ML20094B245

Proceedings of the Workshop on the Role of Natural Analogs in Geologic Disposal of HIGH-LEVEL Nuclear Waste.Held in San Antonio,Texas,July 22-25, 1991

ML20094B245
Person / Time
Issue date:	09/30/1995
From:	Kovach L, Murphy W NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES), SOUTHWEST RESEARCH INSTITUTE
To:
References
CON-FIN-B-6666 CNWRA-93-020, CNWRA-93-20, NUREG-CP-0147, NUREG-CP-147, NUDOCS 9510310349
Download: ML20094B245 (131)

Text

.

NUREG/CP-0147 CNWRA 93-020 Proceedings of the Workshop on tae Role of Natural Analogs in Geologic Disposal of Hig> Level Nuclear Waste Held in San Antonio, TX July 22-25,1991 Edited by L A. Kovach, NRC W. M. Murphy, CNWRA Sponsored by Office of Nuclear Regulatory Research U. S. Nuclear Regulatory Commission Proceedings prepared by U.S. Nuclear Regulatory Commission cnd Center for Nuclear Waste Regulatory Analyses pa atouq, e

's E

i

%,...l /

En A8n; = =

CP-0147 R PDR

AVAILABILITY NOTICE Availabilrty of Reference Materials C4ted in NRC Publications Most documents cited in NRC publications will be available from one of the following sources:

1.

The NRC Pubuc Document Room, 2120 L Street, NW., Lower Level Washington, DC 20555-0001 2.

The Superintendent of Documents, U.S. Government Printing Office, P. O. Box 37082, Washington, DC 20402-9328 3,

The National Technical information Service, Springfield, VA 22161-0002 Although the listing that follows represents the majority of documents cited in NRC publications, it is not in-tended to is exhaustive.

Referenced documents available for inspection and copying for a fee from the NRC Public Document Room include NRC correspondence and internal NRC memoranda; NRC bulletins, circulars, Information notices, in-spection and investigation notices; licensee event reports; vendor reports and correspondence: Commission papers; and applicant and licensee documents and correspondence.

The following documents in the NUREG series are available for purchase from the Governmerit Printing Office:

formal NRC staff and contractor reports, NRC-sponsored conference proceedings, international agreement reports, grantee reports, and NRC booklets and brochures. Also available are regulatory guides, NRC regula-tions in tne Code of Federal Regulations, and Nuctoar Rega:atary Cornmieston 'ssusnces Documents available from the National TechnicalInformation Service include NUREG-series reports and tech-nical reports prepared by other Federal agencies and reports prepared by the Atomic Energy Commission, forerunner agency to the Fuclett Regulatory Commission.

Documents available from public and special technicallibraries include all open literature items, such as books, journal articles, and transactions. Federal Register notices. Federal and State legislation, and congressional reports can usually be obtained from those libraries.

Documents such as theses, dissertations, foreign reports and translations, and non-NRC conference pro-ceedings are available for purchase from the organization sponsoring the publication cited.

Single coples of NRC draft reports are available free, to the extent of supply, upon written request to the Office of Administration, Distribution and Mall Services Section, U.S. Nuclear Regt.'atory Commission, Washington, DC 20555-0001.

Copies of Industry codes and standards used in a substantive manner in the NRC regulatory process are main-tained at the NRC Library, Two White Filnt North,11545 Rockville Pike, Rockville, MD 20852-2738, for use by the public, Codes and standards are usually copyrighted ano may be purchased from the originating organiza-tion or, if they are American National Standards, from the American National Standards institute,1430 Broad-way, New York, NY 10018-3308.

DISCLAIMER NOTICE Where the papers in these proceedings have been authored by contractors of the United States Government, neither the Unitod States Govemment nor any agency thereof, nor any of their employees, makes any warranty, expressed or implied, or assumes any legal liability or responsibility for any third party's use, or the results of such use, of any information, apparatus, product, or process disclosed in these proceedings, or represents that its use by such third party would not infringe privately owned rights. The views expressed in these pro-ceedings are not necessarily those of the U.S. Nuclear Regulatory Commission.

m

NUREG/CP-0147 CNWRA 93-020 Proceedings of the Workshop on the Role of Natural Analogs in Geologic Disposal of High-Level Nuclear Waste Held in San Antonio, TX July 22-25,1991 I

Manuscript Completed: July 1995 Date Published: September 1995 Edited by 4

L A. Kovach, NRC W. M. Murphy, CNWRA Sponsored by

+

Division of Regulatory Applications Office of Nuclear Regulatory Research U.S. Nuclear Regulatory Commission Washington, DC 20555-0001 Proceedings prepared by U.S. Nuclear Regulatory Commission end Center for Nuclear Waste Regulatory Analyses 6220 Culebra Drive i

S::n Antonio, TX 78238-5166 J:b Code B6666 p*" s.,,

__.~.__ _.

a v

ABSTRACT i

A Workshop on the Role of Natural Analogs in Geologic Disposal of High-Level Nuclear Waste was held in San Antonio, Texas on July 22-25,1991. The proceedings comprise seventeen papers submitted by j

participants at the workshop. A series of papers addresses the relation of natural analog studies to the regulation, performance assessment, and licensing of a geologic repository. Applications of reasoning by

~

analogy are illustrated in papers on the role of natural analogs in studies of earthquakes, petroleum, and mineral exploration. A summary is provided of a recently completed, internationally coordinated natural analog study at Poqos de Caldas, Brazil. Papers also cover problems and applications of natural analog l

studies in four technical areas of nuclear waste management: waste form and waste package, near-field

~

processes and environment, far-field processes and environment, and volcanism and tectonics. Summaries

}

of working group deliberations in these four technical areas provide reviews and proposals for natural analog applications.

i i

i I

4 i

i k

l J

l 1

J d

i i

iii NUREGICP-0147

4 TABLE OF CONTENTS ic i

.P.NLe FOREWORD TO THE PROCEEDINGS OF THE WORKSHOP ON THE ROLE OF NATURAL ANA-LOGS IN GEOLOGIC DISPOSAL OF HIGH-LEVEL NUCLEAR WASTE by William M. Murphy and Linda A. Kovach..............................

1

- THE ROLE OF NATURAL ANALOGS IN THE REPOSITORY LICENSING PROCESS -

by William M. Murphy '.........................................

3 U.S. NUCLEAR REGULATORY COMMISSION NATURAL ANALOGUE RESEARCH PROGRAM by Lmda'A. Kovach and William R. Ott.................. 3 NATURAL ANALOG STUDIES: LICENSING PERSPECITVE by John W. Bradbury........

- 15 ROLE OF NATURAL ANALOGS IN PERFORMANCE ASSESSMENT OF NUCLEAR WASTE REPOSITORIES by Budhi Sagar and Gordon W. Wittmeyer.....................

21 LONG-TERM PREDICITONS USING NATURAL ANALOGUES by Rodney C. Ewing.......

29 l

ANALOG EARTHQUAKES by Renner B. Hofmann...........................

37 APPLICATION OFNATURAL ANALOG STUDIES TO EXPLORATION FOR ORE DEPOSITS by Donald L. Gustafson NATURAL ANALOGS IN THE PETROLEUM INDUSTRY by James R. Wood 49 THE POQOS DE CALDAS INTERNATIONAL PROJECT: AN EXAMPLE OF A LARGE-SCALE RADWASTEISOLATION NATURAL ANALOGUE STUDY by Michael Shea..........

61 NATURAL ~ ANALOGUE STUDIES AS SUPPLEMENTS TO BIOMINERALIZATION 67 RESEARCH by M.B.Mc Neil NATURAL GEOCHEMICAL ANALOGUES OF THE NEAR FIELD OF HIGH-LEVEL NUCLEAR WASTE REPOSITORIES by John A. Apps...

75 NATURAL ANALOGS FOR FAR. FIELD ENVIRONMENT / HYDROLOGY by Dwight T. Hoxie.. 101 j

WASTE FORMS, PACKAGES, AND SEALS WORKING GROUP

SUMMARY

by Narasi Sridhar and Michael B. McNeil................................ 105 NEAR-FIELD ENVIRONMENT / PROCESSES WORKING GROUP

SUMMARY

by Wilham M. Murphy......................................... 107 FAR-FIELD ENVIRONMENT WORKING GROUP

SUMMARY

by English C. Pearcy and Ralph E. Cady................................ ' 111 VOLCANISM /rECTONICS WORKING GROUP

SUMMARY

by Linda A. Kovach and Stephen R. Young............................... I15 v

NUREG/CP-0147

4 y

' LIST OF FIGURES

.SE!!*

.P.an -

3-1 Orfatnrmiinn=1 structure of high-level waste research program plan as it relates to regulatory requirements......................................

7

.3-2 Hierarchy of goals and objectives requiring supporting data and analyses............

8 33 Application of analogue studies to provide understandmg for time-temperature ranges for United States repository......................................

9

,3-4 Apphcability of near-field analogues to time-temperature curves for United States repcmiuty.

10

.3-5 Spent fuel analogues versus peak repository temperature.....................

10 3-6 Breakdown of NRC analogue studies with respect to hydrologic conditions...........

11 3-7' CanePal flow diagram of performance assessment showing possible contributions from natural analogue studies..............................

12 5-1 Performance requirements and steps for analysis.........................

22 5-2 Simniatinti of Las Cruces Trench Validationanariment.....................

25 5-3 Evolution of the ore body at the Oklo (natural analog) site (McKmley,1989)..........

26 8-1 Geologic age of world class are deposits..............................44 8-2 Hynnehade=1 mercury-hot springs-gold model developed utihzmg theoretical and field observations............................................

47 8-3 Map showing location of the McLaughlin Gold Mine, Napa and Yolo Counties, California 48 9-1 Andelinal model for gas and oil accumulation showing irregular distribution of gas, oil, and water in an asymmetric anticline..................................

49 9-2 A series of natural analogs showing hydrocarbon trapping in faulted strata with crosional uncomformities (Stewan,1951)..................................

50 9-3 Example of Allan fault-plane for a faulted anticline illustrating structural and stratigraphic geometry..............................................

51 9-4 Porosity versus depth for a series of wells in the Texas Gulf Coast................

52 9-5 Typical geobistory plot for a wellin the southem San Joaqum Valley of Califomia.......

53 9-6 Geohistory plot for reservoir interval of Stevens sarxl at N. Coles Levee 54 9-7 Plot of"Sr/"Sr versus computed time of crystalhzation for strontium data from N. Coles Levee................................................

54 8 Variations in fluid pressure, rock stress, porosity, viscosity, and permeability calculated using Marun equations..........

55 9-9 Initial model for seismic pumping based on concept of rock dilation in vicinity of a fault....

56 NUREG/CP-0147 vi

b r

1 LIST OF FIGURES (Cont'd)

.Figge Page

'l

9-10 Fault-valve model which replaces seismic pumping model;...................

' 57 '

.9-11

. Model for aasin compartments in which nonnally pressured rocks overlie several compartments by supernormally pressured rocks which are separated laterally by faults and vertically by shale or cement seals (Prowley,1990) '.................................

'57 9-12 Model for pressure compartment at Ekofisk field in the Cental Graben of the North Sea....

58 12-1 Schematic diagram to show the decomposition paths of rhyolitic glass when exposed to the aqueous phase...........................................

79 412-2 Schematic diagram to show the decomposition paths of basaltic glass when exposed to the aqueous phase...........................................

80 12-3. Solubility products of silica polymorphs as a function of temperature (Apps,1970)......

80 12 Schematic diagram to show thermodynarme and mass transfer considerations during i

irreversible dissolution and precipitation (Dibble and Tiller,1981)...............

80 i

i 12-5' Observed stability ranges of secondary minerals in hydrothermally altered Icelandic basalts 82 12-6 Activity ratios of major elements as a function of temperature in well water from Icelandic

. geothennal wells (calculated from chemical analyses cited in Arnorsson et al.,1983).....

83 12-7 Activity ratio of [Na*]/[K'] and saturation index of quartz as a function of temperature in well waters from Icelandic geothermal wells (calculated from chemical analyses cited in

Arnorsson et al.,1983) '......................................

84 1

12-8 Variation in log f02 versus fS2, calculated from volatile and noncondensible gas concentrations in stream from various geothermal fields (D'Amore and Gianelli,1984)....

85 i

12-9 Dispersion of yttrium and rubidium in relation to zirconium in Icelandic basalts n

(Wood et al., 1976)................................

85 12-10 Saturation indices of quartz, calcite, pyrite, low albite, potash feldspar, illite, laumonite, wauakite, and stilbite as a function of temperature in Icelandic well waters 88 i

12-11. Log [Na*]/[K*] calculated from compiled thermodynamic data as a function of temperature, compared with the conesponding ion activity products from geothennal well waters 90 12-12 Log K = [Na*][K*]/[Ca**] calculated from compiled thermodynanuc data as a function of temperature, compared with the corresponding ion activity products from geothermal

,~

well waters 92 4........

14-1 ' Classification of the engineered barrier systems.........................

106 i

vii NUREG/CP-0147

,.e, m.

4

,.4

1 LIST OF TABLES Table g

15-1 Matrix of nuclear waste repository near field issues and analog systems that niay be used to address the issuet............................... 109 16-1

. Speciationand solubilityhmits of key radianuclides: present knowledge.......

.... 114 4

e J.

3..

J i

1 4

NUREG/CP-0147 viii

Foreword to the Proceedings of the

]

WORKSHOP ON THE ROLE OF NATURAL ANALOGS IN GEOLOGIC DISPOSAL OF HIGH-LEVEL NUCLEAR WASTE William M. Murphy Linda A.Kovach Centerfor Nuclor Waste Regulatory Analyses U.S.NuclarRegulatory Commission San Antonio, Texas 78228 Washington, D.C. 20555 A Workshop on the Role of Natural Analogs in ogy in a variety of scientific and engineering endeav-Geologic Disposal of High-Level Nuclear Waste ors. Subsequently, separate working groups ad-(HLW) was held in San Antonio, Texas, on July dressed the use of r,itural analogs in four technical 22-25,1991. It was sponsored by the U.S. Nuclear areas of nuclear waste management: waste package Regulatory Commission (NRC) and the Center for and waste form; near-field processes and environ-Nuclear Waste Regulatory Analyses (CNWRA). In-ment; far-field processes and environment; and vol-vitations to the workshop were extended to a large canism and tectonics. Working groups were number ofindividuals with a variety of technical and instructed to define specific technicalissues to which professional interests related to geologic disposal of natural analog studies can contribute, to evaluate the i

nuclear waste and natural analog studies. Participa-status of studies on these issues, and to identify areas tion by over 50 scientists and engineers included staff of additional fruitful research. Conclusions reached s

members of the NRC and CNWRA and repre-by the separate working groups were reviewed in a sentatives from the U.S. Department of Energy closing plenary session.

(DOE); the U.S. National Laboratories; the U.S. Geo-These proceedmgs comprise manuscripts written logical Survey; several uruversities and private or-by plenary session speakers, additional papers con-

)

ganizations; the Nuclear Waste Technical Review tributed by workshop participants, and summaries of Board; the Advisory Committee on Nuclear Waste, results from each working group. Five articles (Chap-Clark County, Nevada; and other orgamzations. Con-ters 2 through 6) address the relation of natural analog tributors to the workshop are identified in the sum-studies to the regu' don, performance assessment, maries of the working groups proceedings (Chapters and licensing of a geologic repository for HLW. A 14 to 17). The objective of the workshop was t series of papers then focuses on applications of rea-examine the role of natural analog studies m perfonn-soning by analogy in other earth science applications, ance assessment, site characterization, and pnontiza-including the effects of earthquakes on engineered tion of tesearch related to geologic disposal of HLW.

structures (Chapter 7) and exploration for ore depos-Expert opinions were informally solicited from its and petroleum (Chapters 8 and 9). In addition, an memben of the nuclear waste management commu-oral presentation at the workshop addressed natural nity and from individuals outside this field. Several analogs studies in the prediction of future volcanic presentations focused on natural analog studies activity and volcanic risk assessment. In Chapter 10, mounted specifically in support of geologic disposal an overview is provided of a recently completed, i

of nuclear waste. In addition, contributions were pro-intemationally coordinated natural analog study at vided by scientists and engineers from other fields Po90s de Caldas, Brazil. Papers are also presented on who routinely construct conceptual and computa-problems and applications of natural analog studies tional models for the evolution of geologic systems in each of the four technical areas addressed by the and who have experience in model validation using working groups (except volcanism / tectonics) (Chap-data from natural systems. A theme of the meeting ters 11 through 13). Finally, the proceedings and was the generality of reasoning by analogy in earth conclusions of the working groups are summarized science applications.

in Chapters 14 through 17.

In an opening session, presentations focused on Divene subjects and points of view were encour-relations between natural analogs and nuclear waste aged and freely aired at the workshop. Both the utility management and applications of reasoning by anal-and limitations of natural analog studies were 1

NUREG/CP-0147

i i

Foreword l

stressed. Debate developed on many topics from the greater definition of the utility of natural analog stud-specifics of the thermodynamm properties of miner-ies in site charactenzation and performance assess-l als to the generality of the range of systems for which ment for geologic disposal of HLW. The editors j

natural analog studies are appropriate. All workshop express their sincere appreciation for the insightful i

participants were invited to provide written contribu-contributions made by workshop participants and the j

tions to these proceedings, and divergent views are serious and time-consuming efforts of speakers, j

respectfully represented here. After review and revi-authors, working group coordinators, reviewers of sion, all submitted manuscripts have been included.

manuscripts, and editorial assistants.

I Although many insights were gained and problems The chapters of this document represent contribu-

]

were clarified at the workshop, the role of natural tions of the individual authors or workshop partici-analogs for geologic disposal of HLW continues to pants. They do not necessarily reflect the views or be an issue of debate and definition.

regulatory positions of the NRC, the DOE, or other l

The organizers and editors desire that the work-orgamntions with which the authors are affiliated.

i shop activities and these proceedings contribute to a

M." one of which is natural analog j

determme their evolutionary cycles. This information studies, be used to support predictive models of per-

?

will then be used to estimate probabilities of mag-formance assessment and site charactenzation activi-matic disruption of the repository.

ties. These studies mpply information conceming k

The draft Regulatory Guide DG-3003 (NRC, e nditions, processes, and events, both anticipated 1990), Format and Content Guide for the License and unanticipated, at the site. Finally, it :hould be Application for the High-Level Waste Repository, recogmzed that " natural analog snidies" can be ap-i can be used as an indication of the range of topics to li P ed to aspects of the repository site characterization which natural analog studies may be applied. In that and performance assessment that are not necessarily i

document, the term " natural analog studies" is used ge chenu,calin nanue.

numerous times in supporting predictive models ap-plied to topics as diverse as geology, hydrology, 4.5 BIBLIOGRAPHY geochemistry, climatology and meteorology, shafts Bonano, E.J., S.C. Hora, R.L. Keeney, and D. von l

and ramps, underground facility, and waste form and Winterfeldt,1990, " Elicitation and Use of Expert

{

packages.

Judgment in Perfonnance Assessment for High-Studies of natural analogs require that two sys-Level Radioactive Waste Repositories,"

I tems/ sites have to be adequately characterized to NUREG/CR-5411, U.s. Nuclear Regulatory show analogous behavior or conditions. Conse-Commission, Washington, D. C.

quently, the initial identification and selection of Chapman, N.A.,1992, Natural Analogues: The State natural analogs is often rudimentary. For example, f Play in 1992, Proceedings of the Tlurd Interna-uranium ore deposits have been assumed to mimic tional Conference on High-Level Radioactive repository chemistry; the unsaturated zone intruded Waste Management. American Nuclear Society, by magma has been assumed to mimic the heated LaGrange Park,IL, p. 1695-1700.

hydrologic system of a HLW repository in the unsatu-Genz, C.P.,1992, Ixtter to T.H. Isaacs entitled Final rated zone; redox fronts mimic corrosion product-Repon of the Natural Analogue Review Group radionuclide interactions. In order to find out how (NARG).

good the analogs are, one must carry out a charac-U.S. Depanment of Energy,1988, " Site Charac-terization program comparable to that of the site.

terization Plan: Yucca Mountain Site, Nevada Re-Only as the characterization progresses can the selec-search and Development Area, Nevada,"

tion of analogs become more refined and specific.

DOE /RW-0199.

NUREG/CP-0147 18

Na: ural Analog Studies: Licensing Perspective U.S. Nuclear Regulatory Commission,1983, Staff radionuclides to the accessible environment as a Analysis of Public Comments on Proposed Rule function of time; and a similar evaluation which 10 CFR Part 60 " Disposal of High-Level Radioac-assumes the occurrence of unanticipated processes tive Wastes in Geologic Repositories," NUREG-and events.

0804.

(D) The effectiveness of engineered and natural U.S. Nuclear Regulatory Commission,1990, Draft barriers, including barriers that may not be them-Regulatory Guide DG-3003 (1990), Format and selves a pan of the geologic repositc,ry operations Content Guide for the License Application for the area, against the release of radioactive material to the High-level Waste Repository.

environment. The analysis shall also include a com-Vaniman, D.T., D.L. Bish, and S. Chipera,1988, "A parative evaluation of alternatives to the major design Preliminary Comparison of Mineral Deposits in features that are imponant to waste isolation, with Faults near Yucca Mountain, Nevada, with Possi-particular attention to the alternatives that would pro-ble Analogs,"LA-11289-MS, UC-70,Los Alamos vide longer radionuclide contamment and isolation.

National Laboratories, Los Alamos, New Mexico.

(E) An analysis of the performance of the major design structures, systems, and components, both

4.6 APPENDIX

SECTIONS OF10 CFR surface and subsurface, to identify those that are PART 60 REFERRING TO NATURAL important to safety. For the purposes of this analysis, ANALOG STUDIES it shall be assumed that operations at the geologic In 660.21(a) Content of Application, the Rule repository operations area will be carried out at the states that "An application shall consist of general maximum capacity and rate of radioactive waste information and a Safety Analysis Report."

stated in the application.

Further, the Rule describes the information con-(F) An explanation of measures used to support the tained in a Safety Analysis Report in l60.21(c) where models used to perform the assessments required in it states,'The Safety Analysis Report shallinclude:

Paragraphs (A) through (D). Analyses and models (1) A description and assessment of the site at which that will be used to predictfuture conditions and the proposed geologic repository operations area is to changes in the geologic setting shall be supported by be located with appropriate attention to those features using an appropriate combination of such methods as of the site that might affect geologic repository op-field tests, in situ tests, laboratory tests which are crations area design and performance."

representative of field conditions, monitoring data, 560.21(c)(1)(ii) states that "the assessment shall con-and natural analog studies."

tain:

In Subpart E-Technical Criteria, 560.101, Pur-(A) An analysis of the geology, geophysics, hydro-pose and Nature of Findings, requires ".. a finding geology, geochemistry, climatology, and meteorol-that the issuance of a license will not constitute an ogy of the site, unreasonable risk to the health and safety of the (B) Analyses to determine the degree to which public. The purpose of this subpart is to set out each of the favorable and potentially adverse condi-performance objectives and site and design criteria tions, if present, has been characterized, and the ex-which, if satisfied, will support such a finding of no tent to which it contributes to or detracts from unreasonable risk."

isolation. For the purpose of determining the presence Finally, {60.101(a)(2) reads, "While these per-of the potentially adverse conditions, investigations formance objectives and criteria are generally stated shall extend from the surface to a depth sufficient to in unqualified terms, it is not expected that complete determice critical pathways for radionuclide migra-assurance that they will be met can be presented. A tion from the underground facility to the accessible reasonable assurance, on the basis of the record be-environment. Potentially adverse conditions shall be fore the Commission, that the objectives and criteria investigated outside of the controlled area if they will be met is the general standard that is required.

affect isolation within the controlled area.

For 560.112, and other portions of this subpan that (C) An evaluation of the performance of the pro-impose objectives and criteria for repository perform-posed geologic repository for the period after perma-ance over long times into the future, there will inevi-nent closure, assuming anticipated processes and tably be greater uncenainties. Proof of the future events, giving the rates and quantities of releases of performance of engineered barrier systems and the L9 NUREG/CP-0147

l Natural Analog Studies: Licensing Perspective i

l geologic setting over time periods of many hundreds will be in conformance with those objectives and or many thousands of years is not to be had in the criteria. Demonstration of compliance with such ob-ordmary sense of the word. For such long-term ob-jectives and criteria will involve the use of data from jectives and criteria, what is reqmred is reasonable accelerated tests and predictive models that are sur assurance, mahng allowance for the time period, ported by such measures as field and laboratory tests, i

hazards, and uncertamties involved, that the outcome monitoring data and natural analog studies."

4 i

i i

E 4

1 s

4 o

i 4

3 i

j h

4 NUREG/CP-0147 20

ROLE OF NATURAL ANALOGS IN PERFORMANCE 5

ASSESSMENT OF NUCLEAR WASTE REPOSITORIES Budhi Sagar and Gordon W.Wittmeyer Centerfor Nuclarr Waste Regulatory Analyses San Antonio, Texas 78228 1

5.1 ABSTRACT temporal scales. Second, the conditions under which repos% is expected m pe&nn long h h Mathematical models of the flow of water and future are hard to define and are, to some extent, transport of radionuclides in porous media will be speculative in nature. The first uncertainty results used to assess the ability of deep geologic repositories prhnarily form the fact that, with the cunent technol-1 to safely contam nuclear waste. These models must, ogy requinng drilling or excavation, it is difficult to in some sense, be validated to ensure that they ade-fully characterize a heterogeneous site without seri-quately describe the physical processes occurring outly impacting its waste isolation capability. The within the repository and its geologic setting. Inas-second factor arises from the difficulty of predicting much as the spatial and temporal scales over which natural events and processes farinto the future, a task these models must be applied m performance assess-that becomes more daunting when the effects of ment are very large, validation of these models human actions must be considered.

against laboratory and small-scale field experiments may be considered inadequate. Natural analogs may Assuming that natural analogs are selected based provide validation data that are representative of on desirable attributes, which include spatial and physico-chemical processes that occur over spatial temporal scales similar to those of repositories [see and temporal scales as large or larger than those Pearcy and Murphy (1991) for further discussion of relevant to repository design. The authors discuss the this aspect), they present unique opportunities for manner in which natural analog data may be used to studying phenomena important to repository per-increase coufic ace in performance assessment mod.

formance at those scales. A sampling ofliterature on els and conclude that, while these data may be suit-natural analogs [e.g., see Commission of European r&le for testing the basic laws governing flow and Communities report EUR 13014 EN (Alexander and transport, there is insufficient control of boundary McKinley,1991), and the literature reviews by and initial conditions and forcing functions to pennit Pearcy and Murphy (1991)] suggests that study of quantitative validation of complex, spatially distrib-natural analogs may be used to:

uted flow and transport models. The authors also (i) identify processes that operate at large scales; express their opinion that, for collecting adequate (ii) determine how processes are coupled so that data from natural analogs, resources will have to be conceptualmodels can be developed; devoted to them that are much larger than are devoted (iii) estimate rates at which various processes op-to them at present, erate so that appropriate constitutive equa-tions can be formulated;

5.2 INTRODUCTION

(iv) validateperformanceassessmentmodels;and (v) ai qua at vec rmborationofrepository The unusually large spatial and temporal scales associated with high-level nuclear waste (HLW) geo-logic repositories present a major challenge to radi-In some of the natural analogs literature, the terms ologic safety assessment.' WhHe the basic scientific validation and verification are applied interchange-laws applicable to geologic waste disposal are the ably to models. For this paper, a model is defined as same as for other engineering projects, there are two an abstract concept representing the complex physio-main features that reduce the level of confidence with chemical processes-the abstraction being specific which the future performance of the repositories can for the purpose for which the model will be used. In

' be determined. First, uncertainties in site and design practice, the concept will be described by an algo-data tend to grow larger with increasing spatial and rithm, for example, complemented through computer 21 NUREG/CP-0147

Role ofNatural Analogs in Performance Assessment code or software. These computer codes are verified United States, there are quantitative performance to assure that the implementation of the underlying measures for both the total system and the subsys-model is correct. However, the model itself is vali-tems. This is shown in the bottom line of the chart in dated to assure that the conceptual abstraction of the Figure 5-1, where five regulatory requirements appli-processes is acceptable. Validation of performance cable to HLW repositories are identified. The three assessment models with data from natumi analogs is left-hand boxes identify the generally applicable en-the most often cited reason for undertaking analog vironmental standards promulgated by the U.S. En-studies. It is this aspect of natural analog studies that vironmental Protection Agency (EPA)(EPA,1985).

will be examined in this paper.

The EPA standards apply to the entire repository system. In contmst, the remaining two boxes identify 5.3 BRIEF OVERVIEW OF the rules developed by the U.S. Nuclear Regulatory PERFORMANCE ASSESSMENT Commission (NRC) (NRC,1983), and these apply to We shall restrict the definition of performance Particular subsystem (e.g., Groundwater Travel Time assessment for the purpose of this paper to quantita.

for the Site subsystem and the Package Life and the tive estimates of measures of future repository per.

Release Rate Rules for the Engineered Barriers sub-formance. The performance measures is usually system). It may be noted that the EPA standard will defined in regulations that vary from country to coun.

eventually be integrated into the NRC rule for the try. Annual risk is probably the most common meas-Purpose ofits implementation.

ure of perfonnance, human dose is another. In the SITE WASTE FActuTY k

---

===-==----e-CHARAC TERISTK:S Pf10PERTIE S OESIGN t DISPOSAL SYSTEW h

f h

f CHARACTER 12A flON ANO RFOKNAL PROCESS CESIGN l

MODEL MODELINO MCC(UNG CONCEPTUAL-I I

SCENARIO l

DEVE Mil 2 SCENARto DEVELOPME N T 1

1 tr ti PROEIA81UTY A

/T ES T1MA TES MOOEUNG l

4 l

"';"g.oJa 1-s I,

(SOUF4CE TERM) l 4

'l OROUNDWATER 3 CONSEQUENCE THANSPORT ANALYSIS g

MOD &UNG l

4 l

N RT 4

8 l DOStMETRY l 4

m m

m m

i UNCERTAaNTY UNCE R T AIN TY

1. RACE RTAINTY UNCERTANTY UNCERTAINTY d'

4 4

4 1

ANALYS15 ANALYSAS A%Ysa3 ANA'.YSt$

ANALY515 4 SENairT1Y'TY 4

4 4

4 4

"4"l'i""

40 CFfi 1 113 40C R 1 44 40 CFR t 115 to so v3(n t0 CF

13rf, I--

e CONTalhMENT

• GROLHOW ATT R e 240tviOVAu

• PACKAGE UFE

• GROUNOWATER 5 REGULATORY m

Pf40 TE C TKIN PRO TECTION

• RELEASE RAff TfuYEL TIME COM AN E f

(10,000 YR)

(Ko0 VFu

( 9002) YR; (1000 VR)

{

t l

l Figure 5-1. Performance requirements and steps for analysis NUREG/CP-0147 22

Rde ofNatural Analogs in Performance Assessment Figure 5-1 also summarizes various analyses that pragmatic view may tolerate a higher degree of con-are thought to be required for assessing performance servatism. In any case, knowledge of the " degree of using the current methodology. It may be seen from conservatism" may be important for regulatory deci-Figure 5-1 that assessing performance requires mod-sion making. This is certainly the case in the United eling of geologS, hydrologic, geochemical, thermal, States, where the regulations acknowledge that strict and mechanical processes. It is known that these mathematical proof of the future performance of the processes are coupled in a complex way, although, repository is impossible and, hence, require " reason-with the current state-of-the-art, all of the couplings able assurance" or " reasonable expectation" in meet-cannot be fully described. The representation of these ing the desired safety goals. Currently, it is common couplings in the performance assessment models is to conduct both detailed realistic simulations for cer-commonly simplified to accomplish the calculations.

tain individual processes or a critical part of a system One of the difficult strategic decisions in performance and simplified conservative simulations for the total assessment is to balance the complexity of models system.

against available data on site and engineered barriers.

We assume that the conceptual abstractions con-It is very tempting for the modeler to use an overly stituting a model will eventually be translated into complex model even if data to support such a model mathematical models. Generally accepted mathe-do not exist It is equally tempting for the data collec-matical forms suitable for simulating physical sys-tor to over-sample one facet of the site or to empha-tems are parametric in nature. The uncertainties in size one aspect of the design and neglect others. Some these model parameters, and also in the form of basic considerations in selecting models for perform-relationships representing constitutive and state ance assessment and their implementation as com-equations, can collectively be called " technical un-Puter codes are:

certainties." Usually, these technical uncertainties are (i) Mechanistic (in contrast to empirical) repre-required to be explicitly represented in performance sentation of processes is preferable. Mecha-assessment models. For example, the risk measure of nistic representation requires explicit use of performance incorporates in it not only the conse-the basic principles of conservation of mass, quence, but also the probability of the causative momentum, and energy at an appropriate event. The United States HLW regulations (EPA, scale. Constitutive laws and state equations 1985) incorporate probabilities explicitly; that is, used in the fonnulation of mechanistic mod-they specify not only a level of performance, but also els are normally derived from observations.

the probability level at which it must be met. This (ii) Models whose complexity is compatible with dictates that the performance assessment models be the complexity of site and design data are probabilistic in nature, preferable (iii) Flexible it iplementations (e.g., numerical so-fact that perfonnance assessment models can vary in their degree of realism (or conservatism) and lutions) are preferable. Flexibility is essential to analyze system behavior under the widely can be either deterministic or probabihstic can have a lar8e impact on how model validation is to be varying conditions that may occur in the fu.

defined and demonstrated.

(iv) Fastandefficientimplementationispreferred so that sensitivity and uncertainty analyses 5.4 BRIEF OVERVIEW OF MODEL can be performed.

VALIDATION Another strategic decision to be made in perform-The concept of validation is generally defined ance assessment modeling is whether to model"real-from the view of realistic, deterministic models. In istically" or " conservatively." Only a degree of this context, model validation requires corroboratica realism or conservatism is implied here, since these that, under site specific conditions, the abstracted terms cannot be defined in an absolute sense. Inclu-model represents " reality" and, therefore, the model sion of greater detail regarding space-time depend-estimates of the (unverifiable) future state of the ence of processes leads to a higher degree of realism system are acceptable. Since the system states can be in the conceptual model A higher degree of realism observed only in the present, there are no experimen-is preferable from a scientific viewpoint, while a tal means to detennire its future states. Therefore, no 23 NUREG/CP-0147

Role ofNatumi Analogs in Performance Assessment means to compare model predictions to actual system involved parties a consistent set of rules which, when states are available, followed, will provide a meaningful estimate of the Therefore, for practical purposes, model valida.

degree of validation. Although it is easier to define tion is sought by comparing model results to experi.

the gmdness-of-fit weight in a quantitative manner, ments conducted by design (laboratory or field this measure is still dependent upon the objective for which the model is to be used. For example, if early experiments) or by nature (natural analogs in the case arrival of low concentrations of contammants is not of the HLW repositories). Natural analogs are sys.

tems whose behavior, at least in certain well-defined of concem, it may not be necessary that the model aspects, is analogous to the system under investiga.

accurately predict the entire shape of a contaminant P ume; it may be sufficient that the model estimates l

tion. In addition, the analogous system has evolved so that many of its states have been observed. If a the migration of the plume centroid and increase in model can be valida:ed against the analog, then this local plume spread in a reasonable manner. Other such measures are discussed in the paper by Ababou model may be assumed to apply to the system of et al. (1992).

interest.

We note that, based on Popper's (1959) philos &

5.5 MODEL VALIDATION WITH phy, the very idea that a theoretical model can b NATURAL ANALOGS validated by any one experiment on any spatial or temporal scale has been criticized on logical grounds.

While most laboratory and field experiments have in the Popperian view, experiments may only refute the advantage of human control, their spatial and (rather than validate) models. Thus, simply because temporal scales a e usually much smaller than those model and nperimental results compare does not ofinterest for the design of the actual repository. Even constitute a proof of model validity. Only when no carefully designed field-scale experiments may still experiment can be found to refute a model may it be lack control at the boundaries or fail to measure declared validated. Therefore, model validation is important state variables with sufficient spatial and impossible in the strict sense. These and other con-temporal resolution. An example of insufficient spa-siderations of model validation are discussed by tial resolution of field measurement locations is Ababou et al. (1992) where it is accepted that a model shown in Figure 5-2. It depicts contoured bromide cannot be absolutely validated, and the concept of concentrations from a model simuhtion overlain by

" degree of validation" is introduced for practical dots which represent the actual solute sampling loca-applications. Admittedly, this concept is subjective in tions for a controlled test at the partially saturated Las nature, but it may be of value in the regulatory envi-Cruces Trench site (Wierenga et al.,1989). It is ronment.

readily apparent that even though the measurement The degree of validation of a model can be esti-Points are closely spaced, the computed plume shape is such that it is difficult to draw a conclusive infer-mated by assigning weights that depend upon both the scope of the experiment and the goodness-of-fit ence regarding the match between the model results and observations. Different conceptual models have between model and observed data. These weights are been developed in which the complexity of both the accumulated over the suite of experiments for which init al saturations and the spatial structure of the i

validation is attempted. With regard to the scope of soil-hydraulic properties are varied. In support of the an experiment, the validation weight assigned to the experiment is proportional to the broadness of the c neept that the model must not, in some sense, range of test conditions. In general, the broader the incorporate more detail than is capable of being re-test conditions, the better test of the applicability of s Ived in the actual experiment, the simple models the model it is. Thus, a field-scale experiment with Produced simulated bromide plumes whose first and second moments most closely matched those ob-

{

many scales of heterogeneity and varying boundary conditions will have a higher validation weight than served (Sagar and Wittmeyer,1991),

an experiment on a small homogeneous sample.

The Las Cmces Trench experiment is specifically Again, a subjective judgment will have to be made designed for validating flow and tansport models in regarding the weight assigned to a validation experi-unsaturated, unconsolidated soils similar to condi-mmt based on its scope. In the regulatory arena, it tions expected at low-level waste sites. The spatial may be possible to specify through agreement of domain of the experiment is about 15 x 15 x 6 m, and NUREG/CP-0147 24

=. -... -.

o

\\

Go00-3.5E-4 3.0E-4 2.5E-4

-260 2.0E-4 e6 1.6E-4 S

t4 3.t E-4 3

5.0E-S

~

T.0E-6

• f

-600 i

g

~

I e

i I

s-i I

e i

250 500 750 1 s X(cm)

[

l S

Figure 5-2. Simulation of Las Cruces Trench Ve!!dation Experiment. Note the difficulty of measuring the contaminant plume even when samples are located in a systematic manner.

~

i Role ofNatural Analogs in Puju,,.;,a Assessment the infiltration and redistribution of water and solute models. Consider the evolution of the Oklo ore body is monitored over approximately 400 days. While (Cowen,1976), an important natural analog to spent there is considerable heterogeneity in soil properties, fuel, since it contains transuranic activation prod-there are no fractures or fracture-like features that ucts, and fission products. The Oklo ore body func-would generally be expected in a repository located tioned as a natural reactor some 2 billion years ago in consolidated materials. While consistent rules for with intermittent operation in different zones (see assigmng ' validation weights' have not been formu-pigure 5-3). Significant site characterization (not too lated yet, it is worthwhile to consider the Las Cruces dissinular from what will be needed for an actual site)

Trench experiment as an example of how such rules would be required to esumate the present state of the may be formulated. The space scale of the experiment system. The present system state, though imperfectly is two orders of magnitude larger than a laboratory known, is but one poirt on the evolutionary curve.

core, and it is about two orders of magnitude smaller What went on m the more than 2 billion years of its than that of a low-level waste (LLW) repository. The time scale of the experiment is about three orders of ev lutionaryhistmy(e.g,geologicchanges,climanc magnitude smaller than that of the LLW repository.

variations, and geochemical and transport processes)

Significant natural heterogeneities similar to those in can only be roughly estimated. As is true of most the LLW repository are present in the experiment.

inverse problems, tracmg the evolution backward in There is considerable confidence in the data because time is fraught with difficulties, such as nonunique-of the level of control over the experimental condi-ness and mathematical instability. Particular care tions. A weighting scale can be assigned to all these must be used with natural analogs so that one does factors for arriving at a composite validation weight not end up simply comparing results of one model for the experiment. Altematively, rather than formu-(that used to generate the evolutionary history) with late explicit rules, an expert panel may be asked to those of another (that is being used for performance consider all of these factors and assign such a weight.

assessment). Therefore, despite the fact that the Note that the validation weight for the same ex}mi-ment would be less for the case of models that are to be applied to the HLW repositories because of the o,oga, go,,o absence of features (e.g., fractures) that may domi-s l

nate the flow and transport in hard trxt. Note also i

that the degree of validation of a particular model will be determined by combining the validation weight of l

the data set (or expenment) and the degree to which

,4 l

the model results (e.g., concentrations, centroid of the j

plume, etc.) compare with that of the experiment.

E

", U depletion in 2

All natural analogs are only approximately analo-3 3 Y okin reactor zones l

gous to the real system, if for no other reason than Rate in l\\

f

\\

e that the future evolution of the real system is not Reactor l sS x

=

ron.3 known and, the:efore, it cannot be said that it will be

[2 _ - - 1 \\

v$u$s similar to the analog. However, considering that sub-p l l [N, jectivity is inherent in determining the degree of W

\\

N Extrapolation P

\\N' to the future validity, even approxime analogs can be of great t,

yq< t'X

/

M#

utility when the degree of approximation is accounted

{

for in the validation weight. Natural analogs may be i

N-() - } s Eca assigned higher validation weights because these 00 """

h l

have space-time scales similar in magnitude to the

' P '"

actual system. But this has to be compensated by the past

-2

-1 present 1

future fact that natural analogs suffer from the lack of ex.

Tirne (bnon years) perimental controls. The initial conditions from which a given analog has evolved are never known Figure 5-3. Evolution of the ore body at the and may at best be arrived at through the use of other Oklo (natural analog) site (McKinley,1989) l l

NUREG/CP-0147 26 l

Rde ofNatural Analogs in Performance Assessment space-time scales of analogs are comparable to those mm/yr (Johnson and Francis,1980). It is generally of the repository system, lack of knowledge ofiuitial accepted that the age dating of the artifacts is prob-and boundary conditions will cause the validation ably correct, but that it is impossible to define the weight assigned to natural analogs to be low. For environmental conditions under which degradation quantitative model validation, it would be useful to occurred. It is also argued that some metal artifacts adequately characterize the analog at this time and may have corroded at a much higher rate and disap-then monitor it (and also its boundary conditions) into peared, thus biasing this estimate. McKinley (1989) the future for, say, many tens (if not hundreds) of argues in the safety assessment of the Swiss program years. Such data, if attainable, would have much that, based on observations from natural analogs, an higher validation weight for HLW repositories.

assumption of corrosion rate of 0.1 mm/yr with a P tting factor of 3 is conservative. Such use of analog i

Even in the absence of the data suggested above, qualitative conclusions of great relevance can be data to estimate the degree of conservatism is worth-drawn, for example, if Oldo has maintained its integ.

while, but it does not provide support to model vali-rity (i.e., the radionuclides have migrated only so far) dation.

for over 2 billion years, then a HLW repository in a From the limited literature study performed by similar environment might be expected to be safe for these authors,it is apparent that qualitative arguments a few million years. However, as indicated by Chap-like that of McKinley (1989) are characteristic of man and Miller (1993), caution should be exerted in natural analog studies more than quantitative model drawing such general conclusions. In support, they validation. Even such conclusions as those of Mckin-cite the following excerpt from nuclear industry lit-ley are limited to the near-field processes. For exam-erature, "The Oklo reactors ran gently at the kilowatt-ple, these are related to the leach rates and power level for millions of years. They never blew radionuclide solubilities. We have not come across up. The radiation and waste from them did not deter many conclusions regarding far-field processes of surrounding life forms. Over immense timescales, the even the qualitative kind.

waste has barely moved away from the reactor site.

Because of the relatively large cost of in-depth As a result, scientists today are confident that waste investigations of natural analogs, some of the these in man-made stores and repositories is likely to move studies in the past have focused on narrow issues even less...." It may be noted that an opposite related, for example, to geochemistry. Alexander and argument can be made. The lack of abundance of McKinley (1991) and McCombie (1991) recommend similar natural reactors can be interpreted to mean that natural analog studies not remain focused on that such a system is, in general, unstable. Of course, narrow issues but adopt a broader prospective. In the the scarcity of such systems may be merely a reflec-context of model validation, we believe that data from tion of the lack ofinvestigations and not ofits inherent natural analog studies will have only low to moderate instability. In addition to qualitative understanding, validation weight. Hence, it may, in fact, be more important quantitative insights regarding elemental productive to plan analog studies to address well-de-speciation and processes governing migration have veloped but narrow specific issues. For example, been gained from the investigation of the Oklo analog inferences regarding average rates of various physi-site.

cal processes under conditions of natural analogs may In the view of the authors, natural analog studies Provide useful bounds for assessing performance of can be effectively used to ascertain, in a broad sense, HLW repositories under similar conditions. Such the applicability of basic laws describing flow and bounds are only marginally useful for model valida-reactive transport processes to greater spatial and tion, but may be invaluable for providing qualitative temporal scales. In addition, as indicated by Pearcy assurances regarding the effectiveness of the reposi-and Murphy (1991), natural analogs may be useful in tory system.

conveying to the public a sense of the long-term stability possible in a geologic environment.

5.6

SUMMARY

AND CONCLUSIONS Similar problems also arise in natural analog stud-We are somewhat pessimistic regarding the use of ies related to engineered barriers. Based on discovery natural analog data for the specific purposc of model of archeological artifacts, for example, the corrosion validation. The primary reasons for this pessimism rate of iron and copper has been estimated to be 0.1 are:

27 NUREG/CP-0147

Role ofNaturd Analogs m Perfonnance Assessment (i) lack of experimental control of the natural Cowan, G.A.1976. A natural fission reactor. Scien-analog; t#ic American 237: 36-47.

(ii) unknown initial conditions from which the Environmental Protection Agency.1985. Environ-analog evolved; mental Radiation Protection Standards for the 4

(iii) need of relatively large resources to ade-Management and Disposal of Spent Nuclear Fuel, quately characterize even the present state of High-Level and Transuranic Wastes; Final Rules, the analog.

Title 40 Part 191. Code of Federal Regulations.

While useful qualitative results can be obtained by Washington, DC. EPA.

studying the natuml analogs at the current level, it is Johnson, A.B. and Francis, B.1980. Durability of doubtful that these would generate data that would metals from archeaological objects, metal meteor-have high validation weight in quantitative model ites and native metals. PNL 3198, Pacific North-validation.

west Lab., Richland, WA.

We also believe that natural analogs alone, even if McCombie, C.1991. Critical uncertamties in safety more resources are committed to their study, will not assessment and how to address them. Proceedings provide the preponderance of evidence that will be of the Fourth Natural Analogue Working Group required for validating performance assessment mod-Meeting. EUR 13014 EN. Brussels, Belgium:

els. Such evidence can be accumulated only by con-EEC: 19-29.

ducting a variety of tests under a variety of conditions.

Mc Kinley, I.G.1989. Applying Natural Analogues in However,if a natural analog site can be exammed in Predictive Performance Assessment, Part 1: Prin-detail and monitored for the long term (at 1 cast tens ciples andRequirements, Risk Analysis in Nuclear of years), then it would be possible to constmct an Waste Management. Brussels, Belgium: EEC:

expenment with high validation weighL 359-375.

Popper, K.R.1959. The Logic ofScientVic Discovery.

5.7 BIBLIOGRAPITY London, England: Hutchinson Co.

Sagar, B., and G. Wittmeyer.1991. Chapter 8: Per-Ababou, R., B. Sagar, and G. Wittmeyer.1992. Test-formance Assessment Research, Report on Re-ing procedures for spatiahy distributed flow mod-search Activities for the Quarter April 1 through els. Advances m Water Resources 15(3): 181-198.

June 30,1991. CNWRA 91-03Q. San Antonio, Alexander, W.R., and I.G. McKinley.1991. Natural Texas: CNWRA.

analogues in perfonnance assessment: Improving U.S. Nuclear Regulatory Commission.1983. Dis-models of radionuclide transport in groundwaten, posal of high-level radioactive wastes in geologic by studying the natural environment. Proceedmgs repositories, technical criteria, Title 10 Part 60.

of the Fourth Natural Analogue Working Group Code offederal Regulations. Washington, DC.

Meeting. EUR 13014 EN. Bmssels, Belgium:

Pearcy, E.C., and W.M. Murphy.1991. Geochemical EEC: 119-151.

natural analogs literature review. CNWRA 90-008.

Chapman, N.A. and W.M. William.1993. Using San Antonio, Texas: CNWRA information fmm natural systems to build confi-Wierenga, P.J., A.F. Toonnan, D.B. Hudson, J. Vin-dence in perfo:mance assessment. Pre-Print of son M. Nash, and R.G. Hills.1989. Soil physical Proceedings of the Fifth Natural Analogue Work-properties of the Las Cruces treneh site.

ing Group Meeting, Toledo, Spain.

NUREG/CR-5441. Washington, DC: NRC.

NUREG/CP-0147 28

=-

l e

1 LONG-TERM-PREDICTIONS USING g

i-NATURAL-ANALOGUES

~

i-Rodney C. Ewing UnimsityofNewMexico Abuquerque,NewMenco 87131 1

1 6.1 ABSTRACT nism of the waste form or sorpuon mecha-One of the unique and scientifically most challeng-nisms during transport through an aquifer)?

@) h h WW physical ud chemical ing aspects of nuclear waste isolation is the ex rapo-parametersaccuratelyknownandayymr te s

j lation of short-term laboratory gta (hours to years) to the conditions in the repository (e.g., acti-3 to the long time periods (10 -10 years) required by nide solubilities in a concentrated brine)?

regulatory agencies for performance assessment.The (iii) Are the scenarios that are selected to describe duect validation of these extrapolations is not possi-the possible release of radionuclides plausible ble, but methods must be developed to demonstrate

& ly intmsinT) compliance with government regulations and to sat-

' isfy the lay public that there is a demonstrable and At this tinae, one of the most detailed and elaborate reasonablebasisforre pdngthelong-termextrapo-efforts at performance assessment is the work at lations.

Sandia National Laboratories on the performance Natural systems (e.g., " natural analogues") pro-assessment of the Waste Isolation Pilot Plant in New r

l vide perhaps the only means of partial "vahdation,"

Mexico (Hunter 1989; Bertram-Howery and Hunter, as well as data that may be used directly in the models 1989; Bertram-Howery and Swift,1989). A general that are used in the extrapolation. Natural systems description of the approach to performance assess-provide data on very large spatial (nm to km) and ment is given by Campb211 and Cranwell (1988).

temporal (10 -10' years) scales and in highly com-Despite the mathematical sophistication of the ap-5

[

plex terranes in which unknown synergisms may proach used in a performance assessment, uncertain-i affect radionuclide migration. This paper reviews the ties can be large in the predicted performance, and the t pplication (and mo st importantly, the limitations) of entire methodology must be " confirmed." At the sim-t data from natural analogue systems to the "valida-plest level, confirmation consists of three activities:

i tion"of performance assessments.

(i) Measurements according to accepted proce-dures of physical and chemical properties

6.2 INTRODUCTION

4 (e.g., actinide solubilities). This is quality as-As societies move forward in their efforts to de-surance i

sign, build, and finally, license and operate a nuclear (ii) Verification of codes by comparing results to waste repository, three phrases " performance as-the results of other codes designed for the p

sessment," " validation" and " natural analogues"-

same purpose and by using generally consis-become part of the common language and concern.

tent sets of data.

Because the construction (no matter how difficult) of (iii) Fmally, validation is obtamed through addi-a repository is not a demonstration that it works, we tional laboratory or even field-scale tests rely on performance assessment to bound or limit the (Geldhart et al.,1988; Kelmers et al.,1987).

- extrapolated long-term behavior of the repository, These tests are usually designed to test a l

The performance assessment must represent a rather proposed mechanism for radionuclide re-elaborate model of extrapolated behavior. There can lease, but the spatial and temporal scales of be considerable uncertainty in such a model. Typical such experiments are extremely limited.

questions conceming the sources of error or uncer-Thus, this is a type of limited hypothesis 1 tainties include:

testing. The length of time and the expense of

]

'(i) Are the physical and chemical processes such tests limit their number. Larger spatial

- modeled correctly (e.g., the cormsion mecha-and temporal scales may be tested against the 29 NUREG/CP-0147 i-

'1

long-Term Predictions lising Nat ral Analogues historic or geologic record. For groundwater 6.3 SCIENTIFIC" PROOF" AND models this is often referred to as calibration.

ANALOGY These processes are perhaps best illustrated by The phrase " natural analogue" conveys the sense experience with groundwater models. The reader is that in natural systems (usually of great age), there referred to the work of Konikow and Bredehoeft are situations or materials that are analogous to the (1992) which discusses these specific limitations.

conditions or materials of interest in a waste reposi-The levels of confirmation-acqmsition of bas,ic tory, it is important to define what is meant by such physical and chenucal data, code verification, field-an analogy, as it leads immediately to the idea that scale validation, and histonc calibration-are all un-there can be a " proof by analogy." In the broadest portant, as well as difficult and time-consuming, but sense, an analogy refers to a simdarity between things these efforts remain trivial in comparison to the otherwise unlike, that is, a partial resemblance. In higher level of confirmation that asks the fundamen-logic, analogy is an inference that certam resem-tal question of whether the predicted or bounded blances imply a probable further similarity. Since the perfonnance will encompass the actual performance.

time of Aristotle, logicians have argued over whether It is to this question that studies of natural systems, the use of analogy is simply another form ofinduction usually referred to as natural analogues, enter into or whether it represents a distinct class of arguments, consideration [Eisenberg,1986; Kovach and and, if so, how the validity of an argument by analogy McCartin,1988; Birchard and Alexander,1983; can be demonstrated (Aristotle,1966; Niiniluoto, Petit,in press (a); Petit,in press (b)]. Natmal analogue 1980). The most useful discussion is still to be found studies examine system behavior on the correct spa, in John Stuart Mill's A System ofLogic (Mill,1874) tial and temporal scales, but there are many limita, in which he argues that the distinction between anal-tions (Ewing and Jercinovic,1987), and there has ogy and induction is an artificial one, that, in practice, been only a limited effort to explicitly use natural both rest oa the collection and cormlation of obser-analogue studies in a performance assessment. In vations. The two approaches, however, differ impor-fact, there appears to be considerable confusion con.

tantly in one's ability to demonstrate their validity.

ceming just exactly how this might be done, particu-Induction is based on a scientific understanding of the larly when data from natural systems are combined causality between events or phenomena, while anal-with the probabilistic analyses which are a common ogy, in the absence of proof, temporarily accepts a part of a performance assessment (Chapman and probable hypothesis. This is a common and legiti-Come,1989). The public and licensing requirements mate approach in the physical sciences where conclu-for validation of performance assessment and the sions based on an analogy are used until a negative increased complexities with resulting delays in these instance has disproved the hypothesis. In the physical efforts have lead finally to a certain note of despair sciences,it is often easy to identify the variables that (Krauskopf,1990) and an effort to revise the expec-are important in the causal relationships of a particu-tations of the public and the regulatory agencies lar phenomenon, and this lends strength to the anal-(National Research Council,1990). There are special ogy. The validity of the inferred relationship rests on difficulties in describing the risks that have small the pertinence of the variables selected to describe the h

P enomenon. Mill's own words best make the point:

probabilities in a long-term risk context (Svenson and Karlsson,1989); here again, natural systems are used "Since the value of an analogical argument... de-to communicate ideas to the public.

Pends on the extent of ascertained resemblance, com-pared first with the amount of ascertained difference In all of these discussions, natural systems or natu-and next with the extent of the unexplored region of nd analogues remain a " bullet on each viewgraph" unascertained properties, it follows that where the which purports to describe the " validation" process.

resemblance is very great, the ascertained difference In this paper, I want to explore how natural systems very small and our knowledge of the subject-matter l

might actually be used in the performance assessment tolerably extensive, the argument from analogy may process. I will consider the use and limitations of approach in strength very near to a valid induc-natural analogues, particularly limitations that are tion... It is hardly necessary to add that... no inherent to the processes of scientific proof and anal-competent inquirer into nature will rest satisfied with 08Y-it when a complete induction is attamable, but will l

NUREG/CP-0147 30

lag-Term Pmlictions lising Natural Analogues i

consider the analogy as a mere guidepost, pointing sumed the constancy of physical laws, 'Their immu-out the direction in which more rigorous investiga-tably constancy alone can enable us to reason from tions should be prosecuted."

analogy, by the strict rules of induction, respecting The question of the pertinence of the selected the events of former ages, or, by a comparison of the variables used in an analogy is one of establishing state of things at two distinct geological epochs, to cause-and-effect relationships. Bertrand Russell arrive at the knowledge of general principles in the 1

(1983) refers to this as a " fruitful analogy." "One of economy of our terrestrial system" (Lyell,1830). In the most important and difficult things about the nuclear waste disposal, we reason from our own inductive method is the discovery of fmitful analo.

geologic epoch to the next. Thus, analogy is (and gies, and the connected problem of the analysis of a must be) a powerfullogical and pedagogical tool. But 4

complex phenomenon into elements that can be stud-what are the special limitations of analogy when l

ied separately. The fruitful analogy is one that dis.

applied to a predicted behavior?

closes a similarity in causation, and the investigator 6.4 LIMITATIONS IN HISTORICAL has to begin by guessing at the cause."

SCIENCE APPLIED TO Russell (1983) goes on to point out that, "we PERFORMANCE ASSESSMENT cannot so easily, if at all, reach a numerical estimate of the probability ofinductions." Unfortunately, this Ewing and Jercinovic (1987) have discussed the is inherent to the process of a probabilistic perform-nature and special " limitations" of proof in historical ance assessment, which requires this " numerical es-sciences. Geo-scale predictions over long periods of timate." We are trapped between our inability to time (e.g., behavior of nuclear waste forms in nuclear assign probabilities, and yet, " ultimately, we must waste repositories) have the same limitations as post-reach a point where we use experience of what is dictive confirmation of geologic hypotheses, only the known as a basis for inferring what is unknown, and direction of the time scale has been changed (the this sort of inference is only valid if induction is future instead of the past).

valid."(Russell,1983). As faras this authoris aware, The scale in time and space of natural systems does there is no rigorous proof available that demonstrates not allow one to carry every hypothesis into the j

that induction is valid.

laboratory. The test of a geologic hypothesis rests on This view of " analogy" immediately limits what its compatibility with selected observations (often one may expect from this approach. Proof or "valida-prejudiced by the hypothesis in hand) and extracted tion" of a hypothesis (e.g., the long-term durability of fmm a geologic record that is woefully incomplete.

a borosilicate nuclear waste glass (Ewing and Jerci-Geologists have wrestled with these limitations, al-novic,1987; Jercinovic and Ewing,1988)] can only ways trying to put some distance between the false be approached (never anived at), and then only to the impression of others that their efforts are "merely i

extent that details of one system correlate to the descriptive," and they rather have strived to practice details of another system (e.g., the correspondence their craft with the same rigor as other physical sci-between corrosion of basaltic glass at a mid-ocean entists. Scientists in the nuclear waste community are 3

ridge and the corrosion of a borosilicate glass in a confronted with the same difficulties when they try bedded-salt repository). The selection of the pertinent to use natural systems to predict long-term behavior.

variables that are used to describe the phenomenon Each geologic system seems disturbingly unique in (e.g., in the case of glass corrosion, these may include its occurrence, with a wide range of variables (e.g.,

temperature, pH, surface area to solution volume solution composi: ion, age, flow rate), many poorly ratio, flow rate) have a direct bearmg on the validity defined or unknown. The importance of these vari-of the conclusions. Finally, a probabilistic assessment ables to the phenomena being described may change of projected or predicted behavior may not be possi-over the course of time, and, most importantly, with ble.

such wide variations and uncertainties between sup-We should also note the historic role of analogy in posedly similar occurrences, there seems to be little developing an understanding of geologic processes hope of arriving at useful generahzations.

over the expanse of geologic time. Charles Lyell, the The distinguished paleontologist G.G. Simpson founder of modern geologic methodologies, was paid particular attention to this situation in an essay, firmly rooted in an actualistic philosophy that as-

" Historical Science," in which he outlines both the 31 NUREG/CP-0147

I long-Term Predictions Using Natural Analogues methodologies and limitations that are inherent in use postdiction to confhm that certain processes and historical sciences (Simpson,1963). Simpson de-mechanisms operate on the scale of earth systems, we fined historical science as the " determination of con-cannot project their configurations into a future suc-figurational sequences, their explanation, and the cession of "configuratiom" or scenarios. We can testing of such sequences and explanations." By confhm large-scale, long-time mechanisms, but we

" configuration," Simpson means to describe all as-can never be sure that these are the currently operative pects of the system as it changed through time during mechanisms. We can also, with cenain limitations, the past. For a geologist, this is a common process of use the rock record to establish probabilities for writing a geologic history. The development of a events in the natural system, but these estimates will geologic history proceeds by the typical accumula-remain largely subjective. This is why the develop-tion of facts or observations welded together by nor-ment of release scenarios is so imponant and, at the mal inductive and deductive processes. The unusual same time, so difficult. We gain " hints" from past aspect of the process is the " testing" of the hypothe-behavior, but finally only experience and expenjudg-sis. There is no experiment that can be done that will ment determine the probabilities and risk; hence, the repeat the activity on a part of the Earth's surface over importance that we ascribe to each scenario.

long periods of time. The test can only be a compari-At this stage, it is important to distinguish between son of deduced results of the hypothesis against the a scenario and an hypothesized future behavior. A actual rock record.

hypothesis has the property that it can be disproved As an example, if one proposes meteorite impact

[in contrast, a hypothesis cannot be proved (Popper,

as the cause oflate Cretaceous extinctions, where are 1961)], either by predictions and experiments or post-the high pressure phases that would have formed dictions and careful observations. There is no part of during such an event? The correlation oflate Creta.

a perfonnance assessment that can be " proved."

ceous extinctions with an iridium anomaly, high.

There is no experiment or observation that can vali.

pressure silica polymorphs, and the graphitized date the predicted performance of a nuclear waste carbon in thin layers of shale at the Cretaceous-Ter-repository. A scenario is instead a proposed perform-tiary boundary is a convincing form of proof (Al.

ance, and the best that can be done with a scenario is varez,1987) that may still be contested by othem that a probability of occurrence (perhaps based on the (Archibald,1982). In the parlance of Simpson, this is geologic record) be assigned and an evaluation made the process cf "postdiction" (as compared with pre.

of the consequence. This is an important point, as diction)in which there is a correlation of the past and scientists are trained to disprove hypotheses by look-present. If the geologic record were complete, the ing for exceptions. As the exceptions mount in num-process of hypothesis testing by postdiction would be ber, the theory or paradigm is disproved (Kuhn, tedious, but always rewarding. That is, we would 1970). One must realize that results of a natural always anive at an answer that could be used. The analogue study may not be consistent with a perform-fact that the geologic record is notoriously incom.

ance assessment, but performance sessments (by plete (Kerr,1991), however, makes the process not definition) cannot be disproved.

only tedious, but often unrewarding. The absence of Natural analogue studies can never be used to shock features at the Cretaceous-Tertiary boundary prove or disprove a proposed performance. State-would not demonstrate that the extinctions were not ments such as, " Natural glasses have existed for caused by meteorite impact, but could simply be the millions of years" or " Natural glasses are easily al-result of a scanty geologic record. The rock record tered" do not speak for or against the long-term seldom provides a complete record of catastrophic stability of nuclear waste glasses. The two statements events (Kerr,1991), as they occur on such a short time only indicate a lack of correspondence between the scale. These are limitations unique to the limitations historical configurations that characterized the corro-of the geologic record, but what are the additional sion environments of the glasses.

Limitations of our predictive efforts embodied in per-formance assessments of the long-term behavior of 6.5

SUMMARY

OF LIMITATIONS nuclear waste repositories?

This discussion has tried to emphasize some of the The first, and obvious, limitation is that there is no conceptual and philosophical issues that complicate rock record for future events. So, even though we can the use of natural analogues. I end by classifying NUREG/CP-0147 32

long4errn Predbtions thing Natural Analogues these limitations into two types: (i) the limitations of Lutze, B. Grambow, J.K. Bates, J. Janeczek, J.C.

models and (ii) the limitations of " confirmation" or Petit, and T. Murakami) over the past ten years. I am

" validation."

grateful for the cor.tmuity of support over many years Significant errors in performance assessment provided by the Svensk Kurnbrunslehantering AB.

models may occur due to Studies of natural systems always require time and

• the selection of the wrong deterministic model patience. Major parts of this paper were presented at

• an incorrect analytical solutions for the model the Third Intemational Symposium on Advanced Nu-

• an incomplete description of the system mod.

clear Energy Research: The Global Environrnent and cled Nuclear Energy (Ewing,1991) and at the Third Inter-

• the presence of nonlinear systems.

national Conference on High Level Radioactive The last item on this list has not been discussed in Waste Management (Ewing,1992).

j this paper; but one should simply note that nonpedo.

Finally, the author thanks the organizers of the dic, irregular behavior may occur, and probably does

" Workshop on the Role of Natural Analogs in Geo-occur, in nonlinear systems. The short-term temporal logic Disposal of High-Level Nuclear Waste" at both evolution of such systems can be accurately pre.

the Center for Nuclear Waste Regulatory Analyses dicted, but the long-term behavior is highly unpre.

and the U.S. Nuclear Regulatory Commission. This dictable, and deviations may evolve rapidly with Paper was written in the spirit of stimulating discus-sion in the informal setting of the workshop. Al-time.

The limitations of " confirmation" or " validation" though many of the issues that I have raised require further, careful development, I hope that I have at include least provoked critical discussion.

• A performance assessment hypothesis, by defi-

"III "' *^"".* D* *** **d'

6.8 REFERENCES

• A hypothesis can only be m. validated, not pmven to be tme. This is particularly true of Alvarez, L.W.: Mass Extinctions Caused by Large argument by analogy orinduction.

Bolide Impacts: Physics Today,84 (1987) 24-33.

Archibald, J.D. and W.A. Clemens: American Scien-

6.6 CONCLUSION

S tist,70,377 (1982).

I should end by saying that this list of limitations Aristotle, Posterior Analytics, Book II., iv, translated does not mean that one should not move fonvard. The by Hugh Tredennick, Harvard University Press, problem exists. Nuclear wastes must be permanently Cambridge, Massachusetts (1966) 185-189.

isolated from the accessible environment. However, Bertram-Howery, S.G. and R.L. Hunter, editors: Pre-one cannot use elaborate mathematical approaches, liminary Plan for Disposal System Charac-such as applied in perfonnance assessments, as a way terization and Long-Term Performance of verifying a predicted behavior. Nor should one Evaluation of the Waste Isolation Pilot Plant, expect the public to accept such a demonstration as Sandia Report 89-0178 (1989).

final or convincing proof. Performance assessment is Bertram-Howery, S.G. and P.N. Swift: Status Report:

simply an organized and sophisticated way of think-Potential for Long-Term Isolation by the Waste ing about the elements of the problem. Natural sys-Isolation Pilot Plant Disposal System, Sandia Re-tems (the phrase " natural analogue" is probably not port 90-0616 (1989).

necessary) provide vital data (Jercinovic and Ewing, Birchard, G.F. and D.H. Alexander: Natural Ana-1988; Finch and Ewing 1989 (a), (b) in press, for logues-A Way to increase Confidence in Predic-examples), as well as providing the " guideposts" of tions of Long-Term Performance of Radioactive Mill for understanding complex systems over long Waste Disposal: Materials Research Society Sym-periods of time. There is no method by which a posium Proceedings, D.G. Brookins, ed, vol.15, long-term prediction can be validated.

323 (1983).

Campbell, J.E. and R.M. Cranwell: Performance As-6.7 ACKNOWLEDGMENTS sessment of Radioactive Waste Repositories, Sci-This paper has benefitted from collaborations with ence,239,1389 (1988).

students (R.F. Haaker, M.J. Jercinovic, R. Cowan, Chapman, N.A. and B. Came: Long-Term Predic-and R.J. Finch) and discussions with colleagues (W.

tion: Making Proper Use of Geological Evidence.

33 NUREG/CP-0147

Long-Term Preda:tions Using Natural Analogues NEA Workshop on Performance Assessment, Grambow, B., MJ. Jercinovic, R.C. Ewing and C.D.

Paris (1989).

Byers: Weathered Basalt Glass: A Natural Ana-Come, B. and N A. Chapman: Fourth Natural Ana-logue for the Effects of Reaction Progress on Nu-logue Workir.g Group Meeting: Poqos de Caldas clear Waste Glass Alteration: in Scientific Basis Project Fimd Workshop, Pitlochry, Scotland for Nuclear Waste M anagement IX, ed. L. Werme, (1990).

(1985) 263-272.

Eisenberg, N.A.: Natural Analogues and Validation Hunter, R.L.: Events and Processes for Constructing of Performance Assessment Models: Chemical Scenarios for the Release of Transuranic Waste Geology,55,189 (1986).

from the Waste Isolation Pilot Plant, Southeastem Ewing, R.C. cid MJ. Jercinovic: Natural Analogues:

New Mexico, Sandia Report 89-2546 (1989).

Th.nr Application to the Prediction of the Long-Janeczek, J. and R.C. Ewing: Heterogeneity and Al-Term Behavior of Nuclear Waste Forms: Materi-teration of Uraninite from the Oklo Natural Fission als Research Society Symposia Proceedings, J.K.

Reactor: Science, submitted.

Bates and W.B. Seefeldt, editors, vol 84, 67 Jercinovic, MJ. and R.C. Ewing: Basaltic Glasses (1987).

from Iceland and the Deep Sea: Natural Analogues Ewing, R.C.: Natural Glasses: Analogues for Radio-to Borosilicate Nuclear Waste-Form Glass: JSS active Waste Forms: In Scientific Basis for Nu-Project, SKB Technical Report 88-01 (1988) 221 clear Waste Management, vol.1, G. McCarthy, ed.,

pages.

Plenum Press,New York (1979) 57-68.

Jercinovic, MJ. and R.C. Ewing: Corrosion of Glass Ewing, R.C.: Natural systems: Prediction of radionu-in the Geologic Environment. In Corrosion of clide migration. Proceedings of the Third Interna-Glass, Ceramics and Ceramic Superconductors.

tional Symposium on Advanced Nuclear Energy D.E. Clark and B.K. Zcitos, eds., Noyes Publica-Research (1991) 167-175.

tions, New Jersey,in press.

Ewing, R.C.: The role of natural analogues in per-Kelmers, A.D., R.E. Meyer, J.G. Blencoe and G.K.

formance assessment: Applications and limita-Jacobs: Radionuclide Sorption Methodologies for tions. Proceedings of the Third International Performance Assessments of High-Level Nuclear Conference on High Level Radioactive Waste Waste Repositories: A Perspective Gained from an Management, ed. James S. Tulenko (1992) 1429-NRC Workshop. Nuclear Safety, 28, 515 (1987).

1436.

Kerr, R.A.: Dinosaurs and Friends Snuffed Out: Sci-Finch, RJ. and R.C. Ewing: Alteration of Natural ence,251,160 (1991).

UO2 Under Oxidizing Conditions From Shink-Konikow, L.F. and J.D. Bredehoeft: Ground-water olobwe, Katanga, Zaire: A Natural Analogue for models cannot be validated: Advances in Water the Cormsion of Spent Fuel. SKB Technical Re-port 89-37,31 pages (1989).

Resources (1992)in press.

Kovach, L.A. and TJ. McCartm: The Role of Natural Finch, R.J. and R.C. Ewing: Uramnite Alteration Products in an Oxidizing Environment and Their Anal gs m, Performance Assessment: Transac-Relevance to the Disposal of Spent Nuclear Fuel:

19 8*

SKB Technical Report,in press (a).

Finch, RJ. and R.C. Ewing: Phase Relations of the Krauskopf, K.B.: Disposal of High-Level Nuclear Uranyl Oxide Hydrates and Their Relevance to the Waste: Is it Possible?: Science,249,1231 (1990).

Disposal of Spent Fuel. In Scientific Basis for Kuhn, T.S.: The Structure of Scientific Revolutions:

Nuclear Waste Management XIV. T. Abrajano, Jr.

University of Chicago Press, Chicago (1970) 210 and LH. Johnson, eds., Materials Research Soci-Pages.

ety, Pittsburgh, Pennsylvania, in press (b).

Lutze, W., G. Malow, R.C. Ewing, MJ. Jercinovic Geldart, R.W., B.P. McGrail, K.C. Rhoads and MJ.

and K. Keil: Alteration of Basalt Glasses: Impli-Apted: Validation of a Nuclear Waste Repository cations for Modeling the Long-Term Stability of Performance Assessment Model: Comparison of Nuclear Waste Glasses. Nature, 314, (198.5) 252 Theory With Experiment, Materials Research So.

255.

ciety Symposia Proceedings, MJ. Apted and R.E.

Lyell, Charles: Principle: of Geology: John Murray, Westerman, eds, vol. I12,341 (1988).

Albemarle-Street (1830).

NUREG/CP-0147 34

Long Terrn Predictions Using Natural Analogues Mill, J.S.: A System of Logic, Ratiocinative and Petit, J.C.: Design and Performance Assessn~at of Inductive: Being a Connected View of the Princi-Radioactive Waste Forms: What Can We Leam ples of Evidence and the Methods of Scientific from Natural Analogues?: Chemical Geology, in Investigation: Harper & Brothers, Publishers, New press (b).

York (1874) 659 pages.

Popper, K.R.: ne Logic of Scientific Discovery:

National Research Council: Rethmkmg High-Level Basic Books, Inc., New York (1%1) 480 pages.

l Radioactive Waste Disposal: a Position Statement Russell, B: The Art of Philosophizing and Other of the Board on Radioactive Waste Management, Essays. Rowman & Allanheld, Totowa, New Jer-National Academy Press (1990).

sey (1983) 119 pages.

Niiniluoto, Dkka: Applications of Inductive Logic:

Simpson, G.G.: Historical Science in Fabric of Geol-edited by.L.J. Cohen and Mary Hesse, Clarendon ogy, edited by C.C. Albritten, Jr., Freeman, Coo-Press, Oxford (1980) 218-234.

per & Company, Stanford, California 1%3) 24.

s Petit, J.C.: Migration of Radionuclides in the Svenson,0. and Gunnar Karlsson: Decision-Making, Geosphere: What Can We Leam from Natural Time Horizons, and Risk in the Very Long-Term Analogues?: Radiochimica Acta,in press (a).

Perspective: Risk Analysis,9,385 (1989).

i 1

4 1

35 NUREG/CP-0147

4

._m ANALOG EARTHQUAKES 7

Renner B.Hofmann Centerfor Nuclear Waste Regulatory Analyses San Antonio, Texas 78228 i

7.1 ABSTRACT or process that cannot be observed directly or that is difficult to observe directly."

Analogs are used to understand complex or poorly understood phenomena for which little data may be Geological phenomena are~ generally observable available at the actual repository site. Earthquakes are only at the surface of the earth or in a one-dunensional complex phenomena, and they can have a large num, bere through the earth's surface to a limited depth.

ber of effects on the natural system, as well as on However, a more complete understanding of a par-engineered structures. Instrumental data close to the ticular geologic feature may be obtamed in regions of

- source of large earthquakes are rarely obtained. The great topographic relief or in regions where geologic rare events for which measurements are available features have been turned on their side, and erosiori i

may be used, with modifications, as analogs for po-has produced a natural cross section. Limited expo-tential large earthquakes at sites where no earthquake sures of a geologic feature, that is similar to a feature data are available. In the following, several examples of ne of these relatively rare natural cross sections, of nuclear reactor and liquified natural gas facility may be used to infer that which cannot be seen.

siting are discussed. A potential use of analog earth-On-going depositional processes produce features quakes is proposed for a high-level nuclear waste reminiscent of those seen in sedimentary rocks.

(HLW) repository.

Therefore, by analogy, these processes are presumed to have occurred in the past. Aheration products of 7.2 THE CONCEPT y mt enna regions, newy, by dogy, atv are used to infer the cause of alteration products m The use of analogs is the first stage of a better geology elsewhere.

understandmg of complex phenomena. Analogs are That earthquakes are observed to accompany fault used to predict future happenings when too little is movements implies that fault offsets observed else-known about the phenomena being predicted or when where were also likely to have been accompanied by theories are not completely accepted. Analogs may earthquakes. InstmmentM recordings of carthquakes be used to add credibility to a theoretically derived accompanying certain types of faulting at specific prediction. A characteristic of an analog is the lack of distances, by analogy, are attributed to earthquakes a mathematically described statistical variability. It is that have a similar origin but for which no instrumen-usually unknown whether an analog represents the tal recordings are available.

mean of a variable of interest or a rare extreme. Often there is only one analog to an anticipated event or Much of the subject of geology has its orii; ins in condition, and it is not likely to be a perfect match.

analogs that were often referenced, although not nec-essarily with the word " analog," by early writers, Among dictionary definitions, an analog is some-such as Hutton (1795). In a sense, all of geology is thing that is similar to something else. Definition of based upon andogs' the word " analogy" yields further insight, for exam-ple, if something is similar to something else in some A past recorded earthquake whose characteristics respect, it is likely to be similar in others. That the or effects are sufficiently similar to those expected at words " identical," "all" or "in every respect" did not a new facility site may be used to predict design loads appear in any definition encountered, is notable. An or the consequences of potential structural degrada-analog may be a model as inferred by the definition tion. Ideally, a strong motion record to be used as an of "model" in Bates and Jackson (1980): "A working analog should have been recorded from an tarth-hypothesis or precise simulation, by means of de-quake having similar characteristics as that expected scription, statistical data, or analogy, of a phenomena at the new site, for example:

37 NUREG/CP-0147

Anabg Earthquakes e nature of the fault and the dynamics ofits move-Geological Survey (USGS) and the U.S. Nuclear ment Regulatory Commission (NRC) concluded that there e distance from fault to facility was sufficient uncertamty in the published location nature of the tectonic stress field and stress drop of this earthquake 60 km offshore that it was possible e

during an earthquake that this earthquake might have occurred on the Hos-fault orientation to the facility and stress field, Sri fault.

e (Rogen et al.,1977; Boatwright and Boore, The Lompoc esrthquake, therefore, became an 1982) analog for a 7.5-magnitude earthquake located off-earthquake magnitude shore,5 miles from the plant. There were no strong e

path geology motion records for the Lompoc earthquake or from e

depth of earthquake focus and depth of facility any other earthquake of this magnitude at this dis-e fault parameters, for example, length, width, tance. Considerable effort was expended in extrapo-e and offset lating the effects of other earthquakes to this location, number, and size of asperities or bar-magnitude and distance. Dr. Robert Page of the e

riers USGS applied prior research on the number of cycles Usually, not all variables will be identical or avail-of strong motion at various levels for the Alaskan able. Consequently, modifications based on other pipeline to this site. Dr. Nathan Newmark, consultant incompletely analogous earthquakes will be required.

to the NRC, interpreted Dr. Page's information in Predictions of earthquake ground motion are based terms of design criteria. Again, the analog earthquake on:

was based on a past earthquake on what might have direct analogs been a similar fault. The analogy was imperfect, e

interpolation or extrapolation from near or par-however, because all desired data concerning the e

tial analogs analog were not available. Consequently, extrapola-formulae or computer codes that reproduced tions fmtn the data of other imperfectly analogous e

ground motion of analogous or nearly-analo.

earthquakes were employed to complete the neces-gous earthquakes.

sary design criteria.

7.3 SOME EXAMPLES Continued research on this pmject has resulted in additional proposals, meluding earthquake source Example 1: The Diablo Canyon Nuclear Power computer modeling to obtain or confirm adequacy of Plant.

design criteria.

In the initial license application, a magnitude 6.75 aftenthock was proposed at a 6-12-mile depth, Example 2: NRC Task Action Plan A46 regarding caused by a magnitude 8+ main shock on the San Andreas fault about 40 miles distant. The assignment componentsfor oldplants.

was made largely on the supposition of an unknown The requirement for seismic quahfication of some but possible fault at depth. Also, a magnitude 7.25 on components was imposed after construction of the the Naciemento fault, about 20 miles distant was plants. These components had been used in conven-proposed. This earthquake was a scaled-up analog to tional fossil fuel power plants throughout the world.

the magnitude 6.3 Long Beach (1933) earthquake.

Studies by the Brookhaven. Lawrence Livermore, Modified Mercalli Intensities of damage were docu-and Sandia National Laboratories, Southwest Re-mented for the 1933 earthquake. An equivalence search Institute (SwRI), EQE Inc., and others docu-between Modified Mercalli Intensities and peak mented that many of these plants had been subjected ground acceleration compiled from many earth-to earthquake shaking and that the components had quakes provided a design basis.

not failed (Yanev,1984; Smith and Dong,1983; Later, an offshore fault, the Hosgri fault, was de-Chang,1987; and Kana et al.,1983).The magnitudes termined to be oflicensing significance. A magnitude and distances of the earthquakes to the plants and 7.3 offshore earthquake in 1927 had caused rock some records were available. These earthquakes be-slides, sand boils, water spurts, a tsunami of 6 feet came analogs for seismic shaking table input and,in and some fallen chimneys in the vicinity of Lompoc, turn, verified that the test shaking levels were analo-Califomia (Coffman and von Hake,1973). The U.S.

gous to real earthquake ground mot on.

i NUREG/CP-0147 38

An&g Ea%aks This example differs from the previous one and analog for a larger more distant canhquake on the suggests a somewhat different use of analogs. It is to Queen Charlotte fault. This value was checked i

directly support the performance of structural or me-against a correlation with acceleration for Modified chanical elements by using analog exposures to vi-Mercalli Intensities observed for the Ms = 8+ earth-bratory ground motion without the use of an quakes on the San Andreas fault in California (1872 intennediate approximate relationship or model.

and 1906) (Coffman and von Hake,1973) and also against Canadan buMng code requima ts. h This type of procedure eliminates the necessity of constructing approximate' models of earthquake anal gs and the code requirements were in reasonable agement (Hofmann et al, B82).

shaking, for example, peak acceleration as a function I

ne os acauna h.qmpd natuml of distance and magnitude and appropriate spectral ga receiving terminal in eastern Canada (Fenco envelopes for test shaking or accepting a white-noise

1. "#"I'#"##

spectral shape for such test shaking. The method The site was located in an area of past seismic depends upon many analogs. Therefore, a statistical ctivity where magnitudes had reached 7 or slightly distribution of parameters may be possible, which higher. These earthquakes became analogs for an might add credibility to the result. A possible negative canhquake near the s, e. Howem, no strong modon a

aspect is that the details of the analogs may not be records were available from these Ms = 7 carth-well investigated and compared to the site or sites in quakes. Consequently, the record from the Kem question. The question of whether foundation types, County 1952 Ms = 7.7 (Coffman and von Hake,1973) fault types, stress drops, etc. are similar to the site in earthquake was slightly adjusted for distance and question may not be possible to answer or may be magnitude and used as design entena for the facility.

largely ignored.

Example 3: The planned Grassy Point, British Co-7,4gy999y9ygggyg9 lumbia, liqwped natural gas compression and stor-At times, only the occurrence of an analog earth-agefacilities, and seaport.

quake is known and no strong motion records may be This planned facility was located 90 km east of the available. The analogy is between the geologic con-offshore Queen Charlotte fault. The Queen Charlotte ditions, for example, fault rupture length, and the fault is a strike slip fault that is analogous to the San magnitude of the earthquake. Under these circum-Andreas fault in California. Both faults have experi-stances, a record is interpolated or extrapolated from enced magnitude (Ms) 8+ earthquakes. There were other imperfectly analogous earthquakes, or a record no strong motion records from magnitude 8+ earth" is derived from formulae or computer codes that were quakes. The location of the facility and the Queen derived from the study of many earthquakes. It can Charlotte fault are in sparsely populated areas. Con-be concluded that analogs of various types, some of sequently, reliable Modified Mercalli Intensities of them imperfect. are useful.

damage also were not available. The San Andreas Ms = 8+ earthquakes, however, had well-docu-7.5 ANALOG EARTHQUAKES FOR mented damage intensities at distances of 90 km from A HIGH-LEVEL NUCLEAR the fault. Further, there was a strong motion record WASTE REPOSITORY closer than 90 km to a Ms = 7.7 earthquake on the White Wolf fault near Bakersfield. Califomia. An A number of variables affect earthquake shaking examination of seismic source theory indicated that, at a site. They are in the source, path, site, and for frequencies of vibration that could affect the fa.

instrumentation categories. Source variables include cility, the vibrations of a magnitude 7.7 earthquake those in Section 7.2 of this paper. Some of these would be similar in amplitude to those of magnitude variables also influence the displacement of faults 8+. Therefore, the Ms = 7.7 Kern County (also known possibly hidden at depth, that is inferred from the as the Arvin Tehachapi) earthquake of 1952 could be occurrence of an eanhquake.

used as an analog to predict ground motion at the Path variables include:

source to facility distance Grassy Point facility. The shaking recorded for the e

path geology Kern County earthquake was scaled down to repre-a sent the greater distance. With modification, the Kern

+ characterization of sedimentary wedges and County earthquake strong motion record became an their effect on surface waves (Herrmann,1978) 39 NUREG/CP-0147

Anabg Earthquakes Site and instmment variables include:

groundwater from an aquifer to higher eleva-e depth of facility and depth of the strong motion tions and reducing the distance of radionuclide instmment pathways to the biosphere.

e facility and instmment sites-foundation geol-

• Faults or fault intersections may open, creating ogy and geomechanics permeable pathways along which groundwater e frequency response of the strong motion seis.

or gases may migrate to the biosphere, mograph

• Faults may intersect a waste package, thereby Knowledge of all of these variables is not usually damaging this engineered barrier, causing available or possible to obtain for either an antici-leaks, which,in tum, reduce the time of radionu-pated earthquake near a site or for potential analog clide migration out of the repository. Such j

earthquakes to be used to predict effects at the site.

faults may also provide a low permeability path-Many of these variables are discussed by Hofmann way directly from the waste to the surface or to (1991). This lack of knowledge may be translated into moving groundwater that may carry radionu-an expected variability in the effects at a site. Often clides to the surface.

there is insufficient information to determine the Finding analogs for all these effects is not likely.

mathematical nature of this potential variability. It For those that can be found, very site-specific char-can be a limitation to the application of analogs if acteristics may be needed to reduce uncertamties.

several of them are not available.

Modification of analog observations may be neces-The qualities of an analog ea'thquake to be applied sary to better reproduce the desired conditions at the to any facility design should fulfill the following site. Modifications for ground motion may be accom-criteria.

plished by determuung the desired maxunum magni-

• The analog should have occurred in an environ-tude for the site and the potential source distance from ment such that the effects to be predicted at the the site. Ground motion amplitudes may then be repository could have been observed.

adjusted using standard curves. Compensation for e The analog earthquake should have occurred in records of the analog being at the surface and the a tectonic and geologic environment much like repository being at depth may be developed by com-that of the repository.

puter modeling or by deternumng ground motion

• If the analog includes a strong motion seismic amplitude ratios from seismic recoidings on the sur-record, it should have been recorded at a depth face and at depth in mines. This would be another and in material that are similar to that of the analog. Groundwater movement from strain changes repository, might be determined from scaling of near-analog Vibratory ground motion effects that may be of effects to the earthquake magnitude and groundwater coneem to a deep repository are:

table at the site. An alternative is finite-element strain

• Shaking of canisters in their boreholes causing groundwater flow modeling.

stress cracks and accelerated corrosion

• Spalling of tunnel walls, thereby enlarging the 7.6 POSSIBLE EXAMPLE ANALOG zone of higher permeability around tunnels and EARTHQUAKES FOR providing added avenues for groundwater or YUCCA MOUNTAIN gaseous ndgration The likelihood of strong shaking at or near Yucca e Shaking (or fault movement) may cause stress Mountain, Nevada, has been estimated, for example, changes which squeeze groundwater from rock by Sommerville et al. (1987) and Rogers et al. (1977).

pores to higher elevations (Wood et al.,1985)

Additional work performed may use other boundary that are closer to or encompass formerly dry conditions. Rogers et al. (1977) limit their hazard repository shtfts, thereby providing faster ra-estimates to magnitude 7 or lower earthquakes. Som-dionuclide pathways to the biosphere.

merville et al. (1987) assigned maxunum magnitude Fault uovement, which usually, but not always, potentials to a number of faults in or near the Yucca causes earthquake shaking, may produce undesirable Mountain site.The maximum magnitudes were about effects at a deep repository such as:

7.4. Higher magnitudes on longer but more distant

• Faults may move, increasing gouge thickness faults were also considered in their hazard estima-and decreasing permeability, thereby diverting tions. The estimates in Table 3 of Sommerville et al.

NUREG/CP-0147 40

- ~

Amhg Eat %mks (1987) correlate Ms = 7.4 with fault lengths of about stances. Analogs do not usually include knowledge 53-57 km in ttus scgien. The nearest fault of that of potential variations in observations caused by the 1

length, in their study, was the Rock Valley fault, statistical or chaotic nature of the processes involved, 20 km or more from the site. The largest fault closer or from overlooked, unknown, or ignored variables.

]

to the site was the Paintbrush Canyon with a pre-Analog earthquakes need not be perfect to be j

sumed maxunum magnitude of 7 based onlength and useful. Perfect analogs are seldom found. The depth maximum Quatemary paleo-offsets. Further field where strong motion information is desired may dif-

. work may better def' e fault length and offset and fer from the depth of a strong motion seismograph m

probable future slip in the current stress situation.

that has recorded an earthquake of the proper magni-

"Ibese efforts may change the maximum magnitude tude and distance. The media within or upon which forparticular faults. As field work at Yucca Mountain information is desired may be different from that on i

progresses, other faults may be found to be near 57 which a stmng motion seismograph is located. 'Ihe km long.

type of fault movement causing the earthquake may

~

Observation; of the effects of the potential exam-differ from that expected at the site ofinterest. A large

)

L ple analogs below may be adjusted if the desired number of differences degrades the utility of an ana-i magnitudes for faults at Yucca Mountain differ.

log, but one or a few differences may be accommo-j These are possible examples of analog earthquakes dated.

for effects that may be possible at Yucca Mountain.

• 1983 Borah Peak, Idaho, Ms = 7.3 7.8 BIBLIOGRAPHY

- Occuned in an extensional environment on Bates, R L. and J.A. Jackson.1980. Glossary ofGe-

[

the border of the Basin and Range Province ology. Falls Church, VA, American Geological

- Dip slip faulting Institute: 749.

I j

Boatwright, J. and D.M. Boore.1982. Analysis of the

~

ed gr und accelerations radiated by the 1980 Liver-

- Shallow groundwater table must be modified mere Valley earthquakes for directivity and dy-for the Yucca Mountain site.

namic source characteristics. Bulletin of the

• 1980 Vieste,1taly, M = 6.5 Seismological Society of America (BSSA) 12:

j

- Occurred in an extensional ennronment 1843-1865.

I

- Dip slip faulting 4

Chang, T.Y.1987. Seismic Qualification of Equip-l

- Ocd in a tufimence ment in Operating Nuclear Power Plants. Unre-l

- Groundwater eifects observed solved Safety Issue A-46. NUREG-1030.

- Shallow groundwater table must be modified for the Yucca Mountain site.

Coffman, J.L. and C.A. von Hake.1973. Earthquake Near-field ground motion records were not avail.

History of the United States. Publication 41-1.

able for the 1983 Borah Peak earthquake. Rovelli et Boulder, CO: NOAA, U.S. Department of Com-merce.

al. (1988) summanze scaled accelerations for earth.

i quakes from extensional areas in Italy. They do not Fenco Consultants Limited.1982. ING Receiving l

verify theoretical results with an analog easthquake, Termmal Seisnic Risk - Gros Caconna. Quebec

. suggesting that no suitable strong motion records are and Melford Point, Nova Scotia - Arctic Pilot available. Another analog may be required for ground Project. Consulting report prepared for Trans motion effects.

Canada Pipelines. Toronto: Fenco Consultants i'

Ltd.

j

7.7 CONCLUSION

S Herrmann, R.B.1978. ComputerPrograms in Earth-l-

Although the entire complexity of an earthquM.e quake Seismology 1, 2. Report to the National is seldom available from analog eart'h. th'y Science Foundation. St. Louis, MO: Earth and b

i remain good examples of natural analogs and their Planetary Sciences Department of St. Louis Uni-i limitations. Analogs are a first step in a better under.

versity.

standing of complex natural processes. Partial ana-Hofmann, R.B., G.E. Rawlings, and N.R. McCam-

. logs are useful, but modifications may be necessary mon.1982. Repon to Associated.Kellogg Ild. on to apply analogs in some, perhaps many, circum-Seismic Analysis for the Proposed ING Site at

~

i 41 NUREG/CP-0147

I An*% Eartiumbs l

Grassy Point, British Colwnbia. Vancouver, B.C.:

Smith, P.D. and LG. Dong.1983. Correlation of Golder Assoc 1stes Ltd.

Seismic Experience Data in Non-Nuclear Facili-Hofmann, R.B.1991. Probabilistic Fault Displace-ties wirk Seismic Equipment Qualification in Nu-ment and Seismic Hazard Analysis Literature Re-clearPlants. A 46.NUREGICR 3017.

view. CNWRA 91-013 prepared for the NRC. San Sommerville, M.R., R.C. Lee, and G.N. Owen.1987.

Antonio,TX: SwRI.

Technical Basis and Parametric Study of Ground Hutton, J,1795. Theory of the Earth. Edmburg, Scot-Motion and Surface Rupture Hazard Evaluations land.

at Yucca Mountain, Nevada. URS/ John A. Blume Kana, D.D., DJ. Pomerening, J.R. Unruh, and E.Z.

Assoc. report. SAND 86-7013. Albuquerque,NM:

Poich.1983. Evaluation of methodology for seis-Sandia NationalLaboratory.

mic qualification of nuclear plant electrical and Wood, S.H., C. Wurts, C.T. Lane, N. Ballenger, M.

mechanical equipment. SwRI 6582 001-02 pre-Shaleen, and D. Totorica.1985. The Borah Peak, pared for the NRC. San Antonio,TX.

Idaho, earthquake of October 27,1983 - hydro-Rogers, A.M., D.M. Perkins, and F.A. McKeown.

logic effects. Earthquake Spectra 2: 127-150.

1977. Prelimmary assessment of the seismic haz-Yanev, P.I.1984. Summary of ground motion inten-ard at the Nevada Test Site. BSSA 77: 1587-1606.

sities from the Coalinga, California, earthquake of Rovelli, A., O. Bonamassa, M. Cocco, M. Di Bona, May 2,1983 -based on observed ground, struc-and S. Mazza.1988. Scaling laws and spectral tural, and equipment response, prepared for Seis-parameters of the ground motion in active exten-mic Qualification Utility Group by EQE sional areas in Italy. BSSA 78: 530-560.

Incorporated. San Francisco, CA.

NUREG/CP-0147 42

APPLICATION OF NATURAL' ANALOG STUDIES TO-EXPLORATION FOR ORE DEPOSITS Donald L Gustafson Consulting Fanomic Geologist Reno,Neosde 89509 8.1 ABSTRACT be applied to ore deposit exploration by replacing Natural analogs are viewed as similarities in nature

" ccurrencesofmaterials"with"knownoredeposit,"

and are routmely utilized by exploration geologists in

" Processes" with "known ore-forming processes, their search for economic nunemi deposits. Ore de-both chemical and mechanical," and " comparable to posit modeling is undertaken by geologists to direct sogne aspect of a systein ofinterest" with "the type of their exploration activities toward favorable geologic deposit de geologist is searching fw and hopre *o

' discover." A key statement in the defindon ic 9 6 environmems and, therefore, successful programs.

Two types of modeling are presented: (i) emg parable to some aspect of a system ofinterev.%ce model development based on the study of known ore we deposits are unique feanares m natum M ae deposit chssidstics, and (ii) concept model devel-nm exacdy me same. There are, howem enough opment based on theoretical considerations and field simdarities tojustify developing a model to aid in the search for new discoveries.

observations that suggest a new deposi; type, not known to exist in nature, may exist and justi5es an 8.3 MODELING exploration program.

L

' Key elements that are important in empirical The first step in ore deposit modeling is the selec-

[

model development are duvei, and examples of tion of the desired commodity to discover. The same

(

successful application of these natural analogs to conunodity may occur in economic quantities in di-i l.

exploration are presented. A classic example of suc.

verse geological environments, necessitating the cessful concept model development, the discovery of study and construction of various model types. Two 4'

the Mci =gMia gold mine in California,is presented.

types of modeling can be undertaken.

The mili*ian of natural analogs is an important Empirical Model Development-This approach facet of mineral exploration. Natural analogs guide involves the study of existmg ore deposits, construct-explorationists in their search for new discoveries, ing a three-dimensional (3D) model of various char-U increase the probability of success, and may decrease acteristics, and applying this knowledge to the overallexploration expenditure.

exploration for a similar deposit.

Concept Model Developraent-This approach is 7

F

8.2 INTRODUCTION

based on theoretical considerations and field obser-vati ns.Conceptdevelopmentinvolvesdevelopment I

Natural analog studies are used by exploration f a new concept or deposit type not known to exist geologists in their search for economic mineral de-m nature,but thought to be possible.

f posits. Geologists commonly refer to their explora-tion approach as ere deposit modeling or model 8.3.1 Empm.. cal Model Development develaamaatiand every economic geologist has a model in mind durmg his search, whether he knows Key elements that are important and studies for it or not, based on studies and observations of known empirical model development are as follows:

ore deposits.

Age-The Proterozoic age hosts the majority of i

E.C. Pearcy and.W.M. Murph" (1991) recently

. World Class deposits around the world. Figure 8-1 L

defined natural analogs as " occurrences of materials demonstrates the importance of the Proterozoic age -

. or processes in nature that may be viewed as compa.

rocks as the host for major deposits.

rable to some aspect of a system ofinterest." Another Host Rock--Certain host rocks are known to con-way of defining natural analogs is that they are sind.

tain specific coc'.modities. An example is that rare larities in nature. Pearcy and Murphy's def'mition can carth elements are hosted by carbonatites.

i L

43 NUREG/CP-0147

.m.

Application of Natural Analog Studies to Expiorationfor Ore Deposits GEOLOGIC TIME TABLE

• WORLD-CLASS DEPOSITS EONS ERAS M.Y.B.P. Y 8

Y I

I X

b Y

+ McLAUGNUN 0

CENOZOICp~ 65 - + CRIPPLE CREEK 230 PALEOZOIC 600 PRECAMBRIAN UPPER (Z) s00 -

MOUNTAIN PASS PALABORA $

PROTEROZOlc X KEY LAKE E

• " " ^

x JABILKUA 0

LOWER (X)

R A g ROXBY E

T LIND y

A

+ WITWATERSRAND E MINAS GERAIS 2500

+ KALGOORUE M

Ma,gy,g, m

+ RHODESIA.SEBAWKE B

+ RHODESIA R

. BARBERTON g

ARCHEAN(W)

A N

4500

• DRAWN TO SCALE Figure 8-1. Geologic age of world class are deposits Structure-Major ore deposits are commonly lo-exploration in a closed basin environment versus a cated in or associated with rift zones, cratonic environment.

Alteration Patterns-Concentric alteration halos Geologic Setting-Depending on the selected are particularly useful in porphyry copper explora-commodity of interest, a geologist may concentrate tion.

NUREG/CP-0147 44

Application ofNatural Analog Studies to Explorationfor Ore Deposits Geochemical Signature-Antimony, arsenic, in Chile. David Lowell was also involved in this and thallium display characteristic geochemical discovery and applied his knowledge of alteration anomalies in gold deposit environments.

pattems, sulfide distribution, and leached capping Sulfide Distribution-Pyritic halos occur above studies to explore this major system. The mine is and laterally to copper deposits and contain increas-currently operated by BHP-UTAH Minerals.

ing amounts of copper bearing sulfides as the central Numerous other examples of empirical model de-portion of the deposit is approached.

velopment are present in the literature, but their dis-Mineral Association-Barite commonly occurs cussion is beyond the scope of this paper. The point as a cap above stratiform lead-zinc-silver deposits, to be made is that empirical model development, or and cinnabar (mercury) is a significant pathfinder application of natural analogs,is an important facet j

mineralin gold exploration.

in exploration for mineral deposits and has been Geophysical Characteristics-The geophysical routinely applied with success.

characteristics associated with a deposit vary depend-ing on host rock, alteration, and mineralization. A 8.5 CONCElrr DEVELOPMENT magnetite deposit generates a magnetic high in con-As discussed above, concept development is the trast to a magnetic low over an altered porphyry development of a model based on theoretical consid-copper depostt.

erations and careful field observations to search for a Proper development of these characteristics and deposit type not known to exist in nature, but thought construction of a 3D model gives the exploration to be a possioility.The development of the mercury-geologist a blueprint of a type of deposit and enables hot springs-gold model that led to the discovery of the him to intelligently explore for and evaluate similar McLaughlin gold deposit in northem California m deposits with a muumum of time and expenditure.

1978 is a classic example of successful application of concept development. The following is a brief de-8.4 EXAMPLES OF SUCCESSFUL scription of the steps that led to the discovery of this APPLICATION OF EMPIRICAL significant gold deposit utilizing concept develop-MODEL DEVELOPMENT ment (Gustafson,1991).

TO EXPLORATION In the mid-70s, Homestake Mining Company was Probably the best documented successful applica-actively exploring for gold deposits, mainly in the tion of natural analogs to mineral exploration is the western United States, that could be mined by surface discovery of the Kalamazoo segment of the San methods. Homestake geologists were encouraged by Manuel porphyry copper deposit. The porphyry cop-management to generate new ideas, concepts, and ore per boom of the 1960s prompted detailed studies of deposit models to use in their search for ore deposits.

the known deposits by mining companies and indi.

The exploration prognm included exammation of old viduals. Foremost in model development was David precious metal districts to determine their potential Lowell, consultant, who applied alteration zoning for open pitable ore left behind by previous miners, pattems and sulfide distribution to determine that a and the writer was involved in these enmmations.

portion of the known San Manuel deposic was down-The fact that gold occurs in geothermal-hot spring faulted along the San Manuel fault (Lowell,1968).

environments is well documented in the literature.

Deep exploration drilling discovered the Kalamazoo The Broadlands Geothermal Field in New Zealand deposit at a depth of approximately 3,000 feet. This and Steamboat Springs in Nevada are examples of is a classic example of applying model development occurrences where gold is reported in limited quanti-or natural analogs to exploration that led to discovery ties and is actually precipitating from geothermal of a major copper resource. Based a comparison of waters today. Physical and chemical conditions in hot observed alteration patterns to those relations e other spring systems are right for the formation of gold deposits, geologists can detennine if a prospee is deposits, but these deposits were largely ignored by located high in the system or deep in the minerah ted explorationists, because economic concentrations system (Lowell and Guilbert,1970).

were not known to exist in presently active systems.

A second successful application of empirical Studies had previously been conducted by model development to porphyry copper exploration Homestake to determine the validity of a hot springs-is the discovery of the world class Escondida orebody gold model, but they revealed discouraging configu-45 NUREG/CP-0147

r Appliantwn ofNatural Analog Studic, to E.xplorationfor Ore Deposits I

rations. Geothermal systems do provide two impor-of these files in mid-1977 identified one area, first tant elements needed to form ore deposits: (i) a exammed by Homestake geologists in 1926, as sig-plumbing system, and (ii) hydrothermal solutions.

nificant, justifying a follow-up exammation. This Therefore, the hot springs-gold model was intriguing project area, Cherry Hill, near Wilbur Springs, at this point from a theoretical standpoint, but, prior Colusa County, California, was described by to the McLaughlin gold discovery, was only a Homestake geologists as containing hot spring de-concept.

posits, mercury minerahzation, and gold minerahza-The basic field observations made by the writer tion. Both mercury and gold had been mined from during the examination of numerous precious metaj this area on a limited scale in the late 1800s and early districts advanced a working hot springs-gold model 1900s.

supported by physical and chemical considerations.

Field examination of the Cherry Hill area con-The occurrence of occasional active or fossil hot firmed the previous observations. The area contains spring deposits, in the fonn of predominately sili-active hot springs, hot spring sinter deposits, and ceous sinter deposits, plus mercury minerahzation in widespread me:cury and gold mineralization in al-the form of cinnabar, was observed in close associa-tered Jurassic Knoxville sediments.

tion with several known precious metal districts.

This property examination produced th necessary Mercury had been used as a geocherical pathfinder working model for the mercury-hot springs-pld con-for years by explorationists in their : arch for gold cept and confirmed the earlier theoretical considera-deposits, but geologists did not consiuer mercury tions and field observations. Additionally, it provided mines as a potentiallocation to look for gold deposits.

an area containing ore grade gold value in association Other common mineral associations recognized in with mercury and hot spring deposits. The hypotheti-known gold camps are sulphur and antimony miner-cal concept now became a working model. The alization.

Cheny Hill occurrence also directed Homestake's Thesc observations were instrumentalin the devel-attention to the mercury districts of the California opment of the mercury-hot spangs-gold concept, and Coast Ranges, which were not known for gold pro-they triggered the thought process of relating actual duction and, therefore, were long ignored by precious field observations to theoretical considerations. Since metal explorationists.

hot spnng deposits and mercury mineralization occur The Cherry Hill target was drill tested and a small, near existing gold mines, why not consider searching currently uneconomic, open pitable inventory deline-for gold deposits near hot springs sud mercury ated. The Cherry Hill project served an important role mines? Metal zoning in minemi deposits is a well-in Homestake's successful gold exploration program documented feature. It was envisioned at this point that followed by providing an area to study and retine that low-temperature minerals, such as mercury, the genetic target model.

would fonn high in the hydrothermal system, and, if gold were present in the hydrothermal fluids, it 8.6 MERCURY-HOT SPRINGS-GOLD should precipitate at a higher temperature and, there-EXPLORATION MODEL fore, lower elevation, below the mercury mineraliza-tion. Examples of gold districts examined that contain A mercury-hot synngs-gold exploration model hot spring deposits and mercury mineralization in was developed and presented to Homestake manage-association with gold mineralization are (i) Bodie, ment by the writer at the Fall,1977 budget meeting.

Califomia, (ii) Borealis, Nevada, and (iii) Goldfields, The model, as presented, is shown in Figure 8-2. It Nevada.

consists of a geothermal system, structurally control-led, containing hot spring sinter deposits with cinna-As part of Homestake's exploration activities in bar (mercury) mineralization at the surface, with the mid-70s, a review of the numerous prospect ex-anomalous to significant gold value changing down-amination reports in Homestake's explomtion files ward to significant to economic gold value within a was undertaken to determine if any prospects exam-permeable host rock. Arsenic and antimony values ined by Hornestake over the past 75+ years justified were envisioned as additional significant geochemi-re-exanunation. This activity is not a pre-requisite for cal indicators of potential gold minerahzation at concept development, but in this case, has important depth. The presence of a permeable host rock is implications to the McLaughlin discovery. A review important to provide dissemination of gold minerali-NUREG/CP-0147 46

Application ofNatural Analog Studies to Explorationfor Ore Deposits HYPOTHETICAL MERCURY-HOT SPRINGS-GOLD MODEL FALL,1977 Hot Springs

~[

Sinter Deposits

~

Zon deih p:ng L.MM2D & g=

mm.

4&.,

,2 %.

w w &:mmm+5$a w:Jf sgVOLCANICS,N 1

mWU %

W,,c.7? O,

• =%

i:e g:.

~

O F=

v M y w @am w ~+

m wy n yn ut W~

.A,m o

oy O

• 5

- ),W 06 o o

, gnw o O ge, e 0

w o

oa O 8SEDIM.ENTS dmo N

oARGILLIZATION po 0,o Fa o e o o o

.gf yHW o c=

0 o

d y

g o

g 0

0 c

syV op og#n y

o VO O

T f

e, a

x B u>

y E Gold Zone y LIClFICATibE f

~

gg i.- l gGold in Quartz Veinlet g Figure 8-2. Hypothetical mercury-hot springs-gold model developed utilizing theoretical and field i

observations zation over a large area that could be mined by surface gram in the Coast Ranges of California near the methods rather than confining mineralization within working model, the Cherry Hill projects. The pm-a vein system that would dictate an underground gram was eventually expanded to include the westem mining situation.

United States.

)

t With the working exploration model developed, The program initially involved an extensive litera-and Cherry Hill presented as an example, a reconnais-ture review of all published mercury-hot spring oc-sance budget was requested to pursue the concept.

currences. Due to the location of Cherry Hill, the Homestake management viewed the concept and re-Clear Lake area received top priority, and reconnais-connaissance program with enthusiasm and provided sance began in this area. Eighty-eight mercury-hot l

sufficient funds to conduct a reconnaissance pro-spring areas were selected for exammation in the field gram.

and prioritized from the literature search.

An initial exammation, consisting of mapping and 8.7 MERCURY-HOT SPRINGS-GOLD sampling, was performed m. each area to determine RECONNAISSANCE PROGRAM geological characteristics and geochemical signature, The first question to answer was where to apply particularly the presence or absence of gold values.

the concept to achieve success and accomplish The exammation approach was to sample as deeply Homestake's corporate goal of discovering a plus as possible in the minemlized system as topography 1,000,000-ounce gold deposit. On a global sale, the would allow, that is, in dramages cutting the system, Pacific Rim, with its active geothermal areas and since gold was envisioned to be deposited at a lower metal pmvinces zoned around the Pacific, was con-elevation, in the higher temperature environments.

sidered potential prospecting territory, but it was Number ten on the list to examine was the Manhattan decided to concentrate the initial reconnaissance pro-mercury mine near the intersection of Napa, Yolo, 47 NUREG/CP-0147

Application ofNatural Analog Studies to E1plorationfor Ore Deposits

' ouun the completion of 409 drill holes and underground uenoocry 'N - ~ ~ ~ ~ ' ' [,7,', testing. The project was tumed over to operations in

,' we

!mmd June 1982, and the first bar of gold was poured in g""7 w.'q.G,"S^ \\,

March 1985. This classic epithermal mercury-hot 8,

c springs-gold deposit is owned and operated by

~~~~

MeLA GHLIN

\\

s Homestake Muung Company and by March 1990, S " " g^ gg ',,

had produced 1,000,000 oz of gold. Cunent produc-u Sa= u,nm he tion is approximately 250,000 oz of gold per annum.

'u

,Q',

The exploration that produced this discovery is an i,

"'",,,' -g- 'j, !,

excellent example of concept development and how natural analogs can be applied to mineral exploration.

', com j; t

2

' ' [,I 8.10 DISCUSSION

~

Figure 8-3. Map showing location of the This paper demonstrates the importance of natural McLaughlin Gold Mine, Napa and Yolo analogs in mineral exploration and the approach ex-P oration geologists use to discover new nuneral re-l Counties, California sources. Application of modeling or natural analogs and Lake Counties, California. Figure 8-3 shows the does have significance in the Yucca Mountain site location of the Manhattan mine, which was later characterization in that, once the Yucca Mountain site changed to the McLaughlin mine in honor of is fully described, the geological characteristics of Homestake's past chamnan, Dr. Donald McLaugh-known mineral deposit models can be applied to lin. The McLaughlin Gold Mine is located 13 airline determine if any potential models fit the geological miles south of the Cherry Hill area.

environment if a reasonable model is suggested, knowledge of modeling can assist in the evaluation 8.8 MANHATTAN MINE INITIAL of potential mineral resources.

EXAMINATION 8.11 BIBLIOGRAPHY The Manhattan mercury mine was examined on February 16 and 17,1978. The extensive alteration Gustafson, D.L.1991. Anatomy of a discovery; the and quartz veining observed on the surface were McLaughlin Gold Mine, Napa, Yolo, and Lake impressive and sampling of outcrops proved the pres-Counties, Califomia. BJ. Skinner, editor. Econ.

ence of gold values. Initially 32 samples were taken Geol. Monograph No. 8:350-359.

and assayed for gold, with a high value of 0.34 oz Au/ ton. Examination of this prospect and the assay Lowell, J.D.1%8. Geology of the Kalamazoo ore-results proved the concept, and a project was born, body, San Manuel district, Arizona. Econ. Geol.

Opumism was high that the Manhattan area had 63:645-654.

potential to become Homestake's first open pit gold mine.

Lowell, J.D., and J.M. Guilbert.1970. Lateral and vertical alteration.minerahzation zoning in por-I'I '

8.9 FOLLOW-UP ACTIVITIES AND OPERATION Pearcy, E.C., and W.M. Murphy.1991. Geochemical Natural Analogs Literature Review. Center for The area was drill tested and an ore reserve of Nuclear Waste Regulatory Analyses. CNWRA 20,000,000 tons @ 0.16 oz Au/ ton calculated after 90-008.

NUREG/CP-0147 48

l l

5

. NATURAL ANALOGS IN THE PETROLEUM INDUSTRY g

James R. Wood Michigan Technological Unitmity Houghton, Michigan 49931

9.1 INTRODUCTION

Numerous generic models for gas and oil accumu-lations were amassed by the 1940s, which could be The use of natural analogs in petroleum explora, tenned " natural analogs." These were catalogued and tion predates the industry itself. The earliest oil fields were discovered by people who had only a rudimen.

taught to succeedmg generations of petmlemn geob tary (at best) knowledge of geology, and they worked gists.This period was perhaps the apex of the use of natural analogs for exploration, which are well exem-solely by analogy, generally looking for surface ex, pressions of oil, such as seeps and tar pits. When the Plified by the diagrams reproduced in Figure 9-2.

Perhaps not coincidentally, this period also coincided anticlinal theory (Figure 9-1) of gas and oil accumu.

with the discovery of most of the world's major oil lation became popular, more geologists were em-fields and virtually all of the world's major petroleum ployed in the search for hydrocarbons, and surface mapping of anticlines became a standard tool in the Pmvinces.

explorationist's toolkit. The advent of seismic tech-From 1950 onward, as hydrocarbons became in-niques, in a way, simply extended this strategy by allowing geologists to "see" deeper into the subsur.

creasingly difficult to find, exploration strategies re-lied less and less on natural analogs and more and face. To a large extent, that is still current practice--

more on increasingly sophisticated technology, prin-one stilllooks for anticlines with closed contours as ci ally seismic technology, but also on such tools as P

potential targets.

remote sensing, as that science developed in the late 1970s. Today, the principal tools are computer mod-els that " integrate" as much of the available data as A

B C

possible and display it in Ways that would be difficult D

"^.

if not impossible without computers. The principal

[

objective is to " reduce risk," and any technique that accomplishes that goal is a candidate for inclusion in an integmtet, model.

To some extent, the transition to computer models reflects the increased capabilities of computers, par-J ticularly individual workstations, but it also reflects the significantly increased difficulty of finding gas and oil Older, once reliable techniques, based pri-marily on mapping subsurface geology have largely fhiled the industry in the 1980s, and the increased l5 Ml K

A,B,C, D s WERS enphasis on computertzed models is an attempt to develop new strategies. However, overlying all of this N

WAm oAs Figure 9-1. Anticlinal model for gas and oil is the increasingly evident fact that most of the acces-accumulation showing irregular distribution of sible hydrocarbons have been discovered and much gas, oil, and water in an asymmetric anticline, of that has been produced. Hydrocarbon exploration This was one of the first natural analogs used by in the lower 48 states clearly peaked about 1984-85 a'xlis now in a state of decline. Thus, the petroleum the petroleum industry and still dominates exploration strategies in some companies industry may serve as an analog itselfin the develop.

ment and utilization of technologies that include natu.

(Stewart,1951).

ral analogs. In particular it may serve as a guide to the 49 NUREG/CP-0147

f Natural Analogs in the Petroleum industry i

hh ~" ~ ' k'p%k'55.E g-SEEP

'@,,, - - ~.,), ssw~. -

p _ _. ~. ~ ~,,,,;,,,,

~ q '- -..

-~-

-~

r s,,e,,,

%,s

~ ~ ~

~,

-f,,, / s. s.'

~ ~~ ~ ~ ~~ ss N'(Q,,,

ss y Y k,',,

N Ng

'W

\\\\N ss

\\\\

'y s':,,,,,

,,s,,,s,\\Ng\\

/

f

=-1

~.

N i

~ ~~5.' V

'e

'. t ~ ~., '- ~-.-

~..

t w.-., ",

,. _:~..-

~ : ~...' ~_-.:': ~,~.~, -

~-3; ~:

~

~

,~.

- ~. - _ -

--.- f:,,,,,,

' ~ ~.,., ~., ~ ~

.-_~_'~-:~~~

~ -

-- ~ -,r.~

-t. ' ' - -

-.-.~~~..-

~

~

~ ' 33'ii~'.3

' [~ ' ' ' -.,~ ' '

~

~~

6, 5--

~~/ S', 9 S.il

-::=.7, ', ', ', ',I l

s

/

s

~.~wi.,.,,'=,,,%~ *, ~ ~

~s s

--e i i i

- ~~**=-

m*

~~/ I I

i i Figure 9 2. A series of natural analogs showing hydrocarbon trapping in faulted strata with erosional uncomformities (Stewart,1951)

" life cycle" of natural analogs in a particular applica-models and combining them with "hard data" from tion.

seismic surveys and well logs. The recent industry h

emP asisonbasinmodelingisagoodexampieof that Natural analogs played a large role in the early tred stages of the petroleum industry, but they came to be supplanted in the later, mature stages by technology.

The present era, which is a time of "old age" in the Specifically, the anticlinal model was joined by stra-petroleum industry, is one in which attention is being tigraphic traps, fault traps, diagenetic traps, and hy-directed toward integrated computer models that at-drodynamic traps as natural analogs, but industry teropt to combine theory with hard data in an effort emphasis shifted toward developing better seismic to simulate the evolution of a sedimentary basin on a technology beginning in the 1970s.

large scale. These models are, for the most part, just The natural analogs were not abandoned by any in their infancy, although the mass-transfer models means; they were by this time a part of the back-are an exception. However, only time will tell if this ground that every petroleum geologist was assumed will be a successful approach; success for the industry to have coming out of school. But by themselves, the being defined as reducing exploration rist.

analogs no longer provided the competitive advan-As an example of the evolution from the use of tage that large companies required to survive. How-natural analogs to models, we can consider two ever, the latest trend in industry research could be cases---the sxalled Allan fault-plane maps and at-described as quantifying the older natural analog tempts to predict porosity "in advance of the drill."

1 NUREG/CP-0147 So

Natural Analogs in the Petrolemn Industry The Allan fault-plane map can be regarded as a bridge FAULT PLANE PROJECTION between a tme geologic analog and a purely theoreti-cal model. Above all, the Allan fault-plane map does reduce exploration risk, and, mainly for that reason, it is widely used in the industry. At the other end of the spectrum, efforts to predict porosity in advance of w

the drill failed to reduce risk and were abandoned by the industry. It is instructive to examine both of these cases, one a success and the other a failure, as exam-

~

ple of the transition from natural analogs to computer models.

9.2 THE ALLAN FAULT-PLANE MAP The Allan fault-plane model has been discussed in detail (Allan,1989). This model essentially relates faults to fluid migration and hydrocarbon entrapment.

Figure 9-3. Example of Allan fault plane for a The basic assumption is that a fault is neither a seal faulted anticline illustrating structural and nor a conduit and thus, that the effect of faulting on stratigraphic geometry. Downthrown block is in migration and entrapment is a function of the proper-front of upthrown block; permeable beds are ties of the rocks that are brought into contact by indicated by blocked ends, impermeable beds faulting and on the geometty of the fault blocks.

by open spaces. In this example, each permeable a C n 8C an er PemeaMe M at As a simple example, consider a faulted anticline Pout As a md, hydroca&ns in which the strata on one side of the fault move down migrate vertically by passing from bed to bed relative to the other. Further assume that the strata act ss me fault %n, M.

consist of interlayered sands and shales. The tech-nique requires mapping the trace of the sands onto the Presented outside of Shell Oil in 1980. However, fault plane for both the upthrown and downthrown work had begun on the model in 1%7, and one can blocks (Figure 9-3). This is the fault-plane map. From infer that Shell Oil used the technique for a number it, geologists can deduce the likelihood that hydrocar-of years before word leaked out to the rest of the bons migrated from source beds to traps by crossing industry. This is an aspect of industry research that the fault where two permeable sands were juxta.

sets it apart from either academic or government posed. They can be done whem hydrocarbons may be research in that results are not communicated in a trapped by juxtaposition of shales opposite sands.

timely fashion and cannot be used by others to build Atlan (1989) discusses several examples of the ef-upon. Generally, one finds many wheels being redis-fects of varying sand / shale thicknesses on migration covered and results not published until either the and trapping. That paper should be consulted for competition clearly has acquired the expertise or it is more detail on the method.

no longer of any proprietary advantage. In addition, What is important here is the fact that a natural the fact that all research is driven by economic objec-analog, the anticline, has been used in conjunction tives seriously limits the directions researchers can with another natural analog, the fault-plane, to infer

pursue, a process, the migration of hydrocarbons. It is no longer sufficient to know that a faulted anticline 9.3 POROSITY PREDICTION exists; it is also necessary to access the probability Another example along these lines is the industry that at least one of the sands is filled with hydrocar-attempt in the late 1970s to predict porosity. Porosity bons. The fault plane taap provides a simple solution is obviously an important reservoir parameter, and to that problem, and, for that reason,it has found wide knowing beforehand if a target has adequate porosity application in the petroleum industry as a technique would provide a significant advantage. Industry in-that reduces exploration risk.

terest in secondary porosity was initially triggered by It is interesting to note that the paper outlining the petrographic observations that showed most sand-fault-plane technique appeared in 1989 and was first stones were subjected to dissolution processes that 51 NUREG/CP-0147

Neturn! Analogs in the Petroleum indus uy could only have occurred at depth. This dissolution TEXAS GULF COAST led to large increases in porosity on a hand specimen PAIRED POROSITY DATA scale, and it was thought that, if the processes oc-curred on a reservoir scale, then targets at depths once o-I thought to contain only completely cemented rocks

~

o

• oe o

IS

• g*..;

• yy**.

o*gJ #ga might in fact have adequate porosity.

g

• v*.

Most oil companies adopted an aggressive re-8&

f.*

o o.

g 4 e'. 1,* #

search policy directed toward quantifying and pre-5' o

dicting porosity trends with depth. At the time these 6

o,,oy.,h,'*,I,

/

studies were being initiated (~1978), diagenetic stud-s-go No b,* ' '.a ies were a low priority in most oil companies. From

[

g [' [=fa 1978 to about 1984, diagenetic studies were in vogue,

y lllo3W o

and geochemists and sedimentary petrographers were in high demand. However, around 1984, it became evident that the pmblem of porosity predic-tion would not have a quick solution, and attention began to wane.

"o io' io POROSITY There were several reasons for this. One was the fact that cementation and lithification processes were Figure 9-4. Porosity versus depth for a series of not well understood in tenns of either the geologic wells in the Texas Gulf Coast. Open circles settings or the basic physics and chemistry. Data indicate maximum porosity over a cored showing the variation in porosity with depth began to interval; open circles indicate minimum show that the porosity variation at any one depth Porosity. Length of the average cored intervals greatly exceeded the variation with total depth (Fig-is 18-20 feet. Thus, there are wide swings in ure 9-4). Thus depth (e.g., overburden load) alone Porosity over a short distance as well as a would not adequately account for porosity loss. The general trend toward lower pc.csity with depth.

second reason for waning interest was that other fairly simple. In contrast, rock alteration is much techniques, such as basin history diagrams and ther-m re complex; key questions remain to be answered, mal maturation models,were beguming to have some and critical experiments remam to be performed. In impact. Resources and attention stufted from porosity addition, the problem of model verification emerged prediction to maturation prediction, and the problem as a centralissue. It was not difficult to model poros-of porosity, cementation, and lithification were sty changes at depth, but proving the validity of any largely relegated to the academic community, a situ-Particular model has, so far, proved to be impossible.

ation that persists to the present.

The ultimate reason for the loss ofinterest was that Thus, the ability to move from natural analogs to 5 years ofintense efforts did not produce a model that theoretical models in an applied setting has proven explorationists could use to " reduce risk." The 5-year enormously difficult in practice. The fault-plane moratorium appears to be a pretty general rule in the model, which could be easily verified and under-petroleum industry. Projects that do not produce use-stood, is an example of a successful transition in the fut results in that length of time are abandoned, or,if petroleum industry. In contrast, the porosity predic-they have certain scientific merit, they are handed tion problem cannot even be stated in a simple, con-over to the academic community for further develop-cise fashion that would be agreeable to most ment. The current research effort in basin compart-geologists. The bottom line is that moving from natu-ments, for example, stems almost entirely from that ral analogs to theoretical models has proven more cucumstance.

difficult and treacherous than previously imagined.

However, several things can be learned from this In addition, it appears that, where such transfer has experience. One is that an adequate theory is neces-been successful, the natural analogs were already in sary in order to develop a useful tool. In the case of place and served to both guide and verify the model.

the fault-plane maps, both anticlines and faults were In a practical or applications-oriented environment, well understood, and,in addition, the modelitself was model verification has to have a high priority; perhaps NUREG/CkO147 52

Natural Analogs in the Petroleum Industry not to the point that models are not developed just or fission track data (Naeser et al.,1990). Assuming because no one can yet think of a way to verify them, conductive heat flow, the bunal history diagram can but verification needs to be considered at the time of be contoured for temperature (Figure 9-5).

initial model development.

These diagrams can be adapted to a variety of uses.

For example, the application of burial history dia-9,4 BASIN MODELS grams to diagenesis lies in being able to relate the Currently, a significant fraction of the research cementation to a specific time and depth. Several effon in the petroleum industry is directed toward ways have been proposed, but one method is to use basin modeling in one form or another. The ultimate O isotopes to estimate cementation temperatures, and aim is to produce a computer code that will simulate then, use the known thermal history to locate the the features in the evolution of a sedunentary basin cementation event in time-depth space (Wood and necessary to access the hydrocarbon potential. Many Boles,1991). The technique is relatively simple; once codes, some produced in academia, have the ability the temperature of cementation is established (via O to model certain aspects, but none are comprehensive isotopes) and the burial history diagram is contoured and none operate in three dimensions (3D). Develop-ment of true 3D basin history codes is a high priority.

f r temperature, sen, knowing the present sam'le depth, one simply backtracks along the burial trajec-Conventional single-well burial history models tory of the sample until it intersects the appropriate take ages (time) and depth to formation tops from is thenn. For example, assume that a sample is taken wells or measured sections and reconstruct a burial history by plotting time of deposition against rock exactly at the boundary between the Stevens sand and thickness (Figure 9-5). The only significant compli-the Reef Ridge /McLure (Figure 9-6) and that the cation is to account for sediment compaction during isotopic temperature is 80 *C. Then, the time and burial, but nearly all available models now do this.

depth of crystalhzation occurred at the point on the Once a bunal history diagram is generated, the ther-diagram where the top of the Stevens intersects the mal history is reconstructed using vitrinite reflectance 80 *C isotherm.

0'

\\ T ( C)

\\

Tulare San Joaquin 60

' 4,.

Etchegoin 80 7

g Reef Ridge /McLure x

w 100

[

Stevens 6 2.25 120 Faultvale Q

140 Devilwater/Gould Media /Rio Bravo

$ Vedder CLA 67-29 Salt Creek j

u Sandstone Tumey/Kreyenhagen l

5.5 Tejon 50 40 30 20 10 0

Age (Ma)

Figure 9-5. Typical geohlstory plot for a well in the southern San Joaquin Valley of California. Plot has been contoured for isotherms and corrected for compaction (Wood and Boles,1991).

53 NUREG/CP-0147

Natural Analogs be the Petroleum Industry

\\

North Coles Levee

%w g,\\,

f g\\

y e

50

-s i

k"

~

4

\\

so E

4x4.X 80

~

g-4 opis$ A ' "*

. san e jgg

+ 9tevens 110 3.0 t

5 8

6 4

2 0

Age (Ma)

-2 i

4i i

i 0.7065 0.7070 0.7075 0.7080 0.7085 0.7090 Figure 94. Geohlstory plot for reservoir ustmsr interval of Stevens sand at N. Coles Levee.

Upper boundary corresponds to the Figure 9-7. Plot of"Sr8Sr versus computed approximate top of Stevens sandstone. Data time of crystallization for strontium data from Points are calculated depths and times of N. Coles Levee. Heavy curves outline enveloM crystallization for carbonate cements.

of values. Light lines connect samples in order An additional application derives from die fact that of crystallization (Wood and Boles,1991).

it is then possible to construct time-series plots for nally, only an internal company memorandum was any parameter that can also be measured in the same written. As a result, when the basin history codes sample; for example, the C and Sr isotopes in the were eventually written, the development team calcite cements. Time spectra (e.g., plots of Sr/"Sr started from scratch and did not include the calcula-against time) are shown for the North Coles Levee tion of formation pressures. The new programs were, (NCL) strontium isotope data in Figure 9-7. Both the thus, more limited than the earlier program and con-embedding of the samples in the burirl history dia-sumed approximately 4 man-years of effort, much of gram (Figure 9-6) and the time spectra are techniques which may have been unnecessary.

that provide a quantitative link between sample meas-urements and basin models.

This brief sketch of the history of an industry research project highlights a recurnng problem in industry re. search; the failure to keep new generations 9.4.1 The Martin Model of researchers in touch with previous results. Corpo-Burial history models that include inorganic di-rate memory has proven to be remarkably short, agenesis are relatively new tools for analysis of min-perhaps only 10-15 years on average. It is consider-eral paragenesis, and only a few papers have appeared ably shorter in areas where all the investigators have retired orleft the company.

on the subject (Woronick and Land,1985; Burley et govd equations for Martin,s model of a al.,1989; Wood and Boles,1991). Ilowever in 1%9, 8

J. Martin at the Chevron Oil Field Research Company conipacting basin may be wntien:

outlined an algorithm that computed pore fluid pres-q - 9S = K/ (--0P/dz + p.)

sure as a function of burial history (Figure 9-8).

(Darcy's Law)

Martin's intention was to model " abnormally high fluid pressures" in sedimentary basins, but in the 4 /dz = &#dt 9

comse oflooking at that problem he also essentially (Fluid conthuity) solved the problem of computing basin history well before the technique became popular outside the in.

-0[S(1 - 9)]/3z =-0(1 - 9)/8t dustry. Unfortunately, nothing was published exter-(Solid continuity)

NUREG/CP-0147 54

. ~

Natural Analogs in the Petrok um Industry PERMEABILITY (DARCYS) 10*

10 107 10*

10 4

4 FLUID PRESSURE AND VERTICAL STRESS I

I I

I I

(THOUSANDS OF PSI)

VISCOSITY 0

6 IU

'6 0

04 08

'0 0

l I

I 8

8 i

3 2

]

4

[

g f6

~

8 O

h

@g n.

O j

P 10 9

4 0

@g g.-

9 6

12 dg 9

• p A

14 p

m g%

i f

1. p 16

'O

^!

0.15 0.25 0.35 0.45 0.55 1

POROSITY I

i 1

1 I

I i

i 1

0 50 100 150 200 250 300 350 400 j

TEMPERATURE (* 6 i

i Figure 9-8. Variations in fluid pressure, rock stress, porosity, viscosity, and permeability calculated using Martin equations discussed in text. Results in Egure are for after 40 My of constant sedimentation at a rate.of 500 feet /My. The abrupt change in permeability corresponds to a change in I

deposition from a clay to sill 1

o = o, + P S = velocityofsolidmatrix t

= time (Effective stress) z = verticaldistancefromtopofbasement

= fluidviscosity a =t [(1 -9) p,+ 9p ] dz pr = density of solids Q=-AVT pw = fluid density 1

o = totalvenicalstress (Fourier' Law) os = stress on solid matrix where:

Q = heat flow

$ = Porosity 12 = thermalconductivity P = fluidpressure T = temperature K = permeability q = fluidflowrate i

55 NUREG/CP-0147

untmiAnalogs u the u,,:. = 1nduery These equations represent six unknowns (q, S, T, with diagenesis and will include time-temperature P, o, and o.). The porosity is assumed to be a known dependence.

function of o., while the fluid viscosity is a known function of temperature, T, and pressure P. The vari-9.5 SEISMIC PUMPING AND ation in permeability with depth, z, depends on the FAULT VALVES -

amount of compaction and the initial permeabilities in the sediments when they are deposited. Thus vari-The seismic pumping model introduced by Sibson

. ations in K and z are c4'ilataA with the solution, et al. (1975) attributed the movement of pore fluid in The boundary conditions require that q=0 at z=O rocks to the dilation / collapse cycle associated with so that.

large-magnitude ear %=be This is, in turn, a vari-ation on the inclusion / collapse theory of deep earth-BP/3z = - p, at z = 0 quakes proposed by Brady (1975). This model 30./&z = -(1 - 9)(p,- p.) at z = 0 (Figure 9-9) assumed that, prior to an earthquake, a At the surface, volume of rock around the fccus dilates in response o.=P=0 for z = h t tect nicstressesbytheopeningofextensioncracks and fractures. The development of this fracture po-where h is the total thickness of sedunent, and the rate rosity decreases the fluid pressure in the dilatant zone of deposition, d(t), at z = h is also a boundary condi-and induces a slow inward migration of pore fluid tion. These equations constitute a moving boundar,i -

from the surroundmg rock. When stress is released value problem that has to be solved numericany.

fonowing an earthquake, the cracks in the dih However,thesolutionsprovidethedependenceof the zone close and expel the fluid rapidly upwards. The fluid pressure, porosity, permeability, and lithortatic expelled fluids will follow a path ofleast flow resis-stress, as well as the variation in rock thickness as a tance, which will generally be the more permeable function of time (Figure 9-8).

Efforts are now under way at several institutions

%g*g (Cornell, Michigan Technological University) to write a burial history code based on these equations, y

q which should result in an improvement in the cunent

(

j ouno,.

practice, as well as an expansion of the capabilities Mo, i

of basin history models to handle situations involving rautt,t m smow geopressured rocks. In general, this study should provide a better understandmg of the chemical and p

physical processes that give rise to fluid movements l(l'

")an zows and rock alteration in basins. The timing and duration

[ (

of cementation events in clastic rocks is of particular s i s=P.,,

s s interest. If a case can be made that fluids move N =

~2 7=t =j episodica'ly through complex flow networks of po-N

• J %#

rous sands connected by faults, then it would be N

worthwhile to try to establish a link between rock

)

cementation and seismic activity in basins.

,e s

A logical follow-up to the Martin burial history model would be to develop a mass-transfer code

. p ar similar to EQ3/6 or CHHlER that would model the geochemical and isotopic evolution of the cements h^"^ = '*n and vein fill. However this step is beyond current capabilities and wouki require either massive rewrit.

Figure 9-9. Initial model for seismic pumping ing of current mass-t:ansfer codes or r, tarting from based on concept of rock dilation in vidnity of a scratch and writing a code specifically designed for fault. This model assutned that pore Duld diagenetic analysis. Some preliminary work is in entering the dilatant zone is rapidly expelled progress on writing a new mass-transfer code based following stress release after a seismic event on the Simplex algorithm that will c: cal specifically NUREG/CP-0147 56

Natural Analogs in the Petroleurn industry 1

horizons encountered. At depth, the flow willlikely liters of fluid per seismic event) rapidly from onc follow faults, but, at shallow depths, high-permeabil-pressure-temperature environment to another. Sibson ity zones in sands and silts will offer altemative et at (1975) calculate that a fault with a displacement 5

8 i

routes.

of 1 km could produce 10 -10 fluid pulses. Sibson Sibson (1981) subsequently introduced the fault (1981) discussed direct evidence (e.g., elevated valve model as an alternative to the seismic pumping groundwater flow, vein fill, etc.) that channel flow of model. The basic tenet of this modelis that overpres-pore fluid accompanies shallow faulting.

sure seals in basins are episodically punctured by earthquakes, and fluid migration occurs until either 9.6 BASIN COMPARTMENTS hydrostatic conditions are reached or until the fault An ther model closely related to the fault valve "self seals" by precipitating cements (Figure 9-10).

m del that has been proposed recently is the basm This version of the seismic pumping model has obvi-c mPanment model.The essence of this model(Fig-ous affinities through the overpressure connection to ure 9-11) is that basins are broken up into compart-l more conventional basin models.

ments or " bottles" separated laterally by faults and Both of these models provide for a cyclic pumping vertically by cement seals (Bradley,1975; Powley, of pore fluid, which could provide explanations for

)

1990; Hunt,1990). The model is based primanly on both the fluid driving force and the epi:. odic nature of measurements of fluid pressures m, wells that suggest the fluid movement.They also provide a mechanism there are abrupt changes or discontinuities in fluid for mineral precipitation / dissolution in that the fluid pressure laterally across faults and venically across i

pumping moves large volumes of fluid (e.g.,10 (top) cement seals (Figure 9-12).

(a) PREFAILURE The vertical seals are postulated to be zones in which large volumes of cements have been precipi-rLuto,REssu"E tated. It has been suggested that the vertical seals are HYOROSTATIC g j FAUL of the 100 *C isotherm and that the cements migrate upwards as the basin subsides such that the cements

,,y remain more or less fixed on the 100 C isotherm

\\

l %

(Hunt,1990). Whether this is exactly true or not, there

]

i 1

• 4 does seem to be a: least some evidence that certam o""'" ESSURE k

I diagenetic reactions, such as albitization :md illitiza-DE 'm i

5 f

tion, also occur near this temperature (Boles,1979, 1984; Land and Milliken,1981; Lundegard et al.,

1984; Morad et at,1990) and may be related to the (b) POSTFAILURE formation of top seals.

rouio PRESSURE

\\ \\ PERMEABLE RUPTURE valve model and the basin compartment idea. At-rowE k

\\\\.j BASIN COMPARTMENT MODEL

\\

mcruRao seu Q

3

/

\\L oe >m

}%

y

1. ' OVERPRESSURED HYDROSTATIC Figure 9-11. Model for basin compartments in Figure 9-10. Fault valve model which replaces which normally pressured rocks overtie several seismic pumping model. In this model, the fluid compartments by supernormally pressured movement following a seismic event is due to the rocks which are separated laterally by faults puncturing of the overpressure seal by the fault and vertically by shale or cement seals (Sibson,1981).

(Prowley,1990) 57 NUREG/CP-0147

I I

Nehmd Analogs in un Petrolaan huhntry 2=

. problem has bearmg on porosity evolution, reservoir NORMAL PRESSURES quality, and hydic# von migration, as well as an 4'

understandmg of one of the most basic geologic processes.

. SEA 1. N Recent 'MaMag about processes occumng in sedi-l6=

mentary basins has led to models that propose long

[s.

Upper Ekofisk EtonsK distance (>l km) transport of fluids and cements.

Compartment Ryk Models based on setsrme pumping (Sibson et al.,

g 10 =

EKoRSK 1975), fault valving (Sibson,1981) and basin com-p sm partments (Powley,1990; Hunt,1990) are good ex-E amples. However, with the exception of the work by 12 '

Boles and Ramseyer (1987), hard evidence support-Lower Ekofisk ing or refuting such ideas is generally lackmg.

14 '

Compartment A previous study of carbonate-cemented sand zones at NCL led to the conclusion that cements in 16 some of the zones were precipitated episod cally and 0

4 8

12 could be related to seismic activity (Boles,1987; l

PRESSURE (P.S.111000)

Boles and Ramseyer,1987; Schultz et al.,1989; Wood and Boles,1991). In addition to the recent Figure 9-12.Model for pressure cc p ; ; t work at NCL, the regional study by Surdam et al.

at Eko5sk Beld in the Cental Graben of the (1988), which modeled the geologic history of the North Sea. Note the paucity of data points the southern San Joaqum Basin, and the work by Cathles interpretation is based upon (Powley,1990).

and Smith (1983) on episodic ore deposition are j

tempts are under way to link these two models into a relevant to this study. Boles (1991) discussed some moir comprehensive basin model, an effort that illus-Petmgraphic observations that suggested hydrocar-trates the current trend toward buildmg large basin bons were involved in the dissolution of plagioclase I

models. The interdisciphnary nature of this type of at NCL These observations would be consistent with research is also evident and partially explains the earlier suggestions by Surdam et al. (1984) and Sur-trend toward research teams composed of experts in dam and Crossey (1985) that organic acids leached i

allied fields, as well as the tendency toward including from hydrocarbons are instrumental in feldspar di-more personneloutside a company.

agenesis.

9.7 MASS TRANSFER AND 9.8

SUMMARY

DIAGENESIS The petroleum industry has used natural analogs practically from its inception at the turn of the cen-Finally, one of the last major areas to be incorpo-tury. Early analogs were simply basic geologic struc-rated into basin models is geochemical mass transfer.

tures, which, although new at the time, were easily Although a number of codes currently exist in this verified and put to practical use. From about 1900 to area, EQ3/6 and CHril ER, for example, to date none after Wodd War II, these types of natural analogs i

have been combined with basin history models or were expanded and refined. However the rise of j

with basin compartment models. This is ahnost cer-seismic techniques in the 1940s and 1950s began the j

tainly a shortcoming that will be addressed in the shift in emphasis from purely natural analogs to geo-future, since the ability to predict diagenetic changes logic interpretation based on seismic data. In the late j

in rocks is still an industry priority.

1960s, computer modeling began to add a new di-Currently, one of the principal problems in di-mension to the industry research efforts, and, by agenesis lies in understandmg rock cementation (as 1980, basin modeling had taken hold.

j distinguished from rock lithification), which requires In its simplest form, basin modeling was the inte-consideration of: (i) the sources of the cementing gration of natural analogs with models from other materials, (ii) the transport of the cements, and (iii) disciplines. Adding hydrocarbon maturation calcula-the deposition (precipitation) of the cements. This tions to basin history models is typical of the direc-NUREG/CP-0147 58

htural Analogs in the Petroleum Industry tion. Today, the greater. research emphasis is on more sultants and will focus on development of coupled comprehensive basin models, with the ultimate aim models for basin processes. Key questions will in-to be able to model the evolution of a sedimentary volve model verification and moving new technology basin from the initial deposition of sediment to the more rapidly and efficiently to practical application.

present state. Progress has been made in this direc-Finally, there is already a noticeable shift from tion, but generally in isolated areas, and the effort is exploration models to work on reservoir management really in its infancy at the present time.

and enhanced recovery techniques. This area offers One major stumbling block has been in the area of considerable opportunity to combine natural analogs model verification. It is obviously impossible to with quantitative models, since the " easy" oil has simulate basin development in the laboratory, and, already been recovered and new techniques will re-thus, answers o key questions, such as, "What is the quire a better understanding of the natural environ-dommant mode of fluid flow in the subsurface?"

ment and the effects of recovery on the rock matnx.

remain unanswered. The issue of verification re-volves around two points: (i) how should " ground

9.9 REFERENCES

tmth" data be incorporated into theoretical models?

Allan, U.S. (1989) Model for hydrocarbon migration and (ii), what degree of verification is necessary?

and entrapment within faulted structures: Ameri-Complete verification is out of the question for a can Association of Petroleum Geologists Bulletin, comprehensive basin model, so the problem of just 73 (7), p. 803-811.

what data need to be acquired and how they should Boles, J.R.1979 Active ankerite cementation in the be integrated into the models is rapidly becoming the subsurface Eocene of Southwest Texas: Contr. to question of the day.

Mineralogy and Petrology 68, p.13-22.

.n addition, the move from purely natural analogs Boles, J.R.1984 " Secondary porosity reactions in the to more theoretical models represented by computer Stevens sandstone," San Joaquin Valley, CA, Bull.

codes has presented a new set of problerns for re-Amer. Assoc. Petrol. Geol.Mem.37,pp.217-224.

search managers in the petroleum industry and has Boles, J.R.1987 "Six million year diagenetic his-required a change in research procedures. Coupled tory," North Coles tevee, San Joaquin Basin, Cali-with the shrinkage of research budgets, program man-fornia in J. D. Marshal (ed.) Diagenesis of agers have begun to rethink the way in-house research Sedimentary Sequences, Geol. Soc. Spec. Pub. 36, is conducted. For most companies, including many of pp.191-200 the larger ones, it is no longer possible to pursue all Boles, J.R.1991 Plagioclase dissolution related to oil areas of potential interest in-house. Currently, the residence time, North Coles Levee Field, Califor-concept of " leveraged" research has arisen, the idea nia- (abs) American Association of Petroleum Ge-being that joint projects with universities will stretch ologists,1991 Annual Convention Official research dollars and allow some projects to remain Program. p. 81.

alive that would have otherwise been tabled. At the Boles, J. R. and Ramseyer, K.1987 Diagenetic car-present time, perhaps as much as 70-80 percent of the bonate in Miocene sandstone reservoir, San basic research is being conducted as joint projects, Joaquin Basin, California. Bull. Amer. Assoc. Pet-particularly the basin modeling efforts, since this is rol. Geol. 71, p.1475-1487.

an area of considerable academic interest anyway.

Brady, B.T. Theory of earthquakes. Part III: 1975 In genemi, the petroleum industry found that out-Inclusion colh,pse theory of deep earthquakes Pure side research programs were most beneficial when (i) and Applied Geophysics 114, p. I19-139.

they provided access to new technologies, and (ii)

Bradley, J.S.1975 Abnormal formation pressure:

kept long-term research efforts alive that would have AAPG Bulletin,59, p. 957-973.

otherwise been abandoned. The main difficulties Burley, S.D., Mullis, J. and Matter, A.1989 Timing were in finding competent personnel who were not diagenesis in the Tartan reservoir (UK North Sea):

already working for a competitor and working on a constramts from combined cathodoluminescence time cycle centered around student thesis projects.

microscopy and fluid inclusion studies. Marine Future trends in long-term research in the petro-and Petroleum Geology,6 (2), p.98-120.

leum industry will probably see even more emphasis Cathles, L.M. and Smith, A.T.,1983, ' Thermal con-on joint projects with universities and outside con-straints on the formation of Mississippi Valley-59 NUREG/CP-0147

Natural Analogs in the Petroleum industry Type lead-zine deposits and their implications for Sibson, R.H.1981 Fluid flow accompanying fanit-episodic basin dewatering and deposit genesis,"

ing: field evidence and models: in D.W. Simpson Econ. Geology, v. 78, p. 983-1002.

and P.G. Richards (Editors), Earthquake predic-Hunt, J.1990 Generation nd migration of petroleum tion: an International Review: Am. Geophysical from abnormally pressured fluid compartments Union, Maurice Ewing Series,4, p. 593-603.

Am. Assoc. Pet. Geol. Bull. 74(1) p.1-12.

Stewart, W.A.1951. Unconformities. In LeRoy, Land, LS. and Milliken, K.L 19S 1 Feldspar diagene.

LW. (ed.) Subsurface Geologic Merliods, Col sis in the Frio formation, Brazoria County, Texas School Mines, Golden, CO, p.,32-70.

Gulf Coast: Geology,9, p. 314-318.

Surdam, R.C., Fischer, K. J. and Heasier, H. P.1988 Lundegard, P.D., Land, LS. and Galloway, W.E.

Hydrocarbon maturation modeling of the Southern 1984 Problem of secondary porosity: Frio Forma-San J aquin basin, California: in Studies in the tion (Oligocene), Texas Gulf Coast: Geology,12, 8e I gy of the San Joaquin basin, ed. S. Graham,

p. 399-402.

Pacific Section SEPM Special Publication, p. 53-64' Morad, S., Bergan, M., Knarud, R. and Nystuen, J.P.

Sur

, R.C. and Cmssey, L.J.1985 Organic-inor-1990 Albitization of detrital plagioclase in Triassic gani interactions during progressive burial: key reservoir sandstones from the Snorre Field, Nor-t Porosity / permeability enhancement and/or wegian North Sea: J. Sed. Pet. 60(3), p. 411-425.

preservation: Philosophical Transactions of the Naeser, N.D., Naeser, C.W. and McCulloh, T.H.

Royal Society of London, series A,315, p.135-(1990) Thermal history of rocks m Southem San 156.

Joaquin Valley, California: Evidence from fission-Surdam, R.C., Boese, S.W. and Crossey, LJ.1984 track analysis, American Association of Petro-The chemistry of secondary porosity in Clastic leum Geologists Bulletin,74, p.13-29.

Diagenesis, Mcdonald and Surdam (eds) AAPG Powley, D.E.1990, Pressures and hydrogeology in memoir 37, p.127-149, petroleum basins: Earth-Science Reviews, 29, Wood, J.R. and Boles, J.R. (1991; in press) Evidence

p. 215 226.

for episodic cementation and diagenetic recording Schultz., J.L., Boles, J.R. and Tilton, G.R.1989 of seismic pumping events at N. Coles Levee, CA.

Tracking calcium in the San Joaquin basin, Cali-Applied Geochemistry.

fomia: a Strontium isotopic study of carbonate Woronick, R.E. and Land, L.S.1985 Late burial cements at North Coles Levee; Geochim. Cosmo-diagenesis, Lower Cretaceous Pearsall and Lower chim. Acta. 53, p.1991-1999.

Glen Rose Formations, South Texas: In N.

Sibson, R.H., Moore, J.M. and Rankm, A.H.1975 Schneidermann and P. M. Harris (Editors), Car-Seismic pumping--a hydrothermal fluid transport bonate Cements. Soc. Econ. Paleontol. Mineral.,

mechanism. J. Geol. Soc. Lond.,131, p. 653-659.

Spec. Publ.,36: p. 265-275.

NUREG/CP-0147 60

I THE POGOS DE CALDAS INTERNATIONAL PROJECT: AN g

EXAMPLE OF A LARGE-SCALE RADWASTE ISOLATION-NATURAL! ANALOGUE STUDY 1

Michael Shea TermCon Inc.

~

Chicago, Illinois

10.1 INTRODUCTION

Ulbrich (1984). The regional geology, geochemistry,

. The proper isolation of radioactive waste is one of and hydrology are discussed in more detail by Hol-mes et al. (1991), Schorscher and Shea (1991), and today's most pressing environmental issues. Re-Shea (1991).

ecarch is bemg carried out by many countries around the world in order to answer critical and perplexing 103 CURRENT STUDIES questions regarding the safe disposal of radioactive 1

waste. Natural analogue studies are an increasingly important facet of this international research effort 103.1 Project History 3

(chnaman et al.,1984; Chapman and McKinley, The Po90s de Caldas Project was conducted for E

1990).

3-1/2 years from June 1986 until December 1989, The Pogos de Caldas International Natural Ana-under the joint s,mnsorship of SKB (Sweden), NA-logue Project was designed to study radwaste-related GRA (Switzerland), the Department of Environment i

processes occurrmg in the natural environment, and (United Kingdom), and the Department of Energy j

j the study area contains many features analogous to (United States), with considerable support from a those anticipated to impact the isolation of radioac-number of orgam7ations in Brazil, most notably Nu-j tive waste. The project was a multidisciplinary study, clebras (now Uranio do Brasil). A first-year feasibil-covering such diverse disciplines as geology, geo-ity study (Smellie et al.,1987) was followed by 2-1/2 chemistry, hydrology, geomorphology, and microbi-years of data collection and interpretation. The U.S.

ology, as well as related modeling-Department of Energy (DOE) was involved in the i

The Pogos de Caldas Project represents a major technical and programmatic planning of the project effort of the intemational technical and scientific from its inception, with members on both the Steering community towards addressing one of modern civi-and Techmcal Committees, as well as in active re-lization's most critical environmental issues-radio-search. The Commission of European Communities active waste isolation.

hosted the fmal symposium of the Pogos de Caldas Project at Pitlochry, Scotland, in June 1990.

10.2 PREVIOUS STUDIES AT The project gem a series of technical reports, POQOS DE CALDES which are pub'ished in 15 volumes and a collection The Pogos de Caldas area has been the subject of of scientific papers that will appear in a Special issue other long-term studies. In particular, these have fo-of the./ournal of Sploration Geochemtrtry. These cused on the Morro do Ferro Th/REE ore deposit. In papers will be beneficial to the scientific members of 1977, an international conference on areas of high the international radioactive waste community, as i

natural radioactivity was held in Pogos de Caldas well as adrnaustrators and government officials re-(Cullen and Penna Franca,1977). Studies have also sponsible for the management of research project::,

l!

been canAM regarding the biological uptake of U, and they will serve as an important guidepost for Th and their daughter products (Eisenbud et al.,

other current and future studies. They will also be of j'

1984; Campos et al.,1986).

interest to earth and environmental scientists for their The geology and geochenustry of the Pogos de technical and scientific information. Finally, applied CnMas area has also been previously studied. Exam-modelers and mathematicians will fmd these papers pies of these earlier studies include those of Ellert et to be a valuable example of a complete and interactive

[

al. (1959), Amaral et al. (1%7), Bushee (1971), and M1ing exercise.

61 NUREG/CP-0147 i

ne Popu De Caldas IntematkmalProjat 103.2 Study Sites microbiological chemical reactions, flow channeling.

The study area, in the province of Mmas Gerais, matnx diffusion, and redox retardation. These proc-Brazil, is located within an 80-million-year old alka-esses and interactims wm be disassed funher be-lic volcanic / plutonic complex comprising a 35-km I *-

diamat-subsidence caldera. A naturally high radio-The reader shald note that this summary paper is activity is associated with the local volcanic rocks highly-distilled and "pechy," in the spirit of the (mainly phonolites and nepheline syenites), geother-Presentation at the San Antonio workshop, with an mal spnngs, and uranium deposits. It contains two absence of pedagogical approach or discussion sec-sites of panicular interest, the Osamu Utsumi ura-tion. Therefore, I strongly encourage the reader to nium mine and theMorm do Ferro thorium / rare-earth read the more thorough and explanatory Summary ore body. The first site is notable for the prominent Chapter of the Pogos de Caldas Report Series (Chap-redox fronts marainaA_ in the rock, while the Morro man et al.,1991), as well as any specifically refer-do Ferm is recogmzed as one of the most naturally enced reports.

radioactive locations on the surface of the Earth. The geology and mineralogy of these sites are discussed 10.4.1 Sorption in more detail by Waber et al. (1991) and Waber Amorphous phases at Pogos de Caldas were ob-(1991).

served to reduce trace element mobility. As they age and crystalle, they may either release sorbed species 10.3.3 Principal Objectives into the groundwater or incorporate some given spe-The project had four principal objectives that were cies from the groundwater into their crystalhne struc-used to focus the research of the project (Smellie et ture (Nordstmm et al.,1991). Therefore, performance al.,1989), such that it remamed as pragmatically assessment models may need to account for the pre-oriented as possible towards the overall goal of study.

cipitation of an amorphous phase evolving to a crys-ing processes as they related to radioactive waste talline phase, rather than only sorption onto isolation issues.1 hose objectives were: (i) assist in crystalhne phases. Some trace element fixation proc-verification / calibration of hydrochemical models, esses appear to be irreversible (MacKenzie et al.,

codes, and d**=hacas used to evaluate rock / water 1991). Thus, the reversible sorption mechanism mod-interactions and the solubilities and spectations of els currently used in performance assessment codes elements; (ii) determine interactions of natural do not appear to be appropriate. These performance groundwater, colloids, radionuclides, and mineral assessment models tend to be too conservative and, surfaces with respect to radionuclide transport proc.

in tum, overly pessimistic, which may have negative i

esses and colloid stability; (iii) produce a model of cost / benefit and siting criteria impacts.

geochemical transport across redox fronts, with spe-l cial attention to redox-sensitive natural series ra.

10.4.2 Solubility and Speciation dionuclides; and (iv) model migration of REE/U-Th The results of trace element solubility and specia-series radionuclides during hydrothermal activity tion modeling for Po90s de Caldas gmundwaters simdar to that anticipated in the near field.

were generously conservative, but several errors were noted in the thermodynamic databases that were 10.4 OBSERVATIONS AND used (Bruno et al.,1991). In particular, it was noted i

IMPLICATIONS that a number of key observed mineral phases were Many of the results of the project have a direct missing in the databases. Therefore, it is important to application, or at least an implication, towards per.

ensure that the thennodynamic d'tahacas used in formance assessment. Some of the natural processes Perfonnance assessment include the appropriate min-and interactions observed at Poqos de Caldas need to erals that control solubility and speciation for any be incorporated into performance assessment mod.

given test case, 1

els, particularly in terms of sorption, solubility /spe-ciation, code / database comparison, colloid transport, 10.4.3 Colloid Transport microbial transport, redox transport, and hydrother-Evidence from the Morro do Ferro study site mal transport. These observed processes / interactions showed that groundwater particulates were not mo-include amorphous to crystalline phase transitions, bile and may actually have aided trace element retar-NUREG/CP-0147 62

De Pogos De Caldas l&a Wm.d Pnyect dation through an element fixation / particle filtration redox " retardation" adds a possible element of con-mechanism (Miekeley et al.,1991a; Miekeley et al.,

servatism to transport models involving redox dis-1991b). These results suggest that there may not be a continuities.

" colloid problem," but where they are broadly appli-cable to other geologic environments is not clear and 10.4.6 Hydrothermal Transport needs further study.

Cathles and Shea (1991) applied physical and chemical codes to model both the natural, intrusive-10.4.4 Geomictubiology driven hydrothermal system at the Osamu Utsumi Microbially mediated chemical reactions at Pogos uraruum mine, as well as a case in which hydrother-de Caldas appear to be very significant, even in deep mal processes were driven by a hypothetical under-groundwaters (West et al.,1991). Therefore,if spe-ground repository. The hypothetical repository was cific microbial reactions and their related enhanced modeled to be simdar to a possible United States kinetics are not accounted for in geochemical models, high. level nuclear waste (HLW) design, and placed performance assessment results will not necessarily within the Osamu Utsumi uranium mine host rock.

be conservative.

The hydrothermal circulation system around the hy-pothetical repository proved remarkably simdar to 10.4.5 Redox Transport that observed at the mine. However, the calculated Modeling of the redox fronts at Pogos de Caldas alteration around the repository was only 0.1 percent (Cross et al.,1991) included simple mass-balance relative to the natural system. The relatively high scoping calculations, chemical equilibrium model.

temperatures operative in the natural system at Pogos ing, kinetic chemical modeling, and modeling of de Caldas, and as modeled for the hypothetical re-specific features of the redox fronts, such as miem-Pository, are only probable for some United States biological processes (West et al.,1991), natural-se.

HLW scenarios and possibly also for the French ries radionuclide profiles (MacKenzie et al.,1991),

HLW repository concept.

and trace element redistribution (MacKenzie et al.,

1991).

10.5 LESSONS LEARNED The simple mass-balance calculations explain the In a more general sense, there were many impor-formation of redox fronts in very general tenns, but tant lessons learned in the Po90s de Caldas Project.

greatly simplify the processes known to be occumng Those outlined below are discussed in more detail by at such fronts. Coupled transport / chemistry models Chapman et al. (1991).

provide a better simulation of some aspects of the Heterogeneities in the physical and chemical e

fronts, but they are really interpretive models without properties of the rocks and waters occur at all any convincing predictive abilities. In particular, they scales and are not always possible to charac-poorly simulate trace element chemistry in either terize adequately.

solution or solid phases' Robust models were required to interpret the e

The process-specific modeling sheds some light sparse and heterogeneous data, and a simdar on the inaccuracies of the chemical modeling. Mic" approach willlikely be needed to characterize a bial catalysis appears to be very significant in redox repository site.

systems, particularly affecting the chemistry of sul-Mthough there has been mobilization and en-e fur. Natural-series measurements clearly show that nehment of uranium and other radionuclides, redox fronts move very slowly. Further, trace element large-scaled rapid transport of radionuclides distributions reveal that the Pogos de Caldas redox has not occurred at Pogos de Caldas.

fronts appear to retard a wide spectrum of trace elements, even many nonredox sensitive elements,

• The project has shown that data collection can-probably due to elemental sorption / fixation or solid not be rushed, no matter how many resources are available at the outset.

solution in precipitating amorphous FeO(OH).

It is particularly important that hydrochemical The above results support the far-field conceptual e

model of advective flow of oxidants along major SPeciation data be collected in situ.

fracture zones coupled with slow, diffusive solute

. Any study site will be more complex, ill-de-exchange with the surrounding rock. The observed fined, and perturbed than expected.

63 NUREG/CP-0147

The Pogos De Cnidas Intemational Prgect

• It is not always possible to constrain a process Bushee, J. (1971). Geochronological and pet-

' adequately in order to study and interpret it.

rographic studies of alkalme rocks fmm==he Fundamental refinements to hydrochemical Brazil. Ph.D. dissertation, University of Califor-e trace element models are theoretically transfer-nia, Berkeley, able and will be usable at any study site, if Campos, M., Franca E. Penna, N. Loba, R. Trindade, applicable.

and L. Sachett (1986). Migration of Ra from the Analogue studies pmvide invaluable experi-Th ore deposit ofMorm do Ferro, Pogos de rama, a

ence for conducting performance assessments.

Brazil. J. Environ. Radiation 3,145-161.

• Modelers and investigators must coordinate Cathles, L, and M.E. Shea (1991). Near-field high-data requirement at the outset.

temperature transport: Evidence from the genesis Some data needs may be impractical or impos-of the Osamu Utsumi uranium mine, Pogos de e

sible to fulfill.

Caldas, Brazil, SKB TR90-22; NAGRA NTB90-31; UK DOE WR90-53.

10.6

SUMMARY

rha,= an N., and I.G. McKinley (1990). Radioac-Data collected at Pogos de Caldas were success-tive waste: Back to the Future? NewScientist 126, fully used to verify / calibrate (first steps toward vali-54-58.

dation?) numerical models and to confirm or correct Chapman, N., I.G. McKinley, and J.A.T. Smellie laboratory measurement. In particular, conections (1984). The potential of natural analogues in as-

' were made to thermodynamic values as used in geo.

sessing systems for deep disposal of high-level chemical codes. Also, investigators identified mate

• waste. NAGRA NTB 85-41, rials and processes that had not previously been Chapman, N.A., I.G. McKinley, M.E. Shea, and identified as related to radionuclide migration and J.A.T. Smellie (1991). Introduction and Summary needed to be mcorporated into performance assess-of the Project, SKB TRB77; NAGRA NTB90-ment models. The most specific example is the dis-77; UK DOE WU4077.

covery of key nuneral phases, observed to be present in the field, that were not included in current thermo.

Cross, J., A. Haworth, P.C. Lichtner, A.B. MacKen-dynanue databases.

zie, L Moreno, I. Neretnieks, D.K. Nordstrom, D. Read, L. Romero, S.M. Sharland, and C.J.

10.7 ACKNOWLEDGMENTS Tweed (1991). Testing models of redox from mi-I thank R. Levich of the DOE, Las Vegas, for his grati n and geochemistry at the Osamu Utsum long-standmg support for the Pogos de Caldas Project nune and Morro do Ferro analogue sites, Po90s de Caldas, Brazil, SKB TRB21; NAGRA NTBB (and natural analogs in general), as well as his assis-30; UK DOE WROSI.

tance in prepanng this summary. I also thank L.

Kovach of the U.S. Nuclear Regulatory Comnussion Cullen, T.L., and E. Penna Franca (1977). Interna.

(NRC) for her interest and invitation to give this

'i8"al S)mPosiaan on Areas of High Namtal Ra-presentation at the NRC Natural Analog Workshop dioactivity. Rio de Janeiro: Acahmin Brasileira in San Antonio, Texas. Support for this paper and de Ciencias.

presentation was under DOE Contract Eisenbud, M., K. Krauskopr, E. Franca Penna, W.

DEAC0887NV10576.

Lei, R. Balaird, P. I ia==Ma and K. Fujimori (1984). Natural analogues for the transuranic acti-10.8 BIBLIOGRAPHY nide elements: An investigation in Mmas Gerais, Amaral, G., J. Bushee, V. Cordani, K. Kawashita, and Brazil, Env. Geo. Water Sci. 6,1-9.

J. Reynolds (1%7). Potassium-argon ages of alka-Ellert, R., A. Bjornberg, and J. Coutinho (1959).

line rocks from southern Brazil. Geock Cosmoch.

Contribucao A geologica do macico alcalino de Acta 31,117-142.

Pogos de Caldas. Eol. Fac. Fil. Cienc, Letras 237, Bruno, J., J.E. Cross, J. Elkenberg, I.G. McKinely, 5-63.

D. Read, A. Sandino, and P. Sellin (1991). Testing Holmes, D.C., A.E. Pitty, and R. Noy (1991). Geo-of trace element geochemical models in the Po90s morphological and hydrogeological features of the de Caldas analogue study, SKB TR90-20; NA-Pogos de Caldas Caldera and the Osamu Utsunu GRA NTBB29; UK DOE NR90-051.

mine and Morro do Ferro analogue study sites, NUREG/CP-0147 64

The Po90s De Caldas Internatwnal Project Brazil., SKB TR%l4; NAGRA NTB90-23; UK phonolites from the Pogos de Calda alkabne com-DOE WR9%45.

plex, Mmas Gerais, Brazil, SKB TR90-13; NA-l-

MacKenzie, A.B., P. Linsalata, N. Miekeloy, J.K.

GRA NTB90-22; UK DOE WRB044 Osmond, and D.B. Curtis (1991). Natural radionu-Smellie, J., Magno L Barroso, N. Chapmna, I.

clide and stable element studies of rock samples McKinley, and Franca E. Penna (1987).'Ihe Po os from the Osamu Utsumi mine and Morro do Ferro de Caldas Project feasibility study: 1986-7. In l

analogue sites, Pogos de Caldas, Brazil, SKB Natural Analogues in Radwactive Waste Dirposal TR@l6;NAGRA NTB90-25;UKDOE WR@

(ed. B. Come and N. Chapman.), pp. 118-132.

1 047.

Graham and Trotman, London.

4 l

Mackeley, N., O. Coutinho de Jesus, C-L Porto da Smellie, J., N. chaaraan L McKinley, Franca E.

Silvera, P. Emmalata, J.N. Andrews, and J.K. Os.

Penna, and M. Shea (1989). Testing safety assess-mond (1991a). Natural series nuclide and REE ment models using n stural analogues in high natu-geochemistry of waters from the Osamu Utsumi ral-series gmundwaters: The second year of the mine and Morro do Ferro analogue study site, Po90s de Caldas Project. Scientific Basisfor Nu-Pogos de Caldas, Brazil, SKB TR90-17; NAGRA clear Waste Management XH, 863-870. Material NTB9026; UK DOE WR94048 Research Society.

}

Miekeley, N. O. Coutinho de Jesus, C-L Porto da Ulbrich, H. (1984). A petrografia, a estrutura e o Silveim, and C. Degueldre (1991b). Chemical and quimismo de nefelina sienitos do macio alcahno de Pogos de Caldas, MG-SP. Livre Docencia The-l physical charactenzation of suspense particles and ses, Universidade de Sao Paulo.

colloids in waters from the Osamu Utsumi mine l

and Morro do Ferro analogue study sites, Po90s de Waber, N. (1991). Minemlogy, petrology, and geo-chemistry of the Pogos de Caldas analogue study i'

Caldas, Brazil, SKB TR90-18; NAGRA NTB90 sites, Minas Gerais, Brazil: II Morro do Ferro.

27; UK DOE WR90-049.

NTIS, SKB TR90-12; NAGRA NTB90 21; UK Nordstrom, D.K., I. Puigdomenech, and R.H. McNutt DOE WR90-043.

i (1991). Gaeharnical modeling of water-rock in-Waber, N., H.D. Schorscher, A.B. MacKenzie, and teraction at the Osamu Utsumi mine and Morro do T. Peters (1991). Mineralogy, petrology, and geo-i Ferro analogue study sites, Pogos de Caldas, Bra-chemistry of the Pogos de Caldas analogue study zil, S KB TRB23; NAGRA NTB90-32: UK DOE sites, Mmas Gerais, Brazil: I Osamu Utsumi ura-l WR94054.

nium mine, SKB TR90-11; NAGRA NTB90-20; I

Schorscher, H.D., and M.E. Shea (1991). Outline of UK DOE WRWO42.

4 the regional geology, mineralogy, and geochemis-West,J., A.Bialta,andI.G.McKinley(1991). Micro-try of Pogos de Caldas, Mmas Gerals, Brazil, SKB biologicalWs&h UM & W TR90-10; NAGRA NTB90-19; UK DOE WR90-Morro do Ferro analogue study sites, Pogos de Caldas, Brazil, SKB TR%l9; NAGRA NTR90-Shea, M.E. (1991). Isotopic geochemical charac-28 UK DOE WR9M50.

4 terization of selected nepheline syenites and t

i l

4

)

65 NUREG/CP-0147 d

m

- - - -.. - +. - - -. - - - -

l l

NATURAL ANALOGUE STUDIES AS SUPPLEMENTS TO y

BIOMINERALIZATION RESEARCH i

- M.B. McNeil U.S. Nucle *:rReguintory Commissum

. Washingtors, D.C. 20SSS i

11.1 ABSTRACT this paper, and the sense in which certam words will be used. Biommerahzation describes biological reac-Chemical reactions can aher the chemistry and i

tim leading m ee cuadon of inganic peducts, crystal structure of solid objects over archeological which need not be natural minerals, either by precipi-or geological times, while preserymg external physi-tation or by chamral rdacaman' of other materials.

cal shapes. The reactions resulting in these stmemm Biocorrosion will refer to biammaratha' ion reactions

' offer natural analogues to laboratory experimaa*= in m which a solid metallic phase ns oxidized. This paper biommernbrarian and to biologically influenced al-will consider ee snady ot'rnineral altendon pm&icts, i

teration of nuclear waste packages, and thus, they generally pseudomorphs, or corroded artifacts, with

- offer the only available way of validatmg models that the intention of developing an understandmg of the I

purport to predict waste package behavior over ar-macdons &at led to 6eir formation. These reacdons i

chaeological or geological times, will be regarded as natural analogues to the reactions

. Potential uses of such analogues m. the construc-that are expected to take place durmg degradation of L

tion and validation of hypaMic=1 mechanisms of a waste package. The intent of research in this field microbiological corrosion and biommerahzation are is to use these natural analogues to validate models reviewed. Evidence from such analogues suggests for weste package degradation and codes based on that biofilms can control materials alteration in ways these models.

usually overlooked. The newly hypothesized mecha-

[

msms involve control by biofilms of the cation flow A pseudomorph is defined mineralagien!!y as a near the solid surface and offer plausible mechanisms specimen that has been reacted so that the intemal for the formation of mixed-cation minerals under stmeture is altered, but the extemal form is preserved.

1 j-conditions that _would lead to dealloying in abiotic For the study of natural analogues, it is useful to experiments; they also can account for the formation broaden the Minition somewhat. In this paper a of unusual minerals [such as posnjakite, P='da'norph will describe an object (natural or an Cu4SO.(OH)6.H20] and mmeral morphologies un-artifact) whose extemal shape has been preserved, but usual in corrosion [malschite, Cu2COf0H)2, rarely whose intemal crystal structure or composition (usu-i.

fonns botryoidally under corrosion conditions and its ally both) has been changed. An example is a crystal j

occasional presence on archaeological objects that of cuprite, cubic Cu2O, that has altered to monoclinic j

appear to have undergone microbiological corrosion malachite, Cu2CO3(OH)2, while retaining its gross may be related to biofilm phenomena).

cubic shape. The term alteration will be used for all chemical reactions involving a change in the compo-i_

11.2 INTRODUCTION

sition of an object, part of an object, or its dissolution.

4 Corrosion will be used for alteration reactions involv-1 The purpose of this paper is to show how studies xidati n f a metal species oripnally m. a solid j

of alteration products, particulady pseudomorphs and j

of corroded artifacts, can provide data that form a useful supplement to laboratory research data on There are three commonly accepted classes of

, biommeralization processes, and to laboratory re-biomineralization and one subclass for which a new j

search on nuerobiological corrosion of materials of term is introduced. Microbiologically intermediated importance to high-level nuclear waste (HLW) pack-reactions are inorganic extracellular reactions caused 4

ages.

by microbiologically influenced changes in the bulk

- Before any data or arguments are gai, it will electrolyte chemistry. Microbiologically induced re-

. be useful to define the limits of the t. abject matter of actions are those in which the reactions are extracel-i 67 NUREG/CP-0147 i -

i 4

Natural Analogue Skiies as Supplements to Biommeralaatwn Resarrdn 1

lular and inorganic in nature but are controlled by the induced reactions has not been performed, except l

local biofilm chemistry. MicrobiologicaHy control-perhaps,in the case of m-wh bactena.

led reactions are intracellular reactions in which the l

composition, crystal structure, or form of the crystals, 11.3 ALTERATION PRODUCTS AND

}

is controlled by the cell. An example would be the PSEUDOMORPHS i

production of hexagonal prismatic crystals of Minerals may undergo reactions (biologically in-greigite, Fe4 (conjecturally pseudomorphic after fluenced or not) with their environment that result in troilite, FeS) in magnetotactic bacteria; this termmol-changes in chemical composition, generally accom-ogy is adapted from Lowenstam (1989). A subclass panied by changes in crystal structure. A crystal that i

of induced reactions that is not usually considered has been subjected to this process without change in separately includes extracellular reactions in which its gross shape is termed a pseudomorph. In a situ-the microorganisms act to control the morphology of ation where one is deslung with metals this definition the alteration products without affecting the compo.

needs to be bmadened to include crystallme aggre-i l

sition or crystal stmeture. Outside the study of natural gates, since metal single crystals are laboratory curi-analogues this type of reaction is oflittle importance, osities. Pseudomorphs have long been treated as but its appearance in archaeological applications ap.

geological curiosities, and some of the most compre-pears to justify a separate designation, and it will be hensive work was done on them m the last century termed microbiologically structured or shaped altera-(Blum,1843-1873). When one is attempting to ex-tion, though, under some circumstances, the mineral tract information from a natural analogue, presum-can be described as a pseudomorph 'of the microor-ably to validate some model of a mineralization ganism (Watterson,1991). The reason why a mineral process, pseudomorphs are particularly useful, be-4 g

9f g,;g; g

, ggy spectmen pseudomorphic after another mmeral is of such importance is that the specimen itself displays known and because there is a lack of gross physical shape change. The same points are valid for altered the nature of both the reactant and of the product. For scheologicalobjeca example, a specimen of malachite may have formed To extract information on corrosion processes by corrosion of copper or by alteration of some other 4

from a pseudomorph or a corroded artifact, one must i

copper mineral. If the specimen has the form of a first observe the nature and morphology of the corro-l bowl, the starting material was almost certainly a sion/ alteration product, and then combine this with copper alloy (though in a completely mineralized reasonable estimates of the local water chemistry specimen it would be difficult to determine the origi-over the period during which the reactions were tak-i nal composition), but if the specimen is a crystalhne ing place, and of the relevant thermodynamic (and j

mass with no pseudomorphic characteristics, deter-sometimes kinetic) data. One considers various plau-mining the reaction path might be a matter of guess-sible alteration mechanisms and excludes those

}

work. Even ifit were known that the original material which, given known thermodynanuc and kmetic data, j

were a sulfide, the identity of the sulfide might be in could not have produced the observed structure.

doubt, and the reaction path might have been any of Sometimes one observes phenomena, such as Li-

}

several.

esegang pattems (Scott,1985), alternating patterns of Microbiologically influenced corrosion (MIC) different minerals which provide periodically layered i

covers corrosion reactions that involve any of these structure for the corrosion products. Such observa-types of biomineralization. In this paper, discussion tions need to be interpreted in terms of models test-of microbiologically intermediated corrosion will be able under laboratory conditions and offer interestag i

limited; microbiologically intermediated corrosion is Prospects for research.

a very common phenomenon (for example, virtually all sulfiding corrosion outside chemical plants is in.

11A MICROBIOLOGICALLY termdM and much carbonate corrosion as well),

INFLUENCED PRECIPITATION OF 1

and generally of only peripheral interest to biopre-MINERALS cipitation research. Microbiologically controlled re-Information on biomineralization reactions can be actions will not be treated extensively, because the extracted from alteration products by comparing the i

detailed research required to disenmmate them from structure and composition of the products to various NUREG/CP-0147 68

Natural Analogue Studies as Supplements to Biommerali::ation Research conjectured reactions. If pseudomorphs of one min-abstracted in Arts and Archeology Technical Ab-eral after another exist, or other altered structures stracts. Gettens (1%3) and Brown et al. (1977) pre-make clear the identities of the reactant and product, pared extensive reviews; a much more detailed one looks for a (biological or nonbiological) path for bibliography on copper alloys, citing references as far the reaction.

back as Pliny, is contained in Sharkey and Lewin The absence of pseudomorphs of one mineral after (1971). McNeil and Little (1990,1991a,1991b) con-another means nothing; the reaction may be a com-tain some more recent citations.

mon one but may not occur pseudomorphically. The argument that a reaction transforming one rtineral 11.6 MICROBIOLOGICAL into another cannot be effectuated microbiologically CORROSION OF ARTIFACTS can be made from the absence of the second mineral The most important class of bacteria for artifact as an alteration product (pseudomorphic or other-mineralization is SRB (Baas-Becking and Moore, wise) of the first, but this line of reasoning must be 1%1; Hamilton,1985). SRB-induced biominerahza-applied with great care to minerals that are found in tion can affect any metal whose thermochemistry is only a few places. For example, it is unknown for such that the sulfides are reasonably stable with re-reevesite [Ni Fe2(OH)aCO3.4H20] to for:n as an al-gard to the oxides, even iridium. The resistant metals 6

teration product ofjosephimte (FeNi) in nature, but it appear to be gold, titanium (Little 1991), end zirco-as possible to alterjosephinite under a olofilm in such nium, metals whose sulfides are stable in the presence l

a way as to produce reevesite; it just happens that of water only under extraordmary conditions. Alumi-conditions at the one place m the world where there r.am, chromium, and magnesium sulfides are more is still significant uncollected josephinite are not right stable, but reactions between ions of these metals and for such alteration.

hydroxyl ions or dissolved oxygen are so favorable Isotope geochemistry may be usefulin the study energetically that sulfiding corrosion in the biosphere of biomineral pseudomorphs. It is known that the seems relatively unlikely.

action of sulfate reducing bacteria (SRB) is isotopi-Surfaces of minerals and artifacts in the biosphere cally preferennal as far as sulfur is concerned, and it are colonized by bacteria, which produce biofilms.

might be possible to discriminate sulfidmg by an Microorganisms in the biofilm are capable of main-mfinite, well-mixed sulfide reservoir (a situation taining an environment radically different from the which might be typical of intermediated reaction) and bulk in terms of Eh, pH, and concentrations of various sulfiding by a hmited sulfide population in a biofilm, species (Walch,1989; Newman et al.,1991); in con-with a long renewal tirae (which would be more sequence, it is possible to form beneath the biofilm typical of microbiologically mduced corrosion), if various compounds that are not stable in the bulk there were some way of estimating the isotopic rati environment. An example of such biofilm-dominated i

m the bulk reservoir.

mineralization is the conversion of metal to a sulfide I

mineral by the action of biofilms contammg both 11.5 ARTIFACTS SRB and oxygen-consuming bacteria (Pope,1986).

Artifacts have been extensively studied from the Since the biofilms are nonuniform, it is possible for corrosion standpoint. One reason is that archaeolo-the local environment to vary chemically on a sub-gists and museum curators need to understand corro-millimeter scale. This can produce complex minera-sion reactions so es to ensure that these reactions do logical assemblages on a fine scale (Gouda et al.,

not accelerate once a specimen has been excavated, 1993).

and to determine to what extent the consequences of Sulfide biomineralization is probably the most past corrosion reactions can be reversed. There also comprehensively studied subset of biomineraliza-may be interest in whether a particular mineral tion, the one where there has been some study of formed as part of a corrosion or aheration process, or microbiologically controlled minerahzation. SRB whether it was deliberately applied (for example, as have been observed to produce sulfide mineralization a pigment) or was formed as part of the artifact's of Cu, Fe, Ni, Cd, Ag, Pb, and Ir (Baas-Becking and manufacture.

Moore,1961). In some cases they can produce a The identity and structure of corrosion products on sulfide of one metal on a substrate of another, or even artifacts has been the subject of extensive literature, on a silicate (Duncan and Ganiaris,1987).

69 NUREG/CP-0147

, Netwat Analogur Stuna a Sq '=2 no Biomamaliutwn Raarnh In the sulfide-rich reducing environment at the Fe*+ HS ce FeSH*

solid surface, minerals are formed that, while they 2 FeSH% Feb 2H*

Persist for long periods at standard temperature and pressure (STP,273 K, I arm.) because of slow kinet-

I ics, they are not stable in the bulk environment for Fe2 2 comackinawite S

thermodynamic reasons.The fonnation of such com-pounds, includmg covellite (CuS) and chalcocite appear to account for the formation of mackinawite, (Cu2S) on copper and mackinawite (FeSt.x) and FeS t.x (Goldhaber and Kaplan,1975; Taylor,1980).

Berner (1970) showed that, when elemental sulfuris greigite (Fen) on iron by SRB-containing consortia preses, se mackinawite een alten to pyrite (FeS2) is well documented (Baas-Becking and Moore,1%1; wih formnim of any persistem imermediates in McNeil and Little,1990) and can, to a degree, be undentmd in tenns of stabuky diagens (Garrels detectable amounts. However, alteration to pyrrbotite (Fei-xS) by way of greigite (Fe3S4) appears to be and Christ,1965). However, simple application of typical of miembiologically induced corrosion of stabHity diagmns prepared for pure metals to alloys iron (McNeil and Little,1990), and this may be or to pure metals m, groundwaters of complex soru indicative that SRB, in the usual corrosion condi-chemistry can lead to misleading predictions because tions, cannot produce significant elemental sulfur.

of metastability or because of stabilization of phases The details of the reaction process and the role of by impurities. Eight different copper sulfides havg smythite (Fe9S11), are not fully understood and ap-

- been observed in SRB-induced corrosion of Cu-Ni alloys (McNeil et al.,1991); and five iron sulfides are pear to depend on the details of the suite of microor-characteristic of MIC (McNe3 and Little,1990).

ganisms in the biofilm (Bazylinski et al.,1993).

The apparent end product of such biominerahza-tion can be any of several compounds. Thermody-11.7 MICROBIOLOGICAL n.mic.uy, pyrite is the stable phase, and this is CORROSION OFIRON frequently observed in geological situations (Morse

" "I"

);

te 8PPeam top mom comma The SRB-influenced sulfidation of iron displays all three of the recogmzed types of biominernbration, in eng neering c rr sion applications (McNeil and and it was the first process m. which microbiological Little,1990), though there are certainly cases when corrosion was correctly identified (Gaines,1910) and the sulfides oxidize to produce " green mst II," an for which a mechanism was proposed (van Wolzogen oxidized sulfate mmeral (Olowe et al.,1989).

l Kuhr,1934).

11.8 MICROBIOLOGICAL l

Iron and carbon steel are subject to SRB-interme-CORROSION OF COPPER j.

diated reactions during use in connection with sour The situation of SRB-induced minernhrmtion of l

gas and related waters. There is also ample evidence of microbiologically mduced corrosion under copper is even more confused. If one performs labo-biofilms (Baas-Becking and Moore,1%1; Hamilton' ratory experiments on pure copper with SRB, the usual product is chalcocite (Cu2S), though covellite 1985; McNeil and Little,1990). Iron also displays the (CuS), can be produced, and minor amounts of other most studied case of a microbiologically contmiled minesis am wmaimes observed. Cormsion in na-reaction, the simultanems formatie of pyrite and ture produces chalcocite as well as covellite (Baas-greigite in anaerobic magnetotactic bacteria i

(Bazylinski et al.,1990).

Becking and Moore,1%1; Gettens,1963). If alloys are corroded, the situation is much more complex, Consider the biommerahzation of an iron surface and a number of other copper sulfide minerals are under a biofilm containing SRB and an oxygen con-generated (McNeil et al.,1991).

sumer. The original, uncolonized iron surface is gen-A model for SRB-induced corrosion ofcopper and erally coated with a very thin layer of dense copper-nickel alloys that accounts for most of the magnetite, overlaid by more porous outer layers of observations is as follows. The copper surface, at the i

. magnetite and/or hematite, dependmg upon surface beginnmg of the reaction, is covered with Cu20.

treatment. Fe" must be solubih7ed perhaps with the When the surface is colonized by a consortium of assistance oflocal acidification due to occluded cells; SRB and oxygen con = ming bacteria, the cuprite is then the reactions:

destabilized and solubilized in some way. This is in l

NUREG/CP-0147 70 i

- ~.- - - ~ - ~ - - - - -. -

Natural Analogue Studies as Supplemen:s to Biommeralization Reuarch contrast to the nonbiological minerahzation of copper cipitation. Biofilms may also be responsible for un-artifacts by waters contaming chloride, carbonate, or usual physical forms observed in natural corrosion i

sulfate. In these cases, large amounts of cuprite are but not in laboratory experiments, such as botryoidal retained in the corrosion products, while in the case malachite (Cu2(OH)2CO3]. Malachite, a familiar cor-I of SRB-induced corrosion, the cuprite is wholly con-rosion product on copper alloys, is common in geo-l sumed. The first sulfide conosion product is chalco-logical setting in botryoidal form; that is, with a cite, Cu2S, which precipitates on the surfaces of grape-bunch-like structure. This type of structure, microbes and forms a very porous layer. If the common in geological situations, is scarce on archeo-i groundwater contains significant iron, the precipitate logical objects (Gettens,1%3; Zycherman,1982) l may be chalcopyrite (CuFeS4) instead.

and is never seen in laboratory tests. Botryoidal l

In the case of pure copper, the first deposits of rosasite [(Cu,Zn)2(OH)2CO3] is also known in nature copper sulfide corrosion product are extremely po-but it has never been formed in laboratory corrosion rous and nonadherent. If the reaction is allowed to and appears to be unknown in archeological corro-proceed for a few months, more and more chalcocite sion studies. This type of morphology may be due to l

is formed. Only when the chalcocite/ copper layer has control of mass transfer by biofilm effects, i

4 receded some distance from the solid surface are compounds richer in sulfur formed in quantity. The

11.9 CONCLUSION

S final compound formed is generally blue-remaining Biominerahzation reactions are important altera-

]

covellite (CuSio). Evidence from abiotic measure-tion mechanisms for a wide variety of solids. Believ-

~

ments (Woods et al.,1992) suggests that, so long as able projections of such reactions and estimations of copper ions are available, all reduced sulfide is con-rates over periods inaccessible to laboratory experi-l sumed as rapidly as it is produced, and alteration of mentation requires the construction and validation of chalcocite to more sulfide nch minerals can proceed models for very long-term corrosion behavior of only when there is no competing rapid sulfide con-g waste packages. Validation of such models, and the i

summg mechanism at the SRB surface.

codes based on them, is possible only through the The interaction of microbiological action with study of the products of natural and archaeological dealloying is significant. When an alloy corrodes in corrosion and alteration and consideration of the re-a natural environment, sometimes corrosion proceeds action processcs leading to these products as test by the formation of a mixed cation mineral, for ex-cases for model validation.

j ample:

{

(Fe,Ni) co (Fe,Ni)9Ss pentlandite 11.10 ACKNOWLEDGEMENTS i

(Fe,Ni) co Ni6Fe2CO3(OH)16.4H2O reevesite (Cu,Zn) <= (Cu,Zn)2(OH)2CO3 rosasite The author desires to acknowledge the support of i

the Office of Naval Research, Code 1141MB.

On the other hand, the corrosion can proceed by dealloying. For example, in a Cu-Sn alloy, the Cu 11.11 REFERENCES may dissolve and reprecipitate as a Cu mineral, or it l

may be transported away in solution (Geilmann.

Baas-Becking, L.G.M., and D. Moore, " Biogenic j

1956; Robbiola and Fiaud,1992), leaving an artifact Sulfides," Econ Geol, vol 56, pp 259-272 (1%1).

j with the approximate shape of the original bronze Bazylinski, D.A., R. B. Frankel, AJ. Garratt-Reed, artifact but composed largely of tin or tm minerals. In l

and S. Mann,1990, '.Biommeralization of Iron j

different environments, Sn can dealloy from very Sulfide in Magnetotactic Bacteria from Sulfidic similar alloys, leaving a copper mineral structure Envhom, in " Iron Biominerals," ed R.B.

(Tylecote,1979)

Frankel and R.P. Blakemore (in press).

Dealloying is commonly observed and studied in i

laboratory environments,but the formation of mixed.

Bazylinski, D.A., AJ. Garratt-Reed, A. Abedi, and cation minerals as corrosion products is generally R.B. Frankel,1993, " Copper Association with observed only under natural conditions. This may be It n Sulfide Magnetosomes in a Magnetotactic due to the biofilm acting to limit the escape ofions of Bacterium," Archives of Microbiology (in press).

the dealloying species, maintaining high local con-Berner, R.A.,1970," Sedimentary Pyrite Formation,"

centrations favorable to mixed-cation mineral pre-Am. J. Sci., vol 268, pp 1-23.

71 NUREG/CP-0147 4

- m www m

.m

_.7-.,.

Natum! Analogue Studies as Supplements to Biommeralaatwn Research Blum, J.R., ! 843-1873, " Die Pseudomorphosen des Archeometry," Matedals Research Society, Pitts-Minera1reichs," E. Schweitzbart' sche Verlagshan-burgh, PA.

dlung, Stuttgart (5 volumes).

Ibid,1991b, " Corrosion Products and Mechanisms in Brown, B.F., H.C. Burnett, W.T. Chase, M. Good-Long-Term Conosion of Copper Alloys,"in 'The way, J. Kmger, and M. Pourbaix, eds.,1977, "Cor-Scientific Basis for Nuclear Waste Management rosion and Metal Artifacts," National Bureau of XIII," Matetials Research Society, Pittsburgh, PA.

Standards Special Publication 479.

McNeil, M.B., J.M. Jones, and BJ. Little,1991, Duncan, SJ., and H. Gamaris,1987, "Some Sulphide

" Mineralogical Fingerprints for Corrosion Proc-Corrosion Products on Copper Alloys and Lead esses Induced by Sulfide-Reducing Bacteria,"

Alloys from London Waterfront Sites," in "Recent Proceeding of Corrosion 91, Paper #580.

Advances in the Conservation and Analysis of Morse, J.W., F.J. Millero, J.C. Cornwell, and Artifacts," ed. EJ. Black, pp 109-118, Summer D. Rickard,1987, 'The Chemistry of the Hydro-Schools Press, University of London, gen Sulfide and Iron Sulfide in Natural Waters,"

Gaines, R.H.,1910, Ind. J. Eng. Ind. Chem, vol 2, pp Eanh Science Reviews, vol 24, pp 1-42.

128ff.

Newman, R.C., Webster, BJ., and Kelly, R.G.,1991, Garrels, R.M., and J.C. Chdst,1965, " Solutions, The electrochemistry of SRB corrosion and related Minerals, and Equilibria," Freeman Cooper and inorganic phenomena, ISI/ Intemational, vol 31, Co., San Francisco.

pp 201-209.

Geihnann, W.,1956, " Leaching of Bronzes in Sand Olowe, A.A., P. Bauer, J.R. Genin, and J. Guezennec, Deposits," Angewandte Chemie, vol 68, pp 201-1989,"Moessbauer Effect Evidence of the Exist-211.

ence of Green Rust 2 Transient Compound from Gettens, RJ.,1%3, 'The Corrosion Pmducts of Met-Bacterial Corrosion in Marine Sediments," Corro-al Antiquities,"in " Annual Report to the Tmstees sion, vol 45, pp 229-235.

of the Smithsonian Institution," Smithsonian Insti-tud n, ashingon.

Pope, D.H.,1986, "A Study of Microbiologically Influenced Corrosion in Nuclear Power Plants and Goldhaber, M.B., and LR. Kaplan,1975, " Apparent a Practical Guide for Countermeasures," Electric Dissociation Constants of Hydrogen Sulfide in Power Research Institute, Palo Alto, CA.

Chloride Solutions," Manne Chemistry, vol 3, pp 83-104 Robbiola, L., and C. Fr.aud,1992, " Corrosion Struc-Gouda, V.K., I.M. Banat, W.T. Riad, and S. Mansour, tures of Long-Tenn Burial Cu-Sn Alloys--Effect 1993, "Microbiologically Induced Corrosion of of the Selective Dissolution of Copper,"in ' Pro-UNS N04400 in Seawater," Corrosion, vol 49, pp ceedings of the Intemational Symposium on Con-63-73 trol of Copper and Copper Based Alloy x at n," Rouen, to be published in "Memoires Hamilton W.A.,1985," Sulphate Reducing Bacteda and Anaerobic Corrosion," Ann. Rev. Microbiol.,

et Etudes Scientifiques de la Revue de Metallur-i S e.

vol 39, pp 195-217.

Little, BJ.,1991, " Resistance of Titanium to Micro-Sc tt, D.A.,1985, " Periodic Corrosion Phenomena biologically Influenced Corrosion" (abstract),

m Bronze Antiquities," Studies in Conservation, Minerals, Metals, and Materials Society Annual v 130, pp 49-57.

Meeting.

Sharkey, J.B., and S.Z. Lewin,1971. Conditions Lowenstam, H.A.,1989, "On Biominerahzation,"

goveming the formation of atacamite and parata-Oxford University Press, Oxford.

camite. Amer. Miner., vol 56, pp 179-192.

McNeil, M.B., and BJ. Little,1990, "Mackinawite Taylor, P.,1980, "The Stereochemistry of Iron Sul-Formation During Microbial Corrosion," Corro.

fides" Am. Mineral., vol 65, pp 1025-1029.

sion, vol 46, pp 599-600.

Tylecote. R.F.,1979, 'The Effect of Soil Conditious McNeil, M.B., and B J. Little,1991a, " Correlation of on the Long-Term Corrosion of Buried Tin Laboratory Results with Observations on Very Bronzes and Copper," 1. Archaeological Sci.,

Long Term Corrosion ofIron and Copper Alloys,"

vol 6, pp 345-368.

in " Proceedings of the 1990 Matedals Research van Wolzogen Kuhr,1934, "De Grafiteedng van Society Symposium on Materials issues in Gietijzer als electrobiochemisch Proces in Anaer-w onewn m n yp

Natural Arsah"ue Studia as Supplements to Biommerain m obe Gronden (in Dutch)," Water, vol 18, pp 147-Alaskan Placer Gold," Geology, vol 20, pp 315-165.

318.

Walch, M.A.,1989, " Spatial Distribution of Biotic Woods, T.L. A. Amos, and M.B. McNeil,1992, and Abiotic Components in the Biofilm," in "Sulfiding Conosion of Copper Nickel Alloy in

" Structure and Function of Biofilms," W. G.

the Absence of Biofilms," Corrosion '92.

Characklis and P. A. Wilderer, eds., J. Wiley and Zycherman, LA.,1982, 'Techmcal Exemmation of Sons, New York.

Two Owl-Shaped Tsun," Ars Orientalis, vol 13, Watterson, J.R.,1991 "Prelimmary Evidence for the pp 59-91.

Involvement of Budding Bacteria in the Origin of d

73 NUREG/CP-0147

NATURAL GEOCHEMICAL ANALOGUES OF M

))

NEAR FIELD OF HIGH-LEVEL NUCLEAR WASTE REPOSITORIES John A. Apps latorence Berkeleylaboratory Berkeley,Cahfornia 94720 12.1 ABSTRACT and 60 years after burial of the waste (Wang et al.,

United States practice has been to design high-1979). The temperature of the waste would still be 50 level nuclear waste (HLW) geological repositories to 100 *C above the ambient temperature m basalt, s me 1,000 years after bun,al, and the region affected with waste densities sufficiently high that repository temperatures surrounding the waste will exceed by elevated temperature would extend up to 500 m 100 *C and could reach 250 *C. Basalt and devitrified fr m the wase contaims. g empects m se vitroclastic tuff are among the host rocks considered Presently proposed repository m the vadose zone at ucca untam wm, not exceed 180 *C in the host for waste emplacement. Near-field repository ther-mal behavior and chemical alteration in such rocks is a4a entt 6ewasmsome15yean&closm (Tsang and Pruess,1987; Tsang and Pruess,1992).

expected to be similar to that observed in many geothermal systems. Therefore, the predictive mod-In other countries, for example, Switzerland and Ja-eling required for performance assessment studies of Pan, designs usually limit the maximum temperature t MC to ense nmmnal alention of de bem the near fleid could be validated and calibated using t nite clay barrier sunounding the waste containers.

geothermal systems as natural analogues. Examples

- are given which demonstrate the need for refinement Various arguments can be advanced for or against of the thermodynamic databases used in geochemical operatmg a waste repository so that temperatures modeling of near-field natural analogues and the significantly exceed 100 *C after closure. Advan-extent to which present models can predict conditions tages are that the repository is more compact and in geothermal fields.

cheaper to build and that the thermal excursion of the repository host rocks will induce the formation of

12.2 INTRODUCTION

secondary minerals that act as radionuclide hosts in the event of container failure. On the other hand, The disposal of HLW from commercial power fractunng and groundwater convection could be en-reactors m subsurface geologic repositories is gener-hanced in water saturated systems, and physico-ally viewed as the most practical means of dealing chemical processes will occur that are more difficult with such material. The manner in which it is or will

• 9"*" D be emplaced underground varies from country to country. Among the variables are the type of waste, Many questions arise regarding the consequences the duration of aging of the waste prior to emplace-of near-field alteration when repository ambient tem-ment, the packing density of the waste containers and peratures significantly exceed 100 *C, How will the the concentration of radioactive waste in the contain-host rock alter and at what rate? How will alteration

- ers, the nature of the host rock, the u:e of engineered affect the permeability and porosity of the surround-barriers, and whether or not the system will be satu-ing rock? Will radionuclides be' retarded more effec-rated with water. All of these variables affect the tively than they would in bentonite backfills below maximum temperature excursion experienced in and 100 *C? How will the groundwater composition be around the repository after burial. Early practice in modified, and how will this affect the concentration the United States was to design for maximum tem-of radionuclides in solution? What will be the effect peratures in water saturated host rock to range from of groundwaters interacting with the waste contain-about 160 'C in salt to 260 *C in basalt (Raines et al.,

ers, and how will the oxidation state of the ground-1981). These maxima would be attained close to the water influence the secondary mineral composition?

container-host rock interface and occur between 35 Hence, there are arguments for the more conservative 75 NUREG/CP-0147

Natum! Geochemi:al Analogues of the Near Field ofHigh-larlNuclear Waste Repositories philosophy of buddag a repository to operate at The implicit assumption that local (metastable) lower temperatures, where confidence in predicting equilibrium is achieved in the far field aumm=*=Uy l

the chenucal behavior of backfill bamers and the implies that the kmetacs of chemical reactions are fast surroundmg host rock is aah=amd by laboratory ex-in relation to groundwater velocity, which enmmanly perim-an and field observations at earth surface tem-ranges from m% meters to meters per year. In conse-3 geuw.

quence, laboratory experunents lasting, at most, a year can be conducted to measure far-field chemical 12.3 PERFORMANCE ASSESSMENT reactions with a reasonable expectation that (metas-l REQUIREMENTS table) equilibrium will be substantially achieved i

within the duration of the expenment. Modeling far-l Performance assessment studies are conducted field phenomena is, therefore, concerned less with a prior to construction of a repository in order to predict comprehension of the chemical processes than with the fate of the radioactive waste, given all conceiv.

their quantification, and with other uncertamties, able release scenarios. These studies involve the use such as the physical heterogeneities of the ock and 2

of mathematical simulators whose purpose is to pre.

their effect on the hydrologic behavior of groundwa-dict repository behavior for as long as the radioactive ter flow, waste remams hazardous. Sometimes time periods of The near field, in contrast, is spatially constrained, up to 100,000 years must be considered, although and physical heterogeneities, while not necessarily shorter durations are more common. In the United quantifiable in every deta 1, will at least be observable States, regulations require consideration of tune pe-during repository construction. The host rock matrix riods of 50,300,1,000, and 10,000 years following alters and, therefore, comprehension and quantifica-I waste emplacement. The first 50 years is the period tion of radionuclide transport is much more difficult of monitored retrievable storage requued by U.S.

under such conditions, particularly if temperatures Nuclear Regulatory Commission (NRC) regulations exceed 100 *C. Rock alteration takes place at ele-10CFR, Put 60. The waste containers are also re-vated temperatures, and the system is not locally at quired under these regulations to retain their integrity equilibrium, because of the relatively slow reaction i

for at least 300 years and up to 1,000 years after kinetics involved in mineral transformations and the repository closure. The 10,000-year time period is the nonisothermal nature of the near field, which varies length of time that the radioactive waste must be both temporally and spatially.

isolated from the accessible environment as required The chemical processes in the near field are un-by Environmental Protection Agency (EPA) regula-likely to achieve equilibrium in times sufficiently

].

tions 40CFR, Part 191. Mathematical models should, short to permit meaningful results to be obtained from therefore, be constructed to determme near-field sys-laboratory or field experiments. Indeed, even if ex-l tem behavior and release scenarios in conformity periments were to be conducted for as long as 100 with design critena unposed by these regulations.

years, the duration would still be short compared to If radionuclides are released in a water saturated that needed to observe chemical phenomena of con-i repository, they migrate through convective and ad-cem in the near field. For obvious reasons, it is vective groundwater transport and through diffusion.

unpractical to consider such experimente as a basis I

But they are also retarded by chemical interactions, for Predicting repository behavior when decisions l

such as adsorption, ion exchange or coprecipitation regarding the construction of a repository must be with the host rock. Simultaneously, tbc radionuclides made in much less time. Therefore, some other means decay to daughter products, some of which could also of acquiring the needed information must be found.

be hazardous. In the far field, where the host rocks The only practical approach is to observe natural are unaffected by thermal perturbations due to radio-analogues of the waste repository. Natural analogues active decay, radionuclide transport can be modeled must, therefore, play a vital role in elucidating near-assuming that the host rock matrix remams unaltered field processes.

over the time interval during which the radionuclides Models accounting for all relevant chemical and remam hazardous. It may also be assumed that the transport phenomena in the near field have not been system everywhere is locally at (metastable) equilib-developed. Even though partial models describing rium and that the temperature remains constant.

either the geochemistry, heat transfer, or hydrology NUREG/CP-0147 76

i.

i l

Natural C=+%udogues of the Near Field ofHigh-12cel Nudaar Waste Reponnoms are available, natural analogues are limited in their mathematically equivalent free energy mmimintum application until comprehensive simulators are de-approach to obtain a solution. An liiir

-4 feature s

veloped. For performance assessments of the near of these codes is that they can also calm

• whether field to be successful, both model development and the aqueous solution is==w=W with respect to vari-natural analogue studies should be integrated in order ous mmerals. Conveu,ely, if a mmeral assemblage is to meet performance assessment needs.

specified, the aqueous phase camposition can be cablaM In rock dominarM systems, such as the 12.4 STATUS OF GEOCHEMICAL reposito;y near-neld environment, this capability is MODELS parncularly useful, as it allows the calculation of the Complex systems comprising many chemical maximum possible concentration of radionuclides in

~

the groundwater. Examples of such codes are EQ3 components and many phases at elevated tempera-i tures require computational methods for their inter-(Wolery, 1983,1992), PHREEQE (Parkhurst et al.,

1980), SOLMINEQ.88 (Kharaka et al.,1988), and pretation and quantification. Sophisticated 4

SOLVEQ (Spycher and Reed,1989). All permit cal-geochemical models and suitable codmg with sup-culation of solution equilibria at temperatures to 300

)

porting thermodynamic and kinetic data are needed r 350 *C along with the water saturation surface.

to interpret natural analogue behavior for perform.

Other distribution of species codes are available that i

ance assessment. Relevant computer codes to solve geochemical models currently fall into four classifi, incorporate adsorption and ion exchange reactions, such as MINEQL (Westall et al.,1976; James and r

cations: (i) calculation of phase equilibria, (ii) calcu-4 Parks,1976), MINTEQ (Felmy et al.,1984; K1upka i

lation of the distribution of species in the aqueous and Morrey,1985), and ECHEM (Morrey,1988).

phase, (iii) reaction progress simulators, and (iv) re.

They might find limited application to near-field active chemical transport simulators.

sorption phenomena, but are better equipped to deal l

12.4.1 Calculation of Phase Equilibria with far-field chemical eqmlibria at 25 C. In Russia, the Gibbs free energy mimmintion approach is pre-Codes to calculate phase equilibria, usually m. the ferred. Example codes include GIBBS (Shvarov, a

form of meshes of univariant curves in P-T space, are 1976; 1992) and SELECTOR ++ (Karpov,1981; Kar-useful in defining the stability fields of secondary pov et al.,1992). See also Mironenko et al. (1992).

minerals in hydrothermal systems of the type likely 1

I to be encountered in the near field. Other features 12.4.3 Reaction Progress and Reaction include the ability to plot phase diagrams as a func-1 kne6cs i

tion of P, T, and X, where X is some independently variable chemical component in the system. An ex*

Reaction progress codes simulate the evolution of ample of such a code is GEO-CALC (Perkins et al.,

a chemical system as a function of reaction progress, 1986; Berman et al.,1987; Brown et al.,1989). An-4, or as a function of time. The reaction progress other, SUPCRT92 (Johnson et al.,1992), permits variable is usually chosen instead of time, because so calculations of the position of univariant curves in little is known regarding the rates of dissolution or P-T space, as well as standard-state thermodynamic precipitation of mineral phases. In using reaction i

properties of phases as a function of pressure and progress codes, it is assumed that the reactants, such temperature. Other computer codes of a similar type as the rock-forming minerals, dissolve according to a j

have been reported in the literature, for example, preset ratio, usually in proportion to their molar con-Powell and Holland (1988).

centrations in the rock. Homogeneous equilibrium in 3

the aqueous phase and reversible equilibrium with 12.4.2 Distribution of Species respect to product (i.e., secondary) minerals and the l

Many computer codes are available to perform the aqueous phase are assumed. With increasing knowl-task of distributing species in solution using mass edge regarding the dissolution kinetics of minerals, action and mass balance laws. Such codes are useful the reaction progress variable may be disregarded in j

in interpreting groundwater analyses in relation to the favor of time as a variable. Indeed, this option has

]

thermodynamic parameters affecting their composi-been included in one reaction progress code, EQ3/6, I

tion. Most codes in use in the United States employ by Bourcier (1985). The principal reaction progress the eqmlibrium constant approach as opposed to the codes available at present are EQ6 (Wolery,1984; 4

77 NUREG/CP-0147 a-w m

_u w

w y-en m,

- ~ - - -.

-. ~. -. -.~---.- - - - -..

1 Natural Geochamcal Analogues of the Narr Field ofHigh-level Nuciaer Wate Jtepceitories Wolery and Daveler,1992), PHREEQE (Parkhurst et that are relevant to both the repository and j '

al.,1980), and CHU RR (Reed and Spycher,1989).

natural analogue environments.

• The natural analogue should be suffriently j

12.4.4 Reactive ChemicalTransport well characterized to pernut semi-quantitative, i

Many reactive chemical transport codes are cur-if not quandtative, estunates of chemical reac-i rently under development, and successful simula.

two mtes.

)

tions reflecting various natural geochemical

• The magnitude of the natural analogue pmcess processes have been reported in the literature. Unfor-and its lifetime should approxunate that of a j

tunately, complex simulations reported so far are waste repository.

isothermal and have been conducted only at 25 *C.

• Sufficient analogues should be available to per-1 While satisfactory for far-field predictions, such mit multiple validations over a range of parame-l modeling is of linuted value for the near field. Non-ter values of significance to the repository 1

isothermal simulations at elevated tempera:ures have near-field environment.

j-been limited to relatively simple systems involving Fmding a perfect analogue to a nuclear waste few chemical components, for example, CHMTRNS repositoryis probably impossible. Most will correlate j

(Noorishad et al.,1987) and THCC (Carnahan, with only some of the design criteria for the reposi-1986). To properly model the near field, significant tory under consideration. For example, although the j

advances in the present capabilities of reactive trans-Oklo reactor in Gabon meets criteria for studying the

{

Port simulators will be required.

fate of extinct radionuclides, the host rocks might 12.5 GEOTHERMAL SYSTEMS AS differ from those of the planned repository. Because the Oklo reactor is extinct, it is not possible to exam-NATURAL ANALOGUES OF THE ine the radionuclide transport mechanisms or the NEAR FIELD solubility of those radionuclides under natural conds-tions at elevated temperatures, in contrast, an active l

12.5.1 Selection of Natural Analogues of hydrothermal system could meet many of the criteria tite Near Field for host rock secondary alteration, but would not 4

For nntural analogues to be relevant and useful to contain many relevant radioelements. Natural ana-perfonnance assessment, several criteria should be I gues of the near field, therefore, are never complete t

met:

m every respect, and will necessarily provide only va adon to Win models. Sam %

• Except fors'=ri=Hved partial validation needs, the analogue should be active, rather than ex-I **".y na anal gues, however, should permit va ati n m a suhiently bad nage to gin tinct, that is, the chemical processes to be char-j acterized should be going on at present.

c nfidence m the predictive capabilities of the mod-

• The temperature should be in the range between 2h m *C Various rock types have been eramineA over the

]

• Water must be present, saturating or partially years as potential hosts for geologic repositories.

l saturatmg the rock and be the primary agent Theyinclude granite (Canada, France, Sweden, Swit-involved in mass transport.

zerland, United Kingdom, United States), clay (Bel-

• The groundwater composition should be con-gium), basalt and vitroclastic tuff (United States),

shale (Japan), and bedded salt (USA). Because of the trolled by the host rock mineral assemblages.

interest shown by the United States Government in j

• The chemical environment of the natural ana-the burial of HLW in basalt and vitroclastic tuff, the logue should be similar to that of the repository, use of natural analogues in such host rocks is empha-j that is, the host rocks should be of a similar type sizedin the remainder of this paper.

j and chemical composition.

Basalts and tuffs are extrusive igneous rocks that

• The natural analogue must be quantifiabic in are frequently associated with volcanirany-driven terms of the parameters used in the predictive hydrothermal systems. Active hydrothermal systems

[

models for performance assessment.

are of econonne interest in many parts of the worid

• The performance assessment simulators must as they can be exploited for geothermal energy. Such mcorporate thermodynamic and kmenc models systems are often drilled to recover steun or heated NUREG/CP-0147 78

Natural Geochemical Analogues of the Near Field ofHigh-lael Nuclear Weste Repositories water, and consequently, both their geology and geo-liquid water, resulting in the formation of palagonite chemistry are sometimes characterized in consider-(Eggleton and Keller,1982). Figures 12-1 and 12-2 able detail These geothermal systems persist for taken from Apps (1987) illustrate schematically the

. hundreds and even thousands of years, because their sequence of steps believed to occur in rhyolitic and heat sarce is usually an underlying magma body of basaltic glasses. It should be noted in passing that the substantial size. Heat transfer is accomplished residual glasses in massive basaltic flows, such as are through thermal conduction and convection of water observed in Iceland or the Columbia River basin, can in the rock fractures. Such mechanisms are also ex-approach rhyolite in composition as a result of crystal pected in HLW repositories, because heat transfer in differentiation.

gcothermal systems is somewhat similar to the per-Natural silicate glasses contain between 45 and 75 sistent heat production from decaying long-lived ra-wt. percent SiOzIf they were to dissolve and achieve dionuclides. The temperature range of most equilibrium with the groundwater, they would super-geothermal systems also falls within the design range sammte the soluh wie mspect M p d of planned waste repositones m the United States, cristobalite. Figure l2-3, modified after Apps (1970),

although many geothermal systems are recorded with hstrates the solubility relationships between vari-temperature maxima exceeding those contemplated us sihca polymorphs. If the glass were to consist in repositories. The composition of geothermal wa-nly f ska, the Inineral cnstobahte would precipi-ters is also determined mamly by miner:1 alteration processes; that is, the system is " rock dominated" tate in preference to quartz because it is HaaticaHy rather than " water dominated." Similar groundwater favored. It is generally tnie that the least stable min-compositional control is expected in waste reposito-eral forms in preference to more stable phases when ries. Although comparisons have not been made, the the solution is supersaturated with respect to all of circulating volume of water in a convecting by.

them. Ostwald (1897), using thermodynanuc argu-drothermal system is probably much larger in scale ments, reasoned that an unstable system would ap-than that in a nuclear waste repository. Therefore, the Proach equilibrium through a sequenee of closest analogues to a potential waste repository are Progressively lower energy states. Although a more probably not of commercial interest and have, there-meaningful approach to explain the observed se-fore, not been e=niaad closely. This divergence is quences, which are essentially kinetic in nature, one area that will require particular attention in mod-would be through the application of classical nuclea-eling studies, awvotme 12.5.2 Alteration Mechartisms a 1> -

"^**

@#!!"u.

l l

!C.1 Volcanic rocks often contain natural glass. The g,,,,,;,"4,,,,,,;

j ("00'cl y,,,,,,,

glass is unstable in geothermal systems and tends to i

Q,25,7 l

l decompose into suites of secondary minerals. Natural wr go t,so

{

glasses decompose in a complex manner that is not 6;,"4"l*"

u-o wt* H,oi g

l fully understood. Observations reported in the pub-

{

osvrrnmeo lished literature show that the alteration mechanisms

(,j',kfd,,;

%,%ss t

q, l

are strongly influenced by compositional variations j

in the glass, which results in at least two distinct gg,Ne

,,,j,,,,,,

y structural types (Apps,1987). Rhyolitic glasses form "i,',%"l,*"

G' Ass an open " stuffed tridymite" type structure (Taylor and

• i"""'l w '*a l

Brown.1978) which can easily hydrate and exchange

• • • j'****

l cations without destruction of the alummosilicate

- ;,[,- g-- 3

,,co,,,,,

vitreous state for tens of millions of years (Forsman,

=YE*'i

., f"88W6W@Hy,'8'7",pNERAL ASSEMBLAGE l

framework. Such glasses sometimes remain in the M ERALs i

Mi 1984). In contrast, glasses of basaltic composition do Mgure 12-1. Schematic diagram to show the not appear to retain their vitreous aluminosilicate decomposition paths of rhyolitic glass when framework upon hydration, but decompose directly exposed to the aquens phase l.

into a proto-smectite structure upon contact with i

79 NUREG/CP-0147

.n

-n.

..-m-,.

.n, w

Nahmd Geodemaal Arulogues of the Naar Field ofHigMnel Nuclarr Waste Repositories tion theory. No satisfactory explanation using the "Egc isu., s.i.n.i latter approach has been derived so far.

croe,o,,n, In hydsi-+ mHy altering vitric volcanic rocks, p*,7;"*f,T*y the socalled "Ostwald Rule of Stages" is amply demonstrated, a sequence of metastable~ phases form-

.y GG;l g,q ing that gradually decompose to more stable assem-uicn g cTune blages over time (Dibble and Tiller,1981). With on s................

mereasing temperatures in geothermal systems, the

1. E*An [...._.. [Ed.dEIS', o.asin, o,71/sOoo rates of dissolution and precipitation are anhan=L I

u,,.gon and, consequently, the glass and early formed least e

stable phases are rapidly destroyed, leaving.nore i

no.W,Qoi stable phase assemblages contaming quartz. T1.e sec-

.t ondary mineral assemblages observed in altering vol-o,,,i,,,,,,n, I

canics are, therefore, metastable assemblages that are

,;,,, o,,,osurion, or.a,piranon sECONDAnY z.ow...un.ts..

energy states represented by more stable mineral t-----

uinenA cm assemblages.

l Dibble and Tiller (1981) have described the con-

.........}..........

ceptual basis for such alteration as illustrated in Fig.

Most sTAste ute 12-4. Here, a more soluble phase, r, is dissolving.

AsEs"tice It releases its chemical components to the aqueous 5--~

phase through a diffusion boundary of thickness, Sr..

At some finite distance, x, from the reactant phase, a Figure 12-2. Schematic diagram to show the product phase, p, a.s precipitating. Phase, p, is also decomposition paths of basaltic glass when surrounded by a diffusion boundary,6,p. Transport

1. 1 Posed k the aqueous phase between the diffusion boundaries shielding the reac-tant and product phases is accomplished in geother-mal systems by aqueous diffusion and convective s C"*.*:::..::i;"'?

flow.

,*.. ?".;;"1

<>. R1 m n,,, x',,

o

.-......v,

'"'~o"'

, # "' ~ "* ! o' v

=

6-..

.n.m

-..a in <,,...........

Amorenous..hea.

g A

..RT in O'

--*8'**---- i f

3

- Ri in O, g

30- Cnsionem,

Quan, 8

}

{a.----$

=.Rf in QD - A9 g

40 1

4............Rf M K0gp 8,

t i

e i

36 30 25 20 is n--*

d 3

(K a 10 )

Figure 12 3. Solubility products of silica Figure 12-4. Schematic diagram to show polymorphs as a function of temperature (Apps' thermodynamic and mass transfer 1970) considerations dudag irreversible dissolution and precipitation (Dibble and Tiller,1981)

NUREG/CP-0147 80

Natural Geochemscal Anakgues of the Natr Field ofHigh-Ine! Nuclear Waste Repositories The affinity, A, of a chemical reaction is defined canics could provide a wealth ofinformation regard-j ing the rates of crystal growth. Together with kinetic as A = RTin K/Q models of glass alteration and grain coarsening (Ost-wald ripening), such information could allow scoping where Q is the ion activity product in the aqueous studies to be made of the actual rates of alteration in phase, which is negative for reactants and positive for geothermal systems, and could predict near-field rock i

products. Hence:

altention in a repository.

^

In basalts, the mineral zonation observed as a j

A' = RTin KyQP function of temperature is very consistent, although l

A* = Rt in KylqP minor variations can be attributed to compositional Field evidence suggests that groundwaters in frac-variations in the basalt (Walker,1960 a,b), or if tures and pores are relatively close to equilibrium seawater rather than meteoric water is present. Fig-with respect to secondary (metastable) minerals ure 12-5 shows the approximate zonation associated j

P with alteration in Icelandic basalts summanzed from forming in a geothermal system, that is, A is small.

several investigations (Apps,1983). Surprisingly, This conclusion can be inferred from the uniform.

very similar zonations are observed, often with iden-ity of groundwater compositions and mineral assem.

tical secondary mineral assemblages, although in dif-blages, as noted earlier, tl.e commonly observed close ferent proportions, in volcanic rocks with approach to equilibrium with respect to quartz above compositions ranging through andesitic to dacitic, 200 *C and the general level of crystal perfection in and even rhyolitic. Similar zonation is even observed many zeolitic assemblages growing in host rock voids. Such perfection is unlikely at high degrees of in hydrothermal systems occupying sedimentary ter-rains, such as those at the Salton Sea, Califomia, and supersaturation. As a first approximation,it may be Cerro Prieto, Baja, California.

assumeo that groundwater in geothermal systems reflects a quasi-equilibrium state with reapect to co-The mineralogical uniformity suggests strongly existing secondary mineral assemblages, provided that the composition of the groundwater coexisting that the sampling point is not in a region of rapid with altering volcanic rock might also show a similar convection, or that the system has not been exploited consistency. This indeed proves to be the case for extensively for geothermal energy. Because the ac-basaltic terrains, as is illustrated in Figure 12-6. In this curacy of thermodynamic data of secondary minerals figure, the calculated log activity ratios of major in geothermal systems is commonly less than that chemical components are plotted as a function of necessary to resolve A, the assumption that secon-temperature from data by Arnorsson et al. (1983) of P

dary minerals in geothermal systems are in equilib-waters taken from wells from several different geo-rium with the groundwater is a reasonable thennal fields in Iceland. The values were calculated approximation at this time.

using the EQ3 code (Wolery,1983). The tempera-At relatively low temperatures, distinctive reaction ture-dependent trends of all illustrated component j

rims penetrating vitric volcanic rock matrix blocks activity ratios are quite systematic, with a scatter of are sometimes observed, suggesting that the altera-about il-1.5 log unit. Most of the scatter is probably tion rate of the rock matrix is the rate controlling step.

due to errors in the estimated pH,(-log (H']), as other But in most geothermal fields, the host rock is perva-element activity ratios not incorporating [H+] show sively altered. Under these conditions, it is more smaller deviations, for example, log [Na']/[K+], and likely that the secondary matrix minerals, rather than log [SiO2(aq)], illustrated in Figure 12-7. Groundwa-the residual glass, are the primary source of reactants ter compositions from other geothermal systems in in the aqueous phase. Because many of these minerals rocks varying from basic to acidic also display similar are fine-gmined phyllosilicates, the excess surface temperature-dependent ion activity ratios. Consistent free energy might provide the affinity, A', for the temperature-dependent trends in the fugacities of crystallization of secondary minerals. Further study various gases from geothermal fields are also ob-of the reaction affimties of altering rocks in geother-served as illustrated by D' Amore and Gianelli(1984).

mal systems is needed to resolve these issues Figure 12-8, adapted from their paper, shows a sym-The isotopic dating of zeolitic assemblages and pathetic variation in the calculated fugacity of oxygen 6"O and 3D measurements in altering vitric vol-and sulfur from several geothermal fields, these 81 NUREG/CP-0147

Netml Geodemcal Analogues of the Near Field ofHigWLeoni Nuciner Wanne Rapositories i

I I

I I

I I

I I

I I

I chabazite I

- - - tV.YM-- - -

...t.h.om.sonii.

gism,qndine i

mesolite/scolecite

~~

......... he,uja t}d,it,e,,,,,,,,,

.....s.1&!12..............

.......teit1M t.......

.........T9t4nWs...............

phillipsite laumontite

~~

......................F.aje,!Tg S----------------

- - ~

' smectile (nontronite) l e

mixed smectite / chlorite j

~~"~

chlorite F

calcite 0

chalcedony b

.. 2............... py a,rt,z,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,

.......tP.idote..

......Rfh".h.8........

actinolite

...................a.lb.it.e............

wolla.ste. nite pyrite l-

.........................P.Y.".h,oge,,,,,,,,,,,,,,,,,,,,,,,,,,,,,

I 1

l l

I I

I I

l l

l 1

0 25 50 75 100 125 150 175 200 225 250 275 300 Temperature (*C)

Figure 12-5. Observed stability ranges of secondary minerals in hydrothermally altered IMannlic basalts l

fugacities beang based on observed H2, CE, CO2, and 12.5.3 Application of Geochemical H2S concentrations in the steam phase from produc-Models ing wells.

Gush-mal systems can be used at several levels

'Ibe remarkable conclations in secondary miner.

for the development of geochemical models of near-alogy and groundwater and vapor phase chemistry field behavior.They can be used:

between gwti--al fields strongly suggest that a for the description of the nuneralogy and distri-e bution of secondary alteration at the observa-geochemical model desen.bm.g the hyu>mummally

_m tionallevel altering volcanic rocks could not only be validated e for the purpose of calibrating equilibrium rock-

. over a wide range of volcanic rock types and tempera-watermodels tures, but could also be used in predicting the t!wrno-e fortheformulationandcalibrationofnonequili-dynamic behavior of rock alteration arounM1W in brium kmenc models obeying the Ostwald Rule similar host rocks.

of Stages NUREG/CP-0147 82

Natural Geochenucal Analogues of the Nar Field ofHigh-lael Nuclear Waste Repositories e

e e,a e a

e *n On C

• e, a

g g

6 5

E 3,

o e

e e

a a

e s

So 100 160 200 260 50 100 160 300 260 Temperature (T)

Temperature (*C) w n

a sh 4

es* 'N'

,u

. of e-no e

n ng.

e

{

E a

V e.

8 7

5 e

A m

+

2 r=

s.

1

[

o e

u a

e e

e, e

so 100 150 300 eso 50 100 150 20c 260 Temperature (T)

Temperature (T)

-=

s

• n f

=

=

h r

q~

s...

7 sn g

s --

=.

. e y

f a

a g

%e

,s,e a

=

es e

.n j

I l

4

=

-u

-u 60 100 18 0 300 aso 50 10 0 150 200 aso Temperature ('C)

Temperature (T)

Figure 124. Activity ratios of major elements as a function of temperature in well water from Icelandic geothermal wells (calculated from chemical analyses cited in Arnorsson et al.,1983) 83 NUREG/CP-0147 i

Natural Gawhament Analogues of the Naur Field ofHigh-Inel Nuciarr Waste Repositories s.s a

m m

=

t y

e-,

2.5

)"

=

W E

?

a 7

E M

a e

i..

... u...

o I

5 g

g m

a a

,I e n l

i i

i 50 100 150 200 250 50 100 150 200 250 Temperature (*C)

Temperature (*C)

Figure 12-7. Activity ratio of[Na1/[K1 and saturation index of quartz as a function of temperature in well waters from Icelandic geothermal wells (calculated from chemical analyses cited in Arnorsson et al.,1983) for the repeated partial validation of reactive cems the behavior of the long-lived hazardous ra-e geochemical transport models simulating near-dionuclides in such an environment. These radionu-field behavior.

clides include actinides, such as

Am, 237Np, 2

'I, h angfission products, such as '"Cs In order to make the modeling efforts tractable, the Se, h, and C. Very little of a quan-development work should be broken down into the Sn, steps enumerated above. In the following sections of titative nature can be derived from the geothermal this paper, the author discusses his experience in anal gues regarding the transpon of the transuranic addressing the second level of modeling, that is, the actinides, as their presence in nature, even under procedures adopted in analyzing and mterpreting favorable conditions, is at vanishingly small concen-samples of geothermal fluids taken from geothermal trations. The chemical behavior of fission radionu-wells, as well as that of developing models of geo.

clide elements can be investigated, however, even thermal waters in equilibrium with secondary mineral though the concentrations in natural systems might assemblages, differ significantly from those in a repository. Of particular interest is variation of the concentration of The principal uncertainty with the application of the radioelement in solution as a function of tempera-such model development to a waste repository in volcanic rocks is the degree of alteration. The cumu-ture and its concentration in coexisting secondary minerals in a geothermal system. Such information, lative quantity of heat released by the waste is rela-in conjunction with geochemical models, could be tively small compared with geothermal fields having used to validate partially a geochemical transport commercial potential. It would be desirable to study model concerning those radioelements.

small geothermal fields with transient heat produc-tion, but, as noted earlier, such fields are of limited Information that is incidental to the mobilization economic interest and are, therefore, unlikely to be of elements can also be derived by studying the loss studied in sufficient detail. Another problem con-of coherence between immobile and potentially mo-NUREG/CP-0147 84

i Natural Geochamcal Analogues of the Narr Field ofHigh-Isoel Nucianr Waste Repossiones 30 80-y The Geysers e

A Mount Amiata A

e Trawe 60-e,

• Larderello ee ee*

i o Cerro Prieto

)

{ 4o-i o

~

+

0 35 e

e e

s i

e i

0 50 100 150 200 250 300 o

9 Zr (ppm) 1 (a) g W

[

Predicted Range at h-the Geysers. Based i

on Host Rock l

-40 ee secondary Mineralogy (AppA1991, 20-e

~

m 16-e e

e

~

- gg.

e e e

\\

5 e..

e e e

ne e

, w w es e

n 0

0 50 100 150 200 250 300 17 16 15 14 13 12 11 10 9

8 Zr (pom)

Logi$ ID*'I 2

i (b) l Figure 12-8. Variation in log f02 versus fS2, calculated from volatile and noncondensible gas Figure 12-9. Dispersion of yttrium and concentrations in stream from various rubidium in relation to zirconium in Icelandic geothermal fields (D'Amore and Gianelli,1984) basalts (Wood et al.,1976) bile elements in volcanic rocks during hydrothermal 12.6 VALIDATION OF AQUEOUS alteration. This is illustrated in Figures 12-9a and b, GEOCHEMICAL MODELS THROUGH WELL WATER ANALYSIS FROM from Wood et al. (1976), which show the relationship ACTIVE GEOTHERMAL SYSTEMS between yttrium and zirconium and rubidium and zirconium in altered Icelandic basalts. Zirconium is 12.6.1 Background extremely insoluble m. geothermal waters, and is, therefore, unlikely to be mobilized in a volcanic rock Geochemical codes necessarily require extensive when altered. Yttrium, a trivalent rare earth element, thermodynamic data, AHi.29s, S298, C (T) for exam-P c, and V* for minerals and the corresponding par-l is likewise imhkely to be mobilized and will remain tial m lal ropenies of aqueous species in o& to P

in aluminosilicate structures. In contrast, rubidium, calculate thermodynamic relations between minerals an alkali metal, is easily solubilized and d.ispersed and the coexisting aqueous phase. Creation of data-during alteration. This is reflected in the substantial bases of the needed scope for such modeling is costly scatter in the correlation plot illustrated in Figure in time and effort. Therefore, existing critically evalu-12-9. Similar comparisons could be made of the ated, internally consistent thermodynamic compila-dispersion of other radioelements versus zirconium.

tions, for example, those by Helgeson et al. (1978),

85 NUREG/CP-0147

Natural Geochemical Analogues of the Near Field of High-Inci Nucim Waste Repositories Robie et al. (1979), Berman (1988), Holland and tion. Questions concerning the damage to the crystal Powell (1990) for minerals, and those by Tanger and stmeture during sample preparation, incongruent dis-Helgeson (1988), Shock and Helgeson (1988) and solution, metastability, nonattainment of reversible Shock et al. (1989) for aqueous species are used, equilibrium, or the extreme duration of experiments together with miscellaneous data of dubious heritage, required to demonstrate reversible equilibrium, the to fill omissions in the primary sources Most of the characterization of microscopic amounts of secon-mineral thermodynamic data in the above cited refer-dary product phases, problems in measuring pH, and ences have been compiled from a systematic analysis difficulties associated with sample collection, preser-of high-temperature, for example, 400 C, phase vation and analysis all degrade the accuracy of the equilibria to define univariant curves in P-T space, resulting data. Problems of data reduction due to the coupled with a smaller data set obtained from low-presence of ill-defined complexation, uncertainties and high-temperature calorimetric studies. The re-regarding the true pH, or correction for activity coef-sulting data are represented in the standard state at ficients in strong mixed electrolytes also increase the 298.15 K and I atm or 1 bar pressure, which neces-difficulties of satisfactory interpretation of experi-sitates significant extrapolations from the tempera-mental results and render such investigations a chal-tures and pressures at which the phase equihbrium lenge to the most dedicated investigator.

measurements were made. To be assured that such Furthermore, the systematic study of mineral solu-extrapolations are meaningful, accurate corrections bilities at elevated temperatures is expensive. There-for C; and V must be made, and the participating fore, it is unlikely that resources will be available in phases must be adequately characterized, for exam-the near future to conduct the experiments needed to ple, order-disorder phenomena, the existence of dis-validate the required thermodynamic databases. An placive phase transitions, variations in chemical interim solution to this dilemma is to use natural composition, crystallographic properties, crystallite analogues as a means of validating and/or calibrating size, evidence for metastable equilibrium or nonat-the thermodynamic database.

tainment of equilibrium, etc., in addition to such frequently observed details as the uncertainties in Calculations of mineral solubilities from chemical temperature and pressure. Unfortunately, many of analyses of geothermal well waters would initially these important factors were not reported in the ear-appear to be an even more daunting task than that of lier literature, from which most of the current thermo-acquiring thermodynamic data by other means, dynamic data have been derived. The data are, Groundwaters might not be in equilibrium with co-therefore, of questionable accuracy for the evaluation existing minerals, mixing and convection or advec-of natural waten, over the temperature range of 0-tion could be taking place, and the collection of 300 *C, the range of interest for perfortnance assess.

samples could be difficult and involve contanunation ment studies.

or the loss or gain of volatiles and gases. The cooling of the sample during and after collection might cause In contrast to the abundance of published phase transient precipitation or adsorption of chemical equilibrium studies at elevated temperatures, studies components before chemical analysis. Yet there are reporting aqueous solubility measurements of miner-also significant advantages. The well water has often als are sparse in the entical 0--300 C range, and they been in contact with the host rock for hundreds or are confined mainly to salts, such as carbonates, even thou sands of years, allowing for closer approach sulfates, and phosphates. With the exception of to (metastable) thermodynamic equilibrium than is quartz and its polymorphs, few solubility studies are achievable in the laboratory. Groundwater samples available for any aluminosilicates commonly found can be collected in relatively larfe sample volumes, in geothermal systems. In principle, it should be and their transit from the host rock environment to possible to measure mineral equilibria in aqueous the surface is short. if collected from a producing solutions quite precisely in the laboratory, as tem-well. As a consequence, groundwaters are often pcrature and pressure can be closely controlled. and closer to equilibrium with respect to secondary min-the chemical components in solution can normally be erals than is achievable in solubility studies in the analyzed with an accuracy ofless than 5 wt. percem.

laboratory, and, if the problems inherent in the col-Unfortunately, many experimental uncertainties lection and handhng of natural samples can be over-make successful detemunations a difficult proposi-come, groundwaters could prove to be better o n ---

--m m a em op

Natural Geochemical Analogues of the Narr Field ofHigh-Level Nuclear Waste Repositorus indicators of thermodynamic equilibrium than labo-perature at which the pH was measured and checking ratory experiments.

for consistency between charge balance, pH, alkalm.

Recognition that the groundwater can be close to ity, pCO2, and the gross chemistry of the sample. If equilibrium with respect to certain minerals is not the conductivity of the sample had been measured,it new, (Garrels and Christ,1965; Browne and Ellis, should be compared with that calculated,by the 1970; Amorsson et al.,1983; Gunlaugsson and Ar.

method of A.P.H.A. (1985). Sometimes it was found norsson,1982; Senderov,1980; Aargaard and Hel.

necessary to compare the chemical analysis with geson,1983), but studies that demonstrated this were those obtained in neighboring wells to identify dis-generally conducted with objectives limited to illus.

crepancies in individual components. Where no ob-tration or the independent prediction of the thermo-vious source of error could be identified, balancing dynamic properties of isolated minerals, rather than on CI' was preferred to achieve electrical neutrality, for the comprehensive validation of a complex geo-as this component does not significantly affect the chemical model. An exception is the already cited calculated ion activity products of minerals. These work of Arnorsson et al. (1983), who not only devel initial comparisons were usually successful in identi-oped their own thermodynamic database and model fying gross errors and the chemical component (s)

(Arnorsson et al.,1982), but also analyzed a large requiring adjustment to achieve electrical neutrality.

number of groundwaters from geothermal wells and With high-quality chemical analyses, only minor cor-hot springs.

rection(s) to one or two species of a magnitude ofless In order to test the above approach, the author than 5 wt. percent, were regmred, which is well within conducted such a study using information in the die rnagnitude of uncertainties due to other causes.

published literature. Chemical analyses of ground-The in situ groundwater composition was then recon-waters covering a range of host rocks and tempera-stituted at die measured down-hole temperature, ad-tures were assembled. Wherever possible, justments being made for the loss of volatiles and mineralogical descriptions of the host rocks were also gases and the amount of steam flashed from the compiled. The quality of mineralogical information sample, when such had occurred prior to sampling, varied greatly. In several cases, mineralogical data electrical neutrality being maintained in all these conections.

were obtained from sources spatially separated from the formation from which the water sample was drawn, but where the temperature profile and rock 12.6.3 Cotnparison of Calculated type was similar. The thermodynamic database used Saturation Indices with Mineralogy for the study was compiled by Berman (1988), aug-mented with data from Helgeson et al. (1978) and the As an n.. ial test of the model, den.ved analyses of ut calculated thermodynamic data for clays from Wol-Producing wells forIcelandic basalts by Arnorsson et ery (1978), corrected using a more generally accepted a

3) wem evaluad Calculated saturatix u16 Berman (1988) value of AHr.29s for corundum. Solu-ce se ec ed minerals as a fede of temperatum bility products for all minerals were calculated using the thermodynamic data for aqueous species com-are illustrated in Figure 12-10. Although the scatter f the data in some cases was greater than destred, piled by Tanger and Helgeson (1988), Shock and Helgeson (1988), and Shock et al. (1989) and incor-m st mine I saturati n indica fen &n me range exp cted, or diey evealed systematic deviations sug-porated in the 1987 version of SUPCRT, a revised version of which is now available (Johnson et al.,

geth of enon etiniated m. die t&nnodynamic Properties of either the minerals or the chemical 1992) Solubility product data at 0,25,60,100,150, 200, and 300 *C were formatted for entry into the Potendal of me or mom o% aqwms compments.

database of a somewhat modified 1986 version of Results were sufficiently encouraging to broaden the se pe f the study to include analyses from deep-EQ3 (Wolery,1983).

seated groundwaters and well waters on a worldwide basis. In the following paragraphs, some illustrations 12.6.2 Evaluation of Chemical Analysis reflecting the thennodynamic validation possibilities The first step in evaluating the groundwater is to are given with respect to such a global data set, and check the quality of the chemical analysis by per-some unresolved issues are raised that require further forming an initial distribution of species at the tem-study.

87 NUREG/CP-0147

Natural W ;,1 Analogues of the Nstr Field ofHigh-level Nuclarr Waste Repositories 2

8 a

e e

u...

.g.

...g...

,u v

e,

a g~atr

]

E

• g g

s, a

a a

S e

5 a m dg. m,. E

]

y a

a u

- =

a' cei ra m

'a a

e a%,',

?.

e a

1. s s

i e

i

- i - ' ' ' '

1. ~,

50 HO 150 200 250 50 100 150 200 250 50 100 150 200 250 Toniperature PC)

Temperature PC)

Teure PC) as

.m W

a 4

=

..e u

a W

e m 's i

.e 7

• e 1

a ea 7

g 3

e g

k,

"~

},

is.,

a

=a ah.

j p

e e

a a

g g

• a

}

=

a

=

=

e e

.i 8 i

i s

50 100 150 200 250 50 100 150 200 250 50 100 150 200 250 Temperature fC)

Temperature (*C)

Temperature (*C) 3 3

o e

g3 g

4 2.

af

.]

a a

su a

k2 g

8, <_

i -

5 e

3 3

5 8

B a

e a

7..

c..

j

p......-

g.

=

s 3

2--

E e =

ie.

, i e

i r

. i i i 50 100 150 200 250 50 100 150 200 250 50 100 150 200 250 Temperature PC)

Terriperature (*C)

Temperature (*C)

Figure 12-10. Saturation indices of quartz, calcite, pyrite, low albite, potash feldspar, Illite,laumonite, walrakite, and stilbite as a function of temperature in Icelandic well waters. Horizontal bars represent temperature ranges of observed occurrences (chemical analyses cited in Arnorsson et al.,1983).

NUREG/CP-0147 88

i l

l Natural Geochemtcal Analogues of the Near Field ofHigh-lzoel Nuclear Waste Repositories 12.6,4 Refinement of the Gibbs Free intersected by boreholes in the Ohaki-Broadlands Energy of Minerals from Field hydrothennal area in New Zealand. On the Kam-Data chatka Peninsula, Naboko et al. (1%5) refer to sec-ondary adularia and albite in dichotic and andestic 12.6.4.1 Potash Feldspar and Low Albite volcaniclastics of the Pauzhetka geothennal field, whereas Truklun and Petrova (1974) describe altera-For many years, the element concentration ratio, tion zones containing secondary albite and adularia Na/K, in the liquid aqueous phase has been used to in andesitic lavas and andesitic and dichotic tuffs of j

estimate the soun:e region temperature of hot springs the Bolshe-Bann geothermal field.

and geothermal effluent. As a concentration ratio, it possesses the advantage that it is unaffected by tran-In contrast, the coexistence of low albite with sient boiling or condensation, and it is affected only Potash feldspar is never mentioned in mineralogical slowly by conductive cooling. Foumier andTruesdell studies of cores penetrating sea floor basalts, where (1973) recommended its use only where source re-the temperature ranges from between - 4 and 70 C.

gion temperatures are greater than 150 *C.

Frequent reference is made instead to the presence of The thermochemical basis for the Na/K geother-Potash feldspar, sometimes replacing plagioclase, mometer is not well defined. Usually, discussion Plagioclase in sea floor basalts is commonly replaced centers on the " exchange" reaction between plagio-by secondary clays instead oflow albite. Analcime is 2

clase (low albite) and potash feldspar (adularia or also observed as an authigenic phase, although it is microcline) rarely found in basalts saturated with meteoric waters NaAISi30s + K+ = KAISi308 + Na+

(1) at similar temperatures. As noted earlier, minor vari-ations m mmeral assemblages can be expected.

where Whether this particular deviation is due to the lower temperatures, the presence of seawater, or some other log K (T) = [Na+1 (2) cause remains to be studied further.

[K+]

ecause f extensive mineralogical observations Pa st att em pts t o r econ cile field ob ser vation s wit h C nfirming the coexistence of secondary low albite calcu lations2 sin gpu b lish edher m od ynamicpr oper-and potash feldspar m geothermal fields and the ties of the participating species have not been particu-e asistency of groundwater compositions, it is rea-larly successful, greater faith being placed on the s nable to assume that the Na/K ratio in the aqueous accuracy of thermodynamic data obtained from cal-P ase reflects a near equilibrium state between these h

crimetric studies and phase equilibrium than on two minerals.

groundwater analyses.

The coexistence of secondary low albite and pot.

In Figure 12-11, the calculated values oflog K(T) ash feldspar in geothermal systems is frequently men.

for Ea O t using AGi;9s for low albite and potash tioned in the literature. Tomasson and feldt

. 3m Helgeson et al. (1978), Robie et al.

Kristmannsdottir (1972), in discussing mineral al.

(19~6), berman (1988), Holland and Powell (1990),

teration in the geothermal area of Reykjanes, Iceland, and Kiseleva et al. (1990), are compared with the refer to the occasional presence of albitized plagio.

analytical determinations of the Na/K ratio in well clase, as well as the sporadic occurrence of newly waters from geothermal fields and deep water-satu-rated formations from around the world. Field tem-formed potash feldspar in all geothermal holes. Me.

hegan and Robinson (19%) and Viereck et al. (1982),

Peratures were measured down hole or computed in describing secondary hydrothermal alteration in from well discharge data. The log K(T) values are the Reydarfjordur drill core from Eastem Iceland, calculated along the water saturation curve, employ-mention that low albite and adularia coexist in most ing the entropies oflow albite and potash feldspar of the intersected volcaniclastic ricks, whereas cited by Robie et al. (1979), accepted by all sources, authigenic albite replaces primary plagioclase crys.

and Maier-Kelley heat capacity functions, cited by tals in some of the interdigitated basaltic flows.

Helgeson et al. (1978), for low albite and potash Browne and Ellis (1970) noted the presence of adu.

feldspar. The thermodynamic propenies for the ionic laria and secondary albite in hydrothermally altered species, Na+ and K*, are given by Shock and Hel-intermediate and acid lavas and volcaniclastic rocks geson (1988). The illustrated uncertainty is estimated from those given in the cited references.

89 NUREG/CP-0147

I Natural Gachenucal Analogues of the Nar Field ofHigh-level Nuciar Waste Repositories 3.5 Uncertainty of thermodynamic Cmmihdogs 3.0 3 *C 1

300 'C FIELD DATA low albite 2.5 p

+g g

t:-'

o v

l

,(v 2.0

~~

u..

8 HDNB(1978) oo g

S 3

~

+,, e KOSK(1990)

o. w 1.0 9

K feldspar B(1988)

HP(1989)

RHF(1979) 0.5 0

50 100 150 200 250 300 Temperature (*C)

+ Bolshe Bann, Ru.

6 Wairakei. N2.

Kettleman Dome, U.S.

e Panzhetka, Ru.

o Cajon Pass, Ca.

a Orakeikorako, N2.

o Iceland v Paratunka, Ru.

x Rotorua, N2.

o Ohaki-Broadlands, N2. m Kawerau, N2.

+ Waiotapu, N2.

o Tauhara,N2.

o Savannah River, U.S.

Figure 12 11. Log [Na+y[K+] calculated from compiled thermodynamic data as a function of temperature, compared with the corresponding lon activity products from geothermal well waters NUREG/CP-0147 90

7 Natmalr " " Analogues of the Narr Field ofHigklevel Nudenr Waste Repositories i

The field data are generally consistent and can be

= NaAISi30s + KAISi30s + Ca* + 4H2O fitted by a univariant curve, positioned as indicated

- low albite potash feldspar (3) on Figure 12-11 to within i0.2 log (Na')/(K*) or 1 U and j

in AG7 for reaction (1) at 25* and 2 U at 300 *C. This CaAl2Si4012.2H2O + 2SiO2 + Na+ + K+

contrasts with an uncertainty of nearly 3.5 U for the waarakite quartz i

log K(T) predacted from calorimetry of phase equi-

= NaAISi30s + KAISi30s + Ca* + 2H2O librium measurements. The fitted curve closely fol-I W albite potash feldspar (4) lows the equation proposed by Foumier (1981) for t

l the Na/K geothermometer and the data reported by j

where 11'in et al. (1979) for the Pauzhetka geothermal field' K-(5) but it does not fit the latter investigators' data from

- [Na+]M the lower temperature Paratunka field in the same An attempt at calibrating the Gibbs fr ec energies of f*88 "'

lau m ontiteand wair akiteu singasimilar schemewas The chemical analyses of geothermal wells below first made by Senderov (1980). Figure 12-12 is a l

200 *C reported by Arnorsson et al. (1983) include comparison between the logarithm ofion activity

)

aluminum concentration. The saturation indices of productscalculatedfromanalysesofgeothermalwa-3 low albite and potash' feldspar can, therefore, be ters and log K calculated from thermodynamic data j

computed for those well waters and their solubility of the participating minerals. The low albite and 3

products compared with those predicted from cal-potash feldspar Gibbs free energies were like those i

J orimetry and high-temperature phase equilibda. The described in the preceding section. Those oflaumon-results, illustrated in Figure 12-10, show the satura-tite and wairakite were calculated from phase equi-tion indices referenced te the data by Berman (1988),

librium and heat capacity measurements. There who used the thermodynamic properties oflow albite appears to be a discrepancy of approximately 0.7 in taken from Hemingway and Robie (1977). The satu-log K between predicted and field values.The causes ration indices of both feldspars show trends towards of the discrepancy have yet to be identified, but may supersaturation below 100 *C, which might be in.

be due in part to attainment of equilibrium with dicative of the sampled waters originating at tempera-respect to cristobalite below 150 *C. Line A-A' rep-1 d

tures higher than those measured. In spite of the resents the position of the reaction:

scatter, the saturation indices suggest that AGE.29s of CaAl2Si4012.4H2O + 2SiO2 + Na+ + K+

low albite is more nearly co:Tect than that of potash laumontite cc-cristobalite feldsparin Bennan's database.

= NaAISi30s + KAISi30s + Ca* + 4H2O A correction to AGi.29s for potash feldspar alone low albite potash feldspar (6) i using the field correlation of Na/K yields a revised which gives better agreement with field data.

i value, which is 6.3i3.7 U.mor' more negative than i

that of Robie et al. (1978) or 3.2i3.7 U.mor more 12.6.6 Other Minerals i

negative than that of Berman (1988), but it is A serious difficulty with such procedures is the i

5.6i3.7 U.mor* less negative than AHi29: of micro-frequent need to know the down-hole pH of the cline determined by Kiseleva et al. (1990).

chemical analysis, as many reactions involve hy-drolysis. Unfortunately, measurements of this pa-12.6.5 Laumontite and Wairakite rameter are normally made only after a sample has i

been retrieved and cooled to surface temperatures.

The correlation of the Gibbs free energy of low The down-hole pH must then be calculated using 1

albite and potash feldspar is but one of several exer-known thermodynamic data for dissolved aqueous ctses m the calibration of mineral thermodynamic species at the surface and down-hole temperatures.

properties using geothermal fluids and groundwaters.

As noted earlier, it appears that the calculated down-Otherexamplesm, eludethecoexistenceoflaumontite hole pH is a major source of inaccuracy in validation or wairakite with low albite and potash feldspar, studies. The cause of the inaccuracy is not well un-represented by the equilibrium reactions:

derstood, but it could be related to the fugitive behav-CaAl2Si40 2.4H2O + 2SiO2 + Na+ + K+

1 laumontite quaitz -

91 NUREG/CP-0147

Natural Geoc}umacal Analogues of the Naar Field ofHigh-Ine! Nuciatr Waste Repositories 4.0 i

i i

i O

Laumontite + Ouartz -

3.0 p

Walrakite + Ouartz O g 3

0 2.0 -

en 1.0 - E Cajon Pass, California 3

V Kettleman Dome, California A

A E Sospe Hot Springs, California g g Y Buried Triassic, South Carolina s

o,o _ O leeland b

Og Q Cerro Prieto, Mexico 8

Q

$ Ohaki Dreadlands, N.Z.

'O QQhD A Tachara, N.Z.

Low Albite + K-Feldspar + Quartz O

O Waitakel, N.Z.

.0 0 Bolshe Bar'n. Russia O Paratunski, Russia 6 Pauzbetski, Russia

-2.0 I

I I

I I

0 50 100 150 200 250 300 350 3

Temperature ('C)

Figure 1212. Log K = [Na+][K+y[Ca++] calculated from compiled thermodynamic data as a function of temperature, compared with the corresponding lon activity products from geothermal well waters ior of dissolved CO2 during sample collection and Another source of error might be the thermochemical analysis, thus:

properties of pyrite, although the unifonnity of ther-HCOi+ H% H20+CO (7) m dynamic values in the literature would suggest 2

otherwise. The thermodynamic properties of ferrous Understanding the inter-relationship of pCO2 with ion, Fe* (Nordstrom,1984; Vasil'ev et al.,1985) pH in brines of different salinities is an important step could also be a source of error. In predicting the in reducing uncertainties associated with pH values groundwater composition from the secondary miner-in geothermal systems.

alogy of geothermal fields where pyrite is present, the author (Apps,1991) found significant discrepancies 12.6.7 Pyrite in the predicted concentrations of Fe" and SOF Calculations of the saturation index of pyrite from when compared with measured values using the dis-s luti n reactiom geothermal well waters in Iceland by Gunnlaugsson and Arnorsson (1982) and those reported in this paper FeS + H O = 0.251r +

2 2

in Figure 12-10, show that they are supersaturated 0.25 SOT + Fe* + 1.75HS-(8) with respect to pyrite by severallog units over the range of temperatures where pyrite is observed. Such The cause of the discrepancy between predicted large discrepancies are quite unrealistic. Of the many and observed pyrite solubility in geochemical sys-potential sources of error, one often proposed is the tems requires urgent attention. It is hoped that natural contamination of the water sample by ferrous iron analogues will help in tracing the source of the pmb-lem and aid in its resolution, due to dissolution of the steel casing in the well.

NUREG/CP-0147 92

i i

3

- Naturn! Geximmcel Analogues of the Near Eidd ofHigMsod Nuciant Waste Itaponnorus 4

12.6.8 Comments on the Observed 12.6.10 Disequilibrium Introduced by Scatterin Field Data Groundwater Transport As is observed in Figure 12-6 and 12-7, the scatter Geochemical systems convect, causing local su-

)

in the trends of potentials of chemical components Persaturation and precipitation with respect to some varies from =i0.2 for log [Na*F[K+] and log minerals whenever groundwater from an elevated

[SiO2(aq)] to -il.5 for the logarithms of activity temperature source migrates into cooler regions of the ratios incorporating [H+]. In computing saturation geothermal system. The kinetics of precipitation un-indices, Figure 12-10, the scatter ranges from 0.3 der falling temperature gradients has not been mod-S.I. units for quartz and calcite in their presumed eled, except in a f cw cases (Verma and Pruess,1988),

stability fields to about +3-0.5 S.L units in the case but it is observable in several geothermal fields. At j

l of pyrite. It should be noted, however, that the mag.

Wairakei, New ZaN for example, wells adjacent j

i nitude of errors in S.I. is a function of the number of to major fractures are supersaturated with respect to j

ttoms in a mineral formula unit. Thus, for example, Potash feldspar (Ellis and Wilson,1960). In Icelandic l

in comparing the S.I.s of quartz and albite, S.I.(albite) wells, many groundwaters less than 150 *C are super-should be divided by three, because it contains three saturated with respect to quartz (see Figure 12-10).

SiO2 units in one formula unit, NaAISi30s.

At 250 *C, substantial equilibration with respect to quartz can be achieved in a few days, but at 25 T,

. The scatter can be attributed to any of several equilibuum is achieved only after several thousand variables in what is sometimes a poorly constrained years (Rimstidt and Bames,1980). Therefore, con-system. In the following discussion, the effect of vecting gmundwaten,initia% saturated with respect several of these variables will be examined.

to quartz, can become supersaturated with respect to chalcedeny upon cooling. It should be noted, how-l 12.6.9 Sampim.g and Analysts ever, that supersaturation with respect to quartz in

]

The principal pmblems are preservation of sample Icelandic groundwaters is also achievable through l

integrity and determination of down-hole or in situ dissolution of natural glasses in hyaloclastites or ba-formation water pH. Loss or gain of CO2 or H2ssalt flows.

during sampling and subsequent handling can cause Thermodynamic analyses of hot spring data usu-l substantial shifts in sample pH, particularly if the ally show significant variations in the predicted TDS content of the water is low. Still more errors are source region temperatures when using different introduced if the pH has been detemuned on cor-geothermometers. Furthermore, calculated S.I.s of rupted samples. Errors of 0.5 pH and even greater many minerals known or inferred to exist at depth i

are not uncommon and would easily account for indicate that the hot spring water is no longer in much of the scatter observed in the log activity ratio equilibrium with the source region. Evidence of j

trends of Figure 12-6. Uncertainties in sample pH chemical modification of the groundwater during i

also arise when the pH of high ionic strength ground-transport and cooling is very evident. Each field waters are measured. Errors of as much as 0.5 pH occurrence must be examined carefully, and the ex-l units can arise due to liquid junction errors in the tent of disequilibrium determined. Further investiga-reference electrode when the pH meter is calibrated tions of these phenomena are needed, and caution using standard buffer solutions.

must be exercised in assuming that local metastable Erron introduced through loss of NH3 and H2S thermodynamic equilibrium is generally valid.

could also contribute to the lack of precision in pH 12.6.11 Solid Solutions measurement and calculation of some mineral S.I.s, for example, pyrite. The author need not elaborate on Where minerals are pure or nearly pure substances a

the problems associated with sample collection from (e.g., quartz, albite, pyrite, and anhydrite), the mineral flashing geothermal wells. It suffices to note that activity product should be fixed at some finite value, calcium carbonate and silica precipitation willintm-and S.L should be zero at any given temperature, duce significant errors unless corrected. The dissolu-provided that the mineral is present and at equilibrium tion of steel casing and its impact on the saturation with the coexisting groundwater. This assumption indices of pyrite and Fe(II)-bearing silicates has al-has been tested by the author and by others and ready been noted, appears to be generally valid for hydrothermal sys.

l 93 NUREG/CP-0147

Naturel Ganhemias! Analogues of the Narr Field ofHigh-Inel Nuclair Waste Repositories tems. Several classes of minerals are not pure sub-tween groundwater and secondary minerals. With the stances, however, and vary in composition depending current state of knowledge, field data can be corre-on the bulk chemistry of the host rock and the altera-lated with thermodynamic data, and the thermody-tion kinetics of the primary minerals and glasses. Of namic data for pure substances can be beneficially greatest importance are the phyllosilicates and zeo-adjusted in certain circumstances to refine model lites. Compositional variability in these classes of predictions.

minerals can be described in terms of solid solutions Predictions of groundwater chemistry on the basis of defined end members. In many cases, the solid of the observed secondary mineralogy in a geother-solutions beha.e ideally, for example, the substitu-mal system generally correlate well with the observed tion of Fe" for Mg" in chlorite (Walshe,1986). In composition of the geothermal fluids, provided that others, such as the substitution of Fe** for Fe" + H*

the concentration of C1' is specified (Apps,1991).

in biotites, the solid solution is far from ideal (Beane, Most aqueous species concentrations can be pre-1972). If the substitutions are minor, the solid solu-dicted to within a factor of five, and some to within tion can be approximated to the composition of the 20 mol. percent. Exceptions are the Fe" and SOT as dommant member component, and the S.I. will devi-noted earlier.

are only incrementally from zero in making this ap-Although the overall level of precision might be proximation. The activity product of minor improved incrementally with the further refinement component end members, however, can be expected of thermodynamic data, present modeling and vali-to vary substantially from one geothermal site t dation of geothermal systems as natural analogues are another. Unfortunately, the quality of thermody-sufficient to make prelimmary predictions of many namic data for phyllosilicates and zeolites is gener-radioclement source term concentrations, establish ally inadequate to correct for solid solution the corrosivity of the groundwater to various waste substitutions. Therefore a significant scatter in the container materials, and initiate prelimmary studies activity products of solid solutions is to be expected of chemical transport in the near-field environment.

when these minerals are approximated to their end member compositions.

Thuhief problem remaining is the observed scat-ter m activity ratios of ionic species and mineral Another feature of minerals displaying solid solu-saturation indices calculated from chemical analyses tion pmperties is that the composition can vary both of groundwaters. The principal causes of this scatter with the state of alteration of the rock matrix and with should be identified, and if possible, accounted for in temperature. As the rock alters, the composition of subsequent field sampling programs. Model refine-the geothermal fluid changes m response to changing ment, including the recognition of solid solution chemical conditions. Succeeding layers of crystals models for clays and zeolites, would also be advan-aggregating on a fracture surface or vesicle wall will, tageous. Enhanced precision might also be achieved therefore, reflect these changes. Only the composi-through the application of more sophisticated electro-tion of the mineral at the water interface could be in lyte models to describe the thermodynarnic properties equilibrium at the time of samphag. Bulk chemical of groundwaters with ionic strengths exceeding 0.1.

analyses of solid solutions are, therefore, not appro-priate for equilibrium thermodynamic calculations in The next step in using geothermal systems as geochemical systems.

anal gues is m the development of kinetic models.

This phase of the work will undoubtedly provide a seri us challenge to the geochemist, but it is essential 12.7 CLOSING COMMENTARY if evaluations of the rates of alteration in the near field Geothermal systems in volcanic rocks can serve as of geologic repositories are to be predicted. The suitable analogues of near-field geothermal proc-problem could be tackled in six steps.

esses. They can be used to calibrate thermodynamic (i) Measurement of glass devitrification kinetics data, and their relatively uniform behavior permits in the laboratory by means of flow-through the validation of groundwater compositions calcu-reactors at elevated temperature.

lated from observed secondary mineral assemblages.

(ii) Measurement of crystal growth kinetics at Although further study is required to resolve a num-elevated temperatures in the laboratory ber of details, it appears likely that natural analogues through small displacements from equilib-provide a sound basis for predicting equilibrium be-rium of selected secondary minerals.

NUREG/CP-0147 94

Natural Gevchemscal Analogues of the Near Field of High-laci Nuclear Waste Berositories (iii) Development of crystal growth models of

12.8 REFERENCES

secondary minerals at low levels of supersatu.

A.P.H.A.,1985. StandardMethodsfor the Examina-

'*U* **

tion of Water and Wastewater,1Sth Edition.

(iv) Integration of glass devitrification /dissolu-American Public Health Association, Washing-tion kinetics and crystal growth in a model ton, DC 20005, p. 33-34.

that includes diffusive and convective trans-Aagaard, P. and Helgeson, H.C.,1983. Activity /com-port, as well as surface reactions of dissolu-position relations among silicates and aqueous tion and precipitation. It is probable that solution II. Chemical and thermodynamic conse-surface reaction rates at the low levels of quences of ideal mixing of atoms among energeti-supersaturation observed in the field are fast cally equivalent sites in montmorillonites, illites, m relation to diffusive and convective trans-and mixed layer clays. Clays and Clay Afinerals,

port. Hence, the latter transport processes v.31,p.207-217.

would control the rate.

Alexander, G.B., Heston, W.M., and Iler, R.K.,1954.

(v) Inclusion of step (iv)into a reactive chemical The solubility of amorphous silica in water. Jour-transport model of a convecting geothermal na! ofPhysical Chemistry, v. 58, p. 453-455.

(vi) dat on of the model in step (v) using Apps, J.A.,1970. The Stability field of analeime.

well-characterized geothermal reservoirs.

Ph.D Thesis, Harvard University,347 p.

Apps, J.A.,1983. Hydrothermal evolution of reposi-The model developed in step (v) could be used to tory groundwaters in basalt. In N.R.C. Nuclear explore the rates of dispersion of soluble and insol-Waste Geochemistry '83. U.S. Nuclear Regulatory uble elements as illustrated in Figure 12-9 in addition Commission Report. NUREG/CP-0052, p.14-51.

to more general validation studies of geothermal Apps, J.A.,1987. Alteration of natural glass in radio-fields. Although the program suggested above is quite active waste repository host rocks; a conceptual demanding, it is not unrealistic given the current state review. Lawrence Berkeley Laboratory Report of the art in model development.

LBL-22871,38 p.

The final step m model development involves the Apps, J.A.,1991. Predictions of the evolution oflocal incorporation of radioelement transport in the near steam phase chemistry and residual brine compo-field. Although some fission product radioelements s tion in the steam reservoir during exploitation of might be correlated with field observations of the The Geysers steam field, California. Draft Report behavior of their nonradioactive counterparts, there for the Geothennal Energy Division, United States will be no such opportunity to observe the behavior Department of Energy,77 p.

of actinides other than uranium and thorium. Rehance Amorsson, S., Gunnlaugsson, E. and Svavarsson, H.,

must instead be placed on laboratory experiments 1983. The chemistry of geothermal waters in Ice-under simulated near-field conditions. This small step land, II. Mineral equilibria and independent vari-beyond the domain of natural analogues will, how-ables controlling water compositions. Geochumca ever, be buttressed by the confidence gained from et Cosmochumca Acta, v. 47, p. 547-566.

model development and validation of the geochemi-Am rsson, S., Sigurdsson, S. and Svavarsson, H.,

cal framework in which the radionuclides would mi-1982. The chemistry of geothermal waters m Ice-gate.

land I. Calculation of aqueous speciation fmm 0-to 370 *C. Geochimica et Cosmochimica Acta, v.

In conclusion, it can be confidently stated that 46,p.1513-1532.

natural analogues in the form of geothermal systems Beane, R.E.,1972. A thermodynamic analysis of the in volcanic rocks are not only desirable, but necessary effect of solid solution on the hydrothermal stabil-for the development of performance assessment mod.

ity of biotite. Ph.D. Dissertation, Northwestem els of the near field of repositories in volcanic rocks, University, Evanstown, Illinois,195 p. plus ap-such as that planned at Yucca Mountain, Nevada.The pendix.

prelimmary results given in this paper suggest that Berman, R.G. and Brown, T.H., and Perkins, E.H.,

such optimism is well founded. Nevertheless, the 1987. GE\\lO(0,/)-CALC: Software for calculation geochemist should not underestimate the challenges and display of P-T-X phase diagrams. American that lie ahead.

j Afineralogist, v. 72, no. 7-8, p. 861-862.

l l

95 NUREG/CP-0147

Natural Geochemical Analogues of the Near Field ofHigh4evel Nuclem Waste Repositories Berman, R.G.,1988. Intemally-consistent thermody-Geochimica et Cosmochimica Acta v.37, pp.

namic data for minerals in the system Na20-K20-1255-1275.

CaO-MgO-FeO.Fe2O3-Al2O3-SiO2-H20-CO2 Foumier,R.O.,1981. Applicationofwatergeochem-Journal of Petrology, v. 29, p. 445-522.

istry to geothermal exploration and reservoir engi-Bourcier, W.L.,1985. Improvements in the solid neering. Chapter 4 in Geothermal Systemst solution modeling capabilities of the EQ3/6 Geo-Principles and Case Histories (L. Rybach and chemical Code. Lawrence Livermore National LJ.P. Muller, Eds.). John Wiley and Sons, Ltd.

Laboratory Report UCID-20587,8 p.

Gardner, L.U.,1938. Reaction of the living body to Brown, T.H., Berman, R.G., and Perkins, E.H.,1989.

different types of mineral dusts with and without PTA-SYSTEM: A Geo-Cale software package for complicating infection. American Institute ofAfin-the calculation and display of activity-tempera-ing and Afetallurgical Engineers Technical Publi-ture-pressure phase diagrams. American Afiner-cation No. 929,15 p.

alogist, v. 74, no. 3-4, p. 485-487.

Garrels, R.M. and Christ, C.L.,1 %5. Solutions, Min-l Browne, P.R.L. and Ellis, AJ.,1970. The Ohaki-erals and Equilibria, Harper Row, Publishers, New Broadlands hydrothermal area, New Zealand:

York,450 p.

Mineralogy and related geochemistry. American Gunnlaugsson, E. and Arnorsson, S.,1982. The Journal ofScience, v. 269, p.97-131.

chemistry of iron in geothermal systems in Ice-Carnahan, C.L.1986. Simulation of uraru.um trans-land. Journal of Volcanology and Geothermal Re-port with variable temperature and oxidation po-search, v.14, p. 281-299.

tential: The computer program THCC. Lawrence Helgeson, H.C., Delany, J.M., Nesbitt, H.W., and Berkeley Laboratory Report LBL-21639,10 p.

Bird, D.K.,1978. Summary and critique of the D'Amore, F. and Gianelli, G.,1984. Mineral assem-thermodynamic properties of rock forming miner-blages and oxygen and sulfur fugacities in natural als. Amnican Journal of Science, v. 278-A, p.

water-rock interaction processes. Geochimica et 1-229.

Cosmochimica Acta. v. 48, p. 847-857.

Dibble, W.E., Jr. and Tiller, W.A.,1981. Kinetic Hemingway, B.S., and Robie, R.A.,1977. Enthalpies f formation of low albite (NaAISi308), gibbsite model of zeolite paragenesis in tuffaceous sedi-(Al(OH)3) and NaAIO2: revised values for and of ments, Clays and Clay Minerals, v. 29, p. 323-s me alumino-silicate minerals. Journal of Re-330.

search of the U.S. Geological Survey, v. 5, p.

Eggleton. R.A. and Keller, J.,1982. The palagoniti-413-429.

zation of limburgite glass--a TEM study. Neue, Jahrbuchfiir Afineralogie, Afonatshefte, No. 7, p.

Hitchen, C.S.,1935. A method for experimental in-l 321-336.

vestigation of hydrothermal solutions, with notes Ellis, AJ. and Wilson, S.H.,1960. The geochemistry on its application to the solubility of silica. Trans-of alkali metal ions in the Waimkei hydrothermal actions of the Institution of Afining and Afetal-system. New Zealand Journal of Geology and lurgy, v. 44, p. 255-280.

Geophysics, v. 3, p. 593-617.

Holland, TJ.B. and Powell, R.,1990. An enlarged Elmer, T.H., and Nordberg, M.E.,1958. Solubility of and updated intemally consistent thermodynamic silica in nitric acid solutions. Journal ofAmerican dataset with uncertainties and correlations: The Ceramic Society, v. 41, p. 517-520.

system K20-Na20-CaO-MgO.MnO-FeO-Fe2O3-Felmy, A.R., Girvin, D. and Jenne E.A.,1984. MIN.

Al2O3-TiO2-SiO2-C-H2 -02. Journal of Afeta-TEQ: A computer program for calculating aque-morphic Geology, v. 8, p.89-124.

ous geochemical equilibria U.S. Environmental II'in, V.A., Kononov, B.G., Polyak, B.G. and Ko-Protection Agency Report EPA-600/3-84-032.

zlovtseva, S.V.,1979. Hydrochemicd estimation Forsman, N.F.,1984. Dumbility and alteration of of deep rock temperatures. Geochemistry Interna-some Cretaceous and pyroclastic glasses in North tional1979, p.119-132.

Dakota. Journal ofNon-Crystalline Solids, v.67, James, R.O. and Parks, G.A.,1976. Application of p.449-461.

the computer program "MINEQL" to solution of Foumier,R.O.andTruesdell, A.H.,1973. Anempiri-problems in surface chemistry. Unpublished ad-cal Na-K-Ca geothermometer for natural waters.

dendum to Westall et al (1976),35 p.

1 NUREG/CP-0147 96 j

Natural Geochemical Analogues of the Near Field ofHigMevel Hudear Waste Repositories Johnson, J.W., Oelkers, E.H., and Helgeson, H.C.,

computation equilibrium in systems involving 1992. SUPCRT92: A software package for calcu-solid, aqueous and gas nonideal solutions. Ab-lating the standard molal thermodynamic proper-stracts, Second International Symposium, Ther-ties of minerals, gases, aqueous species, and modynamics of Natural Processes and Russian reactions from 1 to 5000 bats and 0 *C to 1000 *C.

Symposium, Thermodynamics in Geology, N0-Computers and Geosciences, v.18, no. 7, p. 899-.

vosibirsk, Russia, September 13-20,1992, p. I17.

947.

Morey, G.W., Fournier, R.O., and Rowe, JJ.,1%2.

Karpov, I.K.,1981. Fiziko-Khimicheskoe Modeli-The solubility of quanz in water in the temperature rovanic na EVM v Geokhunii (Physicochemical interval from 25 to 300 C. Geochimica er Cosmo-Modeling by Computerized Geochemistry).

chimica Acta., v. 26, p.1029-IN3.

Nauka, Sibirskoi Otdelenie, Novosibirsk, USSR, Morrey, J.R.,1988. FASTCHEM* Package, Vol-247 p.

ume 4: User's guide to the ECHEM equilibrium Karpov, I.K., Kulik, D.A., and Chudnenko, K.V.,

geochemistry code. Electdc Power Research Insti-1992. Computer technology ofimitation and mod-tute Report EPRI EA-5870-CCM.

eling in physico-chemical processes in geos-Naboko, S.I., Karpov, G.A. and Roznikova, A.P.,

ciences: theory, results, outlooks. Abstracts, 1965. Gidrotermal'nyi metamorphizm porod i

[

n Second Intemational Symposium, Thermodynam-mineralo-obrazovanie, iHydrothermal metamor-ics ofNatural Processes and Russian Symposium, phism of rocks and mineralization]. In Pau:hetskie Thermodynamics in Geology, Novosibirsk, Rus-Goryachie Vody na Kamchatka. Nauka, Moscow, sia, September 13-20,1992,p.25.

p.76--I 18.

Kharaka, Y.K., Gunter, W.D., Aggarwal, P.K.,

Noorishad, J., Carnahan, C.L. and Benson, L.V.,

Perkins, E.H., and DeBraal, J.D.,1988.

1987. A report on the development of the non-Solmineq.88: A computer program for geochemi-equilibrium reactive chemical transport code cal modeling of water-rock interactions. U.S. Geo-CHMTRNS. Lawrence Berkeley Laboratory Re-logical Survey Water-Resources Investigations port LBL-22361,229 p.

Report,88-4227,420 p.

Nordstrom, D.K.,1984. Partial compilation and revi-Kiseleva, I.A., Ogorodova, L.P., Siderov, Yu.I., and sion of basic data in the Water programs. U.S.

Khodakovskiy, I.L.,1990. Thermodynamic prop-Geological Survey WRI Report, #84-4186,40 p.

erties of alkali feldspars. Geochemistry Interna-Okaoto, G., Takeshi, O., and Katsumi, G.,1957, tional 1990, p.93-100.

Properties of silica in water. Geochimica et Cos-Kitahara, S.,1960. The polymerization of silicic acid mochimica Acta, v.12, p.123--132.

obtained by the hydrothermal treatment of quartz Ostwald, W.,1897. Studien uber die Bildung und and the solubility of amorphous silica. Review of Umwandlung fester K6rper. Zeitschrrftfrir Physi-Physical Chemistry ofJapan, v. 30, p.131-137.

kalische Chemie, v. 22, p. 289--330.

Kmpka, K.M. and Morrey, J.K.,1985. MINTEQ Parkurst, D.L., Thorstensen, D.D. and Plummer, geochemical reaction code: Status and applica-L.N.,1980. PHREEQE - A computer program for tions. In Proceedings of the Conference on the geochemical calculation, [ Revised and reprinted, Application ofGeochemicalModels to High-level 1990). U.S. Geological Survey Water Resources Nuclear Waste, (G.K. Jacobs and S.K. Whatley, Investigation Report 80-96,195 p.

Eds.). U.S. Nuclear Regulatory Commission Re-Perkins, E.H., Brown, T.H. and Berman, R.G.,1986.

port NUREG/CP0062, p. 46-53.

PT-System, TX-System, PX-System: Three pro-Lenher, V., and Merrill, H.B.,1917. The solubility of grams which calculate pressure-temperature-com-silica. Journal of the American Chemical Society, position phase diagrams. Computers and

v. 39, p. 2630-2MO.

Geoscience, v.12, p. 749-755.

Mehegan, J.M. and Robinson, P.T.,1982. Secondary Powell, R., and Holland, TJ.B.,1988. An internally mineralization and hydrothermal alteration in the consistent thermodynamic dataset with uncertain-Reydadjordur drill core, Eastern Iceland. Journal ties and correlations: 3. Applications to geo-of Geophysical Research, v. 87, p. 6511-6524.

barometry, worked examples and a computer Mironenko, M.V., Zolotov, M.Yu., Frenkel, M.Y.,

program. Journal of Metamorphic Geology, v. 6, 1992. Algoritlun, computer code and database for p.173-204.

97 NUREG/CP-0147

Natural Geochanical Analogues of the Near Fic!d ofHigh-Lent Nuclear Waste Repositories Raines, G.E., Rickertsen, L.D., Claibome, H.E.,

Thermodynamics in Geology, Novosibirsk, Rus-McElroy, J.L. and Lynch, R.W.,1981. Develop-sia, September 13-20,1992,p.51.

ment of reference conditions for geologic reposi-Siever, R.,1%2. Silica solubility,0-200 'C, and the tories for nuclear waste in the USA. In Scientific diagenesis of siliceous sedunents. Journal of Ge-Basisfor Nuclear Waste hfanagement, (John G.

ology, v. 70, p.127-150.

Moore ed.), Plenum Press, New York, v. 3, p.

Spychalski, R.,1938. Zur Stabilitut von Kie-1-10.

seloxydhydraten definierter Zusammensetzung.

Reed, M.H. and Spycher, N.F.,1989. CHILLER.

Zeitschrift fiir Anorganische und Allgemeine Department of Geological Sciences, University of Chemie, v. 239, p. 317-320.

Oregon, Eugene, OR 97403.

Spycher, N.F. and Reed, M.H.,1989. SOLVEQ: A Rimstidt, J.D. and Barnes, H.L.,1980. The kinetics computer program for computing aqueous-min-of silica water reactions. Geochimica et Cosmo-eral-gas equilibria. Revised Preliminary Edition.

chimica Acta, v. 44, no. I1, p.1683-1700.

Depanment of Geological Sciences, University of l

Robie, R.A., Hemingway, B.S. and Fisher, J.R.,

Oregon, OR 94703.

1978. Thermodynamic properties of minerals and Tanger, J.C. and Helgeson, H.C.,1988. Calculation 5

related substances at 298.15 K and I bar (10 of the thermodynamic and transport properties of Pascals) pressure and at higher temperature. U.S.

aqueous species at high pressums and tempera-Geological Survey Bulletin 1452,456 p.

tures: Revised equation of state for the standard Scheel, L.D., Fleisher, E., and Klemperer, F.W.,

partial molal properties of ions and electrolytes.

1953. Toxicity of silica. I: Silica solutions. Ar.

American Journal ofScience, v. 288, p.19-98.

chives ofIndustrial Hygiene and Occupational Taylor, M. and Brown, G.,1978. Structure of mineral Afc licine, v. 8, p. 564-573.

glasses-I. The feldspar glasses NaAISi30s, Senderov, E.E.,1980. Estimation of Gibbs energy for KAISi30s, CaAl2Si20s. Geochimica et Cosmo-laumontite and wanakite from conditions of their chimica, v. 43, p. 61-75.

formation in geothermal areas. In Proceedings of Tomasson, J. and Kristmannsdottir, H.,1972. High the International Conference on Zeolites, 5th, temperature alteration minerals and thermal j

(Rees, Lovat V.C., Eds.). Leyden: London, Eng-brines, Reykjanes, lecland. Contributions to Min-land, p 56-63.

eralogy and Petrology, v. 36, p.123-124.

Shock, E. and Helgeson, H.C.,1988. Calculation of Trukhm Yu., P., and Petrova, V.V.,1974. Geo-thermodynamic and transport properties of aque-chemical effect of hydrothermal metasomatism ous species at high pressures and temperatures:

and variation in the composition of thermal solu-l Correlation algorithms for ionic species and equa-tions with time,(based on Bolshe-Banny geother-tion of state prediction to 5 kb and 1000 C.

mal deposit of Kamchatka). In Hydrothermal Geochimica et Cosmochimica Acta, v. 52, p.

biineral. Forming solutions in the Areas ofActive 2009-2036.

Volcanism. [Gidrotermal'nye Mineraloo-Shock, E.L, Helgeson, H.C. and Sverjensky, D.A.,

brazuyushchio Rastvory Oblastei Aktivnogo Vul-1989. Calculation of the thermodynamic and kamzma) (S.I. Naboko, Ed.) Nauka Publishers, transport properties of aqueous species at high Siberian Branch Novosibirisk,1974. Translated pressures and temperatures: Standard partial molat from Russian, and published for the U.S. Dep.

properties of inorganic neutral species. Geo-Inter. and Nat. Sci. Found., Amerind Publishing chimica et Cosmochimica Acta, v. 53, p. 2157-Co. Pvt. Ltd., New Delhi,1982, p. 274-285.

2183.

Tsang, Y.W. and Pruess, K.,1992. Personal commu-Shvarov, Yu.V.,1976. Calculation of the equilibrium nication.

composition in a multicomponent heterogeneous Tsang, Y.W. and Pruess, K.,1987. A study of ther-system. Doklady, AkademiiNauk SSSR, v. 229, no.

mally induced convection near a high-level nu-5, p.1224-1226. [In Russian]

clear waste repository in partially saturated Shvarov, Yu.V.,1992. The software for equilibrium fractured tuff. Water Resources Research, v. 23, modeling of hydrochemical processes. Abstracts, p.1958-1966.

Second International Symposium, Thermodynam-Vasil'ev, V.P., Dmitrieva, N.G., Vorob'ev, P.N., Va-ics ofNatural Processes and Russian Symposium, sil'ev, V.N. and Nechaeva, I.I.,1985. The standard NUREG/CP-0147 98 m

Natural Geochemu:al Analogues of the Natr Field ofHigh-lael Nuclear Waste Repositories enthalpy of formation of the iron (II) ion in aque.

Wolery, TJ.,1983. EQ3NR A computer program for ous solution at 298.15 K. Russian Journal of/nor.

geochemical aqueous speciation-solubility calcu-ganic Chemistry, v. 30, p. 957-959.

lations: User's guide and documentation.

Verma, A. and Pruess, K.,1988. Thermohydrological Lawrence Livermore National Laboratory Report conditions and silica redistribution near high-level UCRL-53414,191 p.

nuclear wastes emplaced in saturated geological Wolery, TJ.,1984. EQ6, a computer program for formations. Journal of Geophysical Research, v.

reaction path modeling of aqueous geochemical 93, no. B2, p. ll59-ll73.

systems: User's guide and documentation, Viereck, L.G., Griffin, BJ., Schmincke, H. U. and Lawrence Livermore laboratory Report, UCRL-Pritchard, R.G.,1982. Volcaniclastic rocks of the 51 (DRAFT),251 p.

Reydarfjordur drill hole, Eastem Iceland 2. Altera.

Wolery,TJ.,1992. EQ3NR, A computer program for tion. Joumal of Geophysical Research, v. 87, p.

geochemical aqueous speciation-solubility calcu-lations: Theoretical manual, user's guide, and re-6459-6476.

lated documentation (Version 7.0). Lawrence Walker, G.P.L.,1960a. Zeolite zones and dike distri-Livermore National Laboratory Report No.

bution in relation to the structure of the basalts of UCRL-MA-110662 Pt. III,246 p.

castem iceland, Journal of Geology, v. 68, p. 515-Wolery, TJ., and Daveler, S.A.,1992. EQ6, a com-528 Puter prognm for reaction path modeling of aque-Walker, G.P.L.,1960b. The amygdale minerals in the ous geochemical systems: Theoretical manual, Tertiary lavas of Ireland. III. Regional distribution,

's guide, and related documentation (Version Mmeralogical Magazine, v. 32, p. 503-527 7.0). Lawrence Livermore National Laboratory l

Walshe, J.L.,1986. A six-component chlorite solid Report UCRL-MA-110662 Pt. IV,337 p.

solution model and the conditions of chlorite for-Wood, D.A., Gibson, l.L., and Thompson, R.N.,

mation in hydrothermal and geothermal systems.

1976. Elemental mobility during zeolite facies Economic Geology, v. 81, p. 681-703, tnetamorphism of the Tertiary basalts of eastern Wang, J.S.Y., Tsang, C.F., Cook, N.G.W. and With-Iceland, Contributions to Mineralogy and Petrol-erspoon, P.A.,1979. A Study of the Regional ogy,v.55,p.241-254.

Temperature and Thermodynamic Effects of an Underground Repository for Nuclear Waste in 12.9 ACKNOWLEDGMENTS liard Rocks. Lawrence Berkeley Laboratory Re-The author would like to express his appreciation port LBL-8271 Rev.,60 p.

to Dr. George Birchard of the NRC for his encour-l Westall, J.C., Zachary, J.L. and Morel, F.M.M.,

agement and original support in undertakmg much of l

1976: MINEQL - A Computer program for the the work reported in this paper. The author is also calculation of chemical equilibritun compositions indebted to the NRC Advisory Committee on Nuclear of aqueous systems, MIT Technical Note No.18, Waste for partial support in attending the Workshop 91 p.

on the Role of Natural Analogs in Geologic Disposal Wolery, TJ.,1978. Some chemical aspects of hy-of High-Level Nuclear Waste, San Antonio, Texas, l

drothermal precesses at mid-oceanic ridges - A July 23-25,1991, and in drafting the paper. A critical theoretical study. I Basalt-seawater reaction and review of the manuscript by Dr. William Murphy of chemical cycling between the oceanic crust and Southwest Research Institute, San Antonio, Texas, the oceans,11 Calculations of chemical equilib-was much appreciated. Many of the calculations of rium between aqueous solutions and reinerals. Un-species distributions in geothermal waters were exe-published Ph.D. thesis, Northwestem University, cuted on the author's behalf by Mr. Gee Minn Chang 263 p.

of the Lawrence Berkeley Laboratory.

99 NUREG/CP-0147

i NATURAL ANALOGS FOR FAR-FIELD g

ENVIRONMENT / HYDROLOGY Dwight T. Hoxie 7

U.S. GeologialSurvey Denver, Colorado 80225

13.1 INTRODUCTION

ated by emplaced waste can be regarded to be negli-

- ibly small. Conditions in the far field, therefore, are g

The term," natural analog," most frequently refers -

determined by the natural state and evolution of the to a natural setting m which materials, conditions, and processes are considered to be similar to those that repository geosphere environment.

may occur within an engineered structure or system Because moving groundwater is considered to be i

following the system's placement within the natural the primary transport mechanism for radionuclide l

environment. For example, the geochemistry of a release to the accessible environment from geologic

-naturally occumng uranium ore body may provide repository systems, characterization of the present insight into possible rates and mechanisms of ra-state and expected evolution of the geohydrologic dionuclide transport and retardation that could be system is of major concem. Geohydrologic condi-expected to occur near a geologic repository for the tions in the far field will control the flux of ground-disposal of high-level radioactive waste (HLW). The water reachm.g the repository environment (which, application of analog studies, however, need not be restricted to engineered systems; the concept of therefore, may be available to contact emplaced l

" natural analog" can be extended to include the study waste) as well as the flux of groundwater available l

of natural systems that possess characteristics and for radionuclide transport between the repository and features similar to those of a particular natural system the accessible environment. Changes in far-field geo-ofinterest. As with engineered systems, natural-ana-hydrologic conditions during a mandated repository l,

log studies for application to natural systems tend lifetime of 10,000 years or more are expected to occur primarily to be of heuristic value, allowing insight principally in response to future climatic change. The mto poorly understood processes and conditions that potential geohydrologic consequences that could be may be occurring within the system ofinterest or that may be expected to occur within the system in re-caused by climan.c change m. elude possibleincreases sponse to possible future changes of external environ-in the flux of groundwater reaching the repository and mental conditions. In this context, analog studies are shorter groundwater travel times from the repository applied widely in the natural sciences as part of the to the accessible environment. The significance of long-standing method of scientific induction. Natu-climatic change for waste containment and isolation ral-analog studies also can provide quantitative data will depend on the response characteristics of the by which to test and validate the application of con-geohydrologic system and are likely to be of greatest ceptual and numerical predictive models to a natural potential importance for repositories located in un-system of interest. Clearly, the value of natural-ana-saturated-zone environments. Unsatu-ated-zone sites log studies depends entically on the extent to which information and data are transferrable from the ana-are f major concern, because the United States re-log system to the system ofinterest.

Pository program is presently evaluating the suitabil-ity of unsaturated tuff at Yucca Mountain, Nevada,

. The present discussion considers the applicabih,ty as a host rock for the nation's first repository for of natural-analog studies to characterize far field con-ditions in the vicinity of geologic-repository systems.

commercially generated HLW. Consequently, dis-for the disposal of HLW. The far field is that region cussion here will be directed principally at the appli-within the geologic host medium that is sufficiently cation of natural-analog studies to characterize distant from the repository such that the immediate hydrologic far-field conditions for repositories sited effects of repository construction and the heat gener-within thick unsaturated-zone environments.

101 NUREG/CP-0147

Natural Analogsfor Far-field Environment / Hydrology 13.2 UNSATURATED ZONE FLOW 500 to 750 m thick and consists of a layered sequence AND TRANSPORT of unsaturated, variably fractured, welded and non-welded tuffs of low matrix porosity (generally rang-The quantitative description of groundwater flow ing from 10 to 30 percent). In spite of these m geologic media requires the development of con-differences, however, the present approach is based ceptual models for the flow processes and the trans-on the presumption that the Darcian soils-based lation of these models into an appropriate model is appropriate to describe moisture storage and mathematical framework to pemtit calculation of flow within the unsaturated rock matrix. In this con-flow rates and directions. The putative theory of text, thin unsaturated zones in soils constitute a fun-groundwater flow m saturated media derives from a damental analog for thick unsaturated zones in l

Darcian continuum model of flow. This model is layered rock. Specific issues that remain unresolved l

based on Darcy's law, which states that the flux of include (i) the validity of the macroscopic continuum groundwater at any location within the flow region is model for charactertzmg the possibly complexly in-proponional to the gradient of the hydraulic head at terconnected pore space in unsaturated rock, (ii) that location. With specific regard to characterizing proper accounting for the interaction and coupling of the moisture balance m soils, the Darcian continuum liquid-water flow in the rock matrix and that possibly theory has been generalized and extended for appli-occurring in open fmetures, (iii) incorporating the cation to characterizing the storage and movement of effects of spatial heterogeneity that could lead to the water m unsaturated media. The underlying concep-occurrence of preferential flow pathways and local-tual model is based on the hypotheses (i) that the zed regions of enhanced water saturation, and (iv) medium can be represented as a macroscopic contm-allowing explicitly for high-intensity transient events uum of interconnected pores and (ii) that capillary involving phenomena such as rapid two-phase flow and adsorption forces under gravity control water and wetting-front instabilities. In order to resolve storage and flow within the pores. Because the pores these issues, new theoretical and modeling ap-are occupied by both liquid-phase and gas-phase proaches may be required, and natural systems dis-i fluids, the resulting Darcian flow equation, unlike its playing some or all of these features can be expected saturated-zone counterpart,is highly mathematically to provide analogs for developing and testing new nonlinear, which makes the definition and measure-and refined models.

ment of hydrologic properties difficult in unsaturated In addition to the difficulties m characterizing the media and often poses problems for conceptual and fl W f groundwater tn unsaturated geologic media, numerical model validation in specific applications.

the generally accepted concepts of solute dissolution An additional complication arises under transient and transport as they apply to saturated conditions flow conditions when the dynamics of the gas phase m y n t be straightforwardly applicable to unsatu-within the air-filled pores may influence liquid-phase rated environments. For example, the continuum ap-conditions and flow.

proach may not be appropriate in the presence of Although the Darcian model for groundwater flow geologic heterogeneity that may occur over a range has been successfully applied to both field and labo-of scales and may lead to preferential flow and trans-ratory studies in hydrophilic soils, the direct applica.

pon pathways. The generalconcepts of" retardation" bility of the model to unsaturated-zone environments and " dispersion" may take on new meanings in un-in general has yet to be adequately demonstrated. Of saturated-zone settings. Funhermore, the presence of particular concern for geologie repositories located ia air within unsaturated transport pathways will need thick unsaturated zones is the characterization of to be taken into account not only as it affects the moisture flow in variably saturated, spatially hetero-liquid-water flow field but also as a medium for geneous indurated rocks, such as occur at Yucca potential gas-phase radionuclide transport and as a Mountain, Nevada. The major problems in extending chemical reactant along the transport pathways.

the soils-based model concern scale and composition.

Unsaturated zones in soils typically do not exceed a 13.3 COUPLED FLOW AND few meters m thickness, and soils generally are high-TRANSPORT PROCESSES porosity (3040 pen:ent) media composed of uncon-solidated materials. The unsaturated zone at the Under the most general set of circumstances in the Yucca Mountain site, on the other hand, ranges from far field, coupled processes involving simultaneous NUREG/CP-0147 102

Natural Analogsfor Far-field Environment / Hydrology heat and moisture flow and solute transport in the sidered as long as unsaturated conditions are main-presence of chemical reactions will need to be con-tained in the far field. Consequently, as long as un-sidered. Heat flow will occur in the far field in re-saturated conditions prevail in the far field and sponse to the natural geothennal gradient and may significant fast flow and transport pathways are ab-consist of both conductive and convective compo-sent, equilibrium chemistry under fixed or only nents. In nonisothermal unsaturated-zone environ-slowly varying ambient conditions may suffice to ments, the moisture distribution generally will be characterize the dommant retardation processes.

determined both by liquid-phase flow and advective gas-phase water-vapor transport. Radionuclide disso-13.4 CLIhMTIC CHANGE lution and transport will depend principally on the magnitude and direction of liquid-water flux through Because climate ultimately determines the amount the repository system; although gas-phase transport of water available to infiltrate into the unsaturated i

of some radionuclides, such as carbon-14, may occur zone, future climatic change may pose significant within the air filled pore space. The inclusion of consequences for waste containment and isolation in coupled interactive processes introduces additional repositories sited in thick unsaturated zones in arid complexities, which, however, may be most severe in regions, such as the southwestern United States. Of the near-field repository environment where the ef-major concern is the possible occurrence of signifi-fects of heat generated by emplaced waste may lead cantly wetter conditio is of sufficiently long duration to strong thermal gradients and elevated ambient to increase groundwater flux in the unsaturated zone temperatures. Although they may not be directly ap-c.nd, thereby, to increase the amount of water avail-plicable }o far-field unsaturated-zone environments, able to contact emplaced waste and to transport ra-hydrocarbon and geothermal resource evaluation dionuclides in solution. In particular, depending on problems may provide insights and bounds on condi-the specific geohydrologic setting, increased net in-tions in the repository far field. Petroleum engineer-filtration could ir_itiate water flow in fast pathways, ing is concerned with the general problem of such as fractures and faults, that could reduce appre-multi-phase fluid transport in diverse geologic media ciably the time required for water to reach the reposi-l and settings. Geothermal-resource evaluation is con-tory from land surface, as well as the groundwater i

cerned with multi-phase moisture storage and flow travel time from the repository to the accessible en-i under strongly nonisothermal conditions. In addition vironment.

to their possible heuristic value as analog systems, both of these fields have contributed significantly to In order to resolve the concerns posed by the the development of field and laboratory techniques possibility of future climatic change, information is and numerical-modeling methods that will be appli-needed regarding not only the likelihood for the oc-cable to the problem of repository siting in unsatu-currence of climatic change during thelifetime of the rated-zone environments.

repository but also the expected effects on repository performance that could be induced by climatic The principal geochemical issue with respect to change. Because moving groundwater is expected to radioactive-waste isolation in geologic media con-be the primary mechanism for radionuclide release cerns the potential for radionuclide retardation along and transport, the effects ofclimatic change on waste the transport pathways. In the low-temperature, far-containment and isolation will be mediated by the site field environment, the identification and charac-geohydrologic system. In unsaturated-zone systems, terization of the relevant retardation processes is climatic change will be translated into changes in the straightforward, at least in principle. The major distributions of water content and groundwater flux source of uncertamty derives from the unknown con-within the system. The consequences of climatic J

sequences for radionuclide retardation of possible change for waste contamment and isolation will de-future changes in far-field environmental conditions pend on the magnitude of change and the time re-that could lead, for example, to in situ mineral altera-quired for the effects of change to propagate through tion or to the creation of new transport pathways, the system. Consequently, in order to evaluate the Because of the expected long residence time of effects of future climatic change on repository per-groundwater in an unsaturated rock matrix, the kinet-formance, site-specific models and data will be ics of retardation processes will not need to be con-needed to cuable quantitative prediction of the re-103 NUREG/CP-0147

N2 ural Analogsfor Far-field Environment / Hydrology sponse of the geohydrologic system to specified sent-day climatic environments can provide baseline changes in climatic conditions.

data for qualitative, if not quantitative, extrapolation Because the factors controlling long-tean climatic of geohydrologic-system response to possible future evolution remain incompletely understood and quan, climatic change. The principal geohydrologic-system tified, reliable prediction of future climatic change at response to climatic change will include changes in the rainfall-runoff and groundwater recharge-dis-either global or regional scales is not yet feasible.

Consequently, based on the premise that the past is charge relations as well as possible change in water-the key to the future, the evidence for past climatic table or potentiometric surface altitudes. These conditions is taken to be at least indicative of possible geobydrologic-system variables and their response to climate change will depend complexly on both the future climatic variability and extremes. In this sense, past climate is regarded to be a natural analog for 8,cohydmlogic setting and existing climatic condi-tions.

future climate. The expectation that past climate can serve as a guide to future climate is based, in part, on the evidence for cyclic climatic variability that is 13.5 MODEL VALIDATION implied by presently available data. The data indi-Numerical models for natural or engineered sys-cate, for example, that glacials and pluvials recur with tems are developed in order to quantitatively describe periods of several thousands of years. Climate data the state and predict the evolution of these systems in for the past 30,000 years m, dicate that shorter-tenn response to changing extemal conditions. Model climatic fluctuations of variable amplitude are super-validation consists of demonstrating that the numeri-unposed on the long-term climatic cycles.

cal model and its underlying conceptual model pro-Although the evidence for past climatic change vides an adequate and appropriate representation of indicates that future climatic change occurring within the system. In general, models can be validated only the lifetime of a radioactive-waste repository is for a particular system under a known set of condi-likely, the effects of climatic change on a geohy-tions. Models to be implemented to predict changing drologic system and, especially, the consequences for far-field conditions near a radioactive-waste reposi-waste containment and isolation, cannot be evaluated tory during a 10,000-year or longer repository life-directly. Consequently, both palcohydrologic studies time, however, cannot be validated directly.

and studies of geohydrologic systems in a variety of Consequently, such models must be indirectly vali-different climatic settings can be used to estimae the dated by demonstrating their ability to adequately effects of future climatic change on a particular geo-represent a variety of similar, or analog, systems hydrologic system. The correlation of past climatic subject to differing sets of known external conditions.

change with the palcohydmlogic record provides in-By proceeding in this manner, the limits of model sight into the magnitude and timing of the effects applicability can be established and confidence de-produced by climatic change on geohydrology. The veloped that the model is appropriate for use within l

study of systems in different climatic conditions pro-these limits. Therefore, natural-analog systems are vides insight into the end states produced by a spccific expected to be of indispensable use as part of the climatic setting. The combination of paleoclimatic overall model-validation process, especially for those and palcohydrologic studies together with the study models invoked to predict the long-term performance of analog geohydrologic systems in a variety of pre-of a repository system and its environment.

NUREG/CP-0147 104

m WASTE FORMS, PACKAGES, AND SEALS WORKING g

GROUP

SUMMARY

l Narasi Sridhar (Coordinator)

Michael B. McNeil (Coordinator) i Centerfor Nuciarr Waste Regulatory Analyses U.S. Nuclear Regulatory Commission San Antonio, Texas 78228 Washington, D.C. 20555

Participants:

Rod Ewing (Specialist); Willis Clarke; Bob Izusch, Jim Mazer; Richard Codell, Gustavo Cragnolino, Hersh Manaktala,and Jack Parry 1

l 14.1 GENERAL OBSERVATIONS 14.2 USE OF ANALOGUES FOR I

WASTE FORMS l

The group decided to limit the discussion to waste l

forms and waste packages. The scope of the discus-W hsyes hsd & respect to spent kel anal guesinclude&

sions involved both natural and anthropogenic ana-l e the need to develop a better understandmg of logues. The majority of the group felt that research m.

I the crystal chemistry, phase equilibria, and min-the United States on natural and anthropogenic ana-eralogy of actinide oxide and silicate phases in i

logues has lacked stable funding, with a consequent order to project behavior of spent fuel in terms l

destruction of institutional memory. The group felt of changes in the alteration mechanisms over l

that a stable, regularly assembled body of scientists, time.

perhaps organized under the aegis of a national tech-e the need to understand the relationship between nical society, such as Materials Research Society alteration reactions of natural urarunite and l

(MRS), should exist in the future. As an initial task, oeenninerals to sose of spent fuels.

the need to understand the effects of environ-e the group recommended that a comprehensive litera-mental factors on alteration mechanisms.

l ture survey of natural and anthropogenic systems /re-l actions relevant to alteration of waste glass and spent 14.3 USE OF ANALOGUES FOR fuel, degradation mechanisms of container materials, CONTAINER MATERIALS and mechanisms of alterations of chemical battier Copper has been examined in the Swedish and the backfill should be prepared. There was a consensus United States program using both archeological and l

that issues in analogues should be considered in con-natural analogues. Iron has been examined using nection with model verification and performance as-anthropogenic analogues, such as Roman nails. Po-sessment. In contrast, analogs cannot generally be tential uses of analogues include:

exercise models of corrosion product mineral-e regarded as simulations of waste package perform, ance in a repository.

By and morphology, maximum corrosion depth (if possible), and existence of various The engineered barrier system may be classified corrosion mechanisms against observations on l

as shown in Figure 141. It is apparent from Figure the analogue.

14-1 that the approach to' analogues for addressing

• exammation of materials that can be plausibly containment issues would be different from that for argued to have existed in environments similar release rate. There seems to be a longer history of to the repository environments, but to have been 1 wholly reacted.

research into analogues of vitrified waste and spent fuel than metallic container materials. While some While copper and iron may be natural choices for an analogue program because of their long history container materials, such as copper, lend themselves and, in the former case, because of its presence in to the study of natural analogues, for materials, such certain geologic formations in the native state, more as stainless steels, anthropogenic analogues of more recent (30-100 years old) analogues of these and recent origin (50-75 years old) should be considered.

other alloys should not be ignored.

105 NUREG/CP-0147

l D

E rn g

ENGINEERED y

BARRIERS

{y 9

t g

l I

r r

o NON-METALLIC METALLIC VITRIFIED SPENT A

COMPONENTS COMPONENTS WASTE FUEL

[

k B

m CONTAINER OR

FILLERS, CLADDING FUEL COATING BACKFILL ETC.

u e

I f I f CONTAINMENT AND RELEASE RATE RELEASE RATE ASSESSMENT ASSESSMENT Figure 14-1. Classification of the engineered barrier systems

NEAR-FIELD ENVIRONMENT / PROCESSES WORKING 15 GROUP

SUMMARY

William M. Murphy (Coordinator)

Centerfor Nuciar Waste Regulatory Analyses San Antonio, Texas 78228

Participants:

John Bmdbury (Coordinator), John Apps (Specialist), Robert Finch, Ronald Green, Kenneth Krupka, Brian Marshall, Todd Rasmussen, and Harold Wollenberg l

15.1 GENERAL OBSERVATIONS should be odented to obtain results that can contribute t Pufonnance assasments. Rdnant dg data Several themes generally applicable to natural analog studies were developed in the course of work-may c nsist fidentified processes or properties and their coupling, or they may provide bounds or exam-ing group discussions. A consensus prevailed that ples for reasonable system behaviors. Results are there exists a compelling philosophical rationale for likely to be both qualitative and quantitative. The

- natural analog studies in support of the geologic results of analog studies should be transmitted to disposal of high-level nuclear waste (HLW) largely performance assessment modelers.

because the long time scale of concern exceeds direct laboratory or human experience. Phenomena relevant 15.2 THE NFAR-FIELD to long-term isolation of nuclear waste can be identi-ENVIRONMENT: ISSUES AND fled m natural systems. Analogous natural (and ar-PROCESSES cheological) systems pmvide the only means to study some problems or issues that are truly inaccessible in The near-field environment is generally defined as

. laboratory studies. A broad view of natural analogs the region significantly affected by thermal pertmb-was advocated. Many subjects can be addressed in tions associated with the HLW and by construction natural analog studies, and a variety of approaches and the introduction of exotic materials. Thermal may be appropriate. Literature reviews to discover effects on the geochemical, hydrological, and me-reports of natural occurrences of predicted or specu-chanical properties of repository components are is-lated repository materials or processes can constitute sues to be addressed in natural analog studies relevant a valuable analog study. Research can be conducted to the near field, with regard to an estimation and i

at scales that range from microscopic exammations evaluation of the waste isolation characteristics of the of processes and properties of materials to multidis-system. The consequences of the introduction of ex-ciplinary and integrated studies on a geologic scale, otic materials, for example, the effects of the geologic Much research in the geosciences employs reasoning setting on the materials, and the effects of these by analogy. Recognition of this research and experi-materials on the geologic environment, are also near-ence (e.g., methods and successes) can benefit appli-field analog issues. Coupling of thermal, chemical, cations of natural analog studies to nuclear waste hydrological, and mechanical processes is an impor-issues.

tant characteristic of the near-field environment that A challenge for natural analog studies is integra.

may be addressed by study of analogous natural 2

tion with system modeling and performance assess.

systems. Particularly in the near-field environment, ment. Confidence in the safety of geologic systems the coupling of performance-related processes neces-for disposal of nuclear waste requires an identifica.

sitates interdisciplinary studies.

tion and understanding of significant processes in the Chemical reactions can be promoted by near-field natural and engineered systems and the use of that temperature changes due either to changes in equilib-knowledge to evaluate (i.e., model) repository per-rium conditions or increases in reaction rates. Tlese j

formance. Two objectives of analog studies are to reactions can affect solute concentrations (e.g., by I

develop and test hypotheses and to attempt to validate dissolution or evaporation) and speciation, precipita-both conceptual and numerical models. Research tion or dissolution of minerals, changes in mineral 107 NUREG/CP-0147

Near-Field Ent?ironmenVProcesses Working Group Summary compositions or sorption properties, and the genera-15.3 NEAR-FIELD ANALOGS tion and composition of a gas phase (e.g., by volatili-Having developed a sense of the near-field system zation or consumption of dissolved gases or H2O).

and significant processes and issues, the group as-Geologic environments are typx: ally metastable, and sembled a list of potential near-field analog systems.

thermally activated reactions producing more stable These systems were then divided in six general cate-phases can be anticipated in the near field. The nor-godes, wis mme subcategodes, based on seir natu-mal metastability of natural systems over geologic ral(e.g., geologic) occurrences, time suggests that local chemical equd, ibrium would 1.

Hydrothermal systems not be expected to be achieved during transient near-A. Fossil field thermal perturbations. Particularly significant B. Active near field analog studies could be devoted to reaction C. At elevated temrifiurcs along the geo-

. rates at relevant elevated temperatures over periods thennal gradient at the repository site or that are long relative to laboratory accessible time in an analogous seuing scales, but nevertheless short relative to most ge*

2.

Contact zones around dikes, sills, or plutons logic systems (e.g., tens to thousands of years).

3.

The natural reactors at Oklo 4.

Shallow earth systems under the influence of The chenustry of radioelements can be affected by diumal or seasonal temperature variations near-field phenomena. Solubilities of radioelement-5.

Natural redox fronts, natural redox processes containing solids, ion exchange equilibria and kinet-6.

Special mineral and material deposits ics, coprecipitation relations, colloid generation, and A. Ore deposits other processes controlling radioelement concentra.

B. Cement (e.g., portlandite) deposits tions are temperature dependent. In addition, the oxi.

dation state of the system can influence radionuclide C. Meteorimpact sites concentrations in solution. Variations in the redox D. Archaeologicalmaterials 7.

Modern acalogs:

conditions in the near field can be a consequence of A. Nuclear device test sites reactions promoted by temperature variations, by B. In sn, u vitrification sites introduced materials, or by the thermobydrological C. Conramination, waste, and test sites as-phenomena (i.e., coupled heat transfer and fluid sociated with nuclear research and tech.

flow).

nology D. Biological effects on metallic well cas-Near field thermal loading can cause thermal con-ings vection of gas and liquids. In unsaturated media with E. Mine leaching, mineral beneficiation two fluid phases, the distribution of fluids can be sites strongly altered leading to phenomena such as a F. Modem materials behavior m geologic liquid saturation halo, capillary flow due to saturation envimaments gradients, or to thermal heat pipe effects (i.e., counter A matrix was developed relating the issues and current flow of gas and liquid). In addition, osmotic flow can be induced in the near field by thermally processes to the outline of potential natural analogs (Table 15-1). The relations in this matrix are not driven changes in solution concentrations, for exam-intended to be exhaustive or absolute.

pie by vaporization or water-rock reactions.

15.4 FOCUS: HYDROTHERMAL The near-field environment can provide an energy SYSTEMS ANALOGS OF source for microbial activity either through radioac-THE NEAR FIELD tive decay or irreversible reactions of repository ma.

terials, for example, corrosion. Microbial activity can The group chose to evaluate hydrothermal systems have important effects on such processes as corro-in greater detail as natural analogs of the near-field sion, colloid formation, and complexation of aqueous environment and processes. Many active and fossil species with organic ligands. Microbes can generate hydrothermal systems have been studied in the con-local chemical environments with ramifications for text of geothermal energy resources and hydrother-

' radionuclide isolation. In addition, biomass genera-mal ore deposits.These systems generally record the tion can affect hydrologic characteristics.

effects of interactions of hot water with rock. What NUREG/CP-0147 108

Nar-Field EnvironmenVPrccesses Working Group Summa.y Table 15-1. Matrix of nuclear waste repository near fleid issues (Section 15.2 of text) and analog systems that may be used to address the issues (Section 153 of text). Numbers refer to elements of the outline of analogs given in Section 153 of the text.

IIydrothermal Igneous Natural Shallow Redox Mineral Modern Analoes Systems Contacts Reactor Systems Fronts Deposits Analoes i

Issues Thermal effects on chemistry 1A,IB,1C,ID 2

3 6A,6C 7B Thermal effects on mechanics lA,IB,IC,ID 2

3 6A,6C 7B Thermal effects on hydrology 1A,IB,IC,ID 2

3 4

7B Thermal effects on gases 1A,1B,1C 2

4 Redox telations lA,lB,lD 2

3 5

6A,6C,6D Radionuclide chemistry 3

5 6A,6C 7A,7C Microbial activity 5

6A 7D,7E Construction materials I

6B.6D 7F are the general characteristics of hydrothennal sys-havior, for example, acidic hydrothermal waters as-tems, and how do these characteristics compare to the sociated with magmatic activity, may not be repre-near field of a HLW repository? Hydrothermal sys-sentative of heated groundwater in a repository tems can be characterized according to the duration environment. Pervasive alteration in hydrothermal of hydrothermal activity, the fluid flux, the chemistry systems may leave a sparse record of many phenom-of hydrothermal solutions, fluxes of chemical com-ena relevant to relatively short-term alteration of the ponents through the altered rock, the heating rate and repository near field, integrated heat load, and gas phase evolution and The working group suggested that hydrothermal chemistry. These characteristics should be evaluated events of relatively short time and small space scales to establish the closeness of the analogy of particular bear the greatest analogy to the repository near field.

hydrothermal systems to expected characteristics of For example, distal sectors of active hydrothermal particular repository near-field environments, and, systems may be affected over relatively short time hence, their relevance to evaluations of repository periods. In the outflow zones of active caldera by-performance, drothermal systems, hydrothermal flow paths may In general, the thermal loading and near-field tem-have only recently transected country rock, and the perature evolution can be closely estimated for vari-solutions may be near-neutml to slightly alkaline ous repository designs. Good analogies for the effects waters of meteoric origin at 60' to 120 *C. Evidence of heat and fluid-rock interactions at elevated tem-for the effects of hydrothermal activity on spatial pentures in the near field can be provided by natural scales analogous to a repository may also be found hydroAennal systems. However, many hydrother-near igneous dikes or at the margins of larger by-mal systems or aspects of these systems may have drothermal systems.

little apparent relevance to a particular repository In the context of an unsaturated environment, ther-system. A number of general characteristics of hy-mohydrologic phenomena, such as a heat pipe, may drothermal systems may differ significantly from be significant in the near field. It would be of interest repository near-field environments. In pasticular, to examine natural environments and, particularly, heating in repository systems is of short duration hydrothermal systems for the occurrence or evidence relative to most natural hydrothermal systems. Also, for natural heat pipes and their consequences. Evi-total fluxes of water and chemical mass transfer are dence might be obtained regarding heat-pipe effects likely to be small in a repository relative to larger on redox phenomena, mineral precipitation, perme-hydrothermal systems. Solutions with aggressive be-ability changes, and heat-transfer mechanisms.

109 NUREG/CP-O147

FAR-FIELD ENVIRONMENT WORKING

]g GROUP

SUMMARY

English C.Pearcy (Coordinator)

Ralph E.Cady (Coordinator)

Centerfw Nuclaer Waste Regulatory Analyses U.S. Nuclar Regulatory Commission San Antonio, Texas 78228 Washington, D.C. 20S55 F.ui.u: Dxdght Hoxie (Specialist), Daoid Bish, Deirdre Bonk, Ptrul Cloke, June Fabryka-Martin, Philip Goodell, r

Grace Haggerty, James McCord, Arend Meijer, James Wood, Michael Shat, Budhi Sager, Dernd Turner,and Gordon Wittmeyer 16.1 GENERAL OBSERVATIONS manner? What is the potential for multiple benefits fmm the study?

It was generally agreed among the participants in the Far-field Working Group that transferability of Although this workshop was intended to cover the

' data will determine the usefulness of natural analog general applicability of analog research to the long-studies for development of a geologic high-level term geologic storage of HLW, much of the discus-waste (HLW) repository. The extent to which data sion focused on the proposed United States HLW from any one geologic site may reasonably be trans-repository at Yucca Mountain, Nevada. Conse-ferred for use at any other geologic site is not ade-quently, many of the considerations raised here are quately known. In practice, how does one use data site specific.

derived from an analogous system to establish with reasonable assurance that performance objectives 16.2 THE FAR-FIELD ENVIRONMENT will be metin a repository system?

i Selection criteria were recognized as the first and The working definition of "far field" was found to perhaps most important opportunity to improve the vary among the researchers present. No clear bound-I transferability of analog data to a repository system.

ary between the "near field" and the "far field" exists.

l Obviously, transferability is enhanced by having as The "near field" is generally considered to be that l

close a match between the analog and the system of volume surroundirg the waste which will experience interest as possible. The criteria poposed by Chap-significant thennal effects. In practice, "far field" man et al. (1984) emphasize the necesity of such a appears to encompass the volume beyond that range close comparison and were broadly affumed. How-as well as overlapping substantially with what may ever, in addition to having as close a match as possi-be considered in other forums to be the "near field."

ble, it is important that the question to be addressed by a given analog be well defined. A narrowly defined 16.3 FAR-FIELD PHENOMENA question that is resolvable at an analog site may provide useful (i.e., transferable) data even from an Many issues, processes, and events were identified l

analog site, which in other regards, differs substan-as potentially significant for the far-field environ-I tially from the repository conditions. This line of ment. Broadly, these can be considered to fall into reasoning is especially important for analog sites that two categories-hydrologic and geochemical. Of l

i provide a unique opportunity to test a contentious course, there is considerable overlap and feedback l

hypothesis (e.g., fission product geochemistry at between processes occurring within these two gen-l Oklo)-

eral categories. Nevertheless, such a division pro-In addition to considerations for data transferabil-vides a useful framework in which to consider ity, to be most useful within the context of the HLW relevant phenomena. This summary presents a subset program, selection criteria for an analog should con-of the identified hydrologic and geochemical phe-sider a variety of practical aspects. Can the study be nomena considered by the working group to be of completed using currently available technology? Is particular importance for the performance of a HLW l

the study likely to yield useful results in a timely repository.

111 NUREG/CP-0147

Ear-Field Enaronnunt Workmg Group Summary 163.1 Hydmlogic Processes to other sites. The kind clinvestigations and the level of characterization that will lead to a reconstruction 163.1.1 Fracture Connectivity, Fracture of the past effective flow behavior of a site remams Flow and Large-Scale Effective problematic. Pneumatic testing at an analog site, for Hydraulic Fmpues example, Apache leap, Arizona, may provide insight e PmPenies of, Mmeystem pmcesses.

The Yucca Mountain site is underlain by fractured' It may be possible to assess effective flow properties welded tuffs. 'Ibe degree to which the fracture sYs by performing a detailed mapping of the subsurface tenu are pm tatically and hydrologically connected distribution of several environmental tracers, both at has sigmficance for gas-phase radionuclide transport the proposed repository site and at analog site (s).

and the pessible presence and creation of fast liquid-Careful definition and limitation of the questions to phase flow pathways. Both the degree of fracture be addressed at any analog site are required; other-connectivity and the pneumatic and hydrologic sig-wise, characterization of the analog site may become nificance of interbedded unfractured, non-welded as nvolved as that of the proposed repository site.

tuffs is highly uncertam. On a large scale (space and time), an average moisture field, which is not neces-163.1.2 NetInfiltration sarily conservative from a performance assessment Net infiltration refers to the flux of water entering

. point of view (for unsaturated sites and flow field at saturated sites) develops. In an unsaturated site, de-the unsaturated zone below the plant-root zone and is pendmg on the stratigraphy and location of fault the primary source of water for downward gravity-zones, perched water tables may form.

driven percolation through the unsaturated zone.

Treatment of fracture paths is currently based on Present-day rates of net infiltration at the Yucca the hypothesis that fractures can be treated as con-Mountain site are considered to be small (<1 mm/yr tinua of distinct propenies. Mechanisms of fluid and averaged over the surface area of the site). Future contammant interchange between fractures and ma-rates of net infiltration in response to climatic change trix are not well understood. In modeling large-scale are highly uncertain at present. (Increased net infil-unsaturated flow through thick geological deposits, tration rates could create fast-flow pathways that the fine-scale flow behavior that actually occurs can-could transmit water rapidly to the repository horizon not be simulated by deterministic techniques. This is and subsequently to the water table.) Fast-flow path-because: (i) the entire heterogeneous media propeny ways also could cause the creation of perched-water bodies in the unsaturated zone.

field can not be practically sampled and tested, and (ii) even if one could completely define the media Future changes in net infiltration will depend on heterogeneities, the current generation of flow codes the effects of climatic change, specifically enhanced based on Ri: bards' equation are limited by their Precipitation, on the unsaturated-zone geohydrologic computauonal efficiency in obtaming a solution for system. Limits on future climatic change may be a deep water table problem (based on the fineness of inferred from past paleoclimatic change (i.e., the past the nu merical grid required). In recent years, effective may be an analog for the future). The effect of cli-(average) propenies which can be used for large.

matic change on the geohydrologic system tnay be scale, heterogeneous systems have been developed.

accessible to analog study at sites with sirmlar geo-Unfortunately, little experimental work has been per-logic settings but which are located in wetter environ-formed to assess the validity of these new effective ments than the Yucca Mountain site (e.g., Apache property models, particularly at the flow rates / time Leap, Arizona, and Rainier Mesa, Nevada). Direct scales relevant to the radioactive waste disposal transferability would be problematic, but relation-problem. Determmation of appropriate efSctive flow ships and processes might be transferrable.

properties will strongly affect prediction of flow rates and directions in the vadose zone.

163.2 Far-Field Heating An analog study may help in understanding this What effects will the heating of the host rock by a effective behavior if a method can be found to recon-repository (temperatures up to 100 *C are anticipated struct the flow history of the analog system. Such an in the far field) have on the far-field mineralogy, rock understanding may lead to basic hypotheses that are properties (e.g., rock strength, porosity, permeabil-independent of the analog system and can be applied ity), water composition, and degree of saturation?

NUREG/CP-0147 112

Earsield Environment Working Group Summary Because of the relatively low-temperature nature of limited by assumptions concerning available nutri-this process, laboratory expenmental data, hampered ents, as well as amounts and identities of microorgan-by metastability and kinetic problems, will probably isms.

be of limited use in understanding the effects. An It is very likely that microbes are present at most, appropriate natural analog may be ideal for determin-if no A natural analog sites. Appropriate analog ing some of the long-term effects although adequate sites sidi be able to be studied for the possible definition of past conditions will be challenging.

influenu c cny present microbes on other applicable Several technical issues are raised by the possibil-studied processes at that site. For example, microbial ity of far-field heating including: (i) Will the present-activity plays a key role in the transpon of uranium day minerals be altered (this is a particularly at the Po90s de Caldas natural analog site. Both sulfur imponant question for the unsaturated zone)? (ii) oxidizers and reducers are found to affect the oxida-Will the rock be " plugged" by mineral precipitation?

tion of pyrite, producing intermediate sulfur species (iii) Will an unsaturated repository be dried, inducing (polysulphide or colloidal sulfur), which, in tum, a saturation halo? (iv) Will the water composition form secondary pyrite and sulfate and possibly the change sufficiently to affect the stable mineral assem-reductant needed to form the observed pitchblende blage?

nodales. This mechanism is supported by both sulfur As discussed previously, it is unlikely that labora-isotope analysis of the secondary pyrite and the shape tory studies can answer these issues. Analog investi-of the pitchblende nodules. Due to the lack of ther-gations may be able to provide some answers, but modynamic and kinetic data, this mechanism cannot success is uncertain. Possible analog sites include as yet be analyzed quantitatively.

geothermal systems and the north end of Yucca Mountain itself. Additionally, steam flooding of oil 16333 Speciation and Solubility Limits fields such as the Kern River field could provide of Key Radionuclides imponant information.

Speciation and solubility limits for key radionu-clides are critical aspects for considering repository 1633 Geochenu. cal Processes performance and remain poorly defined. Speciation and solubility limits will determine the forms and 1633.1 Colloidal Transport of amounts of hazardous radionuclides available for Radr.onuclides transpon out of a repository. The complexes formed At present, it is highly uncertain whether or not by the radionuclides will vary in composition and colloids will form and be transported through a Yucca stability according to local groundwater conditions.

Mountain repository flow system. Colloid transport The secondary geochemistry of the waste form con-could be a significant mechanism for the transpon of stituents (i.e., the minerals formed after corrosive elements such as Am, Pu, Cm, Th, etc. Colloid trans-oxidation and dissolution of the waste form and the pon has been proposed as the transport mechanism compositions of the solutions in contact with those for Am and Pu at several U.S. Department of Energy minerals) as they move into the far-field will control (DOE) sites. Colloid transport is addressable through the concentrations of radionuclides in solutions that analog research, and has been studied previously may move away from the repository area into the

[e.g., Pogos de Caldas, Brazil (Mickely et al.,1990),

accessible environment. Table 16-1 provides an out-Alligator Rivers, Australia (Edghill,1988)]. Mean-line of present knowledge.

ingful measurements are difficult, and extrapolation Potential analogs for these radionuclides include of results to a repository system is uncertain.

uranium ore deposits, global fallout from nuclear testing, radionuclide waste disposal sites, and under-1633.2 Microbial Activity ground nuclear test sites. Speciation and solubility of Microorganisms may have a significant effect on the species !isted above (as for others) depends criti-radionuclide transport by either directly changing the cally on the camposition of the groundwaters in the oxidation state of a given radionuclide, or indirectly, analog environoeat. Extrapolation of speciation and by maintaining a steady-state system of mineral solubility measureteents made at an analog site to phases that sorb radionuclides. The imponance of anticipated repository conditions requires judicious such processes is uncertain; present understanding is selection of analog systems and thoughtful applica-113 NUREG/CP-0147

Far-Field Environment WorMng Group Sumrnary Table 161. Speciation and solubility limits of key radionuclides: present knowledge Ievelof Uncertamtyin Thermodynamic Databases Availability of Suitable Species and in Tr=W Models Need to Reduce Uncertainty Analor U

Moderate

_,ow High Pu High ivioderate Moderate Np Low Moderate IAw Tc bw kw Me I-129 Low Low Moderate tion. Such considerations provide the bases for rank-Organization. ANSTO AlligatorRivers Analogue ing the need for and likely usefulness of analog work Project-Progress Report, May 1988-August on individual species.

1988:245-252.

16.4 REFERENCES

Miekely, N., H. Coutinho de Jesus, C.L. Porto da Silveira, and C. Degueldre.1990. Chemical and Chapman, N.A., LG. McKinley and J.A.T. Smellie, physical characterization of suspended particles 1984. Thepotential ofnaturalanalogues in assess-and colloids in waters from the Osamu Utsumi ing systemsfor deep disposal of high-level radio ~

mine and Morro do Ferro analogue study sites, active waste. Swedish Nuclear Fuel and Waste Pogos de Caldas, B razil. Fourth Natural Analogue Management Co. (SKB) and the Swiss Radioac-Working Group Meeting, Pitlochry, Scotland,18

{

tive Waste Cooperative (NAGRA). NTB 84-41.

22 June 1990, Final Meeting Report Pre-Print.

{

KBS 84-16. Eir Ber 545.

B. C6me and N.A. Chapman, eds. Brussels: Com-Edghill, R.1988. Colloids in Koongarra groundwa-mission of the European Communities: n*EUR ter. Australian Nuclear Science and Technology 13014 EN.

i i

i t

4 1

l j

I NUREG/CP-0147 114

t

-VOLCANISMffECTONICS WORKING GROUP p

SUMMARY

i Linda A.Kovach (Coordinator)

Stephen R. Young (Coordinator) i

)

U.S. Nuclaer Regulatory Conmussion Centerfor Nuciatr Waste Regulatory Analyses Washington, D.C.20555 San Antonio, Texas 78228

-l

- Partiapents: Kenneth Foland, Don Gustafson, Renner Hofmann, William heman, Jim Luhr, Bruce Marsh, larry McKague, Bill Ott, John Russell, Ardyth Simmons Gerry Stireuelt, Engelbrecht von Tiesenhausen, and Brian Wernicke 17.1 GENERAL OBSERVATIONS style of Neogene and Quaternary (15 Ma to present) tect nic deformation and volcamsm. A less ambigu-

. The overallimpetus for studying geologic analogs us, more descriptive general tenn for such regions of Yucca Mountain is to improve definition and ts geologic analog. A more specific term may be analyses of geologic hazards and to improve esti-appmpriate for focused studies of certain processes, mates of risk for repository design, performance as-such as volcanic analog or structural analog.

sessment, site characterization, and review of a License application. In_ particular, analog studies Use of a properly chosen geologic analog may should be directed toward: (i) reduction of key tech-Provide added information and additionalinsight into.

nical uncertamties (KTUs) related to site charac-the critical dimensions of depth and time that are only terization; and (ii) refinement of constramts on initial indirectly available through investigations restricted and boundary conditions important for modeling and to YM. That is, the subsurface at YM can only be simulation of waste-isolation performance.

studied induectly. Borehole data (i.e., core samples

. Analog regions should be treated as models, essen.

and analyses, and geophysical logs), seismic surveys tially as one geological system used to represent and geophysical potential field surveys at YM may another. Analogs should be used to improve under.

not yield models of the deep subsurface that are standing of processes and conditions at YuccaMoun.

complete enough for reliable hazard assessment. De-tain (YM) that cannot be investigated directly by site velopment of conceptual tectonic models of YM and characterization. Studies of analog sites are likely to subsequent estimates of related natural hazards de-suggest alternative methods and directions of site Pend on interpretations of geologic structures that characterization by identifying important processes exist as deep as 10-20 kms below the surface at YM.

not directly observable at YM. Analogs provide im.

Analog regions that have been uplifted and eroded portant additional constramts on initial and boundary offer a more ducct view of deeper levels of geologic stmetures sinlar to those at YM.

conditions of mathematical models and laboratory experiments. Investigations and analyses of analog Assessment of risk related to cuthquakes, fault regions are generally considered to be indispensable rupture, and volcame eruption requires analyses of 4

as field-scale experiments to explain and predict vol-the time and space patterns of these processes. Time L

canic and tectonic processes and to assess related series and spatial distributions cannot be reliably waste isolation hazards. This is of particular impor-determmed from only a few events. Probabilistic 1

tance because time-space scales of geological proc-methods especially are degraded by lack of critical esses are difficult to simulate under laboratory data. Even a comprehensive site characterization pro-conditions.

gram may not yield the necessary data; it simply may In general, for crustal-scale (>10 kms') geologic not be available at the site. Indeed, data required to 2

l processes, a natural analog is a selected part of the formulate probability distribution functions for risk earth's crust that is similar to another part of particular assessment models may only be avadable from a interest. More specifically, natural analogs are range of analog regions. In pamcular, studies of mod-i-

viewed here as geographic regions or geologic set-em analogs are essential for charactenzation of tran-

. tings that are substantially similar to YM in age and sitory dynamic processes, such as earthquake and j

115 NUREG/CP-0147

.l.-

Volcamsm/ Tectonics Working Group Summary volcanic eruptions. For example, the probability of the specific problem being addressed by the analog large earthquakes (>6.5) at YM cannot be determined study, some of these conditions may be substantially from the modern seismicity record of faults at YM.

more important than the others.

There have been no large historical earthquakes at General recommendations are: (i) take as quanti-YM that directly correlate to faults with known sur-tative an approach as possible; (ii) structure investi-face traces. Thus, one interpretation is that YM may gations to answer specific questions; (iii) select be essentially aseismic. A reasonable altemative in-analog areas carefully, based on def mition of specific terpretation is that the recurrence period of large problems; (iv) use analogs to develop probability earthquakes at YM may be longer than about 200 distributions for important parameters that are not years, a real possibility in this area. A promising measurable at YM; (v) thoroughly document how partial solution to this problem is to supplement the potential analogs are similar to and different from local seismic record at YM by exammmg the historic YM, and thoroughly consider how differences may and palcoseismic record for similar structures else-bias results.

where within the Basin and Range (BR) region. This approach essentially substitutes a wide spatial sam-17.2 MAGMATIC / VOLCANIC pling (i.e., all of the B R region) for the short temporal PROCESSES record at a specific site (YM). Seismicity records Studies of geologic analogs are important to obtain from potential analog regions must be used carefully, however, because structural style influences scis.

cntical data and gain new insight mto assessment of micity. An important concern is, then, proper defini, hazards related to volcanic processes and in assess-tion of criteria used to determine structural and ment of repository risk. Hazard assessment consists volcanic similarity of potential geologic analogs of f two basic parts: (i) identification of potential yg, sources and (ii) estimation of probabilities of occur-rence. Risk assessment combines hazard assessment Similar problems exist in risk assessment related with estimates of potential effects on a repository.

to disruption of a proposed repository due to potential Site characterization may be adequate to determine future volcanic activity in the vicinity of YM. Data potential sources of eruption (e.g., existing vents and available at the site may not be sufficient to formulate vent alignments, fault zones), but it is not sufficient reliable probability distributions of volcanic disrup-to obtain data to reliably estimate probabilities of tion. Adequate data on eruptive styles and recurrence eruption (eruption recurrence, characteristic peri-periods of mafic vents similar to Lathrop Wells can odicities). There are simply too few eruptive episodes be acquired by studying a suite of geologically analo-and too much uncertainty about age of eruption to gous areas. Certainly, the potential physical, chemi-clearly establish eruption recurrence periods at YM.

cal, and hydrological effects of an eruption similar t Thus, adequate data on pattems of spatial distribution i

Lathrop Wells can only be studied in regions where and timing of volcanic vents is only available by l

modem analogs of the Lathrop Wells cone exist.

investigations conducted at a suite of analog sites.

)

Conditions which should be examined to properly Analyses of potential effects of an eruption on a compare, and to establish some measure of the simi-repository (probability of dismption) should include larity of a volcanic / tectonic analog include:(i)in siru observations and analyses of modern analogs. Stud-stress state; (ii) age of deformation; (iii) structural ies of active vents and Holocene vent systems are style (type of deformation);(iv) age of volcanism;(v) likely to yield information that can be used to estimate eruptive style, composition. source depth, and ascent direct and indirect dynamic effects of intmsion and history (type of volcanism); (vi) finite strain state and eruption on repository systems. Because unsaturated strain rates, local and regional displacement fields, conditions are an important part of the natural barrier fault-slip histories; (vii) surface and subsurface hy-system, studies of the physical and thermal effects of drologic and hydrogeologic conditions; (viii) crustal volcanic activity on local water table elevation and stratigraphic sequence (rock types) and thickness, groundwater flow pattems may be particularly im-and associated physical (e.g., gravity, magnetic, and portant at YM. Historicalinformation on pre-eruption velocity structure) and;(ix) chemical properties [ rare water table elevations may be available from areas of earth element (REE) and isotopic compositions] per-recent basaltic cone eruptions, such as Paricudn tinent to volcanic / tectonic processes. Depending on (Mexico).

NUREG/CP-0147 116

%Icanism/Teaonics Working Group sumnary Significant uncertainties remain in estimates of In general, does a lack of modern extmsive volcan-ultimate source depth, magma ascent rates and ascent ism indicate a lack of subsurface magmatism? Does pathways. Data on patterns and rates of magma ascent this hold for YM?

are critical to explain the occurrence and charac-

• How can the magnitude of potential magmatic teristics of eruptive systems, to predict eruptions, and and volcanic processes be determined?

to estimate probabilities of repository disruption.

• How will magma, magmatic fluids, and addi-However, adequate data on these processes are not tional heat alter the near-field environment?

available from a single locale. Rather, multiple ana-How might these changes influence the chemi-log sites, that encompass a range of models, will be cal and physical stability of canister materials, required.

waste materials, and the potential repository Important questions and issues related to volcan-host rock (Paintbmsh tuff)?

ism which require investigations at analog sites in-KTUs reflected in recent research at Lathrop Wells clude:

include: (i) ages of eruptions; (ii) style, duration, and

• Estimates of suberustal flux of material:(i)Can cyclicity of eruption (e.g., mon > versus polyge-useful estimates be made of the volumes and netic); (iii) effects on groundwater flow and water rates of magma accumulation in the lower crust table elevation, and influence of water table elevation and upper mantle? (ii) Can the Quatemary erup.

on eruptive style; (iv) relationship to existing faults; tion history of BR basaltic systems be related to (V) Potential for coupling of intmsive/ eruptive proc-magma flux?

esses and fault slip; (vi) potential for earthquake-trig-gered cruptions. These issues need to be explicitly

• What is the flux of magma into the nu.dd" and considered to properly formulate probabilistic esti-upper crust? How effective are seismic t*

mates of future volcanic activity and repository dis-mographic and heat flow analyses for determin-mption.

ing instantaneous flux?

Phreatic eruptions are of particular importance

• How complete are models of BR magmatism?

because of potentially widespread effects on both How well do existing models represent coupled near-field and far-field conditions. Interaction of pre-thermal, mechanical, and magmatic processes, existing or synchronous fractures or faults with as-

• Are models of basaltic dike intrusion and cone cending magma and groundwater may result in eruption at YM adequate for reliable estimates highly localized conduits for gas / fluid exchange and of volcanic hazard and consequence.

shallow convection. Physical barrien or fluid-flow pathways may be created by emplacement of dikes.

• What are potential effects of lithospheric con-ditions on magma transport and storage? For 17.2.1 Potential MagmaticNolcanic example, what are relationships of extant stress hal gs and strain states, and finite tectonic extension (strongly extended versus less extended re-Pmposed enteria for selecting Quatemary basaltic gions), on evolution of small basaltic systems, volcanic analogs of YM/ Crater Flat include: (i) evi-l as well as the large Neogene silicic caldera dence of hydrovolcanic processes; (ii) approximately systems.

Hawalite composition-somewhat more differenti-ated; (iii) geochronologic analyses comparable to

• Is magma stored in the crust? Does magma pool Lathmp Wells cone-sufficient isotopic and radio-in a mid-cmstal chamber, or is the source in the genic dates should be available to establish the age of mantle, with no pooling? How are feeder dike the analog; (iv) sinular Neogene sibcic eruptive his.

systems for basaltic vents localized at upper t ry(?); and (v) similar Neogene and Quatemary de-mantle / lower crust depths (30-50 kms)? How f tmati n history, are potential instabilities that localize magma ascent related to extensional strain? Can seis.

Proposed magmatic / volcanic analogs include:

mic tomography be used to detect magma Lumar Crater-Reveille Range (LC-RR). The LC-chambers? How can seismic tomographic data RR volcanic field may be a good analog site for be used to improve models of magma storage investigating the influence of faults on vent location and ascent?

and alignment. Caution should be exercised initially, 117 NUREG/CP-0147

Velnmism/ Tectonics Workmg Group Summary however, until additional studies show how similar drothermal alteration of the surrounding rock will the Reveille Range bounding fault is to faults at YM.

contribute substantially to development of perform-Black Mountains (eastern Death Valley). The ance assessment scenarios of volcanic disruption.

Black Mountains area may offer a view of both silicic Heater tests conducted in G-tunnel, within the and basaltic magmatic plumbing systems.The Black

. Nevada Test Site, may provide useful information on Mountains are a relatively intact block of the upper the far-field effects of magmatic heatmg of the Paint-crust that has been rotated and uplifted by extensional brush Tuff. Additional analog sites will be consid-l faulting, and subsequently denuded by erosion. Im-ered, as needed, depending on emerging priorities in portant investigations that may be possible in the site characterization and performance assessment.

l Black Mountains include studies of patterns and spa-tial distributions of magmatic dike systems and rela-17.3 TECTONIC PROCESSES l

tionships of dikes to surface vents. These studies ar Investigations of structural / tectonic analogs are j

unportant to constrain groundwater flow models used important to predict and analyze a range of future to assess the hydrologic effects of potential magmatic potential deformation states at YM. Specifically, pre-processes at YM. Studies o. the Black Mountains diction of hazards at YM related to coupled earth-

' may provide unproved constraints on orientation' quake and fault-rupture processes is likely to depend spatial extent, spacing, and the detailed patterns of W1 of ishd aseinic dikes represented m groundwater flow simulations.

slip. Constraints on fault geometry, incremental slip, Cima; Coso. Studies of the Cima and Coso vol-and associated distributed deformation should be de-canic fields may provide additional information veloped from both site characterization and studies of about the influence of faults on vent locations and similar structures elsewhere in the BR region. For j

alignments. Studies of the regional tectonic setting of example, measurements of fault slip at YM suggest these fields, in conjunction with recent space geodetic that the system is extended only about 10-60 percent.

i data, may also provide insight into relationships be-Studies of analog systems that are more strongly tween finite strain and strain rate on locahzation of extendal (e.g., >100 percent in the Bullfrog Hills volcanism-area) may substantially improve predictions of poten-Erupting mafic cinder cones do not occur in the tial future deformation due to fault slip. Indeed, it may i

nonhem, primarily extensional BR region. Modern be viable to use multiple analog sites as a forward

}

mafic volcanoes in North America are restricted pri-modeling process to examine how deformation might marily to the Pacific magmatic are region of the progress from the current state at YM to a more i

southern BR. Thus, the modem analogs suggested extended condition. Analogs may also show how the here are within plate-tectonic domains of conver-fault system evolves. Current models include systems 3

gence and active subduction, rather than within the that evolve as active low-angle sliding surfaces, as j

broad continental extensional environment of the BR well as systems that evolve by high-angle cross-cut-

{

province. Locally, however, extensional deformation ting of pre-existing, uplifted, low-angle fault sur-j may characterize areas of active cinder cone eruption.

faces.

Proposed Holocene-modem basaltic eruptive ana-Important questions and issues related to tectonics logs mclude:

and structural deformation which would benefit from 3

Jorullo (Mexico), Paricutin (Mexico), Cerro Ne-investigations at analog sites include:

l gro (Nicaragua). Investigations of these modern ba-

• Do detachment systems evolve as low-angle saltic volcanoes are likely to improve existing faults, or do they evolve into low-angle faults?

concepts of eruptive style, and to provide important data on associated dynamic and hydrothermal proc-What role do high-angle fault elements play in 4

e the overall kinematic evolution of detachment 3

esses. Because magmatic volatile content strongly influences the dynamics of volcanic eruptions, com-systerns?

What surface-geological criteria can be used to

. parative analyses of volatile contents and eruptive e

styles at modem mafic vents may significantly im.

discriminate between high-angle (domino prove conceptual models of potential eruptions at style) fault systems and detachment systems?

YM. Studies of the extent, intensity and longevity of How do high-angle dip slip, low-angle dip slip, 4

diffuse degassing of magmatic volatiles and hy-and strike slip (transform) faults interact?

n NUREG/CP-0147 118

^

_ - ~

Volcanism / Tectonics Working Group Summary e What are the implications of fault geometry for effects of underground nuclear tests on tunnels at the earthquake seismic and fauh rupture hazards?

Nevada Test Site. Such studies may be particularly e What is the neotectonic deformation style at appropriate to predict near-field effects, such as seis-YM? Has the structural style at YM changed mic slip on the Solitario Canyon fault or the Paint-significantly over the last 10 million years?

brush-Stagecoach Road fault.

. How does tectonic extension influence the lo-cation and style of volcanism within the central 17.3.1 Potential Tectonic / Structural BR region? Do eruptive volumes correlate to Analogs firete extensional strain or to strain rate?

In practice, candidate structural analogs of YM How are extant or incipient faults influenced by may be restricted to the BR region. Structures are well e

intmsion of magmatic dikes or by eruption of exposed in the mountain ranges, and alternative Neo-vent systems?

gene structural styles applicable to YM are wide-Answers to these questions, and data necessary to spread within the region. Also, analogs from within develop and adequately constrain prediction of fumre the BR region are more likely to have a similar overall tectonic deformation at YM are not attainable solely Neogene and Quaternary geologic history, as well as by site characterization. Important data that is diffi-comparable in situ stress and strain conditions. How-cult to acquire at YM includes:

ever, depending on acquisition of additionalinforma-direct determination of subsurface fault geome-tion about deep fault geometry at YM (i.e., from e

tries reflection seismic surveys), regions such as the East direct detemunation of depths of potential de-Africa / Red Sea Rift system, the North Sea / Viking e

tachment faults Graben, or the Aegean may provide good faulting-slip and paleoseismic histories of faults.

process analogs.

e Investigations of analog regions is necessary t The YM region is structurally complex, and no supplement data acquired by site characterization to clear consensus has emerged on the detailed struc-estimate earthquake hazards and repository risk at tural style. Generally, YM is considered to be part of YM. Data that must be acquired from regions of an extensional system that includes both strike-slip historic seismic activity and from fault systems with and dip / oblique slip normal faults. Uncertainty exists a discernable paleoseismic record, preferably both, in describing how these different faults interact to include:

accommodate modern regional displacements and in rate and style of seismic strain release within the determining the relative mix of potential seismic and e

region that comprises the tectonic setting of YM aseismic slip. Debates about basic subsurface fault vibratory ground motion (shaking) geometry (planar versus listric) and slip history are e

ground accelerations common. However, YM is mapped in considerably l

e attenuation characteristics more detail than most areas that would qualify as e

e aftershockcharacteristics analogs. Thus, the structure of YM is uncertain, and length-area-magnitude relationships the structure of potential analog areas is likely to be e

segmentation even less certain. Consequently, the primary value of e

associated groundwater effects.

structural analogs is in the opportunity to study simi-e An appropriate approach to identify analog fault lar processes, rather than as "look-alike" models of YM.

systems includes use of existing studies, along with reconnaissance trenching and characterization stud-Proposed criteria for selectmg candidate tec-ies to determine the palcoseismic record at potential tonic / structural analogs of YM include:(i) late Neo-analog sites. In concert with these fault studies, ap-gene and Quaternary geologic history of tectonic propriate seismo-tectonic (i.e., earthquake) analogs extension and associated volcanism (generally satis-may be further defined by examming existing focal fled by potential analogs within the BR region); (ii) mechanism solutions and selectively re-computing evidence of association with regional detachment seismic source characteristics and focal mechanisms.

fault systems and related development of metamor-Information on the dynamic response of underground phic core complexes; (iii) detachment system cross-systems at YM to future potential earthquake seis-cut by high-angle normal faults (i.e., Bare Mountain);

micity may be gained by reviewing studies of the (iv) association with crustal-scale strike slip fault 119 NUREG/CP-0147

Volcannm/ Tectonics Working Group Summary systems; (v) erosion to a deeper structural level than Bullfrog Hills. This area may be a good analog for YM; and (vi) associated basaltic volcanism.

a more strongly extended YM system. Neogene and Proposed *a **/ structural analogs include:

Quaternary tectonic. extension in the Bullfrog Hills Black Mountains (Death Valley area). The Black regions is in excess of 100 percent. Detailed review Mountains may be a reasonably useful general tec-of existing geologic maps and sections, supple-tonic / volcanic analog for YM. Neogene volcanism mented with focused field studies, may provide use-within the Black Mountains may be similar to YM in ful examples of listric fault systems at various that silicw ash flow and rhyolite lava sequences cap mtermediate stages of extension. These may be use-the stratigraphic section and occupy a stmeturally high position on the Black-Mountains fault block.

ful as natural feard models of the YM systern.

However, silicic eruption volumes may be signifi-Dixie Valley, CedarMountain, Borah Peak. These cantly less than at YM. The Black Mountains are in areas may be appropriate to develop and constrain a more strongly extended region than YM, and the models of normal and normal-oblique mamshock fault block has been rotated above a regional detach-sources and coupled models of coseismic fault slip ment fault system. The block thus lies essentially on and ground rupture. These regions are particularly its side, exposing a cross section of the upper crust t useful for investigations of fault segmentation. Pa-a depth of about 6-8 km. Processes and features that may be studied include deep fault geometry, interac-leoseismic records (from trenching studies) may j

tion of sequential generations of faults, and relation-Provide additional information on regional-ships of basaltic dike networks to the extensional fault spatial / temporal clustering and on local recurrence system.

periods.

NUREG/CP-0147 120

~

iTC FORM 335 U.S. NUCLE AR REGULATOR Y CoMMIS$loN

1. REPoHT NUMBER ffRckio2 LM",E*l",.7#'s,$f"~ ""*

22oi 22a BIBLIOGRAPHIC DATA SHEET NUREG/CP-0147 (See instructiorn on the reversel

2. TITLE AND SUBitTLE Proceedings of the Workst>p on the Role of Natural Analogs in Geologic Disposal of High-Level Nuclear Waste 3.

DATE REponT PusLiSHED MONTH YtAR 1995 Hald in San Antonio, Texas

' e'ptember S

July 22-25, 1991

4. FIN oR GR ANT NUMBER B6666 6 AUTHOR (S)

6. TYPE OF REPORT L.A. Kovach, NRC W.M. Murphy, CNWRA

7. PE R lod COV E R E D (/nctusew Dare,)

I 8P F R NG Nll AT hon - N AME AND ADDRESS (I! Nac. orownk O*vwon. orroce ar neewon. v.1 Nucker neouserary commasion. emt maums orwress; te contractor. orovun Southwest Research Institute Center for Nuclear Waste Regulatory Analyses 6220 Culebra Road San Antonio, TX 78228-0510

9. SPONSORING oRGANIZATloN - N AM E AND ADDR ESS III knc. rype *kme as abow" *! contractor, provkk NRc orvoion. ortsco or naamn. v1 Nucker nneuterary commossean.

emt m*Hme oeMress.)

Division of Regulatory Applications i

Office of Nuclear Regulatory Research U.S. Nuclear Regulatory Commission Washington, DC 20555-0001

10. SUPPLEME NT ARY NOTES

11. ABST R ACT (Joo wants or heel A Workshop on the Role of Natural Analogs in Geologic Disposal of High-Level Nuclear Waste was held in San Antonio, Texas on July 22-25,1991. The proceedings comprise seventeen papers submitted by participants at the workshop. A series of papers addresses the relation of natural analog studies to the regulation, performance asse-ssment, and licensing of a geologic repository. Applications of reasoning by analogy are illustrated in papers on the role of natural analogs in studies of earthquakes, petroleum, and mineral exploration. A summary is provided of a recently completed, internationally coordinated natural analog study at Pogos de Caldas, Brazil. Papers also cover problems and applications of natural analog studies in four technical areas of nuclear waste management: waste form and waste package, near-field processes and environment, far-field processes and environment, and volcanism and tectonics. Summaries of working group deliberations in these four technical areas provide reviews and proposals for natural analog applications.

12. K E Y wor DS/DESCR1P 1 oH S (use wons orparws rest =#1asser rewane m socatmo rne report.t i3. Av AtLAes U r V 5i AI EMt NT natural analogs, geologic disposal, high-level nuclear waste unlimited 14, St CURII Y CLASSIF l(,ATiON lines Payel unclassified Irna neport>

unclassified tb. NUMBER of PAGES

16. PRICE NRC FORM 336 42 89)

l I

l l

i i

i

}

l l

I l

I i

4 i

Printed j

on recycled paper i;

l i

Federal Recycling Program i

UNITED STATEO

- NUCLEAR REGULATORY COMMISSION

.speoat rountuctass nart Posraos amo nas eam -

WASHINGTON, D.C. 20555-0001

.T..,7 120555139531 1 1AN1*W fy" i DUBLICATIONS SVCS OFRCIAL BUSINESS

' PENALTY FOR PRIVATE USE. 8300 TPS-PDo-NUDEG s

2WFN-6E7 WASHINGTOM DC 20555 L

t' o

4 9

I k

r 5

e-,s.

_,_---,m.

~e

---

e

.