

Oldenburg, C.M., and K. Pruess, Simulation of propagating fronts in geothermal reservoirs with the implicit Leonard total variation diminishing scheme, Geothermics, in press.

Findings Funding

The LTVD scheme has proven to be robust and efficient for complex multiphase and multicomponent nonisothermal flow problems relevant to geothermal reservoir engineering. In geothermal injection and production problems where boiling occurs, the location of the phase front may be very sensitive to the choice of weighting scheme. Our simulations show that the LTVD scheme is more accurate for the boiling front problem at a given discretization than upstream weighting, but that temperature and saturation front propagation are sensitive to grid resolution for both schemes.

Related

This work has been supported by the Assistant Secretary for Energy Efficiency and Renewable Energy, Geothermal Division, of the U.S. Department of Energy, under Contract No. DE-AC0376SF00098.

Publication

http://www-esd.lbl.gov 112


Earth Sciences Division

Research

Objectives

Energy Resources Program Fluctuations in Elastic Waves Due to Random Scattering From Inclusions

Annual Report 1998-1999

sible to obtain formulas for the average values of attenuation and a phase The problem of elastic wave propashift for the coherent part of the field. gation through heterogeneous media Valeri Korneev and Lane Johnson A differential equation for propagais encountered in nu m e rous discition through a layer of arbitrary thickContact: plines. It is particularly important in ness is derived.The equation gives the Valeri Korneev (510) 486-7214, vakorneev@lbl.gov the discipline of seismology because same solution to the problem as a stathe earth is heterogeneous on a broad tionary phase integral evaluation range of scales, so a variety of approaches to this problem have method. At low frequencies it gives a simple formula for elastic been developed. For a medium that is heterogeneous in only one moduli of a composite medium with inclusions of an arbitrary dimension the problem is essentially solved because exact solucontrast. In similar manner, the effect of fluctuation accumulations exist, although estimating and describing the heterogeneity tion is treated, where a differential equation for associated flucin realistic applications can introduce approximations. For media tuations is derived. Solutions of this equation allow us to obtain heterogeneous in 2- or 3-D, the problem is more profound, as it formulas to estimate standard deviations of point measurements is necessary to combine approximate solutions of the wave of attenuation and phase. For both parameters, the fluctuation equations with approximate descriptions of the media; underlevel is the same and equal to half the fluctuation of the square standing when a set of approximations is valid is not a simple amplitude of the field. matter. Because of the complexity of heterogeneity within the The analytical results of this study provide a method of estiearth, it is typically modeled as a random medium in which the mating the effects of scattering upon a plane wave propagating through a layer of randomly distributed spherical inclusions. effects of the heterogeneity upon elastic waves are treated in a Fo rmulas have been obtained for both the ave rage field and the statistical sense. Various approaches to the problem of wave statistical fluctuations about this ave rage.The ge n e ral results can propagation in heterogeneous media have been successful in be used for inclusions of arbitrary size and contrast and for all frecertain applications, but are accompanied by limitations that can quencies; approximations for small inclusions or low contrast raise questions about the validity and generality of the results. i n clusions have also been included. These analytical results have Examples include the lack of conversions between modes of been validated by comparing them with effective media estimates propagation, the failure to conserve energy and the inability to at low frequencies and with nu m e rical simulations over the entire handle strong contrasts in material properties. Our objective is f re q u ency range. In both cases the agreement is satisfactory. to present a method of handling wave propagation in 3-D hetResults show that the relatively simple expressions do an erogeneous media that avoids some of these limitations. acceptable job over a broad frequency range of describing phase Approach and attenuation effects upon a wave propagating through a region containing scatterers. The fluctuations calculated with We followed the basic approach to treat the wave propagaderived equations also serve as adequate bounds on the statistition process as a series of forward scattering problems. The cal uncertainty of the mean field. Because of the random fluctumedium is described as a random distribution of scatterers, ations, any attempt to reliably estimate the characteristics of the where size, material properties and density can vary. For the case scattering on the basis of a single seismogram may be a difficult task. Only the phase shift at low frequencies shows a reasonable where the scatterers are spherical and homogeneous, exact soluapproximation to the mean field.The non-physical negative valtions for the single scattering process are used; multiple scatterues of attenuation are common in the results for a single seising effects are only partly included. The use of exact scattering solutions allows a complete treatment of mode conversions mogram.This means that in most cases some sort of spatial averbetween P and S waves and arbitrary strong contrasts in material aging of the observational data will be necessary before stable properties. They also provide a starting point for deriving lowvalues of the mean phase and attenuation can be estimated. and high-frequency asymptotic solutions that can be compared Assuming that an averaging process has been applied to with other approximate solutions. Analytical results are comreduce the fluctuations to a acceptable fraction of the mean pared with numerical simulations for elastic waves propagating value, it is of interest to consider how data can be interpreted in through a medium containing random spherical scatterers. terms of material pro p e rties of the scattering medium. Measurements at high frequencies of velocity are only dependResults ent upon the background medium and thus can be used to estimate its elastic parameters. Measurements of velocity at low freTo derive basic analytical results we consider a thin slab quencies can be combined to provide constraints on the propextending to infinity perpendicular to the direction of incident erties of the inclusions.The first peak in the phase curve is a well wave propagation, where both the incidence of compressional defined feature and constraints on the material properties can be and shear waves are treated.Assuming randomness in spatial disobtained by fitting the analytical results to the data. tribution of scatterers and neglecting their interaction it is posThe effect of scattering on propagating plane elastic waves 113


ESD

Fluctuations in Elastic Waves Due to Random Scattering From Inclusions

Figure 1. Attenuation for incident P wave results for the average of all the seismograms. Results are shown as broken lines; heavy solid lines indicate analytical estimates for the mean values; light solid lines indicate analytical estimates for mean plus and minus one standard deviation.

Figure 2. Amplitude of the fluctuations as a function of the non-dimensional distance of propagation c*z/R where R is the radius of the scatterers, c their concentration and z the distance of propagation. Results are shown for various values of the non-dimensional frequency, Wm (m=p,s).

was simulated by using the exact scattering solution for a single elastic sphere (Korneev and Johnson, 1996). A thin slab of a scattering medium was simulated by distributing a large number of spherical inclusions having the same radius and with random spacing. Plane elastic P and S wave pulses containing a broad range of frequencies were propagated in the positive z direction. The calculated transmitted seismograms recorded in linear array of receivers were processed to compute the observed velocity and attenuation. In each case the calculations were performed on a single seismogram and for the average of all 20 seismograms (Figure 1).Also shown are the analytical estimates for the attenuation and the analytical estimates for the fluctuations.The analytical estimates are in reasonable agreement with the numerical results for both the mean and variance. For a single seismogram the statistical uncertainty is so large that only general trends can be identified, but in the case of the averaged seismograms a more quantitative evaluation can be performed. The agreement is slightly better for the attenuation than for the phase, which contains the additional complication of phase unwrapping. Also, due to the statistical fluctuations, it may be impractical to reliably estimate attenuation and phase from a single seismogram.Furthermore, the results exhibit relatively simple behavior at low frequencies. In this range the numerical and analytical results are in good agreement (even for a single seismogram) and the phase shows a linear dependence upon frequency. Regarding the behavior at high frequencies, although there are both long and short wavelength oscillations associated with the dimensions and properties of the scatterers, the trend in the attenuation is a constant non-zero value and the trend in the phase is a constant value of about zero. Due to the competing effects of conversion of energy from the coherent field and the resultant decay of the coherent field, the amplitude of the fluctuations in the transmitted field has a maximum at a particular propagation dis-

tance (Figure 2), and an expression for the maximum is given. It appears that the distance to the maximum decreases as the frequency increases, and the distance to the maximum is greater for S waves than for P waves at low frequencies; the situation is reversed at high frequencies. The phase shift of the coherent transmitted wave can be converted to an effective velocity by interpreting the phase as a time shift. At high frequencies, the velocities approach those of the background.

Significance

of

Findings

Rresults provide useful estimates for fluctuations of the main measured wave parametersâ&#x20AC;&#x201D;effective velocity and attenuation. These estimates can be used for receiving array design in order to achieve a desired accuracy and also as inversion data for evaluating a degree of heterogeneity of real elastic media.

Related

Publications

Korneev,V.A., and L.R. Johnson, Fluctuations in elastic waves due to random scattering from inclusions, Journ. Acoust. Soc. Of Am., submitted. Korneev,V.A., and L.R. Johnson, Scattering of P and S waves by a spheri c a l ly symmetric incl u s i o n , P u re and Ap p l i e d Geophysics, 147, pp. 675-718, 1996.

FUNDING This work has been supported by the Assistant Secretary for Fossil Energy, O ffice of Oil, Gas and Shale Technologies, Federal Energy Tech n o l o gyCenter of the U.S.D e p a rtment of Energy under Contract No.DE-AC03-76SF00098, and by the Defense Special Weapons Agency under Grant No. DSWA-01-97-1-0026.

http://www-esd.lbl.gov 114


Earth Sciences Division

Research

Objectives

Energy Resources Program Network Modeling Of Multiphase Flow Processes in Rock

Annual Report 1998-1999

Single-phase fluid flow in a network is analogous to current in an Porosity, absolute and relative perelectrical circuit.Therefore, Kirchoffâ&#x20AC;&#x2122;s Pingan Hunag meability and capillary pressure relacurrent law is applied to the mass baland Larry R. Myer tionships are very important in modance for each node in the network eling and calculation of fl ow in and Kirchoffâ&#x20AC;&#x2122;s second law is applied Contact: porous media. Considerable effort has to the pressure balance in each loop. Larry Myer (510) 486-6456, lrmyer@lbl.gov been devoted in the petroleum indusUsing a graph theory representation try to developing corre l a t i o n s of network properties (Yang et al., between these parameters.The considerable scatter and uncer1994), the relationship between pressure drop and flow rate in tainty in these correlations is, in large part, attributed to a lack of each element is found by Hardy cross-iteration. Permeability of understanding of the relationship between the microscopic pore the network is then derived using Darcyâ&#x20AC;&#x2122;s law. space geometry and macroscopic flow properties. Simulation of drainage is performed using the breadth first This research project has two objectives.The first is to invessearch graph theory algorithm to determine the interconnectivtigate the relationship between the microscopic pore geometry ity among the pores that are filled with the same phase.The priand the macroscopic flow properties and to advance underority queue search algorithm is implemented to simulate the standing of the physics of fluid flow in porous media. The secnonwetting invasion process. In the process of invasion, at a cerond is motivated by the expense and time required to obtain tain capillary pressure, the distribution of wetting phase and commercial relative permeability measurements needed to supnonwetting phase over the network is determined and the satuport development of petroleum reservoirs. A network model is ration for nonwetting and wetting phase is obtained. Having therefore developed to predict absolute and relative permeabildetermined the distribution of phases, subnetworks for each ity of reservoir rock from measurements of microscopic pore phase are defined. structure.These measurements constitute 2- or 3-D images of the Relative permeability can be calculated for each phase at each pore space, in combination with gravimetric porosity data and capillary pressure increment in the nonwetting phase invasion process. The effective permeability for each phase is calculated mercury porosimetry data. using the same procedure as used in the single-phase problem. Approach Relative permeability is finally obtained by taking the ratio of effective to absolute permeability. The absolute permeability is The network model is based on a regular cubic lattice. The the permeability of the entire network to one phase. nodes of the lattice represent pores while the connections Results (called branches) represent pore throats. Pores are assumed to be spherical in shape. Different pore throat cross-sectional A number of simulations were performed to evaluate the senshapes are incorporated. These are circular, triangular with sitivity of predicted behavior to the various input parameters. straight sides, and triangular with concave sides defined by arc These results provide insight into the sensitivity of absolute and segments.The pore throat cross-section is of constant size over relative permeability to various geometric properties of the pore the length of the connection between nodes.The ratio of throat space.All calculations were performed on a 16x16x16 grid. size to pore size is a variable. The significant effects of the The pore size at each node in va riation in pore and thro a t the network was assigned by a random number ge n e ra t o r sizes are demonstrated by simuassuming either a normal, loglations comparing netwo rk s normal or exponential distribuwith norm a l ly and ex p o n e ntion function. t i a l ly distributed pore sizes. The coordination number of Figure 1 compares absolute pera completely connected cubic meability as a function of mean lattice is six.The option of havp o re diameter while holding ing a lower coordination nump o rosity constant at 20%. ber is incorporated in the C i rcular cross-section thro a t s model. are assumed. It is seen that perThe lattice is further defined meability increases much more by grid spacing. The grid spacrapidly if the pores have a noring is determined once the mal distribution.The large numporosity and average pore size bers of small pores associated Figure 1. Absolute permeability as a function of mean pore size for two distribution types. are specified. with an exponential distri b u115


ESD

Network Modeling of Multiphase Flow Processes in Rock

tion produce flow restrictions which keep the permeability low even though the average size is increasing. Relative perm e ability curves corresponding to a normal, log normal and exponential distribution, all with mean pore diameter of 14.4 m, were generated. Curves for normal or log-normal distributions are nearly identical.The greatest differences are in the nonwetting phase relative-permeability curves with the exponential distribution yielding a curve of different shape.The larger numbers of small pores in the exponential distribution limit the number of connections available to the nonwetting phase at low capillary pressures. Finally, the effect of throat cross-sectional shape was investigated. In this case, the absolute permeability is not very sensitive to the cross-section shape, but the relative permeability is. Figure 2 compares relative permeability curves for circular and curved triangular throat cross-sections. A normal distribution of pore sizes was assumed while keeping other parameters constant for the simulations.At a given water saturation the nonwetting phase relative permeability for a network of curved triangles is greater than for circular throats. The opposite effect is seen in the wetting phase relative permeability. For the curved triangular crosssection, both phases can be present at the same time in the throats. The wetting phase captured in the corners contributes negligibly to the effective permeability but affects the water saturation.

strongly related to bulk porosity, the mean pore size, the ratio of throat to pore size and coordination number. Relative permeability is insensitive to these properties. Relative permeability was found to be sensitive to variability in pore sizes reflected in the type of distribution function and standard deviation. Relative permeability was also strongly affected by the cross-section shape assumed for throats.

Significance

This work has been supported by the Office of Science, L ab o ratory Tech n o l o gy Research Partnership Program and Offi c e of Basic Energy Sciences, and Assistant Secretary for Fossil Energy, O ffice of Natural Gas and Pe t roleum Tech n o l o gies of the U.S. Department of Energy under Contract No. DE-AC03-76SF00098.

of

Related

Publications

Huang, P., G.Yang, L.R. Myer and N.G.W. Cook, Simulation of capillary pressure hysteresis in drainage and imbibition process, in (K. Kim, ed.) Linking Science to Rock Engineering, Proc. 36th U.S. Rock Mech. Symp. and ISRM Intern. Symp., Elsevier, 1997. Huang, P., L.R. Myer and G.Yang, Network modeling of multiphase flow processes in ro ck , in Proceedings of the Ninth International Congress on Rock Mechanics, Paris, 1999. Yang, G., L.R. Myer and N.G.W. Cook, Application of graph theory to network models of porous or fractured media flows, in (H.J. Siriwardane and M.M. Zaman, eds) Proceedings of 8th Intern. Conf. on Computer Methods & Advances, vol. II, pp. 1299-1305, 1994.

Funding

Results

A 3-D pore scale network model has provided insight into the relationship between pore scale microstructure and macroscale flow properties. Simulations show that absolute permeability is

Figure 2. Relative permeability for different types of throat cross sections. Data points are calculated values.

http://www-esd.lbl.gov 116


Earth Sciences Division

Annual Report 1998-1999

ENVIRONMENTAL REMEDIATION TECHNOLOGY PROGRAM

TERRY C. HAZEN CONTACT: (510) 486-6223 TCHAZEN@LBL.GOV

The Env i ronmental Remediation Tech n o l o gy Pro gram (ERTP) conducts multi-disciplinary env i ronmental research on chara c t e rization, monitoring, modeling and remediation technologies.This research is directed primarily at Department of Energy and Department of Defense waste site problems. Since many of the contaminants or cl o s e ly related compounds invo l ved in these sites are also dominant at industrial waste sites, mu ch of this research is also applicable to pro blems faced by the pri vate sector and other gove rnment agencies.These projects are both basic and applied, and include eve rything from molecular studies to full-scale field deployments in all types of media (gas, wa t e r, s e d iment) in all types of env i ronments (wet lands to desert s ) . This year’s major customers have been: DOE Office of Environmental M a n agement, DOE Office of Science,Work for Others (DOD) and Work for Others (Industry/Other Gove rnment Agencies).

Demonstrations

and

vents in groundwater, soil vapor ex t raction of volatile organic contaminants (VOCs) and cryogenic drilling to aid core recovery in fra c t u red rock and coarse overburden. ERTP research on ferrofluids has shown that they have potential to effectively guide reactants and barrier liquids to contaminated target zones in the subsurface using electromagnetic forces.They also allow the use of geophysical methods to trace the movement and position of liquids injected into the subsurface.This research, combined with earlier work on viscous barriers for containment (patent pending), provides a great opportunity for containing, controlling and predicting solvent, metal and radionuclide contamination at waste sites—DOE’s greatest environmental cleanup problem. In partnership with the Savannah River Technology Center and Florida State University, ERTP is assisting the Institute for Ecology of Industrial Areas in Poland in conducting a full-scale demonstration of biopile (aeration) technologies for clean-up of a petroleum refinery waste lagoon. After nearly a year of operation the demonstration has shown that both passive and active aeration strategies can meet clean-up standards for polycyclic aromatic hydrocarbons in the soil and that active aeration cuts the remediation time in half.This demonstration has been so successful that the petroleum refinery is commercializing the process for other refineries and fuel stations in Poland. ERTP also supports EM 50 with technical expertise via the S t ra t e gic Lab o ra t o ry Council, the Oakland Site Technology

Deployment

E RTP supports DOE’s Office of Environmental Management in both the areas of Env i ronmental Restoration (EM 40) and the Office of Science and Technology (EM 50). Earth Sciences Division (ESD) scientists directly supervise characterization, remediation and monitoring, and provide regulatory and permitting support to LBNL’s Environment, Health and Safety Division for all enviro nmental pro blems on site. D u ring 1998 the program demonstrated and deployed tech n o l o gies for trench capture of ch l o rinated sol-

117


ESD

Environmental Remediation Technology Program

C o o rdination Gro u p , the multi-agency DNAPL Te ch n o l o gy Advisory Group, the SUBCON Vadose Zone Book and the Hanford Vadose-Groundwater-River Integrated Program.

for consultation re g a rding bioremediation and characterization. The A rmy Corps of Engi n e e rs re c e i ved an award for use of innova t i ve tech n o l o gy from the state of Nebraska for its use of a patented DOE tech n o l o gy to clean up a naval ammunition depot. ERTP also provides technical advisory support for the Bay A rea Defense Conve rsion Action Team (BA D C AT) and the California Environmental Business Council (CEBC).

Field and Laboratory Studies DOE’s Office of Science provides funding for several ERTP projects. The basic research projects funded in this area take advantage of the unique facilities at LBNL, such as the Advanced Light Source,where researchers look directly at the interaction between contaminants, water and minerals at the microscale. This year ERTP also had a new project funded in the Natural and Accelerated Bioremediation Research (NABIR) program. This project looks at mesoscale biotransformation dynamics as the basis for predicting core-scale reactive transport of chromium and uranium.

NABIR Program Office ERTP has continued to be the NABIR program office for the Office of Science. The NABIR program office maintains the NABIR web home page (www.lbl.gov/NABIR/), with links to investigators, program element managers, science team leaders, recent publications, annual meeting registration, calls for proposals, review documents and other web sites. The NABIR program office organized the first NABIR annual investigators’ meeting in Reston,Va., with more than 120 participants, and produced the NABIR Bioremediation Primer, available in hardcopy or electronically via the NABIR home page. The program office also assisted DOE Headquarters in producing a number of guidance documents for prospective field research centers. This effort enabled DOE-HQ to issue its request for applications for the NABIR field research center in January 1999.

Field Demonstrations for DOD Field tests at McClellan Air Force Base for the Department of Defense have shown that vadose zone monitoring systems (VZMS)— instrument packages consisting of tensiometers , suction lysimeters, gas samplers , pressure transducers and therm i s t o rs — permanently installed at multiple levels, provide better monitoring of waste site contaminants.The field tests at McClellan AFB allowed identification of significant shallow sources of contamination.

Partners and Funding

Demonstrations and Technical Assistance for Industry/Other Agencies

E RTP re c e i ves support from DOE pro grams in the Office of Science and the Office of Env i ronmental Manage m e n t . EM programs include the Environmental Management Science P ro gram, the Subsurface Contaminant Focus Area and the Characterization, Monitoring and Sensor Te ch n o l o gy C rosscutting Pro gram. The Office of Science funds the NABIR P ro gram Office at LBNL and the Office of Science, Office of B i o l o gical and Env i ronmental Research funds two env i ro n m e ntal remediation projects. S u p p o rt is also re c e i ved from DOD, Cal/EPA, other DOE Lab s , U n i ve rsity of Califo rnia at Berkeley and U.S. B u reau of Land Management. Other Wo rk for Others in 1998 incl u d e d : Concurrent Technologies, American Technologies, Radian International, Earthtech, Wo o dwa rd & C lyde and IT Corpora t i o n . Partners include UC Berke l ey, Stanfo rd Unive rsity, Westinghouse Savannah River Company, Idaho National Engi n e e ring and Env i ronmental Laboratory, Law rence Live rm o re National Laboratory, Pa c i fic Nort h we s t National Laboratory, U n i ve rsity of Nevada and Unive rsity of Illinois.

ERTP has re s e a rched selenium tra n s p o rt in the Grassland Water District for many years. Recent research has focused on better methods for compliance monitoring and management. The U.S. Bureau of Reclamation has sponsored this work in an effort to better manage selenium loading in the San Luis Drain.M i c robial studies this year have shown that selenium in-transit losses occur and that the fate of this selenium is bed sediments. Manipulation of the microbial ecology of the drain may stimulate the biore m e d i ation of selenium from the water, thereby reducing the mass loading of selenium to the San Joaquin Rive r. ERTP also provides technical consultation to pri vate industry and other gove rnment agencies on implementing LBNL and DOE patented tech n o l o gies at pri vate and gove rn m e n t - owned sites. P ri vate industry must have a license to the tech n o l o gy for use at p ri vate sites and all ERTP expenses are re i m b u rsed by the company or other agency. Seve ral contracts this year were executed

http://www-esd.lbl.gov 118


Earth Sciences Division

Research

Environmental Remediation Technology Program

Objectives

Experimental Studies Of Ferrofluids for Subsurface Applications

Annual Report 1998-1999

The forces ge n e rated in the fe rrofluid close to the magnet can be quite large, Fe rrofluids are stable colloidal susri valing those of gravity for fe rrofluidSharon E. Borglin, pensions of magnetic part i cles which water density diffe rences on the order George J. Moridis, behave as homogeneous pseudo-sinof 200 kg/m3. The magnetic force on Curtis M. Oldenburg fe rro fluid is stro n g ly dependent on gle-phase fluids. Fe rro fluids can flow and Alex Becker p roximity to the magnet because the through porous media such as natural Contact: gradient of the magnetic field drops sediments or fractured ro ck due to Sharon Borglin off quickly with distance. After being gravitational, pressure gradient, c a p i l(510) 486-7515, seborglin@lbl.gov mobilized by magnetic forces, fe rroflulary and magnetic forces. The objecids are held at steady state in a static and pre d i c t able arc-shaped tive of this research is to investigate the potential for the use of configuration around the magnet.While the path to the magnet is fe rrofluids in novel subsurface env i ronmental engi n e e ring and dependent on the system perm e ability and the initial injection l ab o ratory applications in the areas of (1) controlling the flow of configuration, the final ferrofluid distribution is predictable and is liquid in the subsurface and (2) detecting fe rrofluids using ge onot strongly controlled by these parameters. No significant fi l t raphysical methods. tion effects due to the movement of fe rro fluid through porous Approach media are observed. The movement of fe rrofluid can be seen in Figure 1. Fe rrofluid We carried out two broad lab o ratory-scale investigations using is injected into the top of a ve rtical cell filled with water-saturated commerc i a l ly ava i l able aqueous fe rrofluids and permanent magsand. The injection point is located 20 cm from the magnet. A nets in sand-filled containers and Hele-Shaw cells in a va riety of layer of lower perm e ability water-saturated sand is between the configurations. ferrofluid injection point and the magnet. The ferrofluid experiThe fi rst investigation was designed to demonstrate and quanences both gravitational and magnetic fo rces as it migrates downtify the movement of fe rrofluid in porous media. This included: wa rd .The layer of lower perm e ability sand causes the fe rrofluid (1) measurement of the ge n e ration of magnetic fo rces on a fe rto pool and move laterally before flowing dow n wa rd and accurofluid, (2) mobilization of a fe rrofluid in homogeneous and hetmulating around the magnet. As the fe rro fluid re a ches the vicinerogeneous porous media in response to an ex t e rnal magnetic ity of the magnet, the fe rrofluid fo rms an arc-shaped confi g u rafield created by a permanent magnet, (3) the static hold that pertion. Although the heterogeneities affect the fl ow path towa rd manent magnets can exe rt on ferthe magnet, the accumulation of rofluids, and (4) the fi l t ra t i o n the ferrofluid near the magnet is unaffected by the heterogeneity e ffects of the porous media on of the sand. The time elapsed in the fe rrofluid properties. this picture is four hours . The second investigation A typical result of the inve s t ifocused on the ability to detect gation of the subsurface mag n e t i c fe rrofluid in the subsurface using anomalies produced by fe rroflutraditional geophysical methods. ids is shown in Figure 2.This fi gThis was evaluated by measuri n g u re shows the measured and thethe anomalies produced by o retical signal for a spherical fe rmodel ferrofluid shapes using a rofluid anomaly with a diameter t h ree-axis fluxgate mag n e t o m eof 4 cm at three depths. Similar ter. The magnitude of the anomm e a s u rements we re also peralies was compared to model preformed on lab o ratory-scale anomdictions. These lab o ratory-scale alies in the shape of a thin disk, a anomalies can be upscaled to prerectangular horizontal slab , and a dict detection depths for fieldcylinder. The ex p e riments scale anomalies. showed ve ry good agre e m e n t Results b e t ween predictions and measu rements and therefore confi rm In general, the flow ex p e rithat the anomaly behavior can be ments demonstrated that mag- Figure 1. Flow of ferrofluid through a cell filled with a combination of predicted by geophysical models. netic forces cause fe rrofluid to water-saturated sands. The sand in the white band has a lower per- Upscaling the lab o ratory measmeability. The ferrofluid is the black liquid accumulating near the perflow over distances of about 0.25 manent magnet, which is located on the left side of the cell. Elapsed urements to a field scale determ on timescales of hours to days. time is four hours. mined that the maximum detec119


ESD

Experimental Studies of Ferrofluids for Subsurface Applications

tion depths of a 3 m3 spherical anomaly containing a 10% fe rrofl u i d - water mixture is on the o rder of 5 to 10 m.

monitored. This would provide a sensitive and non-invasive method of detecting the injection plume, or determination of high-permeability Significance pathways. Furthermore, since of Findings the ferrofluid consists of magnetite, water and a small From our experiments on amount of surfa c t a n t , the ferrofluid flow, we find the folenvironmental impacts of its lowing: (1) permanent maguse would be minimal. nets create a predictable presPossible applications of this s u re gradient in a mag n e t i c technique include use as a fluid that leads to fluid flow, s u b s u r face barrier ve rifica(2) ferrofluid can be held in tion tool by either magnetizporous media in predictable ing the barrier and using the final configurations which are Figure 2. Measured and predicted anomalies for a ferrofluid sphere of radius magnetic detectors to locate 4 cm at depths of 0.055 m, 0.08 m and 0.12 m. controlled solely by the magthe extent of the barrier, or netic field and are unaffected by the flow pathway, initial injecusing magnetized liquid to detect leaks or fractures in existing tion shape or permeability of the porous medium, and (3) fernon-magnetic barriers. rofluid exhibits only limited filtration effects in the advancing Limitations of ferrofluid-based tracer techniques include the front during flow through porous media. effect of ambient magnetic noise, which can obscure small anomThe potential of controlling fluid motion in porous media aly signals and significantly increase detection limits. This may without direct physical contact has potentially significant applibecome significant if large injection pumps are utilized in field cations. These include controlled emplacement of subsurface sites. Other interferences could come from natural magnetic barrier liquids or treatment chemicals, as well as emplacement of anomalies in the subsurface, which, however, can be minimized geophysically imageble liquids into particular zones for subseby background surveys of the area prior to injection. quent imaging. While the objective is the development of subRelated Publications surface environmental remediation solutions, our results to date on the movement of ferrofluid have been limited to laboratory Borglin, S.E., G.J. Moridis and C.M. Oldenburg, Experimental studlength scales up to approximately 0.25 m. As such, these results ies of magnetically-driven flow of ferrofluids in porous are particularly relevant to laboratory work where ferrofluids media, Berkeley Lab report LBNL-40126, 1998. may find immediate application in any situation where it is desirBorglin, S.E., G.J. Moridis and A. Becker, Magnetic detection of ferable to control the motion or final configuration of fluid in an rofluid injection zones, Berkeley Lab report LBNL-40127, experimental flow apparatus. 1998. From our experiments on ferrofluid detection, we find that the magnetite in the ferrofluid provides a signature sufficiently Funding strong for conventional magnetic detection methods. The magnetic anomaly created by ferrofluids could be used as a novel This work has been supported by the Laboratory Directed subsurface tracer technique. In field-based application of ferR e s e a rch and Development Program of Law rence Berkeley rofluids as tracers, magnetometer grids could be established National Laboratory under U.S. Department of Energy Contract around the injection site and changes in the magnetic field durNo. DE-AC03-76SF00098. ing injection of the ferrofluid-laden injection liquid could be

http://www-esd.lbl.gov 120


Earth Sciences Division

Research

Environmental Remediation Technology Program

Objectives

Laboratory Monitoring Of VOC Biodegradation in Unsaturated Fractured Rock

Annual Report 1998-1999

obtained from a drillcore from DOE’s Idaho National Engineering and S eve ral Department of Energy Environmental Lab o ra t o ry (INEEL) Jil Geller, Simon Davis, (DOE) sites contain volatile organic were used for these experiments.The Mark Conrad, Hoi-Ying Holman, contaminants (VOCs) within deep drillcore was from a site where VOC Jennie Hunter-Cevera fractured rock vadose zones, originatcontamination threatens the Snake and Karsten Pruess ing from improper handling and disRiver Aquifer. The basalt sample was Contact: posal of non-aqueous phase organic broken into 1-to-2-cm-sized ch i p s , Jil Geller liquids (NAPLs). The nature of fast, sterilized and packed into a glass col(510) 486-7313, jtgeller@lbl.gov channelized flow of liquids through umn. The experimental schematic is rock fractures, ponding at lithological discontinuities and longershown in Figure 1. Water-saturated CO2-free air with 100 ppm toluene was delivered to the column at a rate of approximately term diffusion into less accessible matrix rock presents tremenone pore volume/hour.The abundance and δ13C values of CO2 in dous characterization and remediation challenges. Biological 1-ml gas samples extracted with a syringe from the septum port degradation of VOCs may be a significant factor in containing the at the top of the column were analyzed using an automated prespread of these plumes. Field evidence for biodegradation of concentration system interfaced with a Micromass Isoprime isochlorinated solvents in fractured rock vadose zone settings exists tope ratio mass spectrometer (Isoprime).The δ13C values of the (Conrad et al., 1997), but major questions remain as to how to toluene in separate 1-ml gas samples were analyzed using a gas characterize the extent of naturally-occurring biological activity chromatography-combustion system interfaced to the Isoprime. and how to stimulate and monitor it for site remediation. During After monitoring the influent and effluent vapor phase for our three-year investigation of the fate of NAPLs in fractured changes under abiotic conditions, the rock chips were inocurock vadose zones, we have developed a suite of bench-scale lated with bacteria grown in a liquid medium made from basalt experimental techniques to ch a ra c t e rize multi-phase flow ro ck ex t ra c t . Subsequent experiments monitored biologi c a l (water, NAPL and air) and biotransformation of contaminants in activity due to the basalt rock extract and toluene for both flowthis environment, with a particular emphasis on the coupling of through and recycle configurations. fluid flow dynamics and biotransformation processes (Geller et The use of basalt chips ensures adequate surface area for bacal., 1998). In this summary, we present the results of a "geocosm" teria coverage and access to experiment to assess biological flowing fluid to produce measactivity on rock fractures as a urable changes of fluid phase function of environmental facconstituents. This is a prelimitors. nary step to the use of natuApproach rally-fractured cores, where the accessible surface area will be Geocosms are core-scale flow much lower.The column expercells, designed to incorpora t e iments complement ongoing, some aspects of natural condimicron-scale monitoring of biotions, using site samples, indigetransformations with IR specnous micro o rganisms (mixed troscopy on single basalt chips culture) and maintaining near incubated with organic vapor 100% relative humidity in the in a closed system (Holman et vapor phase.Biological activity is al., 1998). monitored by tracking changes Results in the abundance and stable carbon isotope ratios (δ13C) of carTable 1 presents the δ13C valbon dioxide (CO2) in the vapor phase.The value of δ13C of CO2 ues for the various compounds produced by biological activity of interest. Under abiotic condireflects the isotope ratio of the tions, the shift in δ 13C and in CO2 concentrations betwe e n carbon source utilized by the the column influent and efflubacteria, allowing a distinction ent we re insignificant, with between biological and other δ13C approx i m a t e ly equal to CO2 sources. Vesicular basalt rock samples -15‰. Figure 1. Schematic of geocosm assembly. and indigenous microorganisms Following inoculation, humid 121


ESD

Laboratory Monitoring of VOC Biodegradation in Unsaturated Fractured Rock

CO2-free air was again supplied Sample δ13C(‰)1 The geocosms can be used -12 to -15 to the column, without the CO2 in room air to assess the various environtoluene -24.6 to -25.4 addition of toluene, in order to mental factors that may limit monitor CO2 changes due to biological activity in the 1.δ13C(‰)=103x{[13C/12C]sampleutilization of the basalt rock vadose zone and to test [13C/12C]VPDB/[13C/12C]VPDB; VPDB is Vienna extract. Six days following inochypotheses of contaminant PeeDee Belemnite (standard reference) ulation, δ13C of CO2 dropped to tra n s fo rmation pro c e s s e s a low of -35‰ (significantly derived from field monitoring Table 1. Stable isotope ratio of carbon in various constituents. lower than δ13C of toluene; see with stable isotope ra t i o s . Table 1), then recovered to Given that contaminant "stor16‰ (near pre-inoculation valage" in fractured basalts (vesicular matrix) in the vadose ues) over the next 18 days. zone may be away from accesThirteen days following inocusible flow paths, our ultimate lation, CO2 c o n c e n t ra t i o n s reached a peak of 40 times preobjective is to identify factors inoculation levels, then that can stimulate biological declined to several-times preactivity in the presence of inoculation levels over the next contaminants and test meth13 days. The recovery in CO2 ods of delivering limiting subvalues, as well as carbon mass strates. While this study balance, s u g gested that the focuses on fractured basalt, organic substrate from the the approach can be applied ex t ract had been nearly to unconsolidated or partially exhausted. consolidated porous media. For delive ry of 100 ppm toluene for flow - t h rough condi- Figure 2. Changes in the relative abundance of CO 2 and in the stable car- R e l a t e d bon isotope ratio in the vapor phase of the geocosm for recycle configuraPublications tions, CO2 concentrations in the tion, exposed to an initial vapor phase concentration of 100 ppm toluene. effluent increased by a factor of Conrad, M.E., D.J. DePaolo, B.M. Kennedy and E.C. Miller, Carbon 1.5, while δ13C shifted from -15‰ in the influent to -21‰ in the effluent.To increase the residence time of the vapor phase with isotope evidence for degradation of mixed contaminants in the ro ck chips and CO2 changes, the column was shut in for four the vadose zone, Geol. Soc. Am., Abst. with Prog. 26, no. 6, days and the vapor phase circulated by means of the recycle line A186, 1997. s h own in Fi g u re 1. Figure 2 is a plot of the relative abundance of Geller, J.T., H-Y. Holman, G. Su,T-S. Liou, M. S. Conrad, K. Pruess and CO2, and the δ13C values during recirculation.CO2 concentrations J.C. Hunter-Ceve ra , Flow dynamics and potential fo r increased approximately six-fold, while δ13C values decreased by biodegradation of organic contaminants in fractured rock -4 to -9‰. On the day of the last measured value of CO2, the convadose zone, Berkeley Lab report LBNL-52587,1998, Journal c e n t ration of toluene was 30% of its initial va l u e .The sensitivity of of Contaminant Hydrology, submitted. δ13C values is indicated by the fact that drop in δ13C preceeds the Holman, H.-Y., D.L. Perry and J.C. Hunter-Ceve ra . Surfaceincrease in CO2 abundance due to the degradation of toluene. enhanced infra red ab s o r p t i o n - re f lectance (SEIRA) microspectroscopy for bacteria localization on geologic Significance of Findings material surfa c e s , J. M i c ro b i o l . Methods, 34, pp. 59-71, 1998. For vapor-phase delivery of toluene at near 100% relative Funding humidity, δ13C measurements of CO2 indicated that the mixed culture utilized toluene without the presence of flowing water. This work has been supported by the Office of Science, Office While the numbers of indigenous bacteria in the fractured-rock of Health and Env i ronmental Sciences, Biological and vadose zone in arid regions are low, other experiments have indiEnvironmental Research Program, of the U.S. Department of cated that their activity may be stimulated by the addition of Energy under Contract No. DE-AC03-76SF00098. water, nutrients and organic carbon.

http://www-esd.lbl.gov 122


Earth Sciences Division

Research

Environmental Remediation Technology Program

Objectives

Removal of Uranium (VI) From Contaminated Sediments By Surfactants

Annual Report 1998-1999

uranyl nitrate stock solution. Sorption of U(VI) was studied as functions of Given the significant health risks time and concentration of uranium Fred Gadelle, Jiamin Wan u ranium poses, nu m e rous Depart(10-7 to 10-4 M). The pH of the soil and Tetsu K. Tokunaga solution was adjusted with HNO3 and ment of Energy and mine tailing sites Contact: NaHCO3, and calcium concentration require decontamination. Promising Fred Gadelle varied from 0 to 100 mM. chemical treatments such as H2SO4 (510) 486-2226, fgadelee@lbl.gov and carbonate/bicarbonate (CB) sufIn the desorption step, a selected fer from serious limitations: strong matrix dissolution with the treatment solution (10 mL) was added to the previously conacid and reduced desorption efficiency with CB due to the taminated soil. Following the desorption period, the samples buffering capacity of soils. As an alternative to these chemical were centrifuged and the supernatant was collected, diluted treatments, surface active agents (surfactants) might be used to with deionized water and acidified for determination of the remediate metal-contaminated acidic soils. Surfactant aggregates amount of U(VI) removed from the soil.The treatments included or micelles bind metal ions (e.g., Cd2+, Pb2+, UO22+) in aqueous salt solutions, NTA, bicarbonate, H2SO 4, and surfactants. solutions. However, studies conducted in simple aqueous soluDesorption experiments with deionized water and artificial tions without the presence of contaminated soil are not directly groundwater were used as controls. Desorption times varied amenable for evaluating the effectiveness of surfactants in from 5 to 120 hours. Treatment concentrations ranged from 25 removing strongly sorbing ions from soils. The objective of this to 200 mM. The concentration of sorbed U(VI) ranged from research is to understand uranium sorption onto natural sediabout 10-9 to 10-6 mol/g. U(VI) desorption was also investigated for soil-to-liquid ratios of 0.1, 0.2 and 0.5 g/mL. The effect of ments and determine the efficiency of surfactants in desorbing increasing Ca2+ concentration was also determined. U(VI) from contaminated soils. Uranium was analyzed with inductively coupled plasma mass Approach spectrometry using Bi (25 ppb) as an internal standard. Iron was analyzed with inductively coupled plasma atomic emission specThe soil used in this study was collected at the midslope of troscopy. the Melton Branch Watershed on the Oak Ridge Reservation Results (Tennessee). The watershed has a shallow soil profile underlain by fractured saprolite. The soil texture is 25% clay, 35% silt and Uranium Sorption. The rate of sorption is initially very fast 40% sand. The clay fraction is composed of illite (40%), mixed(96.5% of the added uranium is sorbed in less than one hour). layer illite and smectite (40%), kaolinite (20%) and traces of This fast sorption step is followed by a slow reaction rate (i.e., quartz. Carbonates have been completely weathered and a large sorption levels off after quantity of amorphous Fe a p p rox i m a t e ly 48 hours). oxides (3%) is present as The fast sorption rate is coating. attributed to uranium sorpThe ex p e riments contion onto readily ava i l able sisted of two steps: (1) sorpexternal surface sites, while tion of U(VI) onto the Oak the slow reaction rate correRidge soil and (2) desorption sponds to sorption onto less from the soil. Here sorption accessible sites and is conis operationally equated with trolled by diffusion of urameasured removal of U(VI) from the aqueous phase and nium into the sorbing phase. may include influences from U(VI) sorption onto the Oak precipitation. Likewise, desRidge soil is strongly influorption may include influenced by the pH: one ences from dissolution. In observes an adsorption edge the sorption experiments, a at pH = 2-4, followed by a known amount of sediment maximum sorption at 4.8 < was combined with a backpH < 7.0. The pH dependground electro lyte (1 mM ence of the sorption process NaCl and 1 mM CaCl2 in indicates that protons, H+, deionized water) in cencompete with U(VI) fo r Figure 1. U(VI) and Fe release from the Oak Ridge soil as a function of chemical trifuge tubes. Uranium was treatment. [U(VI)] = 10-8 mol/g. [Treatment] = 60 mM except [NaCl] = 140 mM sorption sites (i.e., surface o added to the mixture from a and [CaCl2] = 100 mM. hydroxyl groups, -SOH); at 123


ESD

Removal of Uranium (VI) From Contaminated Sediments by Surfactants

low pH, H+ is the principal sorbing species, forming positively charged diprotonated sites (-SOH2+).As the pH increases, U(VI) ions displace H+ and bind to OH groups on the surface. U(VI) reaction with those surface hydroxyl groups is similar to the hydrolysis reaction observed in aqueous solution only. Sorption of U(VI)-CO32- complexes is responsible for a large degree of U(VI) removal from the aqueous solution near neutral pH. At higher pH, limited sorption of these complexes causes a desorption edge at pH > 8. Sorption experiments with Ca2+ suggest that a relatively minor fraction of U(VI) might be sorbed onto fixed negative charge sites and that Ca2+ could compete with U(VI) for these sites. The strong pH influence and the relatively minor effect of Ca2+ on sorption suggest that U(VI) is mostly sorbed onto amphoteric hydroxyl groups. The nature of the sorbed U(VI) species was not determined in this study. However, one would expect U(VI) to be strongly sorbed to the Oak Ridge soil, principally as inner-sphere complexes. Uranium Desorption. As shown in Figure 1, mild chemical treatment solutions such as NaCl and CaCl2 are ineffective at releasing U(VI) from the soil matrix. More aggressive treatment by H 2SO4 can desorb U(VI) efficiently. However, the strong matrix dissolution by the acid prevents its application as a useful agent for in-situ removal of U(VI). Bicarbonate can also desorb U(VI); h owever, the desorption efficiency is seve rely reduced by the buffering capacity of the acidic soil and by the presence of calcium. As an alternative to these treatment solutions, two anionic surfactants, AOK and T77, were found most suitable for U(VI) removal from the contaminated soil. AOK is a sodium C14-16 olefin sulfonate and T77 is sodium oleyl n-methyl taurate. The most likely mechanisms responsible for U(VI) desorption using the surfactant approach include cation exchange and, to a lesser extent, dissolution of the soil matrix.The ion-exchange reaction takes place in the electric double layer (diffuse and Stern layers) surrounding the micelles. During ion exchange, U(VI) cations replace the counterions Na+. M o n ovalent cations such as UO2(OH)+ are expected to be "loosely" bound to the micelles, while the divalent cation UO22+ is expected to attach strongly to the micellar surface and is specifically located in the Stern layer. Evidence of some matrix dissolution is illustrated by the

amount of Fe released during the desorption experiments: 38 and 28 mg/L for AOK and T77, respectively.These surfactants are very efficient solubilizing agents at low uranium concentrations: about 100% U(VI) removal for [U(VI)]o, sorbed = 10-9 mol/g. At greater uranium concentrations, the desorption efficiency of the surfactant solutions increases with an increase in surfactant concentration and reaches a plateau of 75-80%. Limitations associated with the surfactant treatments were also identified. Surfactant sorption onto the soil and greater affinity between U(VI) and the soil matrix reduce the efficiency of the treatment solutions at large soil-to-liquid ra t i o s . Due to the cationexchange nature of the desorption process, competitive cations such as calcium reduce the amount of U(VI) removed from the soil.

Significance

of

Findings

Results obtained in this study suggest that the use of surfactants such as AOK and T77 to remove metals/radionuclides from contaminated soils can potentially be developed as an ex-situ or in-situ remediation method. Acidic soils with lower oxide and clay contents would be the preferred candidates for such an insitu surfactant treatment. While competition with diva l e n t cations such as Ca2+ would reduce the treatment efficiency, the non-specificity of the desorption process also suggests that these surfactants could be used to remove other hazardous or radioactive cations (e.g., Cd2+, Pb2+, Co2+, Sr2+, etc.) from contaminated soils.

Related

Publications

Gadelle, F., J. Wan and T.K. Tokunaga, Removal of uranium (VI) from contaminated sediments by surfactants, Environ. Sci. Technol., submitted.

Funding This work has been supported by the Assistant Secretary of Environmental Manage m e n t , Env i ronmental Manage m e n t Science Program, of the U.S. Department of Energy under Contract No. DE-AC03-76SF00098.

http://www-esd.lbl.gov 124


Earth Sciences Division

Research

Environmental Remediation Technology Program

Objectives

Multi-Scale Investigations Of Liquid Flow in the Vadose Zone of Fractured Basalt

Annual Report 1998-1999

parts (fra c t u re zones, vesicular lenses, soil, m a s s i ve basalt, ru bbl e The objectives of this investigation zone) of a single basalt flow finger. Boris Faybishenko, are to introduce a new approach to Large-scale components include mulPaul A. Witherspoon, the multi-scale characterization of tiple basalt flows and their surroundChristine Doughty, T.R. Wood*, flow and tra n s p o rt in the fractured ing network of rubble zones and sedR.K. Podgorney* and Jil T. Geller basalt of the vadose zone and to imentary interbeds. develop phy s i c a l ly based conceptual *Parsons Infrastructure and Technologies, Inc., Several laboratory and field infilmodels on a hierarchy of scales. tration tests conducted in fractured Idaho Falls, ID. basalt have shown that a typical feaContact: Approach ture of flow in fractured rocks is Boris Faybishenko (510) 486-4852, bfayb@lbl.gov channeling that occurs at all scales, We illustrate this new approach i n cluding individual fractures, the using results of field and lab o ratory investigations conducted in intra-basalt fracture network and the inter-basalt rubble-zone netthe fra c t u red basalt near Idaho National Engi n e e ring and work. However, the instrumentation used in lab and small-scale Environmental Laboratory (INEEL). The results were obtained field conditions (in particular, to record dripping phenomena) is from field tests carried out at three different sites: (1) small-scale not practical for use at larger scales. Field measurements of flow infiltration tests (ponded area 0.5 m2) conducted at the Hell's Half characteristics in fractured rocks using single probes (such as Acre site near Shelly, Idaho, (2) intermediate-scale infi l t ration tests tensiometers) are uncertain, because the locations of the probes (ponded area 56 m2) conducted at the Box Canyon site near Arco, in relation to the flow paths are not precisely known and these Idaho, and (3) a large- scale infiltration test conducted near the probes average fracture and matrix hydraulic characteristics. INEEL Radioactive Waste Management Complex (RWMC; ponded An important feature of flow in the basalt vadose zone is that area ~26,000 m2). In addition, laboratory investigations included the hydraulic system includes both unsaturated and saturated measurements on fractured basalt cores and small-scale dripping rocks. During flooding at the surface, the saturated zones have a experiments in fra c t u re models. limited and local extent within flow channels in the fractures and vesicular zones. Single probes crossing the saturated fracResults tures also intersect the matrix; hence, the probes measure an averaged water pressure in the fracture-matrix system. However, We find that, at each scale it is difficult to separate the of inve s t i g a t i o n , diffe rent c o n t ribution of the fra cmodels for flow phenomena t u re from that of the must be used to explain the matrix. Using availabl e observed behav i o r. These monitoring techniques models can be used to under field conditions, we describe the flow processes cannot perfect our knowlon different scales, with no edge of initial conditions; apparent scaling principles we can only measure evident. To ch a racterize approximate values of diffl ow phenomena in fra cfe rent para m e t e rs (pre stured basalt, we recommend sure, m o i s t u re content, that investigations be cart e m p e ra t u re, concentraried out in the fo l l ow i n g tions) or determine the hierarchy: elemental, smallranges of these parameters. scale, intermediate-scale and Fluid flow in fractured large-scale (Figure 1).An elebasalt of the vadose zone mental component is a sincan be considered a nonlinear dynamic process in gle fracture or a block of which the behavior, both h o m o geneous poro u s t e m p o ra l ly and spatially, medium. Small-scale compom ay be chaotic. The nents include one or a few chaotic nature of fl ow fractures and the surroundresults from nonlinear ing matrix. Intermediateprocesses and the strong scale components include a fracture network and other Figure 1. Hierarchy of scales of hydrogeological components in fractured basalts. spatial and temporal varia125


ESD

Multi-Scale Investigations of Liquid Flow in the Vadose Zone of Fractured Basalt

tions in moisture content, hydraulic conductivity and fracture connectivity. As a result, in fractured basalt (that is, a nonlinear system) small variations in flow parameters may lead to significant variations in predicted results.The dissipative nature of the system implies that its phase-space volume decreases with time, leading to the formation of strange attractors characterizing the range in which the flow parameters are expected to change.The dynamics of such systems are sensitive to the initial conditions. The response of such systems may include a stochastic component that it is not necessarily a dominant factor on the system behavior. If the stochastic component is not dominant, then a stochastic analysis will not provide useful information. We also find that under field conditions with a limited number of single-probe measurements, we can detect neither the spatial nor temporal chaotic variations of the fl ow para m e t e rs. Therefore,we must use conventional (i.e., nonchaotic) stochastic or deterministic methods to describe fl ow and tra n s p o rt processes. If the stochastic component is not a dominant factor, then a stochastic analysis will provide incorrect answers and should be replaced by a chaotic analysis. If the stochastic component is significant (or if the phase-space dimension is so large that the dynamics look stochastic), then a stochastic analysis is the most appropriate tool to use. Because the system exhibits sensitivity to initial conditions, we can predict only the range in which the flow rate is expected to change, but not the exact flow rates. Where elemental and small-scale components are involved, the flow data can be analyzed using methods of nonlinear dynamics.Where intermediate and large-scale components are involved,a combination of deterministic and stochastic methods of flow analysis can be used.

Significance

of

account the fracture-matrix interaction. The practical application of the intermediate-scale investigation is understanding the flow processes underneath single tanks like those at the RWMC in Idaho and at Hanford. The practical application of the large-scale investigation is to understand the flow processes and develop models to study the distributed flow and contaminant transport problems that exist under the RWMC and elsewhere at INEEL or in the tank farms at Hanford.

Related

Publications

Faybishenko, B., Evidence of chaotic behavior in flow through f ra c t u red rocks, and how we might use chaos theory in fra ctured ro ck hydro ge o l o gy, in Proceedings of the Intern a t i o n a l Symposium "Dynamics of Fluids in Fra c t u red Rocks: Concepts and Recent A d vances," Berke l ey Lab re p o rt LBNL42718, 1999. Faybishenko, B., C. Doughty, M. Steiger, J.C.S. Long, T. Wood, J. Jacobsen, J. L o re and P. Zawislanski, Conceptual model of the geometry and physics of water fl ow in a fra c t u red basalt vadose zone: B ox Canyon Site, Idaho,Water Resour. Res., s u bmitted. Finsterle, S., and B.Faybishenko,What does a tensiometer measure in fra c t u red ro cks? Berkeley Lab re p o rt LBNL-41454, 1998. Podgorney, R.,T.R. Wood and T. Stoops, 1997 Outcrop infi l t ration experiments, Data Summary Report , 1997. Sposito,G., and S.W. Weeks, Dynamical systems theory and fluid fl ow in subsurface zones, in Proceedings of 1997 Fall AGU Meeting , San Francisco, p. F247, 1997. Weeks, S.W., and G. Sposito, Mixing and stretching efficiency in s t e a dy and unsteady gro u n dwater flows,Water Resour. Res., 34, pp. 3315-3322, 1998. Wood,T.R.,and G.T.Norrell,I n t e grated large-scale aquifer pumping and infi l t ration tests, G ro u n dwater Pa t h ways OU 7-06, Summary Report , INEEL-96/0256, L o ckheed Martin Idaho Tech n o l o gies Company, Idaho, 1996.

Findings

It has become apparent that monitoring technologies, characterization methods and prediction procedures developed for porous media are not sufficient to treat fractured basalt. The results of studies on the elemental level can be used to better understand water dripping from a fracture under field conditions in boreholes, tunnels, caves and other underground openings. However, laboratory studies of fractured rocks have certain limitations, because we cannot determine how to combine the individual processes for characterizing the total system. The most important practical application of small-scale investigations is the study of physics and models of flow taking into

Funding This work has been supported by the Office of Environmental Management, Office of Science and Technology, Characterization, Monitoring and Sensor Technology Crosscutting Program, and the Environmental Management Science Program of the U.S Department of Energy under Contract No. DE-AC03-76SF00098.

http://www-esd.lbl.gov 126


Earth Sciences Division

Research

Environmental Remediation Technology Program

Objectives

Poland Petroleum Refinery Sludge Lagoon Biopile Demonstration

Annual Report 1998-1999

acidic, highly weathered petroleum sludges.This waste has been deposited The U.S. Department of Energy and into three open, unlined pro c e s s Terry C. Hazen waste lagoons, 3 m deep, now cove rthe Polish Institute for Ecology of ing 3.8 hectares. Initial analysis indiIndustrial Areas have been coopera tContact: ing in field studies of environmental cated that the sludges were composed Terry Hazen (510) 486-6223, tchazen@lbl.gov remediation since 1995. One of the mainly of high-molecular-weight parafmajor focuses of this program has finic and polycyclic aromatic hy d robeen the demonstration of bioremediation techniques to clean carbons. The ove rall objective of this full-scale demonstration up soil and sediment associated with a waste lagoon at an operp roject is to test and evaluate a combination of U.S.- and Polishating petroleum refinery in southern Poland. After a thorough developed remediation tech n o l o gies and methodologies. ch a ra c t e rization and tre a t ability and risk assessment studies, a Specifically, the goal of the demonstration is to reduce the env iremediation system was designed that took adva n t age of local ronmental risk from PAH compounds in soil and to provide a m a t e rials to minimize costs and maximize treatment efficiency. green zone (grassy area) adjacent to the site boundary. Passive and active gas injection we re compared and contrasted as A 0.3 hectare site, the smallest of the waste lagoons, wa s a bioremediation technique selected for a modified aerofor low pH soil contaminated bic biopile demonstration. with poly c y clic aromatic Approx i m a t e ly 3,300 m3 of contaminated soil (mean TPH hydrocarbons (PAHs). c o n c e n t ration of 30,000 Approach ppm) was targeted for tre a tment. The biopile wa s Since 1995, DOE has divided into two sections; an worked with the Institute of area of approximately 1,610 E c o l o gy of Industrial Areas m2 passively aerated using (IETU) in Poland to develop Baroballsâ&#x201E;˘ and an area of and demonstrate mu t u a l ly a p p rox i m a t e ly 1,390 m2 actively aerated via direct air beneficial,cost-effective env iinjection from the bottom of ronmental remediation techthe pile. Use of both passive nologies. The Czech ow i c e and active aeration methods Oil Refinery (CZOR), our a l l ows for an accurate assessindustrial partner for this ment of cost and effi c i e n c y, project, was chosen because Figure 1. Polish refinery sludge lagoon before remediation. with the most appro p ri a t e of their foresight and comdesign to be deployed for mitment to the use of new f u t u re lagoon remediation at approaches for env i ro n m e nthe refinery. tal restoration. This pro gram is a precedent for Poland in Results which a portion of the funds necessary to complete the The innova t i ve biopile p roject were provided by the design deployed used a comcompany re s p o n s i ble for the bination of passive and active p ro blem. The Czech ow i c e aeration and injection of Oil Refinery,located in southnutrients to incre a s e e rn Poland, was named by biodegradation of the very PIOS (State Env i ro n m e n t a l acidic soil containing high P rotection Inspectorate of PAH concentrations. Poland) as one of the top 80 S i multaneous lab studies most polluted sites in Poland. using soil columns were used Nearly a century of continuto optimize treatment techous use of a sulfuric acidniques and ve rify field obserbased oil re fi ning method by vations under more conCZOR has produced an estiFigure 2. Polish refinery sludge lagoon biopile during remediation. t rolled conditions. This fullmated 120 thousand tons of 127


ESD

Poland Petroleum Refinery Sludge Lagoon Biopile Demonstration

scale demonstration showed Campaign Average Passive Active and its part n e rs provides the that with minimal cost the basis for international techOC-1 80 44 119 total mass of petro l e u m n o l o gy transfer of innova t i ve OC-2 88 82 94 hy d rocarbons could be remediation tech n o l o gi e s OC-3 < 33 33 0 reduced by more than 75% that can be applied to DOE (82 metric tons) in only nine sites as well as in Poland and Table 1. Biodegradation rate (mg Petroleum/kg soil/day). months. During this time the other sites worldwide. most toxic compounds were reduced to levels acceptable for Related Publications multi-use resource activities.Though a va riety of biodegradation monitoring methods were used, in-situ respiration and dehy d roUlfig, K., G. Plaza,T.C. Hazen, C.B. Fliermans, M.M. Franck and K.H. genase activity were found to be best correlated with rates of Lombard, B i o remediation tre a t ability and feasibility studies at biodegradation in the biopile. In addition, it was found that pasa Polish petroleum refinery, in Proceedings, Warsaw96, sive aeration alone could re a ch the same end point as the active F l o rida State Unive rsity Press, 1997. a e rationâ&#x20AC;&#x201D;it just took longer (Table 1).The rates of biodegradation Lombard, K.H., T.C. Hazen, A. Wo rs z t y n owicz and B. Jagosz, were comparable to other prepared bed studies done using Bioremediation techniques for the cl e a nup of a petroleum p e t roleum contaminated soil; i.e., about 60 mg/kg/day. H oweve r, waste lago o n , In Situ and On-Situ Bioremediation: Vol. 5, gi ven that this material was highly weathered and ve ry acidic, p.467, 1997. these rates are quite high. Much of this increase can pro b ably be a t t ributed to the sawdust added as a bulking agent and the active Hazen, T.C.,Controlled phosphate-enhanced bioremediation a e ration process. tested, EPA Tech Trends, 25:3-5, 1997. Hazen, T.C., A. Ti e n , K.H. L o m b a rd , D.J. Altman, and A. Significance of Findings Worsztynowicz, Czechowice Oil Refinery bioremediation demonstration test plan, WSRC-MS-97-214, Westinghouse The finding that in-situ re s p i ration and dehy d rogenase measSavannah River Company,Aiken, SC, DOE-NITS, 1997. u rements more accurately reflect the biodegradation rate sugFunding gests that these direct measurements can be used to provide realtime control of biopile operation to maximize biodegradation LBNL work on this project has been supported by the rates under a va riety of conditions.The cost savings from passive Westinghouse Savannah River Company under U.S. Department a e ration may gi ve it a significant adva n t age over active aeration of Energy Contract No. DE-AC09-89R180035.The Savannah River when clean-up time is not a primary consideration. Technology Center, Florida State University, and the Polish The remediation strategies that have been applied at the CZOR Institute for Ecology of Industrial areas were all collab o ra t o rs on waste lagoon were designed, m a n aged and implemented under this project. Funding was also provided in part from the Polish the direction of the Westinghouse Savannah River Company for National Science Fo u n d a t i o n , National EcoFund and the DOE, the Institute for Ecology of Industrial Area (IETU) and Czech owice Oil Refi n e ry. F l o rida State University. This collab o ration between IETU, DOE

http://www-esd.lbl.gov 128


Earth Sciences Division

Research

Environmental Remediation Technology Program

Objectives

Aerobic Bioremediation Of a Municipal Solid Waste Landfill

Annual Report 1998-1999

Over the past seve ral ye a rs we have successfully demonstrated and treated The purpose of this aerobic landfi l l gro u n dwater in-situ at two unlined Terry C. Hazen bioreactor pilot study was to demonsolid waste landfills at DOE’s Sava n n a h strate the use of air injection and River Site.We used horizontal wells to Contact: leachate recirculation in promoting inject air and gaseous nutrients Terry Hazen (510) 486-6223, tchazen@lbl.gov a e robic biodegradation within the (methane, n i t rous oxide and tri - e t hyl waste mass and remediation of conphosphate) into the gro u n dwater to taminated leachate in a solid waste landfi l l .The processes include b i o remediate ch l o rinated solvents and reduce metals leaching in (1) air injection into the waste mass, (2) re c i rculation of leachate both the gro u n dwater and the vadose zone. For the past two and (3) addition of gaseous nu t rients into the waste mass.These years, A m e rican Technologies, Inc. (ATI), in cooperation with the author, now at LBNL, has demonstrated that it can safely and ecocomponents are considered individually and in combination to n o m i c a l ly conve rt a conventional lined and capped landfill to aerd e t e rmine their technical, economical and practical aspects of obic conditions utilizing an aerobic landfill bioreactor (ALB).T h ey their use in the study. Technical support was provided to h ave shown that aerobically degrading MSW within a landfill sigA m e rican Technologies, Inc., to test and implement this tech n o lnificantly increases the rate of waste decomposition and settleo gy at a Georgia county landfi l l . ment, d e c reases the production of methane gas, reduces the level Approach of toxic organics in the leachate and signifi c a n t ly reduces the vo lume of landfill leach a t e .The ALB eliminates the source of gro u n dConventional municipal solid waste (MSW) landfills worldwater contamination caused by solid waste landfills, and comwide operate under anaerobic (no ox y gen) conditions in which bined with the pumping and treating of the gro u n dwater as a supslow stabilization of the waste mass occurs, producing methane plemental source of landfill moisture addition, p rovides for the ingas and toxic leachate over long periods of time. The current situ treatment of gro u n dwater.A secondary effect of this strategy designs of most modern landfills utilize clay caps, liners and colat most sites is to stabilize metals in the gro u n dwater.This is due to increases in ambient redox potential and pH caused by the aerlection systems to contain the waste (“dry tomb”) and prevent obic or micro a e rophillic env i ronment created by air sparging and the spread of toxic constituents that could find their way into a biodegradation of high-oxygen-demand organics and reducing landfill. Older, unlined landfills pose a constant threat to the env iagents. ronment. Even with the newer landfill designs meeting the U. S. At the Columbia County Baker Place Landfill, a 100-acre subEnv i ronmental Protection Agency (EPA) requirements (RCRA title-D municpal landfill near Govetown, Ga., we connected Subtitle-D), the caps and liners ultimately fail, potentially releasing blowers to the cleanout ports of the leachate collection system methane gas and leachate to the env i ronment. to provide air and recirculated leachate via drip tubes through Although the dry-tomb approach is an attempt at reducing the temporary cover.An initial pressure test was followed by the t oxic releases from a landfi l l , this appro a ch is a temporary soluplacement of vadose zone piezometers throughout the test cell tion. A c c o rding to the EPA, “liner and leachate collection [systo determine changes in soil gas composition. Soil gas and tems] ultimately fail due to natural decomposition…”(EPA, 1988). The EPA recognizes leachate we re anathat “once the unit is lyzed weekly for 24 closed, the bottom months. See Figure 1 l ayer of the landfill for the design plan. will deteri o rate ove r Results time and consequently, will not preInitial air injection vent leachate transtests in the leachate port out of the collection system of unit”(40 CFR 258). As the landfill showe d a result, leachate colthat air penetrated the lection systems and e n t i re landfill via the impermeable caps do gravel bed in the botnot decrease the ri s k tom of the landfi l l . that toxic conIndeed pre s s u re and stituents, t y p i c a l ly fl ow rate measurefound in aging landfi l l ments around the lanl e a ch a t e , will re a ch Figure 1. Aerobic landfill air injection and leachate recirculation schematic. fill showed that even local gro u n dwater. 129


ESD

Aerobic Bioremediation of a Municipal Solid Waste Landfill

a single blower could probioreactor typical of compostvide substantial air flow ing. O p e rating data collected m o re than 1,000 ft from f rom the ALB at the Bake r the injection point, the fa rPlace Road landfill in thest point in the landfill. Grovetown, Ga., since lateM e a s u rements of soil gas January 1997 provides repres h owed that initially ox ys e n t a t i ve cost-benefits that gen increased and then could be realized by any comd e clined to low levels p a rable solid waste landfi l l . once the microbial popuShort-term benefits incl u d e : lation had adapted to aero(1) increased rate of waste mass stabilization and settlebic conditions (Fi g u re 2). ment, (2) mu ch improve d Both CO2 and CH 4 d e clined immediately to leachate chemistry, (3) signifilow or undetectable levcant reduction in leachate els, f u rther indicating that treatment, (4) signifi c a n t a e robic conditions within reduction in noxous odors, the waste were attained. Figure 2. Changes in soil gas concentrations and temperature in the landfill after and (5) significant reduction injection of air began (February 7, 1997). After aerobic biodegradain methane generation. Longtion became active the term benefits resulting from tempera t u re and CO 2 i n c re a s e d ; waste mass stabilization in less than five however, CH4 never appeared again years include the following savings: (1) except transciently when the air post-closure monitoring reduced from was shut of f. Since methane can 30 to five years ($6,000,000); (2) imperonly be produced by methanogens meable closure cap reduced to reduced under completely anaerobic condipermeable cap or altern a t i ve cove r tions, the production of methane ($2,000,000); (3) Clean Air Act emission was stopped by the air injection. control of methane gas (plus cost of gas Lab o ra t o ry analyses of biochemcontrols, $240,000); (4) leachate collecical ox y gen demand (BOD) and tion and treatment ($1,000,000); and (5) volatile organic compound (VOC) elimination of groundwater treatment concentrations in the leachate indicosts ($1,500,000). cated significant reduction by the Related Publications a e robic process. BOD in the “Sump O n e ” l e a chate samples we re Young, J.D., D.J. Altman, K.H. Lombard, reduced by at least 70%. Orga n i c s A.W. Bourquin, D.C. Mosteller and T.C. s u ch as methy l - e t hyl ketone (MEK) Hazen, Sanitary landfill optimization test and acetone we re reduced signififor remediation of chlorinated solvents, c a n t ly, and fecal colifo rm was elimin In-Situ and On-Situ Bioremediation, inated from the leachate.Total VOC Vol. 5, pp. 315-316, Battelle Press, 1997. concentrations in the many of the Hazen,T.C., Sanitary landfill in situ biorevapor samples collected we re less mediation optimization test final report, than 1 ppm. No leachate wa s WSRC-TR-96-0065, Westinghouse treated during the entire study. Prior to this, the landfill was tre a t- Figure 3. Steaming excavated refuse from aerobic landfill Savannah River Company, Aiken, SC, DOE–NITS, 1996. ing 150,000 gal of leachate each operation (>50°C). month.

Funding Significance

of

Findings This wo rk has been supported by American Technologies, Inc., and the U.S. Department of Energy under Contract No. DE-AC02-76SF00098 and DE-AC0989R180035. Funding was also provided by Columbia County, Apppleton, GA.

This study demonstrated that air injection and leachate recirculation can be used to greatly increase biodegradation of municipal solid waste using a standard subtitle-D configuration, thereby converting the dry tomb to active

http://www-esd.lbl.gov 130


Earth Sciences Division

Project

Objectives

Environmental Remediation Technology Program New Software Tool For Visualizing Fracture Data From Oriented Cores

Annual Report 1998-1999

ules into a network of AVS/Express modules. Geological chara c t e rization of fra cA contractor to DTSC collected Janet S. Jacobsen tured sites relies on analysis of cores cores from 29 wells in a contamiand surface fra c t u re mapping. Fra c t u re nated, fractured site.The fracture data data from cores usually consists of fra cconsist of the following information: Contact: Kenzi Karasaki ture location, staining and filling. In • depth down the borehole to the (510) 486-6759, k_karasaki@lbl.gov some cases, “oriented” c o res are colfracture lected; i.e., not only is the location of a fracture known, but also its • fra c t u re dip: s p e c i fied as angle from horizontal and direction, o rientation in space.This re p o rt describes the development of a either north or south • fracture strike: specified as angle with respect to north and software tool for visualizing oriented fra c t u re data in a 3-D condirection, either east or west text. The fra c t u res are visualized as oriented disks. The software • fracture staining: specified as contaminant stained, iron tool includes the capability to select subsets of fractures based on stained, not stained, or “other” (not specified) angle of dip, angle of stri ke, fracture filling and fra c t u re staining; it • fracture filling: specified as is described in more detail in cl ay fi l l e d , sand fi l l e d , no Jacobsen (in preparation). filled or “other” (not speciThe oriented core fra c t u re fied). data used to develop the softAltogether, the data set ware tool was supplied by contains data for 1,962 fracthe Califo rnia Env i ronmental A ge n c y 's t u res.In addition, DTSC proProtection Department of Toxic Subvided coordinates (in eaststances Control (DTSC), ing and northing) of the which funded and provided wells from which the core s technical guidance for the we re take n , ground surfa c e development of the visualizae l evations of the wells, and tion tool. text well identifi e rs .The well coordinates, depths and eleApproach vations we re all specified in fe e t ; the dip and strike The softwa re tool wa s angles we re specified in A developed using the generald e grees. purpose visualization softAs described above, the ware package AVS/Express, fracture data from oriented from A d vanced Vi s u a l cores consist of strike and Systems Inc. AVS/Express is dip, which together specify widely used for re s e a rch, the orientation in space of a engineering and financial plane containing the fra capplications. It provides a t u re . In the visualizations, visual programming environdisks rather than planes repment in which the user resent the fractures. A C++ builds a data flow network computer program was writconsisting of software modten to convert strike and dip ules that provide various data into the geometric infortechniques and capabilities mation re q u i red by for visualizing field data or AVS/Express to visually repsimulation output. resent the fractures as disks. AVS/Express was chosen not The computer program also o n ly because of its wide includes the logic needed to range of visualization capaselect a subset of fractures B bilities, but also because it is according to the following extensible; that is, user-develcriteria: Figure 1. Visualization of fracture data from oriented core. (a) Fractures colored oped computer pro gra m s by type of filling. (b) Only fractures with an angle of dip between 10° and 40° have • range of dip angles • range of strike angles can be incorporated as mod- been included. Fractures are colored by angle of strike. 131


ESD

New Software Tool for Visualizing Fracture Data from Oriented Cores

• type of fracture filling (clay, sand, none, or “other”) • type of fracture staining (contaminant, iron, none, or “other”). When selecting a subset of fractures, the user may choose one or more criteria. For example, the user could choose to limit the range of dip angles to 0° to 10°, the range of strike angles to N30°W to N30°E, the type of fracture filling to clay and sand and the type of fracture staining to contaminant and iron. Each fracture would then be checked against all of these criteria before including it in the set of fractures to be visualized. A user interface (UI) was developed to facilitate specifying visualization parameters and selection criteria.When the visualization application is loaded into AVS/Express, the user is presented with a menu giving the following options: • name of input file of strike and dip data • radius to use when displaying the fractures as disks • how to color the fracture disks (by angle of dip, angle of strike, type of fracture filling or type of fracture staining) • whether to include all fractures or only a subset of fractures in the visualization. If the user chooses to specify subgroups of fractures to include, then a second menu pops up. The second menu first gives the user the option to specify which criteria (angle of dip, angle of strike, type of fracture filling, type of fracture staining) to use to select fractures. Once the user has toggled on one or more of these criteria, the parts of the UI menu corresponding to the selected options become active and the user may specify range of dip and/or strike and type of fracture filling and/or staining. As discussed above, the user may combine different criteria to specify the characteristics of a subset of fractures.

incorporating the necessary logic in the computer program and also by developing a user interface to facilitate specifying visualization parameters and selection criteria. Figures 1a and 1b provide examples of how the software tool can be used. Figure 1a shows all of the fractures in the data set color coded by type of fracture filling.The visualization in Figure 1b includes only fractures with an angle of dip between 10° and 40°. In this figure, the angle of dip of the fractures is clear, so color has been used to indicate the angle of strike angle in order to show the distribution of strike for the subset of fractures.

Significance

of

Findings

The software tool developed by this project provides the user with the capability to: • include all oriented fracture data from a site in a single view • look for trends in or correlations among the fractures • select subsets of fractures to include in the visualization, thereby reducing the density of fractures and the complexity of the visualization • include in the visualization other types of data (e.g., seismic, water sample, lithology, etc.). Visualization of oriented core fracture data is a powerful analysis tool because it provides practically the only way to include all fracture data from a site in a single view. Unlike other analysis techniques (e.g., Rose diagrams), other types of data also can be included in the visualization.

Related

Publication

Results Jacobsen, J.S., User’s manual for development of a softwa re tool to visualize fra c t u re data from oriented core, in prepara t i o n .

This project resulted in the development of a software tool for visualizing fracture data from oriented cores.The tool consists of a visualization network and user interface developed from AVS/Express modules and a C++ computer program developed to convert strike and dip data into the geometric information required by AVS/Express to visually represent the fractures as disks.The software tool includes the capability to select subsets of fractures, based on angle of dip, angle of strike, fracture filling and fracture staining. This capability has been implemented by

Funding This wo rk has been supported by the Califo rnia Environmental Protection Agency through its Department of Toxic Substance Control, which also collaborated in the research through Integrated Contract Order B347960.

http://www-esd.lbl.gov 132


Earth Sciences Division

Research

Environmental Remediation Technology Program

Objectives

Enhanced Data Analysis For the Vadose Zone Monitoring System

Annual Report 1998-1999

ues. By matching various types of data from the site, we are building a defensible conceptual model for modeling contaminant transport.

Curtis M. Oldenburg, The Vadose Zone Monitoring April L. James System (VZMS) installed at McClellan and Peter T. Zawislanski Air Force Base has been collecting Approach data on pressure, temperature, gasContact: and liquid-phase VOC concentrations Curtis Oldenburg (510) 486-7419, cmoldenburg@lbl.gov A considerable data analysis effort and moisture content over a 100-ft is required to use VZMS data to infer vertical section of alluvial sands and contaminant transport. Our approach is to use all relevant data silts for the last two years.Thirteen levels of instrument clusters including VZMS temporal data to build a plausible conceptual were installed over the vadose zone in each of the two boreholes model and then to use forward and inverse modeling to analyze (VZMS-A and VZMS-B), spaced 10 ft apart at site S-7 in the data and continue to improve the model. We refer to this Investigative Cluster-34 (IC-34).A third VZMS (VZMS-C) was conprocess as â&#x20AC;&#x153;enhanced data analysis.â&#x20AC;? We have taken data from structed in 1998 in between VZMS-A and VZMS-B to collect drilling logs, laboratory measurements of core samples, and the matric potential data to a depth of 25 ft.The instruments are conVZMS, and constructed a one-dimensional conceptual model of nected to a data acquisition system installed in a portable buildthe vadose zone at the site. The 100 ft of vadose zone are dising at the site. Two nearby boreholes for neutron logging were cretized with 0.5-ft gridblocks to represent the many different also constructed to monitor moisture content. Gas and liquid layers of silts, sands and clay present. We use T2VOC, the multisamples are brought to the LBNL Environmental Measurements phase and multicomponent model for the flow and transport of Laboratory for chemical analysis. The primary contaminants are water, volatile organic compounds and air for forward simulatrichloroethylene (TCE) and cis-1,2-DCE. Most of the contaminations, and the inverse modeling code ITOUGH2 for sensitivity, tion is concentrated in the top 6 m (20 ft) of the section, with uncertainty and inverse modeling.These modeling tools allow us some additional contamination near the water table. The main to carry out numerical experiments and inverse modeling to issue to be addressed concerns the rate of downward transport determine key parameters controlling movement of gas and liqof TCE from the shallow vadose zone to the water table.Through uid phases in the subsurface.The goals of enhanced data analysis modeling, we estimate values of sediment permeability, effective are to (1) increase the value of the VZMS data by constructing thermal conductivity and moisture content as a function of models that can be used in a predictive mode and (2) validate depth, comparing estimates to laboratory and site-measured vala

b

Figure 1. (a) Observed and simulated gas-phase pressure at a depth of 30 ft in response to barometric pressure changes at groundsurface. (b) Observed and simulated subsurface temperatures at 11, 18 and 30 ft in

133

response to groundsurface temperatures. Values of permeability and effective sediment thermal diffusivity were estimated using ITOUGH2 inversion techniques.


ESD

Enhanced Data Analysis for the Vadose Zone Monitoring System

Figure 2. Observed formation volumetric moisture content and simulated 1-D moisture profiles using recharge rates of 1, 10 and 50 cm/yr.

the models such that defensible remediation decisions at other similar sites can be made based on the understanding gained at the S-7 site.

of permeability, thermal diffusivity and annual recharge are three key features in the conceptual model to be used to model contaminant transport.

Results

Significance

Using a one-dimensional conceptual model that contains a site-averaged hydrostratigraphy (at resolution of 0.5 ft) defined by lithologic logs and laboratory analysis of core, we estimated values of permeability by simulating gas-phase pressure response at depth in the subsurface to a time-varying barometric signal. Permeabilities were estimated by matching observed and simulated gas-phase pressure at the four shallowest instrument cluster depths (6, 11, 18 and 30 ft), where the response to atmospheric variation was strongest, and where, at site S-7, the shallow VOC contamination has persisted over 25 years or so. Figure 1a illustrates the pressure response (simulated and observed) at 30 feet. Results indicate that the inversion is most sensitive to the permeability of the overlying concrete layer. A similar inversion was performed to evaluate effective thermal diffusivities of sediments by comparing observed and simulated formation tempera t u re (Fi g u re 1b). Simulations matched the observed temperature fluctuations at depths of 11, 18 and 30 ft, below which the effects of surface temperature are not discernible. Inclusion in the simulation of an annual recharge of 10 cm/yr provided a better fit to the observed moisture content profile relative to 1 cm/yr and 50 cm/yr (Figure 2).These parameters

The enhanced data analysis has reproduced multiple types of site data and allowed the estimation of sediment properties critical to the transport of contaminants at the S-7 site.The estimated values of permeability, thermal diffusivity and recharge will be incorporated into the conceptual model used to assess TCE transport in the gas- and liquid-phases.

Related

of

Findings

Publication

James, A.L., and C.M. Oldenburg, Enhanced data analysis for the VZMS: Conceptual model design and initial application for the Vadose Zone Monitoring System (VZMS), McClellan AFB 1998, Berkeley Lab report LBNL-41909, 1998.

Funding This work has been supported by the U.S. Department of Defense under Military Interdepartmental Purchase Request FD2040-96-74020EM to the Ernest Orlando Lawrence Berkeley National Laboratory under U.S. Department of Energy Contract No. DE-AC03-76SF00098.

http://www-esd.lbl.gov 134


Earth Sciences Division

Environmental Remediation Technology Program

Annual Report 1998-1999

Numerical Simulation Of Ferrofluid Flow Research

Objectives

permanent rectangular magnets using analytical ex p ressions that neglect Curtis M. Oldenburg, Ferrofluids are stable colloidal susthe presence of fe rro fluid, an Sharon E. Borglin pensions of magnetic particles in carand George J. Moridis approach considered reasonable for rier liquids. Due to their ability to be the applications we consider.The graContact: held in place by magnetic fields, ferdient of H is calculated by simple Curtis Oldenburg rofluids currently find application in a first-order differencing in each of the (510) 486-7419, cmoldenburg@lbl.gov va riety of industrial processes and coordinate directions.At any value of engineered devices, as well as limited use in the biomedical field. H, we assume that magnetization increases linearly with ferWe have conceived of many potential applications for ferrofluids rofluid saturation.While we neglect the coupling between H and in the field of subsurface environmental restoration.These applithe ferrofluid distribution, we consider explicitly the full coucations fall broadly into two main areas: (1) guiding liquids and pling between flow and transport driven by both the magnetic holding liquids in place in the subsurface through application of force (arising from the M and ∇H terms) and the buoyancy force magnetic fields; and (2) using ferrofluids as subsurface flow trac(when gravity is present), which combine to make this a strongly ers that can be imaged by standard electromagnetic and geocoupled flow problem. Once the magnetic field strength H and physical methods. In order to facilitate the design and experiits gradient are calculated, we calculate magnetization and the mentation of porous media ferrofluid applications, as well as to magnetic force at the interface between each gridblock. This c a rry out numerical ex p e riments of ferrofluid flow phenomena, term is added to the pressure gradient and gravitational body force terms.A diagram showing the calculation procedure is prewe have developed simulation capabilities for fe rrofluids.The simsented in Figure 1. ulation capabilities are built upon the re s e rvoir simulator TOUGH2.

Approach

Results

In the presence of an external magnetic field, the singledomain colloidal magnetite particles suspended in ferrofluid align, causing the ferrofluid to become magnetized.The magnetization of the fluid interacts with the magnetic field to produce attractive magnetic forces analogous to the body force on a liquid due to gravity.The attractive force on ferrofluid per unit volume is given by Fmag = µo M∇H [=] Tesla Ampere-turn m/m3 [=] T A m/m3 [=] N/m3, where µ0 is magnetic permeability of free space in units of T m/A, M is the magnetization in A/m and H is the magnetic field strength in A/m.As the magnetic field strength is increased, the ferrofluid reaches a maximum magnetization, or saturation magnetization. Thus, the magnetization is a function of H and can be approximated by simple arctangent functions. In our development of simulation capabilities, we built upon the T2VOC module of TOUGH2 by (1) substituting ferrofluid for the nonaqueous phase liquid organic compound, (2) adding calculations for external magnetic field stre n g t h around a permanent magnet, (3) adding arctangent function approximations for the magnetization of ferrofluid, and (4) adding the ferromagnetic force term. We calculate the external magnetic field (H) for

The demonstration problem models a laboratory experiment designed to show one potential application of ferrofluids in subsurface environmental engineering, that of pulling fluid underground by application of a permanent magnet. In the experiment, a small volume of ferrof luid (EMG 805TM) is injected through a catheter into the left-hand corner of a fluid-filled gap between one plate of smooth glass and one plate of rough glass with a magnet on the right-hand side.The narrow gap between the plates creates an analog porous medium similar to a HeleShaw cell. The domain is modeled as a porous medium with porosity set to 0.999. The fe rrofluid flow experiment is carried out with the cell in a horizontal position and filled with a colloidal silica fluid that closely matches ferrofluid in density. The gradient in the magnetic field produces the a t t ra c t i ve fo rce that pulls fe rrofluid toward the magnet, while viscous forces resist the motion. Shown in Figure 2 are snapshots at four times in the simulation: (a) initial state; (b) t = 6 min; (c) t = 9 min; and (d) t = 13 min. Results as shown in Figure 2 agr ee qualitat i ve ly with the ex p e riment in terms of the direction and geometry of ferrofluid movement in the system. Note in particular the ten-

Figure 1. Schematic of ferrofluid flow calculation implemented in EOS7M and T2VOCM.

135


ESD

Numerical Simulation of Ferrofluid Flow

Figure 2.

Ferrofluid saturation and flow in Hele-Shaw cell simulation at t = 0 min, t = 6 min, t = 9 min and t = 13 min.

dency for the ferrofluid to elongate in the direction of the magnet just as we observed in the experiment.This occurs because the force on ferrofluid is proportional to both the magnetic field strength (prior to saturation) and to the magnetic field gradient. Thus, ferrofluid that is already closer to the magnet is more strongly magnetized and pulled more strongly toward the magnet than ferrofluid that is farther away. Note also the accumulation of ferrofluid around the magnet. The good general agreement between experiment and simulation serves to qualitatively demonstrate the validity of the new simulation capabilities.

net where it segregates out.This phase saturation minimum tends to be close to the edge of the segregating volume near the magnet. Finally, the fo rces on the ferrofluid become ve ry large near the magnet where the magnetization is strong and there is a high field strength gradient. This pro blem is an example of an ove rs t able problem, w h e re fluid is pulled strongly towa rd an inpenetrable boundary and tends to stagnate. An example application s i mulation we carried out involving barrier ve ri fication shows that the effective distances over which a single magnet can pull fe rrofluid are on the order of 1 m.

Significance

Related

of

Findings

We developed and demonstrated simulation capabilities for ferrofluid fl ow in porous media.A number of the physical processes of fe rrofluid flow shown in the sample pro blem are rather uncommon. Fi rs t , we observe the tendency of the ferrofluid to elongate in the direction of the magnet due to the larger fo rces on the fluid at smaller distances from the magnet. Second, fe rrofluid s t a rts out at nearly full saturation and then in the course of trave rsing the domain becomes a two-phase fe rrofluid–water mixture. Near the magnet, it segregates and becomes pure fe rrofluid (Sff = 1.0). S u ch segregation effects are not particularly common in flow and transport where diffusion and dispersion usually lead to smoothing and mixing of constituents.Third, near the magnet, water must be expelled as fe rro fluid is pulled stro n g ly towa rd the magnet. During this evo l u t i o n , t h e re is a minimum in phase saturation between the initial fe rrofluid injection point and the mag-

Publications

Oldenburg, C.M., S.E. Borglin and G.J. Moridis, Numerical simulation of ferrofluid flow for subsurface environmental engineering applications, Transport in Porous Media, in press, and Berkeley Lab report LBNL-40146, 1998. Borglin, S.E., G.J. Moridis and C.M. Oldenburg, Experimental studies of ferrofluid flow for subsurface environmental engineering applications,Transport in Porous Media, submitted.

Funding This work has been supported by the Laboratory Directed R e s e a rch and Development Program of Law rence Berke l ey National Laboratory under U.S. Department of Energy Contract No. DE–AC03–76SF00098.

http://www-esd.lbl.gov 136


Earth Sciences Division

Research

Environmental Remediation Technology Program

Objectives

Advanced Detection of Exposures and Biological Responses to Organic Toxins

Annual Report 1998-1999

tures of PAHs. HepG2 cells were maintained in Eagle's Minimum Essential Organochlorines (OCs) and polyMedium with non-essential amino Hoi-Ying N. Holman, cyclic aromatic hydrocarbons (PAHs) acids and Earle's BSS supplemented Regine Goth-Goldstein, are ubiquitous environmental contamwith 10% fetal calf serum, 1 mM L-gluMichael C. Martin, inants that are toxic and suspected tamine, 10 mM Hepes and antibiotics. Marion Russell human carcinogens.Traditional assessCells we re sub-cultured, and then and Wayne R. McKinney ment of human exposures to these treated for 2 - 20 hours with TCDD, organic toxins and the subsequent BaP and coal tars at environmentally Contact: Hoi-Ying Holman biological responses relies primarily relevant concentrations. (510) 486-5943, hyholman@lbl.gov on high-dose and short-term animal The effect of TCDD was monitored ex p e riments. A major uncertainty and mapped by the SR FTIR spectroinherent to this approach is the extrapolation from the high-dose microscopy technique in the mid-IR region (4000-400). SR FTIR animal experiments to the low-dose and long-term human exposignals (from individual cells) that are specific to the intracellusures. To overcome this uncertainty, we recently have directed lar response after their exposure to these organic toxins were our research effort towards exploring the use of synchrotron identified.To validate the SR FTIR spectromicroscopy technique, radiation-based (SR) Fourier transform infrared (FTIR) spectroresults from the TCDD experiments were compared with the microscopy for identifying chemical changes in cellular nucleic CYP1A1 transcript levels measured by the widely accepted yet acids and proteins as a result of OC and PAH exposures. more time-consuming reverse transcription-polymerase chain To date, the primary research objective has been to identify SR reaction (RT-PCR). FTIR spectroscopic signals of human cell culture systems that Results are indicative of low-dose exposures to OCs and PAHs and could be used as biomarkers. SR FTIR spectromicroscopy is used Dimensionless SR FTIR spectra were recorded at the proximbecause our earlier work has demonstrated that this is a sensitive ity of a cell nucleus of HepG2 cells that were exposed to TCDD and nondestructive technique capable of providing direct bioof different concentrations (0, 0.01, 0.1, 0.5, and 1.0 nM) for 20 chemical information at molecular levels in a biological system. hours.There are considerable differences in the SR FTIR spectra The fine spatial resolution of 5 - 10 microns and strong signal to associated with the CYP1A1 gene expression and the increase in noise levels of SR FTIR spectromicroscopy allow one to detect the associated enzyme activity, with one difference being the the subtle changes in intracellular biochemical processes as the increased absorption of the vibration band 1180 - 1160 cm-1, cencells are exposed to environmental stimuli. tered at ~1170 cm-1. Here, the normalized absorbance intensity Approach for individual cells increased from 0.007 to 0.21 when the TCDD concentration increased from 10-11 to 10-9 M (Figure 2a).The norThe biomarker considered for OC and PAH exposures is the malized absorbance intensity at ~1170 cm-1 for individual coninduction of the cytoch rome P4501A1 CYP1A1 gene expression trol cells was 0.005.This is a 42-fold increase in the absorbance and the increase in the intensity. This systematic associated enzyme activspectral change might be ity. It is well established related to the alteration that CYP1A1 transcript in the DNA base strucincrease in l evels ture and will be the subresponse to exposure to ject of future investigaOCs and PAHs through tion. their binding to the Ah A comparison of the receptors (Fi g u re 1). dose response described HepG2 (human above with that obtained h e p a t o m a - d e ri ved) cells using the RT-PCR techwe re used as model nique is shown in Figure human epithelial cells 2b.The solid line was the that can metabolize l e a s t - s q u a res fit to the PAHs; 2 , 3 , 7 , 8 - t e t radata.The excellent agreechlorodibenzo-p-dioxin ment (with r2 = 0.99) for (TCDD) modeled OCs; measurements from the BaP modeled PAHs and Figure 1. Intracellular responses to dioxin or aromatic hydrocarbon compounds: CYP1A1 two methods indicated coal tars modeled mix- - a global biomarker of exposure. that the fast and direct 137


ESD

Advanced Detection of Exposures and Biological Responses to Organic Toxins

SR FTIR spectromicroscopy technique was indeed comparable to the more time-consuming and widely accepted RT-PCR technique that specifically measures increases in the CY1A1 transcript levels. The SR FTIR spectra recorded at the proximity of a cell nucleus of HepG2 cells that have been exposed to BaP and coal tars at environmentally relevant concentrations showed similar spectral characteristics at ~1170 cm-1. The recorded doseresponse behavior was also similar to those reported in the literature.

Significance

of

(a)

Findings

The agreement between the SR FTIR spectromicroscopic data for dioxin exposures and the RT-PCR results, and the agreement between the PAH measurements and those reported in the literature indicate that the intracellular biological responses to low-dose exposures to these organic toxins are well represented by our specific spectral changes. These changes are associated with CYP1A1 gene expression and the increase in the associated enzyme activity with different types of damage.The capabilities of SR FTIR spectromicroscopy for the direct detection of intracellular biochemical responses to exposures to dilute concentrations of OCs and PAHs will have significant impacts in future research methodology of environmental toxicology.

Related

Publications

Holman, H.-Y.N., M. Zhang, R. Goth-Goldstein, M.C. Martin, M. Russell, W.R. McKinney, M. Ferrari and J.C. Hunter-Cevera, Detecting exposure to environmental organic toxins in individual cells: Towards development of a micro-fabricated device, Proc. Intl. Symp. Biomed. Optics, 3606, pp. 55-62, 1999. Holman, H.-Y.N., D.L. Perry, M.C. Martin and W.R. McKinney, Applications of synchrotron infrared microspectroscopy to the study of inorganic-organic interactions at the bacterialmineral interface, Application of Synchrotron Radiaiton Techniques to Material Science IV (S.M. Mini, S.R. Stock, D.L. Perry and L.J.Terminello, eds.), 524, pp. 17-23, 1998.

(b)

Funding This work has been supported by the Office of Science, Office of National Petroleum Technology Program, Offices of Health and Environmental Sciences, Biological and Environmental Research Program and Basic Energy Sciences, Materials Science Division, of the U. S. Department of Energy under Contract No. DE-AC03-76SF00098.

Figure 2. Development of a micro-fabricated device for detecting human exposure to environmental organic toxins. (a) Dose response of individual living HepG2 cells after 20 hour exposure to 2, 3, 7, 8-tetrachlorodibenzop-dioxin, as measured by the new IR technique (after hours of work). (b) New IR technique versus RT-PCR assay.

http://www-esd.lbl.gov 138


Earth Sciences Division

Research

Environmental Remediation Technology Program

Objectives

An Algal-Bacterial Treatment System for Drainage Selenium Removal

Annual Report 1998-1999

tion and sand filtration. Past and current studies show a clear need to Selenium drainage water treatment reduce dissolved oxygen and nitrate Nigel Quinn, Tryg J. Lundquist*, has been an active area of research for to low levels before selenate can be Bailey F. Green*, W.J. Oswald, more than a decade since the discovreduced to selenite. Oxygen is elimiery of selenium toxicosis at Kesterson Terrance Leighton, R.B. Buchanan nated from the drainage water by the and Max Zarate Reservoir in the western San Joaquin respiration of microorganisms. Nitrate Va l l ey of California. The Grassland is removed by reduction to nitrogen *University of California at Berkeley Bypass Project established monthly gas during denitrification, or in and annual selenium load targets for another flow scheme, by assimilation Contact: Nigel Quinn tile drainage from agricultural water into algal biomass in addition to deni(510) 486-7056, nwquinn@lbl.gov districts within the Grassland Basin— trification. Selenium removed from exceeding these could lead to fines of up to $500,000 per year. the water column accumulates in the algal-bacterial biomass and This has proved a powerful incentive for investment in irrigation inert materials in the bottom of the RPs. Because the biomass is s o u rce control and dra i n age reduction tech n o l o gy. These continu o u s ly decomposing, the volume of solid re s i d u e s monthly and annual load-based targets have helped to change increases very slowly. Removal and disposal of the solids in a the economics of drainage treatment—no longer are expensive landfill should be required only after many years of accumulaunit processes required to achieve the 5 ppb effluent standard. tion. Hence, technologies such as the Algal-Bacterial Selenium The 0.1-acre RPs were designed to promote the growth of Removal System (ABSR), developed by Professor Oswald and his nitrate- and selenate-reducing bacteria. Floating covers on the co-workers at the Applied Algae Research Group (AARG) at the RPs were installed to reduce wind-induced mixing.The 0.1-acre University of California, Berkeley, are well suited for this applicapaddle-wheel-mixed HRPs were designed to cultivate microalgae tion.The primary goal of this project is to demonstrate the techin high concentrations and at high productivities.The low speed mixing of HRPs enhances the selection of algal species that nical and economic feasibility of the ABSR system at a larger bioflocculate and settle in quiescent algae settling ponds. scale, up to 20,000 gallons per day. Carbon dioxide supplementation to the algae is provided by Approach sparging the gas into a carbonation sump near the paddle wheel in each HRP. A baffle in the carbonation sump forces the flow of The current collaborative demonstration project involves sciwater downward. Against this downward current, the carbon entists in LBNL’s Earth Sciences Division, the AARG and the dioxide bubbles are held in suspension as they dissolve into the departments of Cell Biology and Plant Biology at UC Berkeley. water.ASPs provide a quiescent zone for the algae grown in the The ABSR system was constructed in July 1996 and consists of high rate ponds to settle.The launders in the ASPs improve algae two parallel systems, each having a reduction pond (RP), a padsedimentation by removing supernatant from the surface of the dle-wheel mixed algal high rate pond (HRP) and an algae settling pond at a very low overflow velocity.A sloped floor and internal pond (ASP); see Figure 1. By operating two systems simultanesump in each ASP enables the harvesting of the algal biomass ously, the operational parameters of one system can be varied using a diaphragm pump. while the other system is operated as a control. Two fl ow modes are possible in the ABSR facility in the The basic concept of the ABSR system is to grow microalgae Panoche Dra i n age District, each having its adva n t ages in terms of on drainage water and to utithe economics of the system. lize the algal biomass as a In Mode 1 (South ABSR syscarbon source for native bactem), drainage water is teria such as Acinetobacter brought into a HRP where 15 and Psuedomonas, w h i ch to 30 mg/L of nitrate-nitrogen reduce nitrate to nitro ge n can be re m oved thro u g h gas and selenate to selenite. assimilation by algal cells.The Selenite combines with remaining nitrate is removed metal ions and precipitates in the RP via heterotrophic or is further reduced to denitri fication to nitro ge n i n s o l u ble elemental selegas.By removing some nitrate nium.The insoluble forms of t h rough assimilation, less selenium are then separated algal biomass is re q u i red as a from the effluent by sedicarbon source for the denitrifying bacteri a . The disadva nmentation in the ponds folFigure 1. The Algal-Bacterial Selenium Removal System (ABSR). t age of this mode is that carlowed by dissolved air flota139


ESD

An Algal-Bacterial Treatment System for Drainage Selenium Removal

bon dioxide and phosphate must 1 mg/L in the RP effluent by be added to the dra i n age water to September 1997, when the South i n c rease algal growth in the HRP. ABSR was switched to Mode-1 In Mode 2 (North ABSR sysconfiguration. With the gradual tem), the drainage water passes reduction in nitrate concentrathrough the RP first, where most tion in the South RP, selenium of the nitrate is removed by deniremoval has increased. The total tri fication fo l l owed by soluble soluble selenium re m oval has selenium removal.The RP effluent averaged 25%, from an influent containing algal decomposition concentration of 413 ug/L to a concentration of 308 ug/L in the products such as ammonia, phosSouth RP effluent. Selenium phate, and dissolved carbon dioxre m oval decreased dra m a t i c a l ly ide passes to the HRP. Algae grow utilizing the decomposition prod- Figure 2. Selenium concentration (soluble and total T-4) in the North during the winter months (starting in October 1998), when algal ucts, thereby reducing the need ABSR System. assimilation of selenium is at its lowest and the selenium-reducfor supplemental carbon dioxide and phosphate. ing bacteria in the reduction pond are least active and limited in Results their supply of carbon. For start-up, the RPs were fed algae collected over six months f rom the AIWPS Facility at the UC Berke l ey Richmond Field Station.These algae were pretreated with heat, steam or by drying and milling. Subsequent lab o ra t o ry ex p e riments have shown that algae harvested directly from the Panoche and Richmond HRPs and used as substrate directly without pretreatment are two to 10 times more effective for nitrate and selenium reduction than the algae collected over months in the Richmond ASPs. The â&#x20AC;&#x153;freshâ&#x20AC;? algae reduced at least 0.2 g NO3-N per gram volatile solids introduced. Steam pretreatment of the "fresh" algae did not signifi c a n t ly i m p rove nitrate reduction. Molasses was also evaluated as a subs t rate and found to reduce 0.22 g NO3-N per gram volatile solids. Acting on the lab o ratory results, animal-feed-grade molasses has been used to supplement the algal substrate in the North ABSR system, and fre s h ly harvested algae have been exclusively the subs t rate in the South ABSR system. From April to October 1997, the molasses-fed (North ABSR) system has consistently reduced nitrate nitrogen to less than 10 mg/L NO3-N and total soluble selenium 70 to 80% from a mean of 413 ug/L in the influent down to a mean of 87 ug/L (Figure 2).The fl ow to each system was 3,500 gallons per day during this peri o d , gi ving a hydraulic residence time of 58 days. A shorter residence time of 25 days gave similar removal rates when internal baffles were installed in the RPs to prevent flow short circuiting of influent and the flow was increased to approximately 7,000 gals/day. In October 1998 some breakthrough of selenium was evident as selenium removal kinetics decreased with colder winter temperatures. Figure 2 shows the impact that variations in molasses feeding had on soluble selenium removal; the thinner lines in the graphs with the (T-4) suffix represent total selenium. The South ABSR system has been fed "fresh" algae, harvested by dissolved air flotation from wastewater treatment ponds since April 14, 1997. The mean influent nitrate concentration has been 82 mg/L NO3-N. Nitrate removal by the system has improved, reaching less than

Significance

Findings

The molasses-fed North ABSR pilot system perfo rmed well with an ave rage removal of 70-80% of influent selenium. Although the c u rrent pilot plant removes less than 10% of the total selenium mass load ge n e rated by the field dra i n age sump, a full-scale facility, sized to treat the total dra i n age fl ow, would likely be 10 times as l a rge, o c c u pying a land area of between fi ve and 10 acres. The project goal over the next six months is to adjust operational parameters to achieve the greatest selenium mass reduction at the lowest operational cost. Selenium removal is greatest for the Mode-2 process, but the cost of adding molasses, an external carbon source, detracts from the cost-effectiveness of the system. Real-time telemetered sensors and a SCADA system for remote operation and control of the plant will be installed by summer 1999 in order to optimize plant operation and minimize chemical additions.

Related

Publications

Lundquist,T.J., M.B. Gerhardt, F.B. Green,R.B.Trasan, R.D. Newman and W.J. Oswald, The algal-bacterial selenium removal system: M e chanisms and field study, in Selenium in the Environment (W.T. Frankenberger and S.M. Benson, eds), pp. 251-278, Pergamon Press, New York, NY, 1994. Quinn, N.W.T., J.C.McGahan and M.L. Delamore,Innovative strategies reduce selenium in Grasslands drainage, California Agriculture, vol. 52, no. 5, pp, 12-18, 1998.

Funding This project, which was funded initially by a Challenge Grant from the U.S. Bureau of Reclamation, is now supported by a three-year grant from the CALFED Bay-Delta Program.

http://www-esd.lbl.gov 140

of


Earth Sciences Division

Research

Environmental Remediation Technology Program

Objectives

Characterization Of VOC Biodegradation On Rock Surfaces By SR FTIR Imaging

Annual Report 1998-1999

Results

Designing stra t e gies for using Global features of SR FTIR spectra intrinsic microorganisms to successfor the intrinsic micro o rganisms, Hoi-Ying N. Holman, Jil Geller, fully biodegrade and detoxify VOCs toluene and toluene metabolites were Jennie C. Hunter-Cevera ( volatile organic contaminants) consistent with those reported in the and Karsten Pruess requires a thorough characterization literature. Detailed comparative analyContact: of key reactions at a molecular level sis of spectra recorded on microbial Hoi-Ying Holman on mineral surfaces occupied with surfaces of the coated-gold mirrors (510) 486-5943, hyholman@lbl.gov microorganisms. This study show s and on the cleaved basalt samples that synchrotron radiation-based (SR) Fourier transform infrared indicates the existence of distinct SR FTIR absorption bands as (FTIR) spectromicroscopy — with a spatial resolution of 5–10 markers indicative of the presence of biomolecule markers (biomarkers) and of toluene and toluene metabolites. microns and strong signal to noise levels — can readily be Fi g u re 1 shows the spatial distribution of IR absorption bioapplied to imaging characterization of the complex progress of m a rkers associated with protein Amide I of intrinsic micro o rg a nVOC biodegradation on basaltic rock surfaces, especially the isms (1687 cm-1) .The mapping was re c o rded on one of the basalt progress of the initial biodegradation. specimens after it was exposed to 100-ppm toluene vapor at 100% relative humidity for five days.The high intensity of the IR Approach absorption band at 1687 cm-1 implies that the intrinsic micro o rganisms in the basalt ro ck from INEEL could grow extremely Vesicular basalt rock samples and intrinsic microorganisms quickly and form biofilms under our experimental conditions were taken from a site where VOC contamination has threatened within five days.The striking similarities between the IR absorpthe Snake River Aquifer near the Department of Energy’s Idaho tion and the optical images of the biofilms on the surface of the National Engineering and Environmental Laboratory (INEEL). ro ck chip confi rmed that SR FTIR spectro m i c ro s c o py is adequate Toluene vapor was used as a model VOC that can be directly for detecting the presence of biofilms on ro ck surfaces. metabolized by intrinsic microorganisms. The SR FTIR spectra The spatial distributions of IR absorption associated with for films of microorganisms, toluene and possible toluene toluene (1029 cm-1), and metabolites benzyl alcohol (1022 cmmetabolites were measured and validated by comparison with 1) and catechol (1096 cm-1) were mapped and are presented in those in the spectral literature. Basalt specimens were incubated Figure 2.The presence of these metabolites implies that intrinsic microorganisms probably degraded toluene via the pathway in a closed system in the dark for a number of days.The biodegradescribed in Figure 3.The absence of benzyl aldehyde and bendation of toluene on specimen surfaces was monitored using SR zoic acid in our sample implies that the bottlenecks in the FTIR spectromicroscopy.

Figure 1. SR FTIR mapping of intrinsic microbes on vesicular basalt.

141


ESD

Characterization of VOC Biodegradation on Rock Surfaces by SR FTIR Imaging

Figure 3. SR FTIR spectromicroscopy showing possible toluene biotransformation pathways (in red).

down the degradation pathway.This lack of significant link will be the subject of a future study.

Significance

of

Findings

The results revealed that SR FTIR spectromicroscopy is a powerful nondestructive microprobe for direct characterization of toluene biotransformation on the surfaces of geologic materials. This novel surface analytical technique will have significant impacts in future research methodology in the field of environmental research and management.

Related

Publication

Holman, H.-Y.N., D.L. Perry, M.C. Martin and W.R. McKinney, Applications of synchrotron infrared microspectroscopy to the study of inorganic-organic interactions at the bacterialm i n e ral interfa c e , Application of Synch ro t ron Radiation Techniques to Material Science IV (S.M. Mini, S.R. Stock, D.L. Perry and L.J.Terminello, eds), 524, pp. 17-23, 1998.

Figure 2. SR FTIR mapping of microbial transformation of toluene vesicular basalt surfaces.

degradation process were the breakdown of benzyl alcohol and benzoic acid. One-way analysis of variance was conducted to reveal the link between spatial distribution of the intrinsic microorganisms (1687 cm-1), toluene (1029 cm-1) and its metabolites. Results show that the link is significant for the intrinsic microorganisms (1687 cm-1) and toluene (1029 cm-1), with P<0.03. Similarly, the link is significant for the intrinsic microorganisms (1687 cm-1) and the first metabolite benzyl alcohol (1022 cm-1). However, the link is less significant as the degradation proceeds further

Funding This work has been supported by the Office of Science, O ffi c e of Health and Env i ronmental Sciences, Biological and Environmental Research Pro gram and Materials Science Division of the U.S. D e p a rtment of Energy under Contract No. DE-AC0376SF00098.

http://www-esd.lbl.gov 142


Earth Sciences Division

Research

Environmental Remediation Technology Program

Objectives

Stimulation of Methyl Tert-butyl Ether Biodegradation Using A Co-Substrate Approach

Annual Report 1998-1999

can also transform MTBE by “accident.” It is further hypothesized that if the proper co-metabolic additive can be found, the co-metabolite can be used to stimulate MTBE degrading microorganisms; thus, reactor start-up time can be shortened and removal efficiency can be improved.

Methyl tert-butyl ether (MTBE) is an “oxygenate” that is added to gasoline to promote cleaner combustion. William T. Stringfellow The use of MTBE has become wideContact: s p read as a result of U. S . William Stringfellow Environmental Protection Agency pro(510) 486-7903, wstringfellow@lbl.gov grams targeted at reducing carbon monoxide and ozone air pollution.An Approach unintended consequence of using oxygenates to control air pollution has been that service stations, fuel transfer terminals, The approach taken was to conduct field and laboratory farms and other places where gasoline is used or stored are now research in parallel. Laboratory studies consisted of batch reacsources for the release of MTBE into the environment. Not surtions in sealed vials where MTBE removal under different condiprisingly, there are many instances where gasoline spills and tions was measured over time. Samples of full-scale, up-flow, leaks have resulted in the contamination of groundwater with MTBE-degrading biological reactors containing bacteria grown MTBE. This problem is particularly acute in California, where as a biofilm on activated carbon were transported to the laboraMTBE-containing fuels have been used since the early 1990s and tory and tested for stimulation of MTBE-degrading activity. groundwater resources are highly valued. Compounds tested for stimulation included gasoline compoThere is currently no well-established technology for treating nents (such as toluene), suspected MTBE biodegradation prodMTBE-contaminated groundwater.The objective of this research ucts (such as tert-butyl alcohol) and contaminants commonly is to develop a biological treatment technology. It has been found in industrial grade MTBE (such as methanol). It was shown that MTBE is biodegradable, therefore, biological treatobserved that the reactor had “iron-bacteria” in the biofilm, so ment is a promising technology; however, the mechanism by conditions conducive to iron-bacteria were also tested for stimwhich microorganisms degrade MTBE is unknown. Without an ulation of MTBE degradation activity. understanding of the physiology of MTBE biodegradation, the When a potential co-substrate was identified, field studies operation of MTBE biodegrading treatment systems remains hapwere conducted at the Sparks Solvent Fuel Site (SSFS) in Sparks, hazard and unreliable. Nevada. MTBE-contaminated groundwater at SSFS is now being The objective of this ongoing research is to increase the effitreated by a pair of six-ft-diameter Envirex/U.S.Filter up-flow, aerciency of MTBE biodegradation in a “fluidized-bed” bioreactor. obic, fixed-film bioreactors. Controlled studies can be conducted Research is being conducted to develop methods for rapid reacin the field because the reactors are operated in parallel, allowtor start-ups and for lowering MTBE effluent concentrations ing one reactor to be treated from the reactor. A typical and the other to serve as a reactor start-up can take control. In the field tests, more than 100 days before compounds found to be MTBE removal is achieved. s t i mulants we re added to Furthermore, once MTBE is one of the two reactors and being removed, the removal the effect on the added comwill only be 99% efficient, w h e reas new re g u l a t i o n s pound on MTBE treatment will re q u i re re m ovals of was evaluated. 99.9% or greater. Results In order to increase reactor treatment efficiency, it is L ab o ratory studies we re necessary to understand the used to evaluate the potenfundamental processes tial of more than two dozen re s p o n s i ble for MTBE compounds as potential cobiodegradation. The hypothmetabolic stimulants of esis for this research is that MTBE biodegra d a t i o n MTBE is driven by a co-meta( Fi g u re 1). Organic combolic process; i.e., a process pounds (fatty acids) known in which enzymes produced for an unrelated purpose Figure 1. Testing of potential co-metabolites for stimulation of MTBE biodegra- to support the growth of i ro n - b a c t e ria stimu l a t e d (such as energy metabolism) dation. 143


ESD

Stimulation of Methyl Tert-butyl Ether Biodegradation Using a Co-substrate Approach

MTBE biodegra d a t i o n , more fully evaluate the effiwhereas alcohols and gasocacy of this co-metabolic line components we re a p p ro a ch for stimu l a t i n g found to have no effect on MTBE biodegradation. MTBE degradation when Significance compared to contro l s . of Findings However, compounds used as growth substrates fo r This re s e a rch is being iron-bacteria are also subconducted in an effort to strates for other bacteri a , provide innovative technoland the mechanism by o gy for the treatment of which these compounds are MTBE-contaminated effecting MTBE degradation groundwater. It is particuis not well understood. l a r ly important that the Studies are currently undertechnology be applicable to way to define the role of the treatment of contamiiron-bacteria, if any, in MTBE nated water in Califo rnia biodegradation. In addition, and other we s t e rn states, co-metabolic compounds w h i ch have incre a s i n g ly are being used to enri ch Figure 2. Addition of a co-metabolite to full-scale reactors at SSFS resulted in stimbacteria and fungi from the ulation of MTBE biotreatment, giving lower effluent MTBE concentrations in the strict MTBE discharge limitations.We have found a coMTBE degrading re a c t o r treated reactor compared to the control reactor. metabolite stimulant that biofilm, and these isolates may allow the improved treatment of MTBE to lower levels than are being screened for MTBE biodegradation. Isolation of MTBEhas been previously achievable. In addition, the discovery of this degrading pure cultures is desirable if the fundamental mechaco-metabolite has given us a key with which to unlock the nisms of MTBE bio-degradation are to be understood. secrets of MTBE biodegradation. Only through a better and more Field experiments were conducted to test the hypothesis that complete understanding of how microorganisms degrade MTBE the addition of co-metabolites would stimulate MTBE biodegrawill we be able to develop a more efficient and robust, yet ecodation in a full-scale reactor.A review of reactor operations data nomical, treatment technology. demonstrated that the reactors behaved in a reproducible manner and that MTBE effluent data between the two reactors for Related Publication the three months prior to testing were not statistically different. One of the two parallel reactors at SSFS was selected for treatStringfe l l ow,W.T., and S.T. Kilkenny, Biodegradation of ethers using ment while the other served as a control.The treated reactor was fatty acid enhanced microbes, submitted to U.S. Patent Offi c e . conditioned with a high concentration of fatty acids for approximately 12 hours and then given continuous feed of a fatty acid Funding solution over the next two days. MTBE effluent concentration was compared between the two reactors. It was found that the This work has been supported by Kinder Morgan Energy treated reactor out-performed the control reactor (Figure 2).The Partners, L.P., Vista Canyon Group, L.L.C., Envirex/U.S. Filter Co. response was dose-dependent (data not shown). Based on these and Varian Analytical Co. tests, longer term studies at multiple sites are now underway to

http://www-esd.lbl.gov 144


Earth Sicences Division

Annual Report 1998-1999

CLIMATE VARIABILITY AND CARBON MANAGEMENT PROGRAM SALLY M. BENSON

CONTACT: (510) 486-5875 SMBENSON@LBL.GOV

Climate variability and carbon management research is a new addition to the Earth Sciences Division. Over the past two years we have added new staff with expertise in global and regional scale climate modeling, marine geochemistry and soil carbon cycling to provide a strong scientific foundation for addressing concerns related to carbon emissions and climate varability. We have also expanded the research focus of our existing staff to include issues related to deciphering the isotopic composition of ice-cores, developing methods for modeling production of methane from hydrate formations, and predicting the reactivity of CO2 in deep geologic formations. Central to making an important scientific contribution in these important areas is a strong link to the newly formed Center for Atmospheric Sciences at UC Berkeley. In addition to recruiting the core staff for this effort and getting our research program under way, we have been successful in establishing a number of major new initiatives. Our regional climate research group competed successfully to establish the NA S A - s p o n s o red Califo rnia Water Resourc e s Regional Atmospheric Sciences Application Center (RESAC).This Center will provide state-of-the-art real-time and forecast information (observation and simulation) on hydroclimate, water quantity and quality, and runoff - related hazards to wa t e r resources managers. It will also provide support for the ongoing

145

regional and national assessment process by improving our understanding of specific regional features of the climate system and its impacts. We have also been successful in establishing DOCS, the DOE Ocean Carbon Sequestration Center, in partnership with Lawrence Livermore National Laboratory, Scripps Institute of Oceanography, Massachusetts Institute of Technology, Rutgers University, Moss Landing Marine Laboratories and the Pacific International Center for High Technology Research (PICHTR). The purpose of DOCS is to conduct, focus and advance the research needed to evaluate and improve the feasibility, effect i veness and env i ronmental acceptability of ocean carbon sequestration. This has been a very exciting two years and we believe that this area will be a growing part of our research activities in the years and decades to come.

Funding Climate Va rability and Carbon Management research is funded by the U.S. Department of Energyâ&#x20AC;&#x2122;s Office of Science, O ffice of Basic Energy Sciences, Department of Engineering and Geosciences; and Office of Biological and Environmental Research; and the National A e ronautics and Space Administration.


Earth Sciences Division

Annual Report 1998-1999

This page is intentionally left blank

HTTP://WWW-ESD.LBL.GOV

146


Earth Sciences Division

Research

Climate Variability and Carbon Management Program

Objectives

Regional Climate Simulation For the Western United States Using the RCSM

Annual Report 1998-1999

features. The Soil-Plant-Snow (SPS) model The main objectives of this calculates land-surface processes in regional climate simulation are to Jinwon Kim and Norman L. Miller the RCSM. The RCSM predicts soil evaluate the Regional Climate System moisture content, soil temperature, Contact: Model (RCSM) and to investigate the wa t e r - e q u i valent snow cover and Jinwon Kim (510) 495-2375, jinwon_kim@lbl.gov hydroclimate and its variability within canopy water content. It also calcuthe we s t e rn United States. Results lates skin temperature and surface from this study will be used to compute the effects of future wetness, and drainage of soil water into deep ground. It needs global climate variation in the western United States. In addition, data for soil texture, vegetation characteristics and initial guesses evaluation of this multi-year simulation (also known as a hindfor soil moisture and temperature. The initial guesses for soil cast) will be used to identify the strengths and weaknesses of the moisture and temperature were obtained from the NCEP global RCSM for future improvements. In parallel with the simulation analysis data. The land-surface and vegetation characteristics experiment, development of a high performance version of the were obtained from satellite-based monthly-mean data for leafRCSM (RCSM.hp) is under way in collab o ration with the area index and green-leaf fraction data.This recent monthly vegNational Energy Research and Scientific Computing Center etation data from the NCEP/NASA has eliminated much uncer(NERSC) at LBNL. tainty and resulted in a significant improvement in the RCSMThe focus of this hydroclimate simulation experiment is to simulated surface energy and water budget. study precipitation, snow budget and soil moisture variations Precipitation and atmospheric forcing simulated by the MAS that significantly affect human activities and natural environand SPS are processed to compute the area-averaged forcing data ment in the western United States. for the RCSM hy d ro l o gic models (spatially - d i s t ributed TO PMODEL and the spatially-lumped Sacramento model) using an Approach are a - m a t ching method. This method pre s e rves the total water/energy between the MAS model and hydrologic models in a way consistent with the basic assumptions of the models. To generate the western U.S. hydroclimate, we downscaled Accurate area-matching is important for quantitative precipitathe global analysis data from the National Center fo r tion simulations in the western United States, where steep terEnvironmental Prediction (NCEP) at 2.5o x 2.5o resolution to a 36 x 36 km2 resolution (Figure 1). This intermediate-resolution rain generates a large spatial gradient in precipitation. climate data will be further downscaled to a 12 x 12 km2 resoThe western U.S. climate simulation starts in 1979 and covers lution for a California-Nevada domain when the RCSM.hp is comthe next 10-15 year period to capture several wet/dry periods in pleted. The NCEP global analysis data, together with the global recent years. Due to an extensive requirement for computational analysis data from the European Center for Medium-Range resources, the actual period of this simulation will be deterWeather Forecasts (ECMWF), is regarded as the most accurate mined by the available computational resources. large-scale data ava i l able. This pre l i m i n a ry study Howeve r, spatial resolution examined the atmospheric of both analysis data is too physics fo rmulations for a coarse for detailed regional long-term climate simu l aimpact assessments. tion. Different schemes for The Mesoscale Atmospheric atmospheric phy s i c s , most Simulation (MAS) model is importantly for cumulus conthe limited-area atmospheric ve c t i o n , in the MAS can model of the RCSM.The MAS cause a significant difference has an accurate adve c t i o n in the simulated regional clischeme and comprehensive mate. The MAS is equipped physics including short- and with two convection longwave r adiation, convecschemes: the simple Anthes tion, cloud microphysics and scheme and the Simplified cloud-radiation intera c t i o n . A ra k awa - S ch u b e rt Scheme The MAS is being used to (SAS). The SAS scheme, even s i mulate re gional hy d ro cl ithough it is more physically mates of the western United based than the Anthes States and eastern Asia. It is scheme, includes many tuncontinu o u s ly improved fo r Figure 1. Domain and orography (height of terrain above mean sea level) of the ing para m e t e rs. Most of better physics and numerical western U.S. domain at 36x36 km2. The color contour intervals are at 250 m. these parameters are empiri147


ESD

Regional Climate Hindcast for the Western United States Using the RCSM

cal and were obtained from observational studies taken in the Great Plains.As the storm e nv i ronment and the resulting storm stru c t u re widely va ry according to location, existing parameters may not be suitable for the western United States. Hence, ri gorous evaluation of model results and sensitivity tests are important for a successful simulation of long-term re gional climate simulation. A preliminary evaluation of the RCSM is presented below.

January and February while the maximum snowmelt was in April and early May. The simulated snow c over wa s totally depleted at the end of June. Such variation of the snowcover is consistent with the observation in which snow remains only at high elevations after June.

Significance of Findings

The agreement betwe e n the observed and simu l a t e d climate fe a t u res in this preResults liminary experiment indicates that the RCSM is capaWe have completed an inible of long-term climate simFigure 2. Observed (solid line) and simulated (bars) daily-mean precipitation over tial evaluation of the RCSM California during the first 150 days of the simulation. ulation. Observed pre c i p i t afor a long-term climate simution events and the amount lation. In this study, we used the SAS scheme with the parameter of precipitation was well- simulated by the MAS. The surfa c e values currently employed by the NCEP for medium-range global s n ow budget by the SPS is also consistent with the observa t i o n . forecasts. U n d e restimation of spring precipitation could be part i a l ly corFigure 2 illustrates a comparison between the observed and rected by using the Anthes convection scheme instead of the SAS scheme. H oweve r, it needs more investigation to select a simulated daily-mean precipitation within California during the c o nvection scheme most suitable for the we s t e rn United States first 150 days of 1979.The observational data was from the co-op re gi o n . Long-term climate and streamfl ow simulation and deve lstations within California. For each month, 400-500 stations were opment of a high-performance ve rsion of the MAS are curre n t ly available for comparison.We interpolated the simulated precipiu n d e r way. tation to match the location of each station, as precipitation varies rapidly over a mountainous terrain. Related Publications The MAS has well-simulated the daily precipitation events in California. Figure 2 shows that eve ry observed precipitation Kim, J., N. Miller, K. Georgakakos and A. Guetter, River flow event is present in the simulated precipitation, except the last response to precipitation and snow budget in California durevent where the MAS has underestimated precipitation.This indiing the 1994-1995 winter, J. Climate, 11, pp. 2376-86, 1998. cates that the dynamic framework of the MAS is capable of accuKim, J., and N. Miller, Hydroclimate modeling of the western rately handling the tendency of the large-scale flows imposed through time-dependent lateral boundary conditions. United States: A hindcast and 2xCO2 impacts, Conference on detection and modeling of regional climate change, Trieste, The amount of precipitation is also generally well-simulated. Italy, June 1999. The agreement between the observed and simulated values are especially good for January and February (Julian days 1-60). Funding Precipitation was underestimated during the spring time from late March to May. This underestimation was partially corrected This work has been supported by the Laboratory Directed in the experiment using the Anthes scheme instead of the SAS R e s e a rch and Development Program of Law rence Berkeley scheme (not shown). National Laboratory under U.S. Department of Energy Contract The simulated snow budget was also consistent with the No. DE-AC03-76SF00098. observations. The simulated maximum snowfall occurred in

http://www-esd.lbl.gov 148


Earth Sciences Division

Climate Variability and Carbon Management Program

Annual Report 1998-1999

The Regional Climate Center Norman L. Miller and Jinwon Kim

(510)

Research

Contact: Norman Miller 486-2374, nlmiller@lbl.gov

• Southwestern streamfl ow simulations with the National Weather Service’s California Nevada River Forecast Center (CNRFC) and U.S. Geological Survey (USGS). • Runoff contaminant monitoring and real-time management in the San Joaquin Basin with the U.S. Bureau of Reclamation and the San Joaquin River Management Program. • Contaminant identification and monitoring from mine sites in the Sierra foothills with the University of California Space Sciences Laboratory, USGS and California Department of Conservation. • Development and testing of dynamic sediment transport and landslide hazards prediction with UC’s Department of Geology and Geophysics. • Identification of flooded central California valley and other sensitive regions via remotely sensed data from the National A e ronautics and Space Administration and the National Oceanic and Atmospheric Administration. • Snow cover area and snow water equivalent maps for California regions with UC Santa Barbara’s Earth Science System Workbench. • Development of a shared information distribution system with DOE/ACPI collaborators. • Annual Southwestern Significant Results Conference with the University of Arizona at Tucson’s Institute for the Study of Planet Earth.

Objectives

The objectives of the Berkeley Lab Regional Climate Center (RCC) are to provide downscaled climate research and applications to a specialized and ge n e ral user community. Understanding climate variability and change at the scale of impacts requires analyzed simulations at a range of scales and comparison to past climate observations.

Approach The core capability of the RCC is the Regional Climate System Model (RCSM). The RCSM (Figure 1) consists of pre- and postprocessors that nest a suite of process models capable of producing hydroclimate products at short-term (two-three day), seasonal and long-term (downscaled 2 x CO2 scenarios) time scales. RCSM output is used for research, climate and weather forecasts, sensitivity analyses and impact assessments. We have a growing group of collaborators (regional, national and international) researching topics related to hydroclimate.The Regional Climate Center is curre n t ly wo rking towa rd the advancement of mesoscale atmospheric simulations, distributed hydrologic simulations, landslide initiation, water quality, and sediment transport. Established partnered RCC collaborations include:

Figure 1. The Regional Climate System Model (RCSM) consists of a pre-processor (input data), process models and post processors (output data and analysis, visulizations and assessments).

149


ESD

The Regional Climate Center

Results Global Energy and Water Cycle. Components of the Berkeley Lab RCSM have been implemented at the Korean Meteorological Administration and the Au s t ralian Department of Natura l Resources. We view these activities as part of a growing climate research and applications collaborative initiative at Berkeley Lab with UC Berkeley, other UC institutions, Department of Energy labs and other government, university and private groups.

During 1998 we have reported on hydroclimate studies in the southwestern United States and eastern Asia (Kim et al., 1998). In 1999, results from ongoing climate research, a new deep groundwater flux parameterization and our northeastern Australian collaboration have been submitted for publication. We simulated California precipitation and streamflow during the 1997-1998 El Nino winter with good model skill for 48-hour forecasts. The experimental seasonal forecasts still require further refinement. The new deep groundwater flux parameter has improved the hydrologic model calibration and verification. Downscaled hindcast (control) and projected double-CO2 simulations have started as part of the U.S. National Assessment Report.

Significance

of

Related

Publications

Kim, J., N. L . Miller, K.P. Georg a k a kos and A.K. Guetter, R i ver fl ow response to precipitation and snow budget in California during the 1994-1995 winter, J Climate, 11, pp. 2378-86, 1998. Miller, N.L. and J. Kim, Optimization and analysis of a deep groundwater flux parameter within a physically-based surface hydrology model: TOPMODEL, Hydrol. Processes, submitted.

Findings

The establishment of the Berkeley Lab Regional Climate Center represents an important advancement for the Berkeley community of climate researchers, as well as the California and southwestern region of climate information users. Our results are part of the U.S. National Assessment Report, the California Assessment Report, and the Southwestern Assessment Report of the U.S. Global Climate Change Research Program, and the Climate Change and California Ecosystems report of the Union of Concerned Scientists and the Ecological Society of America. Through our Eastern Asian project, we are representing hydroclimate issues related to sustainability in China at the U.S./China Water Workshop and the Third International Conference on the

Funding This work has been supported by the Laboratory Directed R e s e a rch and Development Program of Law rence Berkeley National Laboratory under U.S. Department of Energy Contract No. DE-AC03-76SF00098; by the Administrator of the National A e ronautics and Space Administration, NASA Earth Science Enterprise; and by the University of Califo rnia CampusLaboratory Collaboration Hydrology Project.

http://www-esd.lbl.gov 150


Earth Sciences Division

Climate Variability and Carbon Management Program

Annual Report 1998-1999

Climate Change and Wildfire Severity in California Research

Objectives

models. The analysis produced a geographically - s p e c i fic estimate of the Margaret S. Torn, Insurers are acutely aware that 85 potential effect of climate change on Evan Mills and J. Fried* p e rcent of catastro p h e - related paywildfires, including the effectiveness outs are due to natural disasters, with of fire-fighting infrastructure. *Department of Forestry, Michigan claims averaging about $10 billion per To capture some of the complexity State University year worldwide over the past decade of California’s landscape, this study (Mills, 1998). A recent study by the examined three climatically-distinct Contact: Insurance Services Office (ISO), entiand geographically-separated areas of Margaret Torn tled “The Wildland/Urban Fi re northern California: Santa Clara, near (510) 495-2223, mstorn@lbl.gov Hazard,”concluded that wildfires are a San Francisco Bay; Amador-El Dorado, pervasive insurance risk, consuming an average of five million in the Sierra foothills; and Humboldt (on the northern coast acres per year across the United States at an average insured cost (Figure 1).The regions studied contain substantial areas of wildof about $300 million per year. Insurers feel the effects of wildland/urban interface conditions on the margins of the San fire in several ways. Insured property is at risk, and in some Francisco Bay area, the Sacramento metropolitan area and the cases, the costs of fire-fighting or lost timber are underwritten. redwood region’s urban center of Eureka. El Dorado is the fastest Wildfire-related injuries or loss of life also exact a cost from growing county in California and Amador is the sixth-fastest insurers. M o re ove r, in the aftermath of wildfi re , secondary growing county in the state. events,such as landslides, flooding and water quality impairment The analysis was accomplished with an innova t i ve coupling can all impose additional costs.The insured damages are only a of California Department of Fo re s t ry wildfi re models with the component of the total economic loss and do not reflect the full G o d d a rd Institute for Space Sciences GCM, plus site-specifi c human hardship that wildfires can cause. data on actual fi re - s t a rts over a six-year historical peri o d . The The powerful impact that nu m b e r, location and timing climatic anomalies can have of fi re starts was not on wildfire was demonch a n ged in the analy s i s ; strated last year after rather fi re behavior and fi re s u p p ression we re simu l a t e d droughts, linked to El Niño, with present climate (1 x we re fo l l owed by wideCO2) and future climate (2 x spread, devastating fires in CO2) GCM scenarios. Florida, Indonesia and elsewhere.The latest predictions Results suggest that global warming may also create conditions According to our analyses, that intensify wildfire danger climatic change would cause by warming and drying out fi res to spread faster and vegetation and by increasing burn more intensely in two windspeed.This project asks of the three re gions. The the question:what effect will biggest impacts were seen in climate ch a n ge—as pregrassland, where the fastest dicted by climate simulation spread rates already occur. In studies performed by genfo re s t s , w h e re fi res move e ral circulation models much more slowly, modeled (GCMs)—have on the magniimpacts were less severe.The tude of wildfires in northern California? response of chaparral brush and oak woodlands fe l l Approach between that of grass and forest. The reason that faster To explore this question, fuels respond more is that Figure 1. Map of vegetation types and three regions analyzed in this simulator we combined Califo rnia study. This map was created by Jeremy Fried on October 22, 1998, using Fire fire behavior in these fuels is weather and fire data and Management Analysis Zone boundary and attribute data that were prepared for more sensitive to wind validated fire and fire sup- use with the California Fire Economics Simulator, version 1, in the early 1990s. speed, and elevated wind The authors gratefully acknowledge the provision of this data by Jim Spero of pression models and state-of- the California Department of Forestry and Fire Protection's Fire and Resource speed during fire season was the-art ge n e ral circ u l a t i o n Assessment Program. a striking fe a t u re of the 151


ESD

Climate Change and Wildfire Severity in California

ch a n ged climate we a t h e r San Francisco Bay area and data. Summarizing over all the Sacramento metropolivegetation types, predicted tan area as well as forest global warming results in a resources in Humboldt and greater number of fast fires Del Norte counties. Our and fewer slow fires. results showed that global Future ch a n ges in fuel warming predictions lead to moisture and wind speeds a greater risk of wildfi re also cause modeled fires to damage in Califo rnia, by burn with greater intensity, causing more flammable tri g ge ring more intensive fuel conditions and faster suppression efforts. The utifi re spread rates. As indilization of extra fire supprescated by the models, in most sion resources at high discases climate change would patch level, such as air lead to dramatic increases in tankers and bulldozers, can both the land area burned lead to large increases in by California wildfires and suppression costs. In addithe number of potentially Figure 2. Average frequency of escaped wildfires under present and future tion, even increased dispatch catastrophic fi res—more (double CO2) climate scenarios, by region. of the ava i l able fi re-fighting than doubling these losses equipment and personnel could not always compensate for the in some regions. Several important climate-wildfire interactions effect of wa rming on the number of acres burned and the nu mnot currently captured by these models would amplify the ber of “escape” fi res (those that exceed size or time limits, s u ch as expected growth in wildfires. The growth in wildfire damages those that burn more than 300 acres in grasslands; Figure 2). would occur despite deployment of fire suppression resources at The faster, hotter fires caused by climatic change outran fire the highest current level, suggesting that climatic change could suppression and many more acres were burned than in the curcause an increase in both fire suppression costs and economic losses due to wildfires. rent climate scenario in two out of three regions. In the Santa Clara region, for example, contained fires in grass and brush Related Publications burned 41% and 34% more area, respectively, under climate change than they did in the present climate. The number of Mills, E.,The coming storm—global wa rming and risk manageescaped wildfires increased by 53% and 21% in grass and brush. ment, Risk Management, May, pp. 20-27, 1998. In the Sierra foothills, the effect of climatic change was even m o re seve re . H e re , the number of potentially catastro p h i c Torn, M.S., and J.S. Fried, P redicting the impact of global wa rm(escaped) fires predicted went up dramatically—143% more ing on wildfi re , Climatic Change , 21, pp. 257-274, 1992. each year in grassland and 121% more in brush.With the number Torn, M.S., E. Mills and J.S. Fri e d , 1998. Will climate ch a n ge of escaped wildfires more than doubling, climatic change could s p a rk more wildfi re damage? Berkeley Lab report LBNLlead to a serious jump in fire damage in this region. 42592. Climate change had little impact in California’s Humboldt re dFunding wood region,thanks to predictions of a wetter, less windy climate.

Significance

of

Findings

This work has been supported by the U.S. Environmental Protection Agency, Atmospheric Pollution Prevention Division, via the Assistant Secretary for Energy Efficiency and Renewable Energy, Office of Building Technologies and State Community Programs, of the U.S. Department of Energy under Contract No. DE-AC03-76SF00098.

This report describes a geographically specific estimate of the potential effect of climate change on wildfires and the effectiveness of fire-fighting infrastructure in California.The regions studied contain substantial areas of wildland/urban interface near the

http://www-esd.lbl.gov 152


Earth Sciences Division

Research

Climate Variability and Carbon Management Program

Objectives

A Simple Physical Model Of the Global Hydrologic Cycle: Implications for Paleothermometry

Annual Report 1998-1999

tra n s p o rt mechanism. H oweve r, this model is still significantly less complex than GCMs allowing simple and direct tests of climate parameters.

S i g n i ficant env i ronmental policy decisions rest on our ability to predict f u t u re climate conditions. Because Melissa Hendricks, Approach water vapor is one of the most imporDon DePaolo and Ron Cohen tant greenhouse gases it is important Contact: We start with a relatively simple that we understand the global hydroMelissa Hendricks c o n servation statement describing logic cycle and its connection to the (510) 642-9116, mbhendricks@lbl.gov the essential transport processes that earth’s climate. Ice cores contain long affect isotopic fractionation. The concentration of water in the records of the hydrologic cycle through oxygen and hydrogen atmosphere, w, is described by: isotopic ratios (normally expressed as δ18O and δD). Most climate reconstructions assume that isotopic ratios are a unique ∂w p roxy for tempera t u re and use the relationship observe d (1) = ∇ ⋅ (D∇w ) − v∇w + E − P ∂t between δ18O of snow and mean annual temperature in the present climate to extrapolate to temperatures throughout the ice The first term on the right hand side is transport due to eddy (difcore record. However, isotopic ratios may be dependent on clifusive) processes, and the second term is tra n s p o rt due to advecmatic parameters other than temperature, in which case, the tion.The third and fo u rth are source/sink term s : evaporation from present temperature-isotope correlation may not apply to past the ocean and land surfaces (E) and precipitation (P), respectively. climates. The objective of this research is to (1) develop a relaBy treating isotopes as diffe rent species, analogous equations for tively simple model of latitudinal moisture transport and use it to the isotopic ratios are deri ved. evaluate the physical mechanisms that play dominant roles in Isotopic fractionation ultimately is caused by the fact that cold determining δ18O and δD of precipitation, and (2) reevaluate the air holds much less water than warm air. As moisture is transported f rom wa rm to cold areas, precipitation must exceed evaporation isotopic data from polar ice cores in terms of paleotemperature and other climatic parameters. and the net “rainout”of atmospheric moisture causes isotopic fracThe heavy isotopes, 18O and 2H (D), systematically decrease in tionation. Isotopic gradients in precipitation follow temperature abundance in precipitation falling in re gions of decreasing mean gradients, which are primarily a function of latitude. Consequently, a n nual surface tempera t u re because of isotopic fractionation that the δ18O values of precipitation are high near the equator and very l ow at the poles.The simplest description of the earth, therefore, is o c c u rs at every eva p o ration and condensation step in the hy d rol o gic cycle. Heavy isotopes are most depleted in Antarctic pre c i pa one-dimensional version of Equation 1 in spherical polar coordiitation, in which there is also a nates, with the single spatial n e a r ly linear relationship va ri able being latitude. This is between δ18O and surface temvalid for the Southern perature, T. The linear relationHemisphere because Antarcship is typically attributed to tica is a large landmass cenRayleigh distillation, which is a t e red approx i m a t e ly at the simple conceptual model that South Pole. Since we are examining long-term ave rages, we neglects all complexities of also use the steady state a t m o s p h e ric water vapor transapproximation ( w/ t = 0). port. Simple Rayleigh models Equation 1 therefo re implies ove r - p redict the depletion of that three processes control heavy isotopes with decreasing fl u xes and isotopic ratios in temperature. By contrast, ge nprecipitation: eva p o ration eral circulation models (GCMs) from the ocean surface, eddyproduce re a s o n ably good fits to diffusive transport, and advecthe modern δ18O –temperature relationship, but are too comtive transport. plex to clearly elucidate the Results c o n t rolling mechanisms. Our model expands on Rayleigh In the Southern models by including the phy s ical processes of eva p o ra t i ve Figure 1. Model results for present-day climate in the Southern Hemisphere. Hemisphere, at latitudes poleThe model is run twice to find results for transport solely by eddy-fluxes (blue rech a rge during water va p o r line) and solely by advection (red line). Data are from Dahe et al. and ward of 45°S (average annual s u r face tempera t u res less transport and va riations in the Rozanski et al. 153


ESD

A Simple Physical Model of the Global Hydrologic Cycle: Implications for Paleothermometry

than 10°C), eva p o ration ra t e s importance of including the drop significantly and δ 18O and transport mode in atmospheric δD of precipitation decrease models. quickly as air loses moisture due Two important general results to decreasing temperature. are represented in Figure 2.The Because eva p o ration rates are fi rst is the diffe rence in temporal low, the relative amount of diffurelationship between δ18O and local surface temperature for the sive ve rsus advective transport t h ree locations.This implies that d e t e rmines the rate of decrease in a unive rsal relationship between δ18O and δD since transport by eddy-fluxes induces less fractionaisotopic composition and local tion than transport by advection surface tempera t u re does not (Figure 1). δ18O of Antarctic surexist and that the present-day face snow generally falls between spatial relationship is not generthe predicted results for transport ally applicable to paleotemperature estimates.The second is the by eddy-fluxes only and the increased sensitivity to temperaresults for tra n s p o rtby advection, ture changes with distance from the coast. In general, the isotopic indicating that isotopic ratios in precipitation might be useful for d e t e rmining transport mecha- Figure 2. Model results for changes in δ18O as local surface temperature ratio is most sensitive to tempernisms in Southern high latitudes. decreases for three locations in Antarctica. The green zone is the model ature at locations far inland, such change in δ18O at 70°S in Western Antarctica. The red zone is the preResults of sensitivity tests dicted change at the South Pole and the blue zone is for Vostok (coldest, as the South Pole and Vostok, and reveal that isotopic composition most remote location). Model results for the South Pole and Vostok over- is ve ry weakly dependent on of precipitation in polar regions is lap in the gray region. The solid black line is the relationship in modern temperature at coastal locations, precipitation, which has been used to interpret ice core records. very dependent on the amount of such as the Antarctic Peninsula. water vapor over the continent and on the surface temperature. Our results suggest that, at locations other than Vostok, temperaOther climatically driven changes in the va ri ables that affect the tures during the last glacial maximum were substantially lower hydrologic cycle,such as changes in the location of the evaporation than has been previously estimated based on δ18O data and the modern spatial relationships. source, decreases in sea surface temperature or va riations in the total annual precipitation, do not have a large effect on the values Related Publications of δ18O and δD,but do affect the relationship between δ18O and δD in precipitation, which is very accurately known from observaDansgaard, W., Stable isotopes in precipitation, Tellus, 16(4), pp. tions. 436-468, 1964. Significance of Findings Jouzel, J., et al., Validity of the temperature reconstruction from water isotopes in ice cores, Journal of Geophysical Research, Using this model, we estimate changes in δ18O and δD for 102(C12), pp. 26471-88, 1997. changes in global temperature, including temperature changes Dahe, Q., et al., Distribution of stable isotopes in surface snow that are thought to be representative of the transition from the along the route of the 1990 International Trans-Antarctica last glacial maximum to the Holocene. As expected, the model Expedition, Journal of Glaciology, 40(134), pp. 107-18, 1994. shows that significant changes in δ18O and δD have occurred in Rozanski, K., L. Araguás-Araguás and R. Gonfiantini, Isotopic response to changes in the annual average surface temperature Patterns in Modern Global Precipitation, in Climate Change in and in response to changes in the temperature gradient between Continental Isotopic Records, American Geophysical Union, the equator and the South Pole. Model trials with linear changes pp. 1-36, 1993. in the equator-to-Vostok temperature gradient were used to estiFunding mate likely ranges of temporal δO–T relationships at three locations: 70°S in Western Antarctica near the coast, the South Pole This work has been supported by the Lab o ra t o ry Directed and Vostok (the coldest, most remote location in Antarctica R e s e a rch and Development Program of Law rence Berke l ey (Figure 2). The resulting ranges in δ 18O are bounded by pure eddy-diffusive transport for the smallest slope and by advective National Lab o ratory under U.S. D e p a rtment of Energy Contract No. transport for the largest slope, demonstrating again the DE-AC03-76SF00098.

http://www-esd.lbl.gov 154


Earth Sicences Division

Annual Report 1998-1999

EARTH SCIENCES DIVISION PUBLICATIONS

1997-1999

Ahlers, C.F., S. Finsterle and G.S. Bodvarsson, Characterization and prediction of subsurface pneumatic pre s s u re va riations at Yucca Mountain, Neva d a , in Proc. TOUGH Workshop ’98, Berke l ey, Calif., May 4-6, 1998, pp. 222-227, LBNL-41995, Lawrence Berke l ey National Lab o ratory, Berke l ey, Calif., 1998. Ahlers,C.F., S. Finsterle and G.S.Bodvarsson, Characterization and prediction of subsurface pneumatic response at Yucca Mountain, Nevada, J. Contaminant Hydrology – Special Issue, 38, 47-68, 1999. Al Mahamid, I., C.F. Novak, K.A. Becraft, S.A. Carpenter and N. Hakem, Solubility of Np(V) in K-CI-CO3 solutions to high concentrations: measurements and thermodynamic model predictions, Radiochimica Acta, 222,1–9, 1997. Al Mahamid, I., and B.M. Smith, Litera t u re search on the use of resins for treatment of radioactive wastes, LBNL-40939, Lawrence Berke l ey National Lab o ratory, Berke l ey, Calif., 1997. Anderson, S.P., W.E. Dietrich, D.R. Montgomery, R. Torres, M.E. Conrad and K. Loague, Subsurface flow paths in a steep unchanneled catchment, Water Resources Research, 33, 2637-2653, 1997. Bandurraga,T.M., and G.S. Bodvarsson, Calibrating hydrogeologic parameters for the 3-D site-scale unsaturated zone model of Yucca Mountain, Nevada, J. Contaminant Hydro l o gy – Special Issue, 38, 25-46, 1999. Battistelli, A., C. Calore and K. Pruess, The simulator TOUGH2/ EWASG for modeling geothermal reservoirs with brines and non-condensible gas, Geothermics, 26, 437–464, 1997. Becker,A., B.Wang, K.H. Lee and M.Wilt, Subsurface electromagnetic measurement through steel casing, in Expanded Abstracts of the 67th Annual Int’l Meeting, Soc. Exploration

Geophysics, pp. 1965–1968, 1997, LBNL-42375, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1998. Becker, A., K.H. Lee and L. Reginato, Field test of a wideband downhole EM transmitter, LBNL-43762, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1999. Bennett, D.H., A.L. James, T.E. McKone and C.M. Oldenburg, On uncertainty in remediation analysis: variance propagation from subsurface transport to exposure modeling, Reliability Engineering and System Safety, 62,117–129, 1998. Benson, S.M., Capturing and sequestering carbon by enhancing the natural carbon cycle: preliminary identification of basic science needs and opportunities, LBNL-40813, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1997. Benson, S.M., Influence of nitrate on the mobility and reduction kinetics of selenium in groundwater systems, E nv i ron. Chem. Selenium, 22, 437–457, 1997. Benson, S.M., W. Chandler, J. Edmonds, J. Houghton, M. Levine, L. Bates, H. Chum, J. Dooley, D. Grether, J. Logan, G.Wiltsee and L. Wright, Assessment of basic research needs for greenhouse gas control tech n o l o gi e s , in Pro c . 4th International Conference on Greenhouse Gas Control Technologies,Aug.30 – Sept.2, 1998,Interlaken,Switzerland,LBNL-42398,Lawrence Berkeley National Laboratory, Berkeley, Calif., 1998. Birkholzer, J.T., and C.-F.Tsang, Solute channeling in unsaturated heterogeneous porous media, Water Resources Research, 33(10), 2221–2238, 1997. Birkholzer, J.T., and Y.W. Tsang, Pretest analysis of the thermalhydrological conditions of the drift scale test at Yucca Mountain, LBNL-41044, L aw rence Berkeley National Laboratory, Berkeley, Calif., 1997. 155


ESD

Publications 1997-1999

Birkholzer, J.T., G. Li, C.-F. Tsang and Y. Tsang, Numerical experiments on the probability of seepage into underground openings at Yucca Mountain, in Proc. XII International Conference on Computational Methods in Water Resources, C re t e , Greece, June 1998, Computational Mechanics Publications, Boston, Mass., 1998. Birkholzer, J.T., G. Li, C.-F.Tsang and Y.Tsang, Modeling studies and analysis of seepage into drifts at Yucca Mountain, J. Contaminant Hydrology – Special Issue, 38, 349-384, 1999. Bishop, J.K. B., S.E. Calvert and M.Y.Y. Soon, Spatial and temporal variability of POC in the northeast Subarctic Pacific, Deep Sea Research II, in press. Bodvarsson, G.S.,T.M. Bandurraga and Y.S.Wu (eds.),The site-scale unsaturated zone model of Yucca Mountain, Nevada, for the viability assessment, Yucca Mountain Site Characterization Project Milestone SP24UFM4, LBNL-40376, L aw rence Berkeley National Laboratory, Berkeley, Calif., 1997. Bodvarsson, G.S., Geologic disposal of nuclear waste – Progress made and lessons learned, in Proc. Dynamics of Fluids in Fractured Rocks: Concepts and Recent Advances, Berkeley, Calif., Feb. 10-12, 1999, pp. 181-183, LBNL-42718, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1999. Bodvarsson, G.S.,W. Boyle, R. Patterson and D.Williams, Overview of scientific investigations at Yucca Mountain – The potential repository for high-level nu clear wa s t e , J. Contaminant Hydrology – Special Issue, 38, 3-24, 1999. Boitnott, G.N., and A. Kirkpatrick, Interpretation of field seismic tomography at The Geysers Geothermal Field, California, in Proc. 22nd Workshop on Geothermal Reservoir Engineering, Stanford, Calif., SGP-TR-155, pp. 391–398, 1997. Borglin, S., and G.J. Moridis, E x p e rimental studies of magnetically d ri ven fl ow of fe rrofluids in porous media, LBNL-40126, Lawrence Berke l ey National Laboratory, Berkeley, Calif., 1998. Borglin, S., G.J. Moridis and A. Becker, Magnetic detection of ferrofluid injection zones, LBNL-40127, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1998. Borglin, S., G.J. Moridis and C.M. Oldenburg, On magnetic fluid emplacement: laboratory experiments of ferrofluid flow, LBNL-42203, L aw rence Berke l ey National Lab o ra t o ry, Berkeley, Calif., 1998. Chang, F.-R.C., NT. Skipper and G. Sposito, Monte Carlo and molecular dynamics simulations of interfacial structure in lithium-montmorillomite hy d ra t e s , Langmuir, 13(7), 2074–2082, 1997. Chang, F.-R.C., N.T. Skipper and G. Sposito, Monte Carlo and molecular dynamics simulations of electrical double layer structure in potassium-montmorillonite hydrates, Langmuir, 14(5), pp. 1201–1207, 1998. Chang, F.-R.C., N.T. Skipper, K. Refson, J.A. Greathouse and G. Sposito, Interlayer molecular structure and dynamics in Li-, Na-, and K-montmorillonite-water systems, in Mineral-Water Interfacial Reactions, Kinetics and Mechanisms (D.L. Sparks and T.J. Grundl, eds), 715, 88-106,American Chemical Society, Washington, D.C., 1999.

Cheney, M.A., J.Y. Shin, D.E. Crowley, S. Alvey, N. Malengreau and G. Sposito, Atrazine dealkylation on a manganese oxide surface, Colloids and Surfaces, 137, 267-273, 1998. Cohen, A.J.B., and A.M. Simmons, Development of the sub-sitescale three-dimensional saturated zone flow model: evaluation of flow processes,Yucca Mountain Site Characterization Project Milestone SP33SM4, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1997. Cohen, A.J.B., Analysis of pressure disturbances in unsaturated rock from installation of new boreholes in Alcove 4, ESF, LBNL-41484, L aw rence Berke l ey National Lab o ra t o ry, Berkeley, Calif., 1998. Cohen, A.J.B., Simulation of pumping tests to characterize faults at Yucca Mountain, Nevada, LBNL-42084, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1998. Cohen,A.J.B., and C.M. Oldenburg, Effects of faulted stratigraphy on saturated zone flow beneath Yucca Mountain, in Proc. TOUGH Workshop ‘98, Berkeley, Calif., May 4–6, 1998, pp. 95–100, LBNL-41995, L aw rence Berkeley National Laboratory, Berkeley, Calif., 1998. Cohen, A.J.B., C.M. Oldenburg, A.M. Simmons, A.K. Mishra and J. Hinds, S4Z: Sub-site-scale saturated zone model for Yucca Mountain, in Pro c . Int’l High-Level Waste Management Conference and Exposition, Las Vegas, Nev., May 11–14, 1998, LBNL-41773, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1998. Cohen,A.J.B., and C.M. Oldenburg, Effects of faulted stratigraphy on saturated zone flow beneath Yucca Mountain, Nevada, in Proc. Field Testing and Associated Modeling of Potential High-Level Nuclear Waste Geologic Disposal Site (FTAM) Conference, Berkeley, Calif., Dec. 15-16, 1997, pp. 93-96, LBNL-42520, L aw rence Berke l ey National Lab o ra t o ry, Berkeley, Calif., 1998. Cohen,A.J.B.,Three-dimensional numerical modeling of the influence of faults on groundwater flow at Yucca Mountain, Nevada, Ph.D Thesis, LBNL-43377, L aw rence Berkeley National Laboratory, Berkeley, Calif., 1999. Cohen, D., T.W. Patzek and C.J. Radke, Network model of foam trapping and mobilization,Transport in Porous Media, 1233, 1-32, 1997. Cohen, D.,T.W. Patzek and C.J. Radke, Onset of mobilization and the fraction of trapped foam in porous media, Transport in Porous Media, 28, 253–284, 1997. Compton, J.S., M.E. Conrad and T.W. Vennemann, Stable isotope evolution of volcanic ash layers during diagenesis of the Miocene Monterey Formation, California, Clays and Clay Minerals, 47(1), 84-95, 1999. Conrad, M.E., P.F. Daley, M.L. Fischer, B.B. Buchanan,T. Leighton and M. Kashgarian, Combined 14C and δ13C monitoring of in situ biodegradation of petroleum hy d rocarbons, Environmental Science Tech n o l o gy., 31(5), 1463–1469, 1997. Conrad, M.E., A.S. Templeton, P.F. Daley and L. Alvarez-Cohen, Isotopic evidence for biological controls on migration of petroleum hydrocarbons, Org. Geochem., in press. 156


Earth Sicences Division

Annual Report 1998-1999

Conrad,M.E., D.Thomas, S. Flexer and T.Vennemann, Fluid flow and water-ro ck interaction in the East Rift Zone of Kilauea Volcano, Hawaii, J. Geophys. Res., 102(B7), 15021–15035, 1997. Conradson, S.D., I. Al Mahamid, D.L. Clark, N.J. Hess, E.A. Hudson, M.P. Neu, P.D. Palmer,W.H.Runde and C.D.Tait,Oxidation state d e t e rmination of plutonium aquo ions using x-rayab s o r p t i o n spectroscopy, Polyhedron, 17(4), 599–602, 1998. C o n s t able, S.,A. Orange, G.M. Hove rsten and H.F. Morri s o n , Marine magnetotellurics for petroleum exploration, Pa rt 1:A sea-floor equipment system, Geophysics, 63(3), 816–825, 1998. Cox, B.L.,A. Sweet and E.L. Majer,Application of electrical methods to measure microbial activity in soils: preliminary microcosm results, LBNL-41047, Lawrence Berke l ey National Laboratory, Berkeley, Calif., 1997. Daley,T. M., Single well seismic imaging in a deep borehole using a piezoelectric orbital vibra t o r, LBNL-42673, Law re n c e Berkeley National Laboratory, Berkeley, Calif., 1997. Daley, T.M., Single well seismic imaging acquisition tests: N ovember 1997 at Bayou Choctaw Site, LBNL-42672, Lawrence Berke l ey National Lab o ratory, Berke l ey, Calif., 1998. Daley, T.M., Borehole source comparison – Vertical hydraulic vibrator and orbital vibrator using multi-component crosswell data, LBNL-44022, L aw rence Berke l ey National Laboratory, Berkeley, Calif., 1999. DeFlaun, M.F., C.J. Murray,W. Holben,T. Scheibe,A. Mills,T. Ginn,T. Griffin, E.L. Majer and J.L. Wilson, Preliminary observations on bacterial transport in a coastal plain aquifer, FEMS Microbiol. Rev., 20, 473–487, 1997. Dickens, G., and B.M. Kennedy, Noble gases in hydrate from the Black Ridge, in Proc. Ocean Drilling Program, Scientific Results, 164, in press. Dicke, M., Seismicity and crustal structure at the Mendocino Triple Junction, Northern California, B.A. Thesis, LBNL42723, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1998. Dodson,A., B.M. Kennedy and D.J. DePaolo, Helium and neon isotopes in the Imnaha Basalt, Columbia River Basalt Group: Evidence for a Yellowstone plume source, Earth Planet. Sci. Lett., 150, 443–451, 1997. Dodson, A., D.J. DePaolo and B.M. Kennedy, Helium isotopes in lithospheric mantle: Evidence from tertiary basalts of the Western U.S., Geochimica, 62, 3775-3788, 1998. Dong, Q., and J.W. Rector, Bi-domain modeling of borehole wave propagation, in Expanded Abstracts with Au t h o r ’s B i o graphies, Society of Exploration Geophysicists Int’l Exposition and 67th Annual Meeting, Dallas,Texas, Nov. 2–7, 1997, pp. 262–265, 1998. Doughty, C., Numerical modeling of field tests in unsaturated fractured basalt at the Box Canyon site, in Proc. TOUGH Workshop ‘98, Berkeley, California., May 4–6, 1998, pp. 210–215, LBNL-41995, L aw rence Berke l ey National Laboratory, Berkeley, Calif., 1998. Doughty, C., Mathematical modeling of a ponded infiltration test in unsaturated fractured basalt at Box Canyon, Idaho, LBNL-

40630, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1998. Doughty, C., Investigation of conceptual and nu m e rical appro a ches for evaluating moisture, gas, chemical, and heat transport in fractured unsaturated rock, J. Contaminant H y d ro l o gy – Special Issue, 38, 69-106, 1999. Doughty, C., C.M. Oldenburg and A.L. James, Site S-7 VOC transport modeling for the vadose zone monitoring system (VZMS), McClellan AFB – 1999 semi-annual report, LBNL-43526, Lawrence Berke l ey National Lab o ra t o ry, Berke l ey, Calif., 1999. Evans, W.C., M.L. Sorey, R.L. Michel, B.M. Kennedy and L.J. Hainsworth, Gas–water interaction at Mammoth Mountain, California, USA, in Proc. 9th Int’l. Symposium on Water–Rock Interaction – WRI-9,Taupo, New Zealand, pp. 443–446, 1998. Fan, J., Overlap domain decomposition technique for modeling wave propagation, Ph.D. Thesis, LBNL-42881, L aw re n c e Berkeley National Laboratory, Berkeley, Calif., 1999. Faulques, E., J.Wery, B. Dulieu, C. Seybert and D.L. Perry, Synthesis, fabrication, and photoluminescence of CaF2 doped with rare earth ions, J. Fluores, 8, 283, 1998. Faybishenko, B., T.R. Wood, T.M. Stoops, C. Doughty and J. Jacobsen, A conceptual model of tracer transport in fractured basalt: Large scale infiltration test revisited, in Proc. 1997 GSA Annual Conference, Salt Lake City, Utah, 1997. Faybishenko, B., C o m p a rison of lab o ratory and field methods for determination of unsaturated hydraulic conductivity of soils (to appear in Proc. I n t e rnational Conference Characterization and Measurement of the Hydraulic Properties of Unsaturated Porous Media) LBNL-42022, Lawrence Berke l ey National Laboratory, Berkeley, Calif., 1998. Faybishenko, B., C. Doughty, J. Geller, S. Borglin, B.L. Cox, J.E. Peterson, Jr., M. Steiger, K.Williams,T.Wood, R. Podgorney,T. Stoops, S. Wheatcraft, M. Dragila and J. Long, A chaoticdynamical conceptual model to describe fluid flow and contaminant tra n s p o rt in a fra c t u red vadose zone, 1997 progress report, LBNL-41223, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1998. Faybishenko, B., P. Holland, M. Mesa, D. Burgess, C. Knutson and B. Sisson, Lithological conditions at the Box Canyon Site: Results of drilling, coring and open borehole measurements, 1995–1997 data report, LBNL-40182, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1998. Faybishenko, B., R. Salve, P.T. Zawislanski, C. Doughty, K.H. Lee, P. Cook, B. Freifeld, J. Jacobsen, B. Sisson, J. Hubbell and K. Dooley, Ponded infiltration test at the Box Canyon Site: Data re p o rt and pre l i m i n a ry analy s i s , LBNL-40183, L aw re n c e Berkeley National Laboratory, Berkeley, Calif., 1998. Faybishenko, B., and S. Finsterle, On the physics of tensiometry in fractured rocks, LBNL-43864, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1999. Faybishenko, B., P.A. Witherspoon, C. Doughty and J.T. Geller, Multi-scale investigations of liquid flow in a fractured basalt vadose zone, LBNL-42910, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1999. 157


ESD

Publications 1997-1999

Fay b i s h e n ko, B. (ed.), Proc., International Symposium on Dynamics of Fluids in Fractured Rocks: Concepts and Recent advances, LBNL-42718, L aw rence Berke l ey National Laboratory, Berkeley, Calif., 1999. Faybishenko, B., C. Doughty, M. S t e i ge r, J.C.S. Long, T. Wood, J. Jacobsen, J. Lore and P.T. Zawislanski, Conceptual model of the ge o m e t ry and physics of water flow in a fra c t u re d basalt vadose zone: B ox Canyon Site, Idaho, LBNL-42925, Law rence Berke l ey National Lab o ra t o ry, Berke l ey, C a l i f. , 1999. Feighner, M.A.,T.M. Daley and E.L. Majer, Results of vertical seismic profiling at Well 46-28, Rye Patch geothermal field, Pershing County, Nevada, LBNL-41800, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1998. Feighner, M.A.,T.M. Daley, R. Gritto and E.L. Majer, Report on CDP reflection image and subsurface structure along lines B and C: UZ-16 VSP, Yucca Mountain Project Report, LBNL-40926, L aw rence Berkeley National Lab o ra t o ry, Berke l ey, Calif., 1997. Feighner, M.A.,T.M. Daley, R. Gritto and E.L. Majer, Results of VSP analysis in P#1, Yucca Mountain Project Level-4 Milestone, SP3B6AM4, LBNL-41799, L aw rence Berke l ey National Laboratory, Berkeley, Calif., 1998. Finkel, R.C., and K. Nishiizumi, Beryllium 10 concentrations in the Greenland Ice Sheet project 2 ice core from 3-40 ka, J. Geophys. Res., 102(26), 699–26,706, 1997. Finsterle, S., and B. Faybishenko, Inverse modeling of a radial multistep outfl ow experiment for determining unsaturated hydraulic properties, Advances in Water Resources, 22(5), 431-444, 1997. Finsterle, S., ITOUGH2 command reference, Version 3-1, LBNL40041, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1997. Finsterle, S., ITOUGH2 sample problems, LBNL-40042, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1997. Finsterle, S., C.M. Oldenburg, A.L. James, K. Pruess, and G.J. Moridis, Mathematical modeling of permeation grouting and subsurface barrier perfo rmance, in Pro c . 1997 Int’l Containment Technology Conference and Exhibition, St. Petersburg, Florida, Feb. 9-12, 1997, pp. 438–444, Florida State University, Tallahassee, Fla., 1997. Finsterle, S., and K. Pruess, Development of inverse modeling t e chniques for ge o t h e rmal applications, in Proc. DOE Geothermal Program Review 15,DOE/EE-0139, pp. 2-47 to 254, U.S. Dept. of Energy,Washington, D.C., 1997, LBNL-40039, L aw rence Berkeley National Lab o ra t o ry, Berke l ey, Calif., 1997. Fi n s t e r l e , S., K. Pruess, D.P. B u l l i vant and M.J. O ’ S u l l i va n , Applications of inverse modeling to geothermal reservoir simulation, in Pro c . 22nd Workshop on Geothermal Reservoir Engineering, Stanford, Calif., 1999, pp. 309–316, SGP-TR-155, LBNL-39869, L aw rence Berke l ey National Laboratory, Berkeley, Calif., 1997.

Finsterle, S., and B. Faybishenko,What does a tensiometer measure in fra c t u red ro ck? in Pro c . I n t ’l Wo rkshop on C h a ra c t e rization and Measurement of the Hydra u l i c Properties of Unsaturated Porous Media, Riverside, Calif., Oct. 22–24, 1997, LBNL-41454, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1998. Finsterle, S., ITOUGH2 V3.2 verification and validation report, LBNL-42002, L aw rence Berke l ey National Lab o ra t o ry, Berkeley, Calif., 1998. Finsterle, S., and J. Najita, Robust estimation of hydrogeologic model parameters,Water Resources Research, 34(11), 29392947, 1998. Finsterle, S., A.L. James, J.S.Y. Wang, J.T. Fabryka-Martin, L.E. Wolfsberg, A. Flint, and W.R. Guertal, Model prediction of local plume migration from the cross drift,Yucca Mountain Project Milestone SP33S3M4, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1998. Finsterle, S., Multiphase inverse modeling: An overview, in Proc. Geothermal Program Review XVI, Berkeley, Calif., April 1–2, 1998, LBNL-41638, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1998. Finsterle, S., Pa rallelization of ITOUGH2 using PVM, LBNL-42261, Lawrence Berke l ey National Laboratory, Berkeley, Calif., 1998. Finsterle, S., C. Satik and M. Guerrero, Analysis of boiling ex p e riment using inve rse modeling, in Proc.TOUGH Workshop ‘98, Berkeley, Calif., May 4–6, 1998, pp. 281–287, LBNL-41995, Lawrence Berke l ey National Laboratory, Berkeley, Calif., 1998. Finsterle, S . ,T.O. Sonnenborg and B. Faybishenko,I nve rse modeling of a multistep outflow ex p e riment for determining hysteretic hy d raulic properties, in. Pro c .TOUGH Workshop ‘98, May 4–6, 1998, Berkeley, Calif., pp. 250–256, LBNL-41995, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1998. Finsterle, S., ITOUGH2 user's guide, LBNL-40040, L aw re n c e Berkeley National Laboratory, Berkeley, Calif., 1999. Finsterle, S., and K. Pruess, Automatic calibration of geothermal reservoir models through parallel computing on a workstation cl u s t e r, LBNL-42384, L aw rence Berke l ey National Laboratory, Berkeley, Calif., 1999 Finsterle, S., G. Bjornsson, K. Pruess and A. Battistelli, Evaluation of geothermal well behavior using inverse modeling, presented at the International Symposium on Dynamics of Fluids in Fractured Rocks: Concepts and Recent Advances, Berkeley, Calif., Feb. 10-12, 1999, also to appear in AGU Monograph, LBNL-43331, L aw rence Berke l ey National Lab o ra t o ry, Berkeley, Calif., 1999. Finsterle, S., and J. Najita, Robust estimation of hydrogeologic model parameters,Water Resources Research, in press. Frangos,W., and A. Becker, Magnetic field of AM band radio broadcast signals at the Richmond Field Station, LBNL-42654, L aw rence Berkeley National Lab o ra t o ry, Berke l ey, Calif., 1998. Frangos,W., and S.Ter-Saakian, Resistivity and induced polarization survey at a Russian nuclear waste site, Geophysics, in press.

158


Earth Sicences Division

Annual Report 1998-1999

Fung, P., H. Chen, G.G.Touchard and C. J. Radke, A nonlinear corrosion double-layer model for laminar flow electrification of hydrocarbon liquids in long metal pipes, J. Electrostatics, 40–41, 445–453, 1997. Gabuda, S.P., S.G. Kozlova,V.V.Terskikh, C. Dybowski, G. Neue and D.L. Perry, 207Pb NMR study of novel Pb-Pb chemical bonding in lead monoxides, -PbO and - PbO, Chemical Physics Letters, 305, 353, 1999. Geller, J.T., Laboratory studies of groundwater degassing in replicas of natural fractured rock for linear flow geometry, LBNL41386, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1998. Geller, J.T., H.-Y. Holman, G. Su, T.-S. Liou, M.S. Conrad, K. Pruess and J.C. Hunter-Cevera, Flow dynamics and potential for biodegradation of organic contaminants in fractured rock vadose zones, LBNL-42587, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1998. Geller, J.T., M.B. Kowlasky, P.K. Seifert and K.T. Nihei, Acoustic detection of immiscible liquids in unconsolidated sand, LBNL-42791, L aw rence Berke l ey National Lab o ra t o ry, Berkeley, Calif., 1999. Greathouse, J., and G. Sposito, Monte Carlo and molecular dynamics studies of interlayer structure in Li(H20)3-smectites, J. Phys. Chem., B102, 2406–2414, 1998. Gritto, R.,V.A. Korneev, and J.E. Peterson,The effect of source and receiver coupling on travel time and amplitude inversion results, EOS Trans., 78(46), 462, 1997. Gritto, R., V.A. Korneev, and L.R. Johnson, Nonlinear 3-dimensional inversion of low-frequency scattered elastic waves, Pure and Applied Geophysics, 155, 1-33, 1999. Gritto, R.,A.E. Romero and T.M. Daley,VSP analysis at Long Valley Caldera, Eastern California, LBNL-39108, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1999. Gritto, R., V.A. Korneev and L.R. Johnson, Inversion of scattered waves for material properties in fractured rock, LBNL-39109, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1999. Grossenbacher, K., K. Karasaki and D. Bahat, Curved scanline theory, Math. Geol., 29(5), 629–651, 1997. Gu, B., K.T. Nihei and L.R. Myer, Numerical investigation of fracture interface waves, J.Acoust. Soc., 102(1), 120–127, 1997. Hakem, N., I. Al Mahamid, J. Apps and G. Moridis, Sorption of cesium and strontium on Savannah River soils impregnated with colloidal silica, presented at 1997 International Containment Tech n o l o gy Conference & Exhibition, St. Petersburg, Fla., Feb. 9–12, 1997, LBNL-39498, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1997. H a u k wa , C.B., Y.S. Wu and G.S. B o d va rsson, Thermal loading studies using the Yucca Mountain unsaturated zone model, J. Contaminant Hydro l o gy – Special Issue, 38, 217-255, 1999. Hildenbrand,K., and T.N. Narasimhan,Aquifer characterization by passive monitoring: A case study, LBNL-41086, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1997.

Holman, H.-Y.N., Y. Mao and R. Goth-Goldstein, Bioavailability of organic compounds from solid environmental matrices, in Proc. 18th SETAC Symposium, 335, Society of Environmental Toxicology and Chemistry, Pensacola, Fla., 1997. Holman, H.-Y.N., D.L. Perry and J.C. Hunter-Cevera, Surfaceenhanced infra red ab s o r p t i o n - re flectance (SEIRA) microspectroscopy – A chemical/biological probe for bacteria localization in geologic materials, J. Microbiol. Methods, 34(1), 59, 1998. Holman, H.-Y.N., D.L. Perry, M.C. Martin and W.R. McKinney, Applications of synchrotron infrared microspectroscopy to the study of inorganic–organic interactions at the bacterialm i n e ral interface, Application of Synch ro t ron Radiation Techniques to Materials Sciences, MRS Symp. Series, 54, in press. Hoversten, G.M., H.F. Morrison and S. Constable, Marine magnetotellurics for petroleum exploration, Part 2: Numerical analysis of subsalt resolution, Geophysics, 63(3), 826–840, 1998. Huang, K., Y.W. Tsang and G.S. Bodvarsson, Simultaneous inversion of air injection tests in fractured unsaturated tuff at Yucca Mountain, Water Resources Research, 35(8), 23752386, 1999. Huang, P., G.Yang, L.R. Myer and N. Cook, Simulation of capillary hysteresis in drainage and imbibition process, Int. J. Rock Mech. & Min. Sci., 34(3–4), 135, 1997. Hubbard, S.S., J.E. Peterson, E.L. Majer, P.T. Zawislanski, J. Roberts and K. Williams, Estimation of permeable pathways and water content using tomographic radar data, The Leading Edge, 16(11), 1623–1628, 1997. Hubbard, S.S.,Y. Rubin and E.L. Majer, Estimation of hydrological model parameters from spectral analysis of high-resolution geophysical data, EOS Trans., 78(48), American Geophysical Union,Washington, D.C., 1997. H u bb a rd, S.S., Y. Rubin and E.L. Majer, Ground-penetratingradar-assisted saturation and permeability estimation in bimodal systems, Water Resources Research , 3 3 ( 5 ) , 971–990, 1997. Hubbard, S.S.,Y. Rubin and E.L. Majer, Estimation of hydrogeological parameters and their spatial correlation structure using geophysical methods, in Groundwater Quality: Remediation and Protection (E.M. Herveret and K. Kovar, eds), IAHS Publication 250, IAHS Press,Wallingford, UK, 1998. Hubbard, S.S., Stochastic characterization of hydrogeologic properties using geophysical data, Ph.D. Thesis, LBNL-42557, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1998. Hubbard,S.S.,Y. Rubin and E.L.Majer, Spatial correlation structure estimation using ge o p hysical and hy d ro ge o l o gical data, Water Resources Research, 35(6), 1809-1825, 1999. Hunter-Cevera, J.C., The value of microbial diversity, Current in Microbiology, 1999, in press. Hunter-Cevera, J.C., and A. Belt, Isolation of cultures, industrial microbiology and biotechnology,ASM Press, in press. 159


ESD

Publications 1997-1999

Ingram, B.L., P. De Deckker, A.R. Chivas, M.E. Conrad, and A.R. Byrne, Stable isotopes, Sr/Ca, and Mg/Ca in biogenic carbonates from Petaluma Marsh, Northern Califo rnia, USA, Geochimica et Cosmochimica Acta, 62(19/20), 3229-3237, 1998. James, A.L., and C.M. Oldenburg, Linear and Monte Carlo uncertainty analysis for subsurface contaminant transport simulation,Water Resources Research, 33(11), 2495–2508, 1997. James, A.L., and C.M. Oldenburg, Enhanced data analysis for the VZMS: Conceptual model design and initial application for the vadose zone monitoring system (VZMS), McClellan AFB,1998 semi-annual re p o rt, LBNL-41909, L aw rence Berkeley National Laboratory, Berkeley, Calif., 1998. James, A.L., and C.M. Oldenburg, Site S-7 representative model and application for the vadose zone monitoring system (VZMS), McClellan AFB, LBNL-42643, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1998. Kaelin, B., and L.R. Johnson, Dynamic composite elastic medium theory. Part I. One-dimensional media, J. Appl. Phys., 84(10), 5451-5457, 1998. Kaelin, B., and L.R. Johnson, Dynamic composite elastic medium theory. Part II. T h ree-dimensional media, J. Ap p l . Phys., 84(10), 5458-5468, 1998. Kaelin, B., Seismic imaging of the shallow subsurface with high frequency seismic measurements, Ph.D. Thesis, LBNL-42058, Lawrence Berke l ey National Laboratory, Berkeley, Calif., 1998. Kaelin, B., and L.R. Johnson, Using seismic crosswell surveys to determine the aperture of partially water saturated fractures, Geophysics, accepted. Kennedy, B.M., and D.J. DePaolo, Isotopic and noble gas geoch e mistry in ge o t h e rmal research,in Proc.DOE Geothermal Rev i ew 15, San Francisco, Calif., M a rch 25–26, 1997, DOE/EE- 0139, pp. 2-7 to 2-13, U.S. Dept. of Energy,Washington, D.C., 1997. Kennedy, B.M.,Y.K. Kharaka,W.C. Evans, A. Ellwood, D.J. DePaolo, J. Thordsen, G.Ambats and R.H. Mariner, Mantle fluids in the San Andreas fault system, California, Science, 278, 1278-1281,1997. Kennedy, B.M.,A. Ellwood, D.M. DePaolo,Y.K. Kharaka,W.C. Evans and J. Thordsen, Isotopic evidence for mantle fluids in the San Andreas fault, EOS Trans., 79, S223,American Geophysical Union,Washington, D.C., 1998. Kennedy, B.M., C. Janik, D. Benoit and D. L. Shuster, Natural ge ochemical tra c e rs for injectate fluids at Dixie Va l l ey,in Proc.24th Workshop on Geothermal Reservoir Engineering,Stanford, Calif., Jan. 25-27, 1999, pp. 108-115, 1999. Kharaka,Y.K., B.M. Kennedy, J.J.Thordsen and W.C. Evans, Role of fluids in the dynamics of the San Andreas fault system, California, USA, in Proc. Int’l Conference on Fluid Evaluation, Migration, and Interaction with Rocks, Geofluids Belfast, N. Ireland, pp. 107–110, 1997. Kharaka, Y.K., J.J. Thordsen, W.C. Evans and B.M. Kennedy, Fluids and faults: The chemistry, origin and interactions of fluids associated with the San Andreas fault system, California, USA, in Proc. 9th Int. Symposium on Water–Rock Interaction – WRI-9,Taupo, New Zealand, pp. 781–784, 1998.

Kim, H.J., Y. Song and K.H. Lee, High-frequency electromagnetic inversion for a dispersive layered earth, Geomagnetism and Geoelectricity, 49, 1439–1450, 1997. Kim, J., N.L. Miller, A.K. Guetter and K.P. Georgakakos, River flow response to precipitation and snow budget in California during the 1994-95 winter,Journal of Climate,11,2376-2386,1998. Kim, J., Eastern Asian hydrometereorology simulation using the re gional climate system model, Global and Planetary Change, 19, 225-240, 1998. Kirkpatrick,A., J.E. Peterson, Jr., and E.L. Majer,Three-dimensional compressional and shear-wave seismic velocity models for the Southeast Geysers, in Pro c . 22nd Workshop on Geothermal Reservoir Engineering, Stanford, Calif., January 27–29, 1997, SGP-TR-155, pp. 399–410, Stanford University, Stanford, Calif. 1997. Kneafsey,T.,and K.Pruess,Experimental studies of va p o rizing flows in unsaturated fractures, EOS Trans., 78(46), F308 (supplement),A m e rican Geophysical Union,Washington, D.C., 1997. Kneafsey,T., and K. Pruess, Preferential flow paths and heat pipes: Results from laboratory experiments on heat-driven flow in natural and artificial rock fractures, LBNL-40467, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1997. Kneafsey,T., and K. Pruess,Thermohydrological laboratory tests – Insights into processes and behavior, in Proc. Int’l High-Level Radioactive Waste Management Conference, Las Vegas, Nev., pp. 261–263,American Nuclear Society, 1997. Kneafsey,T., and K. Pruess, Preferential flow paths and heat pipes: Continued laboratory experiments on heat-driven flow in natural and artificial rock fractures and scaling relationships, LBNL-42262, L aw rence Berke l ey National Lab o ra t o ry, Berkeley, Calif., 1998. Kneafsey, T., and K. Pruess, Laboratory experiments on heatdriven two-phase flows in natural and artifical rock fractures,Water Resources Research, 34(12), 3349-3367, 1998. Korneev, V.A., K.T. Nihei and L.R. Myer, Nonlinear interaction of plane elastic wave s , LBNL-41914, L aw rence Berkeley National Laboratory, Berkeley, Calif., 1998. Korneev, V.A., Time domain solutions for nonlinear elastic 1-D plane wave propagation, LBL-37411, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1998. Korneev, V.A., and L.R. Johnson, Attenuation and fluctuations of elastic waves due to random scattering from inclusions, LBNL-43341, L aw rence Berke l ey National Lab o ra t o ry, Berkeley, Calif., 1999. Kov s c e k , A.R., T.W. Patzek and M. Johnston, E valuation of rock/fracture interactions during steam injection through vertical hy d ro f ra c t u re s , SPE Production and Fa c i l i t i e s , 100–105, 1997. Kovscek,A.R.,T.W. Patzek and C.J. Radke, Mechanistic foam flow simulation in heterogeneous and multidimensional media, SPE 39102, SPEJ, 2(4) 511–526, 1997. Lee, K.H., High-frequency electric field measurement using toroidal antenna, LBNL-39894, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1997. 160


Earth Sicences Division

Annual Report 1998-1999

Lee, K.H., H.J. Kim and Y. Song, E l e c t ro m agnetic method for a n a lyzing the pro p e rty of steel casing, LBNL-41525, Law rence Berke l ey National Laboratory, B e rke l ey, C a l i f. , 1998. Lee, S.K., H.J. Kim,Y. Song, H.F. Morrison and K.H. Lee,A new formulation of magnetic field integral equation for 3-D EM modeling, LBNL-43940, L aw rence Berkeley National Laboratory, Berkeley, Calif., 1999. Li, J., and R.P. Srivastav, Computing the singular behaviour of solutions of cauchy singular integral equations with coefficients, Applied Mathematical Letters, 15(3), 57-62, 1997. Lippmann, M.J.,A.H.Truesdell and H. Gutierrez Puente,What will a 6 km deep well at Cerro Prieto find? in Proc. 22nd Workshop on Geothermal Reservoir Engineering, Stanford, Calif., Jan. 27–29, 1997, SGP-TR-155, pp. 19–28, Stanford Univ., Stanford, Calif., 1997. Liu, H.H., C. Doughty and G.S. Bodvarsson, An active fracture model for unsaturated flow and transport in fractured rocks, Water Resources Research, 34, 2633–2646, 1998. Liu, H.H., G.S. Bodvarsson and L. Pan, Determination of particle transfer in random walk particle methods for fractured porous media,Water Resources Research, in press. Lundy, D.Z., J.C. Hunter-Cevera and G.J. Moridis, Susceptibility of polysiloxane and colloidal silica to degradation by soil microorganisms, LBL-38024, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1997. Majer, E.L., R. Gritto, T.M. Daley, V. A. Korneev, M.A. Feighner and J.E. Peterson, Full scale tomographic imaging of the potential repository hori z o n , Milestone Report SP3B2FM4, LBNL42360, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1998. Majer, E.L., L.R. Johnson D.W. Vasco and P. Parker, Milestone SP3B6DM4 – Results of gravity modeling of the Paleozoic Basement, LBNL-42578, L aw rence Berkeley National Laboratory, Berkeley, Calif., 1998. McCauley, S.E., and D.J. DePaolo, The marine 87Sr/86Sr and δ180 records, Himalayan alkalinity fluxes and Cenozoic climate models,in Tectonic Uplift and Climate Change (W.F.Ruddiman, ed.), pp. 427–467, Plenum Press, New York, N.Y., 1997. McCullough, J., T.C. Hazen and S.M. Benson, Bioremediation of metals and radionuclides: What it is and how it works, LBNL42595, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1999. Miller, N.L., and J. Kim, The regional climate system model, in Mission Earth: Modeling and Simulation for a Sustainable Global System (M.C. Clymer and C.R. Mechoso, eds) pp. 43–52, Pub. Soc. Comp. Sim., 1997. Miller, N.L., J. Kim and R. Hartman, Downscaled climate and s t re a m fl ow study of the Southwe s t e rn United States, American Water Resources Association, in press. Mizunaga, H., K.H. Lee and H.J. Kim, Three-dimensional electromagnetic modeling in the Laplace domain, LBNL-42677, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1998.

Moore, J.N., D.I. Norman, B.M. Kennedy and M.C. Adams, Origin and chemical evolution of The Geysers, C a l i fo rnia, hydrothermal fluids: Implications from fluid inclusion gas compositions, Geothermal Res. Counc. Trans., 21, 635–641, 1997. Moses, W.W., M.J. Weber, S.E. Derenzo, D.L. Perry, P. Berdahl and L.A. Boatner, Prospects for dense, infrared emitting scintillators, IEEE Trans. Nucl. Sci., 45, 462, 1998. Moridis, G.J., A. James and C.M. Oldenburg, Development of a design package for a viscous barrier at the Savannah River Site, in Proc. 1997 Int’l Containment Technology Conference and Exhibition, St. Petersburg, Fla., Feb. 9–12, 1997, pp. 452–458, 1997. Moridis, G.J., S.E. Borglin, C.M, Oldenburg, K. Das and A. Becker, Theoretical and experimental investigations of ferrofluids for guiding and detecting liquids in the subsurface, LBNL41069, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1997. Moridis, G.J., S. Finsterle and J. Heiser, Evaluation of altern a t i ve designs for an injectable barrier at the Bro o k h aven National Lab o ra t o ry Site, Long Island, N ew Yo rk , LBNL-41763, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1998. Moridis, G.J., TO U G H 9 0 : A FORTRAN90 implementation of TOUGH2, in Proc.TOUGH Workshop ‘98, Berkeley, Calif., May 4–6, 1998, pp. 300–306, LBNL-41995, Lawrence Berkeley National Lab o ratory, Berke l ey, Calif., 1998. Moridis, G.J., and K. Pruess, T2SOLV: An enhanced package of solvers for the TOUGH2 family of reservoir simulation codes, Geothermics, 27(4), 415-444, 1998. Moridis, G.J.,Y. - S.Wu and K. Pruess, EOS9nT:A TOUGH2 module for flow and solute/colloid transport, in Proc. TOUGH Workshop ’98, Berkeley, Calif., May 4-6, 1998, pp. 142-147, LBNL-41639, L aw rence Berkeley National Lab o ra t o ry, Berkeley, Calif., 1998. Moridis, G.J., Semianalytical solutions for parameter estimation in diffusion cell experiments,Water Resources Research, 35(6), 1729-1740, 1999. Moridis, GJ., S. Finsterle and J. Heiser, Evaluation of alternative designs for an injectable subsurface barrier at the Brookhaven National Laboratory Site, Long Island, N.Y., Water Resources Research, 35(10), 2937-2953, 1999. Myneni, S.C.B., T.K. Tokunaga and G.E. Brown, Abiotic selenium redox transformations in the presence of Fe (II, III) oxides, Science, 278, 1106–1109, 1997. Myeni, SC.B., J.T. Brown, G.A. Martinez and W. Meyer-Ilse, Imaging of humic substance macromolecular structures in water and soils, Science, in press. Nadeau, R.M., and T.V. McEvilly, Seismological studies at Park field V: C h a ra c t e ristic micro e a rt h q u a ke sequences as fault-zone drilling targets, Bull. Seismol. Soc.Am., 87(6), 1463–1472, 1997. Nadeau, R.M., and L.R. Johnson, Seismological studies at Parkfield VI: Moment release rates and estimates of source parameters from small repeating earthquakes, Bull. Seismol. Soc. Am., 88(3), 790–814, 1998. 161


ESD

Publications 1997-1999

Najita, J., and C. Doughty, Using TRINET for simulating flow and transport in porous media, LBNL-42158, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1998. Nakagawa, S.,Acoustic resonance characteristics of rock and concrete containing fra c t u re s , Ph.D Thesis, LBNL-43182, L aw rence Berkeley National Lab o ra t o ry, Berke l ey, Calif., 1998. Nakao, S., J. Najita and K. Karasaki, Sensitivity study on hydraulic well testing inversion using simulated annealing, LBNL41131, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1997. Narasimhan,T.N., Hydraulic characterization of aquifers, reservoir rocks and soils:A history of ideas,Water Resources Research, 34(1), 33–46, 1997. Narasimhan, T.N., A note on setting up the diffusion equation, LBNL-41234, L aw rence Berke l ey National Lab o ra t o ry, Berkeley, Calif., 1997. Narasimhan, T.N., Quantification and groundwater hydrology, Ground Water, 36(1), 1-1, 1998. Narasimhan,T.N., Something to think abou ... Darcy Buckingham Law, Ground Water, 36(2), 194-195, 1998. Narasimhan, T.N., Basic postulates of groundwater occurrence and movement, LBNL-41235, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1998. Narasimhan, T.N., Fourier's heat conduction equation: history, influence, and connections, Reviews of Geophysics, 37(1), 151-172, 1999. Narasimhan, T.N., OHM's Law, Fick's Law, Joule's Law, and gro u n dwater fl ow, in Proc.Theory Modeling and Field Investigations in Hydrogeology: A Special Volume in Honor of Shlomo P. Neuman, Tucson, Ariz., Oct. 17, 1998, LBNL-42824, Lawrence Berkeley National Laboratory, Berkeley Calif., 1999. Nikravesh, M., M. Soroush, M. Johnston and T.W. Patzek, Design of smart wellhead controllers for optimal fluid injection policy and producibility in petroleum reservoirs: neuro-geometric approach, presented at International Thermal Operations and Heavy Oil Symposium, Feb. 10–12, 1997, Bakersfield, Calif., SPE 37557, Society of Petroleum Engineers, Dallas, Texas, 1997. Nikravesh, M., M. Soroush, M. Johnston and T.W. Patzek, Design of smart wellhead controllers for optimal fluid injection policy and producibility in petroleum reservoirs:A neural networkmodel predictive approach, presented at SPE Production Operations Symposium, March 9–11, 1997, Oklahoma City, Okla., SPE 37445, Society of Petroleum Engineers, Dallas, Tex., 1997. Nikravesh, M., A. Garg, A.R. Kovscek, L.M. Castanier and T.W. Patzek, Neural network and its application to CT scanning, in Proc. of Workshop on High Resolution X-Ray Computed Microtomography, Berkeley, Calif.,Aug. 12-13, 1996, 1997. Novak, C.F., I. Al Mahamid, K.A. Becraft, S.A. Carpenter, N. Hakem and T. Prussin, Measurement and thermodynamic modeling of Np(V) solubility in aqueous K2CO3 solutions to high concentrations, J. Solution Chem., 26(7), 681, 1997.

Oldenburg, C.M., and K. Pruess, Higher-order differencing for geothermal reservoir simulation, in Proc. 22nd Workshop on Geothermal Reservoir Engi n e e ring, Stanfo rd , C al if. , Jan. 27–29, 1997, SGP-TR-155, pp. 303–308, 1997. Oldenburg, C.M., Comparison of scale analysis and numerical simulation for saturated zone convective mixing processes, LBNL-40365, L aw rence Berke l ey National Lab o ra t o ry, Berkeley, Calif., 1998. Oldenburg, C.M., S.E. Borglin and G.J. Moridis, On modeling flow and transport of magnetic fluids in porous media,LBNL-40146, Lawrence Berke l ey National Lab o ra t o ry, Berke l ey, Calif., 1998. Oldenburg, C.M., and G.J. Moridis, Ferrofluid flow for TOUGH2, in Proc. TOUGH Workshop ‘98, Berkeley, Calif., May 4-6, 1998, pp. 263–268, LBNL-41995, Lawrence Berke l ey National Laboratory, Berkeley, Calif., 1998. Oldenburg, C.M., and K. Pruess, Higher-order differencing for front propagation in geothermal systems, in Proc. TOUGH Workshop ‘98, Berkeley, Calif., May 4-6, 1998, pp. 13–18, LBNL-41995, L aw rence Berke l ey National Lab o ra t o ry, Berkeley, Calif., 1998. Oldenburg, C.M., and K. Pruess, Higher-order differencing for phase-front propagation in geothermal systems, in Proc. 23rd Workshop on Geothermal Reservoir Engineering, Stanford, Calif., Jan. 26–28, 1998, pp. 127–134, SGP-TR-158, 1998. Oldenburg, C.M., and K. Pruess, Layered thermohaline convection in hypersaline geothermal systems,Transport in Porous Media, in press. Oldenburg, C.M., and K. Pruess, Simulation of propagating fronts in geothermal reservoirs with the implicit Leonard total variation diminishing, Geothermics, in press. Oldenburg, C.M., and K. Pruess, Plume separation by transient thermohaline convection in porous media, Geophysical Research Letters, 26(19), 2997-3000, 1999. Parker, P., Genetic algorithms and their use in geophysics, Ph.D. Thesis, LBNL-43148, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1999. Parra, J.O., L.R. Myer, K.T. Nihei, V.A. Korneev and A.W. Gorody, Guided seismic waves for re s e rvoir ch a racterization, GasTIPS, 3(3), 11–18, 1997. Patzek,T.W., Fluid Injection into low permeability reservoirs, DOE OPA Review, Reston, Va., Jan. 19, 1997, U.S. Dept. of Energy, Washington, D.C., 1997. Patzek, T.W., Optimal fluid injection policy and producibility in fractured flow permeability reservoirs, in Proc. Petroleum Technology Transfer Council Meeting, Univ. of Southern California, Los Angeles, Calif., May 15, 1997. Patzek,T.W.,A.R. Kovscek and P.J. Sullivan, Inferring aqueous contaminant history with an inverse model, in Proc. Seventh Annual West Coast Conference on Contaminated Soils and Groundwater, Oxnard, Calif., March 10–13, 1997. Patzek,T.W., and P.J. Sullivan, Statistical characterization of metalcontaminated fi l l s , p resented at SPE Western Regi o n a l Meeting, Long Beach, Calif., June 25–27, 1997, SPE 38307, Society of Petroleum Engineers, Dallas,Texas, 1997. 162


Earth Sicences Division

Annual Report 1998-1999

Patzek,T.W., Problems and opportunities in the development of hydrocarbon resources in the diatomites, in Proc.West Coast Petroleum Technology Transfer Center, Bakersfield, Calif., Nov. 10, 1996, 1997. Patzek, T.W., and A. De, Lossy transmission line model of hy d ro f ra c t u red well dynamics, presented at SPE 1998 Western Regional Meeting, Bakersfield, Calif., May 10–13, 1998), SPE 46195, Society of Petroleum Engineers, Dallas, Texas, 1998. Patzek,T.W., and D.M. Silin, Control of fluid injection into a lowpermeability rock. 1. Hydrofracture growth, presented at SPE/DOE Improved Oil Recovery Symposium, Tulsa, Okla., April 19-22, 1998, SPE 39698, Society of Petroleum Engineering, Dallas,Texas, 1998. Pellerin, L., and D.L.Alumbaugh,Tools for electromagnetic investigation of the shallow subsurface,The Leading Edge, 16311638, 1997. Pellerin, L., D. Reichhardt,T.M. Daley, F. Gilbert, S.S. Hubbard, E.L. Majer, J.E. Peterson, W. Daily and A. Ramirez, Verification of s u b s u r face barriers using integrated ge o p hysical techniques, in Proc. SAGEEP 1998, Chicago, Ill., March 22-23, 1998. Persoff, P., G.J. Moridis, D.M.Tuck and M.A. Phifer, Laboratory testing of cl o s u re cap repair tech n i q u e s , in Pro c . I n t ’l Containment Technology Conference and Exhibition, St. Petersburg, Fla., Feb. 9–12, 1997, pp. 319–326, 1997. Persoff, P., G.J. Moridis, J.Apps and K. Pruess, Effect of dilution and contaminants on strength and hydraulic conductivity of sand grouted with colloidal silica ge l , in Pro c . I n t ’l Containment Technology Conference and Exhibition, St. Petersburg, Fla., Feb. 9–12, 1997, pp. 578–584, 1997. Persoff, P., J. Apps and G.J. Moridis, Effects of dilution and contaminants on Ludox colloidal silica grout, LBNL-42385, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1998. Persoff, P., J.A.Apps, G.J. Moridis and J.M.Whang, Effect of dilution and contaminants on sand grouted with colloidal silica, LBNL-40129, L aw rence Berke l ey National Lab o ra t o ry, Berkeley, Calif., 1998. Persoff, P., G.J. Moridis, J.Apps and K. Pruess, Evaluation tests for colloidal silica for use in grouting applications, Geotech. Testing J., 21(3), 264–269, 1998. Peterson, J.E., and K.H.Williams, Ground penetrating radar results at the Box Canyon Site: 1996 survey as part of infiltration test, LBNL-40915, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1997. Peterson, J.E., S.S. Hubbard, K.H. Williams, E.L. Majer and P.T. Zawislanski, Moisture content estimation using crosshole radar measurement, EOS Trans., 78(17), S166, American Geophysical Union,Washington, D.C., 1997. Pili, E., B.M. Kennedy, M.S. Conrad and J.P. Gratier, Isotope constraints on the involvement of fluids in the San Andreas fault, EOS Tra n s . , 79, S229, American Geophysical Union, Washington, D.C., 1998.

Pozdniakov, S., and C.-F.Tsang,A semianalytical approach to spatial averaging of hydraulic conductivity in heterogeneous a q u i fe rs, LBNL-40839, L aw rence Berke l ey National Laboratory, Berkeley, Calif., 1997. Pruess, K., On vaporizing water flow in hot sub-vertical rock fractures,Transport in Porous Media, 28, 335–372, 1997. Pruess, K., S. Finsterle, G.J. Moridis, C.M. Oldenburg and Y-S. Wu, General-purpose reservoir simulators:The TOUGH2 family, CRC Bulletin, 26(2), 53–57, 1997. Pruess, K., C.M. Oldenburg, G.J. Moridis and S. Finsterle, Water injection into vapor- and liquid-dominated reservoirs: Modeling of heat transfer and mass transport, in Proc. DOE Geothermal Program Review 15, San Francisco, Calif., March 25–26, 1997, pp. 2-55 to 2-62, DOE/EE-0139, U.S. Dept. of Energy,Washington, D.C., 1997. Pruess, K . ,Two-phase unsaturated flow at Yucca Mountain,N eva d a – A report on current understanding, LBNL-42075, Lawrence Berkeley National Laboratory, Berkeley, Calif. 1998. Pruess, K., C.M. O l d e n b u rgand G.J. Moridis, Ove rv i ew of TOUGH2 ve rsion 2.0, in Proc. TOUGH Workshop ’98, Berkeley, Calif., May 4-6, 1998, pp. 307-314, LBNL-41995, Law rence Berkeley National Lab o ratory, Berke l ey, Calif., 1998. Pruess, K., On water seepage and fast preferential flow in heterogeneous, u n s a t u rated rock fra c t u re s , J. Contaminant Hydrology, 30, 333–362, 1998. Pruess, K. (ed.), Proceedings of TOUGH Workshop ‘98, LBNL41995, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1998. Pruess, K., TOUGH2 – An update, in Proc. 19th New Zealand Geothermal Wo rk s h o p , Au ck l a n d , N. Z , N ov. 1997, pp. 211–216, New Zealand Geothermal Assoc.,Auckland, 1998. Pruess, K. Multiphase flow in fractured rocks – Lessons learned from mathematical models, in Proc. Dynamics of Fluids in Fractured Rocks: Concepts and Recent Advances, Berkeley Calif., Feb. 10-12, 1999, pp. 25-29, LBNL-42718, Lawrence Berkeley National Laboratory, Berkeley, CA, 1999. Pruess, K., B. Faybishenko and G.S. Bodvarsson, Alternative concepts and approaches for modeling flow and transport in thick unsaturated zones of fractured rocks, J. Contaminant Hydrology – Special Issue, 38, 281-322, 1999. Pruess, K.,A mechanistic model for water seepage through thick unsaturated zones in fractured rocks of low matrix permeability,Water Resources Research, 35(4), 1039-1051, 1999. Quinn, N.W.T., and J. Karkoski, Prospects for real-time management of water quality in the San Joaquin River Basin, California, LBNL-40513, L aw rence Berkeley National Laboratory, Berkeley, Calif., 1997. Quinn, N.W.T., J. McGahan and M. Delamore, Innovative drainage management techniques to meet monthly and annual selenium load targets, Calif.Agriculture, in press. Ritcey,A.C., and Y. S.Wu, Evaluation of the effect of future climate change on the distribution and movement of moisture in the unsaturated zone at Yucca Mountain, NV, J. Contaminant Hydrology – Special Issue, 38, 257-279, 1999. 163


ESD

Publications 1997-1999

Romero,A.E.,T.V. McEvilly and E.L.Majer,Three-dimensional attenuation tomography at the Northwest Geysers Geothermal Region, California, Geophysics, 62(1), 149-167, 1997. Rubin, H., D. Dveyrin, J.T. Birkholzer and G. Rouve,Advection and dispersion of contaminant in a permeable medium embedding fractures in which advection velocity is comparatively slow, J. Hydrology, 179, 135–162, 1997. Rubin, Y., S.S. Hubbard and S. Ezzedine, Conjunctive use of geophysical or tracer data with hydrogeological testing for enhanced site characterization, presented at ASCE IAHR Conference, San Francisco, Calif., 260, 1997. Rubin, Y., S.S. Hubbard, A. Wilson and M.A. Cushey, Aquifer characterization, in The Handbook of Groundwater Engineering, Chapter 10, pp. 10.1–10.68 (J.W. Delleur, ed.) CRC Press, New York, N.Y., 1998. Rubin, H.,A.M. Soliman, J.T. Birkholzer and G. Rouve,Transport of a tracer slug in a fra c t u red permeable fo rmation, J. Hydrology, 176, 153–180, 1997. Rumynin,V.G.,V.A. Mironenko, L.N. Sindalovsky,A.V. Boronina, P.K. Konosavsky and S.P. Pozdniakov, Evaluation of conceptual and physical and chemical models for description subsurface radionuclide transport at the Lake Karachai waste disposal site, LBNL-41974, L aw rence Berkeley National Laboratory, Berkeley, Calif., 1998. Rutqvist, J., O. Stephansson and C.-F.Tsang, Hydraulic field measurements of incompletely closed fractures in granite, Int’l J. Rock Mech. Mining Sci., 34(3-4), 1997. Rutqvist, J., C.-F.Tsang, D. Ekman and O. Stephansson, Evaluation of in situ hydromechanical properties of rock fractures at Laxemar in Sweden, in Proc. First Asian Rock Mechanics Symposium: ARMS ‘97, Seoul, Korea, pp. 619–624, A.A. Balkema Publishers, Rotterdam,The Netherlands, 1997. Rutqvist, J., J. Noorishad and C.-F. Tsang, Determination of fract u re storativity in hard ro cks using high pre s s u re injection testing, Water Resources Research, 34(10), 2551–2560, 1998. Rutqvist, J., J. Noorishad and C.-F. Tsang, Coupled thermohydromechanical analysis of a heater test in unsaturated clay and fractured rock at Kamaishi Mine, LBNL-44203, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1999. Salve, R.,T.Tokunaga, J.S.Wang, R. Solbau and J. Clyde, Hydrologic monitoring in unsaturated fractured tuff boreholes, EOS Trans., 78 (supplement), American Geophysical Union, Washington, D.C., 1997. Salve, R., Near-surface wetting of a ponded basalt surface: Observations using time domain reflectometry, Journal of Hydrology, 208 (3-4), 249–261, 1998. Schlueter, E.M., R.W. Zimmerman, P. A. Witherspoon, and N.G.W. Cook, The fractal dimension of pores in sedimentary rocks and its influence on permeability, Eng. Geol., 48(3–4), 199–215, 1997. Seifert, P. K., Effects of pore fluids in the subsurface on ultrasonic wave pro p agation, Ph.D. Thesis, LBNL-41781, Law re n c e Berkeley National Laboratory, Berkeley, Calif., 1998.

Seife rt , P.K., J.T. Geller and L.R. Johnson, E ffect of P-wave scattering on velocity and attenuation in unconsolidated sand saturated with immiscible liquids, Geophysics, 63(1), 161–170, 1998. Seifert, P.K., B. Kaelin and L.R. Johnson, Effect on ultrasonic signals of viscous pore fluids in unconsolidated sand, J. of the Acoustical Society of America, in press. Shan, C., H.-Y. Holman and I. Javandel,A pitfall and its solution in monitoring and cleaning up groundwater contamination by petroleum products, EOS Trans., 78 (supplement),American Geophysical Union,Washington, D.C., 1997. Shan, C.,An analytical solution for the capture zone of two arbitrarily located wells, J. Hydrology, accepted. Shan, C., and I. Javandel, Approximate equations for gas flow through porous media,Water Resources Research, accepted. Shan, C., I. Javandel and P.Witherspoon, Characterization of leaky faults: Study of air flow in faulted vadose zone, Water Resources Research, accepted. Simmons, A. (ed.), Environmental management science pro gram awards: Fiscal year 1997 annual progress report, LBNL-41192, Lawrence Berke l ey National Laboratory, Berkeley, Calif., 1997. Simmons, A., Fracture mineral evolution affecting flow around a potential re p o s i t o ry, in Pro c . I n t ’l High-Level Waste Management Conference and Exposition, Las Vegas, Nev., May 11-14, 1998, LBNL-41057, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1998. Sisson, J.B., and B. Faybishenko, Box Canyon hot air injection study and vadose zone instrument development result,INEL95/0597, Idaho National Engineering and Environmental Laboratory, Idaho Falls, Idaho, 1997. Smith, J.T., and K.H. Lee, Controlled-source magnetotellurics: source effects, LBNL-43121, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1999. S o l o d ov, I.N., A.V. Z o t ov, A.D. K h o t e ev, A.P. Mukhamet-Galeev, B.R. Tagi rov and J.Ap p s ,Geochemistry of natural and contaminated subsurface wa t e rs in fissured bed rocks of the Lake Karachai Area, Southern Urals, Russia,Ap p . Geochem., accepted. Song, Y., and K.H. Lee, Electromagnetic fields due to a loop current in a cased borehole surrounded by uniform whole space, LBNL-42371, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1998. Sonnenthal, E.L., and G.S. Bodvarsson, Modeling strontium geochemistry and isotopic ratio in the unsaturated zone, in The site-scale unsaturated zone model of Yucca Mountain, Nevada, for the viability assessment (G.S. Bodvarsson, T.M. Bandurraga and Y.S. Wu, eds), Yucca Mountain Site Characterization Project Milestone, LBNL-40376, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1997. Sonnenthal, E.L., and G.S. Bodvarsson, Modeling the chloride geoch e m i s t ry in the unsaturated zone, in The Site-Scale Unsaturated Zone Model of Yucca Mountain, Nevada, For the Viability Assessment (G.S. Bodvarsson, T.M. Bandurraga and Y.S. Wu, eds), Yucca Mountain Site Characterization Project Milestone, LBNL-40376, L aw rence Berke l ey National Laboratory, Berkeley, Calif., 1997. 164


Earth Sicences Division

Annual Report 1998-1999

Sonnenthal, E.L., D.J. DePaolo and G.S. Bodvarsson, Modeling the strontium geochemistry and isotopic ratio in the unsaturated zone at Yucca Mountain,Neva d a ,EOS Trans.,78 (supplement), A m e rican Geophysical Union,Washington, D.C., 1997. Sonnenthal, E.L., and G.S. Bodvarsson, Percolation flux estimates from geochemical and thermal modeling, in Proc. Eighth International Conference on High-Level Radioactive Waste Management, pp. 130–132, American Nuclear Society, La Grange Park, Ill., 1998. Sonnenthal E.L., and G.S. Bodvarsson, Constraints on the hydrology of the unsaturated zone at Yucca Mountain, Nevada, from three-dimensional models of chloride and strontium geochemistry, J. Contaminant Hydrology – Special Issue, 38, 107-156, 1999. Sorey, M.L.,W.C. Evans, B.M. Kennedy, C.D. Farrar, L.J. Hainsworth and B.Hausback, Carbon dioxide and helium emissions from a reservoir of magmatic gas beneath Mammoth Mountain, California, J. Geophys. Res., 103(15), 15,303–15,323, 1998. Sorey, M., B. Evans, B.M. Kennedy, J. Rogie and A. Cook, Magmatic gas emissions from Mammoth Mountain, California, geology, pp. 4-16, Sept./Oct. 1999. Soroush, M., and M. Nikravesh, Shortest-prediction-horizon nonlinear model pre d i c t i ve control, Chemical Engineering Science, 53(2), 273-292, 1998. Sposito, G., N.T. Skipper, R. Sutton, S.-H. Park, A.K. Soper and J.A. Greathouse, Surface geochemistry of the clay minerals, Proc. Natl.Acad. Sci. USA, 96, 3358-3364, 1999. Su, G.W., J.T. Geller, K. Pruess and F.Wen, Experimental studies of water seepage and intermittent flow in unsaturated, roughwalled fractures, Water Resources Research, 35(4), 10191037, 1999. Thapa, B.B.,P. Hughett and K.Karasaki, Semi-automatic analysis of rock fracture orientations from borehole wall images, Geophysics, 62(1), 129-137, 1997. Tokunaga,T.K.,A tensiometer for measuring hydraulic potentials on surfaces of porous rock, Water Resources Research, 33(6), 1509-1513, 1997. Tokunaga,T.K., and J.Wan,Water film flow along fracture surfaces of porous rock,Water Resources Research, 33(6), 1287-1295, 1997. Tokunaga, T.K., G.E. Brown, Jr., I.J. Pickering, S.R. Sutton and S. Bajt, Selenium re d ox reactions and tra n s p o rt betwe e n ponded waters and sediments, Environmental Science and Technology, 31(5), 1419-1425, 1997. Tokunaga,T.K., S.R., Sutton and S. Bajt, Mapping of selenium concentrations in soil ag gregates with synch ro t ron x-ray floure scence microprobe, Soil Science, 158(6), 421-434, 1997. Tokunaga,T.K., S.R. Sutton, S. Bajt, P. Nuessle and G. Shea-McCarthy, Selenium diffusion and reduction at the water–sediment boundary: MicroXANES spectro s c o py of reactive transport, Environmental Science Technology, 32, 1092–1098, 1998. Torn, M.S., E. Mills and J. Fried, Will climate change spark more wildfire damage? LBNL-42592, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1998.

Truesdell,A.H., M.J. Lippmann and H. Gutierrez Puente, Evolution of the Cerro Prieto reservoirs under exploitation, in Proc. 1997 Annual Meeting of the Geothermal Resources Council, Burlingame, Calif., Oct. 12–15, 1997, 21, pp. 263–270, Geothermal Resources Council, Davis, Calif., 1997. Truesdell,A.H., and M.J. Lippmann, Effects of pressure drawdown and recovery on the Cerro Prieto Beta Reservoir in the CPIII Area, in Proc. 23rd Workshop on Geothermal Reservoir Engineering, Stanford, Calif., Jan. 26-28, 1998, 23, pp. 90-98, SGP-TR-158, LBNL-41532, L aw rence Berkeley National Laboratory, Berkeley Calif., 1998. Truesdell, A.H., M.J. Lippmann, H. Gutierrez-Puente and J.V. De Leon, The importance of natural fluid recharge to the sustainability of the Cerro Prieto resource, Geotherm. Resour. Coun.Trans., 22, 529-536, 1998. Truesdell, A.H., M.J. Lippmann, J. de Leon and M.H. Rodriguez, Cerro Prieto cold water injection: effects on nearby production we l l s , LBNL-44025, L aw rence Berkeley National Laboratory, Berkeley, Calif., 1999. Tsang, C.-F., A discussion of scientific issues in hydrogeology of l ow perm e ability stra t a , in Pro c . of 30th International Geological Congress, 22, pp. 224–228, 1997. Tsang, C.-F., Linking therm a l , hy d ro l o gical and mech a n i c a l processes in fractured rocks, in Annual Review of Earth and Planetary Sciences, 27, in press. Tsang, C.-F., and I. Neretnieks, Flow channeling in heterogeneous fractured rocks, Rev. Geophys., 36(2), 275–298, 1998. Tsang,C.-F.,and J.Rutqvist,Progress in coupled analysis of a thermohydro-mechanical ex p e riment in fractured rock, in Computer Methods and Advances in Geomechanics (J.-X.Yuan, ed.),A.A. Balkema Publishers, Rotterdam,The Netherlands, 1997. Tsang, Y.W., and J.T. Birkholzer, Interpreting the thermal-hydrological response of the ESF Single Heater Test, Yucca Mountain Site Characterization Project Level 4 Milestone SP9267M4, L aw rence Berke l ey National Lab o ra t o ry, Berkeley, Calif., 1997. Tsang,Y.W., and P. Cook,Ambient characterization of the ESF drift scale test area by field air permeability measurements,Yucca Mountain Project Level 4 Milestone SP9512M4, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1997. Tsang,Y.W., and J.T. Birkholzer,Thermolhydrological response of the Yucca Mountain single heater test, in Proc. 8th International High-Level Nuclear Waste Conf., May 1998, Las Vegas, Nev., pp. 140–142, American Nuclear Society, La Grange Park, Ill, 1998. Tsang,Y.W., and B.M. Freifeld, Role of fractures at different scales in underground heater experiments, in Proc. Dynamics of Fluids in Fractured Rocks: Concepts and Recent Advances, Berkeley, Calif., Feb. 10-12, 1999, LBNL-42718, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1999. Tsang,Y.W., and J.T. Birkholzer, P redictions and observations of the therm a l - hy d ro l o gical conditions in the single heater test, J. Contaminant Hydro l o gy – Special Issue, 38, 385-425, 1999. 165


ESD

Publications 1997-1999

Tsang, Y.W. et al., Yucca Mountain single heater test final report: Yucca Mountain site chara c t e rization project, LBNL-42537, Lawrence Berke l ey National Laboratory, Berkeley, Calif., 1999. Tsang,Y.W., J. Apps, J.T. Birkholzer, J.E. Peterson, E. Sonnenthal, N. Spycher and K.H. Williams, Yucca Mountain drift scale test progress report, LBNL-42538, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1999. Vasco, D.W., A. Datta-Gupta and J.C. Long, Integrating field production history in stochastic reservoir characterization, presented at SPE Annual Technical Conference and Exhibition Denver, Colo., Sept. 1997, SPE 36567, Society of Petroleum Engineers, Dallas,Texas, 1997. Vasco, D.W., Groups, algebras and the nonlinearity of geophysical inverse problems, Geophysics J. Int., 131, 9–23, 1997. Vasco, D.W., K. Karasaki and L. Myer, Monitoring of fluid injection and soil consolidation using surface tilt measurements, J. Geotech. Geoenviron. Eng., 124(1), 29–37, 1997. Vasco, D.W., J.E. Peterson and K.H. Lee, Ground-penetrating radar velocity tomography in hetero geneous and anisotro p i c media, Geophysics, 62(6), 1758–1773, 1997. Vasco, D.W., J.E. Peterson and E.L. Majer, Resolving seismic anisotropy: sparse matrix methods for geophysical inverse problems, Geophysics, 63(3), 970–983, 1997. Vasco, D., and L.R. Johnson,Whole earth stru c t u re estimated from seismic arri val times, J. Geophys. Res., 103(B2), 2633-2671, 1998. Vasco, D.W., Intersections, ideals, and inversion, LBNL-42397, Lawrence Berke l ey National Laboratory, Berkeley, Calif., 1998. Vasco, D.W., and A. Datta-Gupta, Asymptotic solutions for solute transport: A fo rmalism for tracer tomogra p hy, Water Resources Research, 35(1), 1-16, 1999. Wan, J., and T.K. Tokunaga, Film straining of colloids in unsaturated porous media: conceptual model and experimental testing, E nv i ronmental Science Te ch n o l o gy, 31(8), 2413–2420, 1997. Wan, J., and T.K.Tokunaga, Measuring partition coefficients of colloids at air–water interfaces, E nv i ronmental Science Technology, November 1998. Wang, J.S.Y., P.J. Cook and R.C. Trautz, Field testing and observation of flow paths in niches, in Proc. Int’l High-Level Radioactive Waste Management Conference, Las Vegas, Nev., May 11-14, 1998, 1998. Wang, J.S.Y, R.C. Trautz, P.J. Cook, S. Finsterle, A.L. James and J.T. Birkholzer, Field tests and model analyses of seepage into drift, J. Contaminant Hydrology – Special Issue, 38, 323-347, 1999. Weeber, P.A., and T.N. Narasimhan, Slug test in an unconfined aquifer: A Rich a rds' Equation perspective, LBNL-40966, Lawrence Berke l ey National Laboratory, Berkeley, Calif., 1997. Wilt, M., P. Kasameyer, S. Takasugi, K.H. Lee and M.J. Lippmann, Fracture mapping in geothermal fields with long-offset induction logging, in Proc. 22nd Workshop on Geothermal Reservoir Engi n e e ring, S t a n fo rd, C a l i f. , SGP-TR-155, pp. 229–232, Stanford, Calif., 1997.

Wrixon, R., and G.A. Cooper,Theoretical and practical guidelines for using electrokinetics to improve casing support in soft marine sediments, presented at 1998 SPE/IAD.C. Drilling Conference, Dallas,Tex., March 3–6, 1998, SPE/IAD.C. 39299, Society of Petroleum Engineers, Dallas,Tex., 1998. Wu, Y.S., J.B. Kool, P.S. Huyakorn and Z.A. Saleem, An analytical model for nonlinear adsorptive transport through layered soils,Water Resources Research, 33(1), 21-29, 1997. Wu, Y.S., A.C. Ritcey, C.F. Ahlers, A.K. Mishra, J.J. Hinds and G.S. Bodvarsson, Providing base-case flow fields for TSPA-VA: Evaluation of uncertainty of present-day infiltration rates using DKM/Base-Case and DKM/Weeps parameter sets, Yucca Mountain Site Characterization Project Report, YMP Level 4 Milestone SLX01LB2, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1997. Wu, Y.S., A.C. Ritcey, E.L. Sonnenthal, T.M. Bandurraga, C.B. Haukwa, J.P. Fairley, G. Chen, J.H. Li and G.S. Bodvarsson, Incorporation of perched water data into the UZ site-scale model, Yucca Mountain Project Level 4 Milestone SP24UCM4, L aw rence Berkeley National Lab o ra t o ry, Berkeley, Calif., 1997. Wu,Y.S., and K. Pruess, Flow of non-Newtonian fluids in porous media,Adv. Porous Media, 3, 87-184, 1997. Wu,Y.S., and K. Pruess, Several TOUGH2 modules developed for site characterization studies of Yucca Mountain, in Proc. TOUGH Workshop ’98, May 4-6, 1998, Berkeley, Calif., LBNL41995, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1998. Wu,Y.S., C.F. Ahlers, C. Haukwa, A. Ritcey, E. Sonnenthal and G.S. Bodvarsson, The 3-D UZ site-scale-model-calibrations and predictions, in Proc. Int’l High-Level Radioactive Waste Management Conference, Las Vegas, Nev., May 11-14, 1998, 1998. Wu,Y.S., and A.K. Mishra, Modifications and additions to selected TOUGH2 modules,LBNL-41870, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1998. Wu, Y.S., and K. Pruess, A numerical method for simulating nonNewtonian fluid flow and displacement in porous media, Adv.Water Res., 21, 351–362, 1998. Wu,Y.S., K. Pruess and P. Persoff, Gas flow in porous media with Klinkenberg effects,Transport in Porous Media, 32, 117–137, 1998. Wu,Y.S.,A.C. Ritcey, C.F.Ahlers, J.J. Hinds,A.K. Mishra, C. Haukwa, T.M. Bandurraga, H.H. Liu, E.L. Sonnenthal and G.S. Bodvarsson, 3-D UZ site-scale model for abstraction in TSPAVA, Yucca Mountain Site Characterization Project Milestone SLX01LB3 Report, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1998. Wu,Y. S., On the effective continuum method for modeling multiphase flow, multicomponent transport and heat transfer in fractured rock, in Proc. Dynamics of Fluids in Fractured Rocks: Concepts and Recent Advances, Berkeley, Calif., Feb. 10-12, 1999, pp. 375-379, LBNL-42718, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1999. 166


Earth Sicences Division

Annual Report 1998-1999

Wu, Y.S., A.C. Ritcey and G.S. Bodvarsson, A modeling study of perched water phenomena in the unsaturated zone at Yucca Mountain, J. Contaminant Hydrology – Special Issue, 38, 157184, 1999. Wu,Y.-S., C.B. Haukwa and G.S. Bodvarsson,A site-scale model for fluid and heat flow in the unsaturated zone of Yucca Mountain, Nevada, J. Contaminant Hydrology – Special Issue, 38, 185-215, 1999. Xie, G., and J. Li ,A new 3D parallel high resolution electromagnetic nonlinear inversion based on new global magnetic integral and local differential decomposition (GILD), LBNL40265, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1997. Xie, G., J. Li and E.L. Majer, New 3D parallel SGILD modeling and inversion, in Proc. 18th Annual International Conference, P re d i c t ability Quantifying Uncertainty in Models of Complex Phenomena, May 11–15, 1998, Los Alamos, N.M., 1998, LBNL-42252, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1998. Xie, G., J. Li, E.L. Majer and D. Zuo, New 3D parallel GILD electromagnetic modeling and nonlinear inversion using global magnetic integral and local differential equation, in Proc. SEG International Exposition and 67th Annual Meeting, Nov. 2-7, 1997, Dallas, Texas, LBNL-42105, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1998. Xie, G, and J. Li, New parallel stochastic global integral and local differential equation modeling and inversion, Physica D, 133, 477-487, 1999. Xie, G., J. Li, E.L. Majer, D. Zuo and M. Oristaglio, New 3-D electromagnetic modeling and nonlinear inversion, Geophysics, in press. Xu,T., F. Gerard, K. Pruess and G. Brimhall, Modeling non-isothermal multiphase multi-species reactive chemical transport in geologic media, LBNL-40504, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1997. Xu,T., and K. Pruess, Coupled modeling of non-isothermal multiphase flow, solute transport and reactive chemistry in porous and fractured media: 1. Model development and validation, LBNL-42050, L aw rence Berke l ey National Laboratory, Berkeley, Calif., 1998. Xu, T., E. Sonnenthal, N. Spy ch e r, K. Pruess and G. Brimhall, Coupled modeling of non-isothermal multi-phase fl ow, solute tra n s p o rt and re a c t i ve chemistry in porous and f ra c t u red media: 2 . Model applications, L B N L - 4 2 0 5 1 , Law rence Berke l ey National Laboratory, B e rke l ey, C a l i f. , 1998. Xu,T., S.P.White and K. Pruess, Pyrite oxidation in saturated and unsaturated porous media flow:A comparison of alternative mathematical modeling approaches, LBNL-42049, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1998. Xu,T., K. Pruess and G. Brimhall,An Improved equilibrium-kinetics speciation algorithm for redox reactions in variably saturated subsurface flow systems, Computers & Geosciences, 25, 655-666, 1999.

Xu,T., K. Pruess and G. Brimhall, Oxidative weathering chemical migration under variably saturated conditions and supergene copper enrichment, LBNL-43129, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1999. Zawislanski, P.T., and A.B. Crawley, Selenium cycling study in the San Francisco Bay:A partnership involving federal and state gove rnment and industry, presented at SPE/EPA 1997 E x p l o ration and Production Env i ronmental Conference, Dallas, Texas, SPE 37885, Society of Petroleum Engineers, Dallas,Texas, 1997. Zawislanski, P.T., H.S. Mountford, R. Dahlquist, S.J. Rodriguez and R. Salve, Monitoring and data analysis for the vadose zone monitoring system (VZMS), McClellan AFB - Quarterly Status Report 8/15/97-11/15/97, LBNL-41147, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1997. Zawislanski, P.T., R. Salve, B. Freifeld, H.S. Mountford, R. Dahlquist, A. James, S. Rodriguez and B. Faybishenko, Monitoring and data analysis for the vadose zone monitoring system (VZMS), McClellan AFB quarterly status report, LBNL-40377, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1997. Zawislanski, P.T., R. Salve, B.M. Fre i feld, H.S. Mountfo rd , R. Dahlquist, S.J. Rodriguez and B. Faybishenko, Monitoring and data analysis for the vadose zone monitoring system (VZMS), McClellan AFB: Quarterly status report, May 15–Aug. 15, 1997, LBNL-41009, L aw rence Berke l ey National Laboratory, Berkeley, Calif., 1997. Zawislanski, P.T., A.E. McGrath, S.M. Benson, H.S. Mountford,T.M. Johnson, E. Gabet, S.C.B. Myneni,T.K.Tokunaga, S. Chau and H.Wong, Selenium fractionation and cycling in the intertidal zone of the Carquinez Strait, Annual Report, Oct. 1, 1995 – Dec. 31, 1996, LBNL-40993, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1997. Zawislanski, P.T.,T.K.Tokunaga, S.M. Benson, H.S. Mountford,T.C. Sears, H. Wong, D. King and J. Oldfather, Hydrological and ge o chemical investigations of selenium behavior at Kesterson Reservo i r : Progress report (10/194–9/30/96), LBNL-41027, L aw rence Berkeley National Lab o ra t o ry, Berkeley, Calif., 1997. Zawislanski, P.T., H.S. Mountford and R. Dahlquist,Monitoring and data analysis for the vadose zone monitoring system (VZMS), McClellan AFB Quarterly Status Report (8/20/98 –11/20/98), LBNL-42666, L aw rence Berke l ey National Laboratory, Berkeley Calif., 1998. Zawislanski, P.T., H.S. Mountford, R. Dahlquist and A.L. James, Monitoring and data analysis for the vadose zone monitoring system (VZMS), McClellan AFB Quarterly Status Report (5/20/98–8/20/98), LBNL-42326, L aw rence Berke l ey National Laboratory, Berkeley, Calif., 1998. Zawislanski, P.T., H. S. Mountford, R. Dahlquist and S. J. Rodriguez, Monitoring and data analysis for the vadose zone monitoring system (VZMS), McClellan AFB Quarterly Status Report (11/15/97–2/20/98), LBNL-41767, L aw rence Berke l ey National Laboratory, Berkeley, Calif., 1998. 167


ESD

Publications 1997-1999

Zawislanski, P.T., H. S. Mountford, R. Dahlquist and S. J. Rodriguez, Monitoring and data analysis for the vadose zone monitoring system (VZMS), McClellan AFB quarterly status report (2/205/20/98), LBNL-41959, L aw rence Berke l ey National Laboratory, Berkeley, Calif., 1998. Zawislanski, P.T., and A. E. McGrath, Selenium cycling in estuarine wetlands: Overview and new results from the San Francisco Bay, in Env i ronmental Chemistry of Selenium (W.T. Frankenberger and R.A. Engberg, eds), Marcel Dekker, New York, N.Y., 1998. Zawislanski, P.T., and C.M. Oldenburg, Data analysis for preliminary conceptual model design, vadose zone monitoring system (VZMS), McClellan AFB: 1997 annual report, LBNL-41262, Lawrence Berke l ey National Laboratory, Berkeley, Calif., 1998.

Zawislanski, P.T., H.S. Mountford and R. Dahlquist, Monitoring and data analysis for the vadose zone monitoring system (VZMS), McClellan AFB Quarterly Status Report (11/20/98â&#x20AC;&#x201C;2/20/99), LBNL-43084, L aw rence Berke l ey National Lab o ra t o ry, Berkeley, Calif., 1999. Zawislanski,P.T.,and B.Faybishenko,New casing and backfill design for neutron logging access boreholes, Groundwater J.,in press. Zwahlen, E.D., and T.W. Patzek,A comparison of mapping schemes for reservoir characterization, p resented at SPE We s t e rn R e gional Meeting, Long Beach, C a l i f. June 25â&#x20AC;&#x201C;27, 1997, SPE 38264, Society of Petroleum Engineers, Dallas,Texas, 1997. Zwahlen, E.D., and T.W. Patzek, Linear transient flow solution for primary oil recovery with infill and conversion to water injection, In Situ, 21(4), 1997.

168


Earth Sicences Division

Annual Report 1998-1999

Division

Director

Sally M. Benson

Division

Deputy

Director

EARTH SCIENCES DIVISION STAFF

Norman E. Goldstein

Scientists/Engineers Ecology

Tomutsa, Liviu Vasco, Donald W. Williams, Kenneth H. Xie, Ganquan

Hazen,Terry C.* Holman, Hoi-Ying Hunter-Cevera, Jennie C. Quinn, Nigel W. Stringfellow,William T.

Geosphere Dynamics

Geochemistry

Kim, Jinwon Miller, Norman L.

Apps, John A. Bishop, James K. Conrad, Mark S. Christensen, John N. Guerin, Marianne Kennedy, B. Mack Myneni, Satish Perry, Dale L. Simmons,Ardyth M. Tokunaga,Tetsu K. Torn, Margaret S. Wan, Jiamin Waychunas, Glenn A.** Wo l l e n b e rg Jr., H a rold A. Zawislanski,Peter T. Geophysics and Geomechanics

Clyde, John R. Dougherty, James R. Feighner,Mark A. Geller, Jil T. Gritto, Roland Haught, John R. Hoffpauir, Cecil Hoversten, Gary M. Hubbard, Susan S. Korneev,Valeri A. Lee, Ki-Ha Majer, Ernest L.* Myer, Larry R. Nihei, Kurt T. Peterson, John E. Sell, Russell W. Smith, Jeremy T. Solbau, Ray D.

1998-1999

Hydrology and Reservoir Dynamics

Bodvarsson, Gudmundur S. Doughty, Christine A. Faybishenko, Boris A. Finsterle, Stefan A. Freifeld, Barry M. Gonzalez Jr., Emilio Guo,Yonghai Hestir, Kevin F. Javandel, Iraj Karasaki, Kenzi Kneafsey,Timothy J. Li, Guomin Lippmann, Marcelo J. Liu, Hui-Hai Liu, Jianchun Moridis, George J. Mukhopadhyay, Sumit Oldenburg, Curtis M. Pan, Lehua Persoff, Peter Pruess, Karsten Rutqvist, Jonny Silin, Dmitriy Sonnenthal, Eric Spycher, Nicolas Stepek, Jan W. Tsang, Chin-Fu* Tsang,Yvonne T. Unger,Andre Wang, Joseph S. Wu,Yu-Shu Xu,Tianfu Zimmerman, Robert W.

*Department Head **Deputy Department Head

169


ESD

Earth Sciences Division Staff

Faculty Ecology

Geosphere Dynamics

Buchanan, Bob B. Leighton,Terrance J. Oh, Keun-Chan

Boering,Kristie A. Fung, Inez Hydrology and Reservoir Dynamics

Geochemistry

DePaolo, Donald* Doner, Harvey E. Doyle, Fiona M. Ingram, B. Lynn

Brimhall, George Cooper, George A. Narasimhan,T.N. Patzek,Tadeusz W. Radke, Clayton J. Rector, James W. Rubin,Yoram Shen, Hsieh W. Sposito, Garrison Witherspoon, Paul A.

Geophysics and Geomechanics

Becker,Alex Glaser, Steven Johnson, Lane R. McEvilly,Thomas V. Morrison, H. Frank

Post

Doctorates

Geochemistry

Geosphere Dynamics

Bryce, Julia G. Davis, Simon Fischer,Tobias Kong, Ping Lin, Jo Chiu Fang Veerapaneni, Srinivas Vrdoljak, Gordon A. Welton, Kees C.

Guay, Christopher Kyriakidis, Phaedon C. Randerson, James T. Hydrology and Reservoir Dynamics

Borglin, Sharon E. De,Asoke Haukwa, Charles Hu, Qinhong

Geophysics and Geomechanics

Cheah, Sing-Fong Keers, Henk Nadeau, Robert Nakagawa, Seiji Park, Sung-Ho

Research

Associates,

Specialists

and

Technicians

Owens,Thomas L. Shuster, David Smith, Donna S.

Ecology

Castro,Grace M. Geochemistry

Alusi,Thana Carpenter, Scott A. Daley,Thomas M. Gatti, Raymond C. Joyner, Dominique C. Mountford, H. Scott Olson, Keith R.

Geophysics and Geomechanics

Kirkpatrick, Elizabeth A. Scott, Clark L. Geosphere Dynamics

Robitaille, Daniel 170


Earth Sciences Division

Annual Report 1998-1999

Research Associates, Specialists and Technicians (cont.) Hydrology and Reservoir Dynamics

Link, Suzanne M. Menendez-Barreto, Melani Salve, Rohit Shan, Chao TerBerg, Robert Trautz, Robert C. Zhang,Winnie W.

Ahlers, C. Fredrik Bandurraga,T. Mark Cohen, Andrew J. Cook, Paul J. Hedegaard, Randall F. Hinds, Jennifer Jordan, Preston D.

Graduate

Student

Research

Assistants

Ecology

Zarate, Max A.

Hildenbrand, Keary L. Kowalsky, Michael Sutton, Rebecca A. Tseng, Hung-Wen Yang, Jeong-Seok Zhen,Tao

Geochemistry

Hammersley, Lisa Johnson, Kathleen Song, Donald Winter, Stacey J.

Hydrogeology and Reservoir Dynamics

Geophysics and Geomechanics

Benito, Pascual H. Fairley, Jerry P. Garcia, Julio Liou,Tai-Sheng

Anderson, Heidi L. Bessinger, Brad A. Bhatt, Divesh Das, Kaushik K. Frangos,William

Technical, Administrative and Aden-Gleason, Nancy Bell,Andre R. Bradley, Rachel M. Cushey, Mark Cuzner, Marlene C. Fink, Maria Fissekidou,Vassiliki A. Harris, Stephen D. Hodge, Sheryl A. Kaszuba, David Kramer, Bridget R. Lau, Peter K. Lippert, Donald R. Lucido, Nina Mangold, Donald C.

Management Staff

McClung, Ivelina A. Miller, Grace A. Montgomery, Lizette Oberlander, Phil L. Pratt, Mary G. Quarrie,Avril L. Reen, Lorraine Saarni, Marilyn E. Seybold, Sherry A. Stover, Richard C . Swantek, Diana M. Taliaferro, Carol L. Villavert, Maryann Wentworth, H. Katherine Wuy, Linda D.

171



1998-1999 ESD Annual Report