Characterize Framework for Seismicity and Structural Deformation at Yucca Mountain, Nevada Rev 00, ICN 00 ANL-CRW-GS-000003 April 2000 1. PURPOSE The purpose of this Analyses and Models Report (AMR) is to summarize the probabilistic seismic hazard analysis (PSHA) for the Yucca Mountain site such that it can be used in preparing the Disruptive Events WE) Process Model Report (CRWMS M&O 2000a). The PSHA, Probabilistic Seismic Hazard Analyses for Fault Displacement and Vibratory Ground Motion at Yucca Mountain, Nevada (Civilian Radioactive Waste Management System [CRWMS] Management and Operating Contractor w&0] 1998a), was a 4-year multidisciplinary project that assessed both ground motion and fault displacement hazards. The DE process model report is one of a group of documents summarizing models and data that will form the technical basis for the total system performance assessment (TSPA) supporting the Site Recommendation. The DE process model report will describe analyses of seismic and igneous events and their effects for use in the TSPA for Site Recommendation. In addition to supporting the DE process model report directly, this AMR also furnishes indirect support to: Analyses addressing the effects of ground motion and fault displacement for the postclosure period An analysis of seismic-related features, events, and processes (FEPs) to determine whether they should be included in the TSPA for Site Recommendation An evaluation of the extent to which the U.S. Nuclear Regulatory Commission's (NRC) subissues and acceptance criteria associated with the Seismicity and Structural Deformation and other key technical issues have been addressed (NRC 1999). Results of the PSHA also support development of seismic inputs for preclosure design. Seismic hazards potentially affecting the Yucca Mountain site consist of vibratory ground motion and fault displacement. Consideration of these hazards in TSPA depends on their likelihood of occurrence and their effects on engineered items and the natural system. Likelihood of occurrence is addressed in this AMR. Analyses to assess ground motion effects are being carried out in support of Engineered Barrier System, Waste Package, and Waste Form process model reports. Fault displacement effects are addressed in the analyses Fault Displacement Eflects on Transport in the Unsaturated Zone (CRWMS M&O 2000b) and Eflects of Fault Displacement on Emplacement Drifts (CRWMS M&O 2000c), which support the DE process model report. Results on likelihood of occurrence and effects will then form part of the basis for an analysis, Disruptive Events FEPs (CRWMS M&O 2000d), which will determine whether seismic-related FEPs need to be included in TSPA for Site Recommendation. Therefore, one purpose of this AMR is to summarize the PSHA results such that they can be used in these subsequent analyses to evaluate effects of seismic hazard and their impact on performance. For postclosure, the seismic hazard results are being used to evaluate whether future ground motions or fault displacements contribute to any repository events that occur with a probability greater than 1 in 10,000 in 10,000 years, and that have significant effects on overall performance. Consistent with U.S. Department of Energy (DOE) interim guidance (Dyer 1999, Section 63.2), for the purpose of postclosure performance assessment, an event in this regard is considered in this report to be the failure of a structure, system, or component (SSC) to perform its fbnctional goal. A seismic event then is the failure of an SSC to perform its intended fbnctional goal under ground shaking or fault displacement loading during the postclosure period of interest. Although simplified analyses are used to support TSPA for Site Recommendation, ultimately analyses to determine and describe seismic ground motion events for input to performance assessment will involve the development of fragility curves (frequency of failure as a function of ground motion level) for critical SSCs. Depending on final design choices, such SSCs might include emplacement drifts, drip shields, and waste packages. The hazard results can be used to develop a suite of ground motion time histories for different magnitude earthquakes that can be used in analyses to develop the fragility curves. The hazard curves then can be convolved with the fragility curves to determine a probability distribution for the frequency of failure of the SSC to perform its fbnctional goal. Convolution involves multiplication of the hazard and fragility curves and integrating over an appropriate range of ground motion levels to describe the probability of unacceptable performance. For SSCs designed to NRC seismic design requirements, as will be the case for the Yucca Mountain facility (YMP 1997a), integration to about 1.5 to 3 times the seismic design ground motion level appropriately captures the probability of unacceptable performance (Kennedy and Ravindra 1984). These analyses will support updated evaluations of seismic-related FEPs for inclusion or exclusion in TSPA. If FEPs are included, the analyses will also provide the basis for how they are treated. Thus, the ground motion hazard curves summarized in this AMR provide the basis for postclosure analyses. Ground motion hazard results he also being used to develop preclosure seismic design inputs for the potential geologic repository at Yucca Mountain. The ground motion hazard results form the basis for identifLing controlling design earthquakes and controlling ground motion spectra appropriate for the proposed Geologic Repository Operations Area (Dyer 1999). The locationspecific inputs for seismic design of specific SSCs are developed from the controlling design ground motions by taking into account the effects of near-surface soil and rock. The inputs are then used in analyses to design SSCs that accommodate the ground motion such that safety and waste isolation fbnctions are preserved to the extent that the health and safety of the public is maintained. Although supported by this AMR, preclosure seismic issues are not part of the scope of the DE process model report. Rather, these issues are addressed in a series of three topical reports (see Section 6.1 for firther discussion of these reports). Fault displacement hazard curves are also being used for postclosure and preclosure analyses. For postclosure performance assessment, the fault displacement hazard curves can be used in a fashion similar to that for ground motion to identify repository events for input to FEP evaluation and postclosure performance assessment. The fault displacement hazard curves thus provide input to response analyses of the engineered barrier system, waste package system, and natural system in response to fault displacement. Convolving the responses of these repository components and the natural system with the fault displacement hazard curves describes event frequencies for input to the performance assessment. For preclosure seismic design, the fault displacement hazard curves will be used as a basis for determining in which portions of the repository waste emplacement area the probability of fault displacement is so low that it does not need to be considered in design. Also the fault displacement hazard results are being used as input to analyses to determine appropriate setback distances of SSCs from primary faults in order April 2000 to provide reasonable assurance that they will maintain thCir intended waste isolation and containment functions. For any safety-related SSCs that cannot be located such that they avoid significant faults, the hazard curves will serve as a basis to determine the level of fault displacement that must be accommodated in preclosure engineering design. A find purpose of this AMR is to summarize how tectonic models for the Yucca Mountain site were considered and evaluated in the PSHA. The PSHA employed an expert elicitation process to evaluate the extensive database of relevant information pertaining to the characterization of seismic sources, fault displacement, and ground motion at Yucca Mountain. The experts were provided with available information on tectonic models proposed for the Yucca Mountain vicinity. They used that information in weighting the likelihood of different models applying to the site. Thus, for seismic hazard -assessment, no single model is selected for Yucca Mountain Project use, but rather uncertainty in our understanding of the tectonic framework for the site is quantitatively assessed as part of the hazard analysis. This AMR provides a summary of those tectonic model assessments for use in the DE process model report and in performance assessment. To accomplish the above purposes, this AMR has the following objectives as outlined in Characterize Framework for Seismicity and Structural Deformation at Yucca Mountain, Nevada (Development Plan) (CRWMS M&O 1999a): Summarizing the PSHA process and the use of experts to define inputs and characterize uncertainties for seismic hazard computation . e Summarizing the conceptual framework for seismicity and structural deformation at Yucca Mountain based on the interpretations of the PSHA experts and other availabIe information (with emphasis on how uncertainty . in tectonic models and other relevant input parameters was incorporated in the hazard results through the PSHA process) e Summarizing vibratory ground motion and fault displacement .hazard results at Yucca Mountain. As a synthesis of the PSHA, this AMR has no limitations except those implicit in the PSHA; these are discussed in Section 6.1.5. 2. QUALITY ASSURANCE Development of this AMK was evaluated in accordance with QAP-2-0, Conduct of Activities, and was determined to be quality affecting and subject to Quality Assurance Requirements and Description (QARD) (DOE 2000). This is documented in activity evaluations for Work Packages 14016105M5 (CRWMS M&O 1999b) and 1401213DMl (Wemheuer 1999). Accordingly, this AMK was developed in accordance with AP-3.10Q, Analyses and Models. Other applicable DOE Office of Civilian Radioactive Waste Management Administrative Procedures are identified in the Development Plan (CRWMS M&O 1999a). In accordance with QAP-2-3, ClassiJication of Permanent Items, there are no permanent items directly associated with this AMR. ANLCRW-GS-000003 REV DO 11 of 89 \ ~ ~ ~ ~ ~ ~ ~ ~ ~ \ ~ - ~ ~ ~ ~ ~ E C T S \ W C C A M R J \ A M R U ~ - ~ C ~ ~ ) ~ ~ ~ ~ ~ I ~ L W C WlB100 April 2000 3. COMPU~ER SOFTWARE AND MODEL USAGE No software was used to directly support information reported in this AMR. Nonqualified software (e.g., Microsoft Word, Version 97; Aldus Freehand, Version 8) was used to perform support activities, such as document and figure preparation, but output from this software was not used in quality affecting work. Thus, this software is exempt from qualification requirements of AP-SLlQ, Soware Management, Rev. 2, ICN 4. Refer to Section 6.2 and Attachment I1 for discussion of software used for PSHA calculations. 4. INPUTS 4.1 DATA AND PARAMETERS This AMR contains two basic types of input data that all came from the PSHA (CRWMS M&O 1998a). These include the experts' interpretations and the results for the fault displacement hazard assessment (Table I), and the experts' interpretations and results for the ground motion hazard assessment (Table 2). For the ground motion assessment, the experts' interpretations included both the seismic source characterization and the ground motion attenuation characterization (Table 2). The experts' interpretations are based on an extensive database (Appendix B of CRWMS M&O 1998a). Both the experts' interpretations and the hazard results as described in the PSHA are qualified by virtue of the expert elicitation process. All of these inputs are discussed further in Section 6, including explanation of terms and parameters. The Q status of the input data is provided in the Document Input Reference System. 4.2 CRITERIA This AMR complies with Part 63.2 DOE interim guidance (Dyer 1999). Subparts of the interim guidance that apply to this analysis are those pertaining to the characterization of the Yucca Mountain site (Subpart B, Section IS), the compilation of information regarding geology of the site in support of the License Application (Subpart B, Section 21 (c) (1) (ii)), and the definition of geologic parameters and conceptual models used in performance assessment (Subpart E, Section 114 (a)). At this time, criteria in the form of specific System Description Documents have not been identified that apply to this AMR. The PSHA Project was conducted following the guidance provided in Methodology to Assess Fault Displacement and Vibratory Ground Motion Hazards at Yucca Mountain, Nevada (YMP 1997b) and the process was generally consistent with NUREGICR 6372 (Budnitz et al. 1997) and NUREG 1563 (Kotra et al. 1996). This PSHA approach has been accepted in principle by the NRC (Bell 1996). 4.3 CODES AND STANDARDS No specific formally established codes or standards have been identified that apply to this AMR or the PSHA for Yucca Mountain. ANLXRW-GS-000003 REV 00 12 of 89 X:\rwCfMROlECmWCCAMfMAMR~-rrvahamOOrrvahamOO~WC Wlm April 2000 Table 1. lnput Data Used from the PSHA Fault Displacement Hazard Assessment Note that these inputs (the seismic source characterization and the ground motion attenuation characterization) are linked together in the same logic tree data files submitted to the Automated Technical Data Tracking System. 5. ASSUMPTIONS Description I Data Tracking Number Fault Dis~lacement Characterization I M00004MWDRIFM3.002 for Nine erno on strati on Points I Fault Dis~lacement Hazard Results 1 M00004MWDRIFM3.002 I for Nine emo on strati on Points I I Table 2. Input Data Used from the PSHA Ground Motion Hazard Assessment 1 I Data Tracking Number I Description I Experts' lnterpretatlons Feeding into the PSHA I 1 zonesy I Ground Motion Attenuation Characterization (in- I M00004MWDRIFM3.002 ~abulations: - Seismic Source Characterization (includes tectonic models. fault sources, and areal source cludes median ground motions and uncertainhes for normal and strike-slip earthquakes)' PSHA Ground Motion Hazard Results: Peak Ground Accelerations (PGAs) M00004MW DRI FM3.002 I (including mean and fract;'le values for I M00004MWDRIFM3.002 I horizontal and vertical motions) Peak Ground Velocities (PGVs) (includina mean and fractile values for As the analysis for this AMR consists of the translation, summary, and documentation of the PSHA Project, no assumptions were made in this summary. Key assumptions were made, however, in the PSHA process and they are summarized in Section 6.4. M00004MWDRIFM3.002 horizonti and vertical motions) Spectral Accelerations (SAs) (including mean and fractile values for horizontal and vertical motions) The analysis for this AMR is the translation, summary, and documentation of the PSHA Project. There were no models developed for this AMR. The following sections first briefly describe the necessary background information on the PSHA Project, computer software, description of input, assumptions, methodology, and finally results for calculating both the ground motion and fault displacement hazard assessments in the PSHA Project. Sensitivity analyses for the PSHA are also discussed in this section. Substantiation of models in the PSHA was carried out by experts through the expert elicitation and review process during the project. M00004MWDRIFM3.002 April 2000 6.1 BACKGROUND FOR THE PSHA PROJECT Yucca Mountain is located about 160 lun northwest of Las Vegas, in the southern Great Basin portion of the Basin and Range province. The region is tectonically active at a low to moderate strain rate (CRWMS M&O 1998b, Section 12.3; Savage et al. 1999) and is characterized by the presence of abundant late-Quaternary faults and a moderate level of historical seismicity (Figure I), with the occurrence of relatively infrequent earthquakes up to about moment magnitude (M,) 7.5. The Nuclear Waste Policy Act of 1982, as amended in 1987, assigns to the DOE the responsibility of evaluating Yucca Mountain as a potential geologic repository to site the nation's first permanent disposal facility for spent nuclear fuel and high-level radioactive waste. As part of the seismic performance aspect of this effort, the U.S. Geological Survey (USGS) and the CRWMS M&O jointly performed the PSHA Project. The PSHA Project was initiated in August 1994, was on hiatus from October 1995 to September 1996, resumed in October 1996, and concluded in June 1998 with the issuance of the PSHA report (CRWMS M&O 1998a). The overall approach that the DOE has undertaken to address potential seismic hazards at Yucca Mountain is documented in three topical reports: Methodology to Assess Fault Displacement and Vibratory Ground Motion Hazards at Yucca Mountain (Topical Report I ) (YMP 1997b) and Preclosure Seismic Design Methodology for a Geologic Repository at Yucca Mountain (Topical Report 11) (YMP 1997a). A planned third seismic topical report will document 'the results of both the PSHA Project and the Seismic Design Project. The methodology adopted and used in the PSHA Project is described in Topical Report I. This methodology is generally consistent with state-of-the-practice guidance provided by the Senior Seismic H ~ a r d Analysis Committee (Budnitz et al. 1997). The methodology and acceptance criteria used by the DOE to determine the preclosure seismic design of repository SSCs is described in Topical Report II. 6.1.1 . Objectives The objective of the PSHA was to provide ground shaking and fault displacement hazard results for both preclosure (100 years) design determination and postclosure (10,000 years or more) performance assessment. The governing guidance (Dyer 1999) specifies consideration of two categories of preclosure events in the seismic design of the repository SSCs and radiological safety. Category-1 events are expected to occur one or more times during the preclosure operational period of the facility. Design for these events is required to provide for protection of worker safety and for seismic design for ground shaking, with the target hazard established at annual exceedance probability (YMP 1997a). Design for Category-2 events is required to provide for radiological safety protection of the public during the preclosure period. Groundshaking seismic design parameters for these events will be based on hazard at an annual exceedance probability of 10" (YMP 1997a). As experience with the design of engineered structures for fault displacement is limited, corresponding target hazard levels for determining fault displacement preclosure designs are and 10" for Category-1 and Category-2 events, respectively (YMP 1997a). Although criteria are defined for fault displacement design, the primary approach will be to avoid faults capable of significant movement in laying out a repository. For additional discussion of how the PSHA Project relates to ongoing preclosure and ANLCRW-GS-000003 REV 00 140f89 \ \ D A K ~ L ~ - W C ~ M R O J U ~ ~ W C C A M M \ A M R U ~ - I C ~ ~ ~ \ ~ ~ ~ ~ ~ ~ ~ C X . D O C 04118100 April 2000 postclosure seismic design efforts, see CRWMS M&O (1998a, Section 1.2) and Topical Report 11 (YMP 1997a). For postclosure, the objective of the PSHA was to provide hazard results for evaluating whether future ground motions or fault displacements contribute to any repository events that occur with a probability greater than 1 in 10,000 in 10,000 years, and that have significant effects on overall performance. Note that an event as applied here to postclosure assessment does not refer to an earthquake or ground motion event. Instead, an event is the failure of an SSC to perform its functional goal (Dyer 1999, Section 63.2). 5.1.2 Overview As shown on Figure 2, the main activities comprising the PSHA process were (1) multiple-expert evaluation and characterization of seismic sources, including the characterization of potential fault displacement; (2) multiple-expert evaluation and characterization of vibratory ground motion attenuation, including earthquake source, wave propagation path, and rock site effects; and (3) computation of hazard results for both fault displacement and vibratory ground motion. The experts provided alternative evaluations to characterize uncertainties in their interpretations. Based on these alternative interpretations, hazards were calculated and expressed as the annual probability at which levels of ground motion or fault displacement will be exceeded. The fault displacement hazard was evaluated at nine locations within the Yucca Mountain site area (Figure 3). These locations were selected to span the range of known faulting conditions within the area of the proposed repository and the associated surface facilities, based on various surface (e.g. Day et al. 1998) and subsurface studies (e.g. Beason et al. 1996). Faulting conditions ranged from primary block-bounding faults to unfaulted rock, and included sites on secondary faults, and on fractures that were assigned displacement histories . representing displacement conditions encountered in the Exploratory Studies Facility (ESF).. The specific conditions at the nine sites are described in Section 6.3.4. The vibratory ground motion was evaluated at a "reference rock outcrop" located near the center of the proposed repository (Figure 3). The reference rock outcrop is defined as free-field ground surface, at the elevation of the proposed repository, 300 m below the repository ground surface at Point A on Figure 4. The basis of selection and specific properties of this reference site are discussed further in Section 6.3.3.1.1. The assessment of seismic hazards at Yucca Mountain relied upon the results of scientific investigations carried out over the past 20 years. Building upon earlier investigations of the Nevada Test Site region, studies of the Yucca Mountain site have included (1) evaluations of faults within about 100 krn for evidence of Quaternary activity; (2) detailed paleoseismic faulttrenching studies of active faults near Yucca Mountain to determine the history and characteristics of past earthquakes; (3) monitoring of contemporary seismicity; (4) compilation of a catalog of historical and instrumentally recorded earthquakes in the Yucca Mountain region; (5) development of ground motion attenuation relationships for extensional tectonic regimes, which includes the Yucca Mountain region; (6) investigation of local site attenuation characteristics; (7) numerical modeling of ground motion from scenario earthquakes; (8) evaluation of the tectonic stresses from hydrofracture measurements and earthquake focal mechanisms; (9) collection and analysis of geophysical data to assess tectonic models and ANLCRW-GS-000003 REV 00 15 of 89 \UIAK~~~&~~~~\~'UE~~\PROIECTS\WCCAMM\AMRU~~-~~~~~~~)~~CX.DOC 04l18100 April 2000 identify subsurface faults; and (10) collection and analysis of geodetic data to measure ongoing crustal deformation. Results of many of these studies are summarized in Whitney (1996). This extensive database, in addition to the numerous studies performed by nonproject scientists and the already existing literature and information, formed the basis for the Yucca Mountain seismic hazards analyses. For a complete bibliography of material supplied to the PSHA experts, see Appendix B of CRWMS M&O (1998a). The method to calculate ground-shaking hazard at a site is well established in the literature (Cornell 1968; McGuire 1978, 1995). Basic inputs required to perform a ground-shaking hazard calculation at a site are (1) an interpretation of seismic sources that contribute to the site hazard, from which conditional probability distributions of distance of earthquakes from the site can be obtained; (2) an interpretation of earthquake recurrence for each source, including the mean annual rate of occurrence and magnitude distribution of earthquakes; and (3) an evaluation of ground-shaking attenuation for the site region, including the mean and standard deviation of ground-shaking amplitude as a function of magnitude and distance. These inputs constitute an interpretation of the seismotectonic environment of the site. $Given the input evaluations, the hazard calculation method integrates over all values of the variables and obtains an estimate of the mean probability of exceedance of any ground-shaking amplitude at the site. A plot of these results is the well-known seismic hazard curve. The hazard curve quantifies the randomness or aleatory uncertainty of the earthquake occurrence and ground-shaking attenuation. The calculation method can thus be thought of as an aleatory seismic hazard model. 111 addition to the aleatory uncertainty of the seismic hazard, however, is epistemic uncertainty about the seismotectonic environment of a site. Epistemic uncertainty is due to scientific uncertainty about earthquake processes and ground-shaking attenuation and the incompleteness of available data for evaluating these processes. Significant advances in development of methodology to quantify epistemic uncertainty in seismic hazard have been made in the past 20 years (EPRI [Electric Power Research Institute] 1986, Volume 1; Budnitz et al. 1997). Thesc. . . . advances involve the development of alternative interpretations of the seismotectonic ~nvironment of a site -by multiple experts and the structured characterization. of epistemic uncertainty. Evaluations by multiple experts are made within a structured expert elicitation process designed to minimize uncertainty due to uneven or incomplete knowledge and understanding (Budnitz et al. 1997). The weighted alternative interpretations are expressed by use of logic trees (EPRI 1986, Volume 1). Each pathway through the logic tree represents a weighted interpretation of the seismotectonic environment of the site for which an aleatory seismic hazard curve is computed. The result of computing the hazard for all pathways is a distribution of hazard curves representing the full aleatory and epistemic uncertainty in the hazard at a site. For further discussion of aleatory and epistemic uncertainty, see Section 6.5.2. Elements of this methodology as applied in the PSHA for the proposed Yucca Mountain facility are shown on Figure 2. Epistemic uncertainty was evaluated by six teams of three earth science experts, who characterized seismic sources in the Yucca Mountain site region (within a distance of about 100 km) and fault displacement potential at the nine demonstration points, and by seven ground motion (GM) experts, who characterized ground motion attenuation in the site region (see next section for additional discussion of experts and project personnel). Details on the criteria and process for selecting PSHA experts are provided in Section 2.3 of CRWMS M&O (1998a). April 2000 Interpretations for hazard assessment were coordinated and facilitated through a series of workshops. Each workshop was designed to accomplish a specific step in the elicitation process and to ensure that the relevant data were being appropriately considered and integrated. An important goal in the elicitation process was to reduce variability in interpretations that is due to a lack of common understanding of the available data and probabilistic models that are used in the analysis. The integrity of the seismic hazard results rests principally on the scientific quality and thoroughness of interpretations of the seismotectonic environment input to the hazard calculation. It is, therefore, important that the methodology should not constrain the experts' input interpretations. To satisfy this important requirement, the methodology was modified to accommodate interpretations specific to the Yucca Mountain site as required. These modifications were incorporated into the computer code FRISK88 (FRISK 88 V2.0, 10139-2.0- 00) (Section 6.2) that was used to compute the ground-shaking hazard at the Yucca Mountain site. In addition, the code was modified to compute fault displacement hazard using multiple interpretations of fault displacement potential as input. The final PSHA results were presented as mean, median, and fractile hazard curves representing the total uncertainty (epistemic and aleatory) in input interpretations. 6.13 Project Organization The major components and personnel of the PSHA Project organization are shown on Figure 5 i and Table 3. Four technical teams were formed to plan, organize, and lead the technical workshops, facilitate the experts interpretations, and perform hazard calculations: (1) Seismic i Source and Fault Displacement (SSJD) Facilitation, (2) GM Facilitation, (3) Data Management, and (4) PSHA Calculations. A Review Panel was formed to provide technical review and I guidance to the project. Members of the Review Panel were selected to provide expertise in . PSHA methodology and the required input evaluations. The Review Panel provided ongoing review throughout the performance of the PSIIA. Panel members attended all workshops, made .informal comments during the workshops, and made formal recommendations following each workshop. This participatory review allowed the project to make adjustments and take I corrective actions throughout the performance of the work. The Panel also reviewed the draft final report and made formal recommendations. Table 3. SSFD and GM Experts -- SSFD Expert Teams Team AAR: Walter J. Arabasz R. Ernie Anderson - Alan R. Ramelli 1 earn ASM: Jon P. Ake D. Burton Slemmons James McCalpin Team DFS: Diane I. Doser Christopher J. Fridrich Frank H. (Bert) Swan University of Utah US. Geological Survey Nevada Bureau of Mines & Geology 1 US. Bureau of Reclamation Consultant GEO-HAZ ConsuWng, Inc. --I University of Texas, El Paso U.S. Geological Survey Geomatrix Consultants, Inc. April 2000 6.1.4 Quality Assurance -- SSFD Expert Teams Team RYA: Albert M. Rogers James C. Yount Larry W. Anderson Team SBK: Kenneth D. Smith Ronald Bruhn Peter L. Knuepfer Team SDO: Robert B. Smith Craig dePolo Dennis W. O'Leary OM Experts John G. Anderson David M. Boore Kenneth W. Campbell Arthur F. McGarr Walter J. Silva Paul G. Somerville Marianne C. Walck Note: Teams are named by using the first letter The PSHA Project was performed under the USGS Quality Assurance (QA) Program developed during that time (August 1994 to June 1998) for the Yucca Mountain Project. This included YMP-USGS-QMP-3.16, Rev 0, Scientijk Expert Elicitation (USGS 1996), and a latter modification (Rev 0-M1; USGS 1998). The key elements of the program applicable to PSHA were personnel qualifications and training, scientific expert elicitation, software controls, records management, and data management. Records were submitted to the Records Processing Center as per YMP-USGS-QMP-3.16. Personnel qualifications files consisting of position descriptions, resumes, and verification statements were collected for PSHA Project members including the Affiliation GeoRisk Associates, Inc. U.S. Geological Survey US. Bureau of Reclamation University of Nevada, Reno University of Utah Binghamton University University of Utah Nevada Bureau of Mines & Geology U.S. Geological Survey Affiliation University of Nevada, Reno U.S. Geological Survey EQE International Inc. U.S. Geological Survey Pacific Engineering & Analysis URS Greiner Woodward Clyde Sandia National Laboratoties of each member's last name. Management Team, Review Panel, and technical teams. Training iiexpert elicitation and in the applicable procedures was provided via workshops and reading assignments. At the time that the PSHA was performed, the QARD did not specify requirements applicable to scientific expert elicitation. However, the USGS developed YMP-USGS-QMP-3.16, which did include appropriate requirements for scientific expert elicitation. Revision 8 of the QARD (DOE 1998) became effective June 5, 1998 at the end of the PSHA project and it included requirements for scientific expert elicitation in Appendix C. These requirements were based on NUREG-1563, Branch Technical Position on the Use of Expert Elicitation in the High-Level Radioactive Waste Program (Kotra et al. 1996), and implementation of the PSHA Project was generally consistent with NUREG-1563. One potential difference in the PSHA project was the approach to documenting changes in the experts' interpretations prior to finalizing the results. The QARD and the NUREG require the experts to document the reasons for any modifications to the& interpretations. within-the struct&d expert elicitation process implemented for the PSHA Project, this NUREG requirement was considered to be implicitly met by the workshop summaries (CRWMS M&O [1998a], Appendices C and D, workshop summaries). The summaries contain descriptions of preliminary evaluations by experts. During the PSHA project, April 2000 additional specific requirements to justify evolving evaluations were considered by the PSHA Management Team to have the unacceptable consequence of anchoring and biasing the experts' evaluations. During a QA audit of the USGS, DOE'S QA Office compared the USGS procedure, the PSHA Project Plan, and implementation of the NUREG guidance to the then-draft QARD requirements (DOE 1997). The audit team accepted this position and justification by the PSHA Management Team on explicitly documenting changes to the experts' evaluations. 6.1.5 Limitations The PSHA is based on the information and data available at the time of the evaluations (June 1998). The geology and seismology of Yucca Mountain have been well-studied and the PSHA experts were provided an extensive and comprehensive information base fiom which to derive their interpretations. As part of their evaluations, the experts provided not only their best estimates of seismic source and ground motion parameters but also possible ranges in these parameters and their associated uncertainties as allowed by the available data. Thus, these interpretations should capture the fill range of models and uncertainty in their evaluations considering the extensive available data. As data uncertainty was specifically considered by the experts, any new data are not expected to significantly impact the experts' interpretations. Additionally, a procedure is currently in place, AP-AC.lQ, Expert Elicitations, by which new data that might have significant relevance to the PSHA will be evaluated and its potential impact assessed. The ground motion hazard has been computed for a location at the approximate center of the site area on a reference rock outcrop where a kappa of 0.0186 sec and shear-wave velocity of 1,900 mlsec was assumed based on limited site-specific data (Sections 6.3.3.1.1. and 6.4). Thus, the hazard results are strictly only valid for this reference site condition. An analysis is ongoing to develop preliminary seismic design ground motions for the repository, the top of the tuff above the repository, and the Waste Handling Building (CRWMS M&O 2000e). The control point hazard is generally applicable for any location within the area of the proposed repository, with this site condition, although small differences in the hazard are expected near the edges of the proposed repository area. 6.2 PSHA COMPUTER SOFTWARE Within the PSHA process, software QA requirements were applicable only to the computer codes used for calculating the ground shaking and fault displacement hazards, as per YMPUSGS- QMP 3.16 (USGS 1996). Any software used by the experts in developing their interpretations was considered part of the expert elicitation process and was thus exempted fiom QARD software requirements. Following the experts' evaluations, the calculations performed as part of the PSHA Project were executed using a modified version of FRISK88, a software package, developed by Risk Engineering that is accessible through Yucca Mountain Configuration Management (CM). The appropriate software configuration identifiers are shown in Table 4. All of the software used for seismic hazard cal~ulations was originally documented and verified in compliance with YMPUSGS- QMP 3.03, Software (USGS 1997). The fill life cycle plan activities and documentation included completion of the following: ANL-CRW-GS-000003 REV 00 19 of89 X:\x\x~wRoIEmYucCAMM\AMRUmrm00\amr00m00\amr00taDOC OU19/00 April 2000 Software identification form Software CM form a Requirements specification Validation plan 6 Design description User documentation a Software validation report a Review and verification documentation a Documentation for operations and maintenance activities Table 4. Software Used in the PSHA Project Calculations DPREP (code) 1 1 .O 1 ESP0026.01 1 10141-1 .0-00 Computer Code or Subroutine PREP88 (code) POST88 (code) MRE88 (code) The software was appropriate for the application and was used only within the range of validation. Recently, the qualification status of all of the computer codes has been reverified under AP-SI.lQ, Sofhvare Management, Rev. 1, ICN 0. The verification of the two subroutines, MEAN and CMB-FRAC, can be found within Toro (1998). Attachment IJ contains further description and explanation of the software used in the PSHA calculations. Version 1 .O 1 .O 1 .O DRlSK (code) FRISK88 (code) MEAN (subroutine) CMB-FRAC (subroutine) 63 PSHAINPUTS There are three basic types of input developed by the experts for the PSHA: (I) seismic source characterization for the ground motion hazard assessment, (2) ground motion attenuation characterization, and (3) fault displacement characterization. There is overlap between the seismic source characterization for the ground motion assessment and the fault displacement characterization; these are discussed separately below for clarity. USGS Tracking Number ESP001 9.01 ESP0020.01 ESP0021 .O1 1 .O 2.0 1 .O 1 .O 6.3.1 Tectonic Setting and Quaternary Faults at Yucca Mountain CM Software Tracking Number or Reference 10138-1.0-00 10136-1 .O-00 101 40-1 .O-00 The Quaternary stratigraphy and tectonic setting of Yucca Mountain provides the framework necessary for characterizing seismic sources for both the ground motion and fault displacement hazard assessments. Therefore, the following section briefly describes the tectonic setting, with emphasis on the Quaternary aspects, to provide the needed context for understanding the experts' seismic source characterization. A more detailed discussion of the tectonic setting, stratigraphic framework, and Quaternary paleoseismicity is included in Whitney (1996) and Section 12.3 of the Yucca Mountain Site Description Report (CRWMS M&O 1998b). ESP0025.01 ESP001 8.01 ESP5.42 - ESP5.43 April 2000 b 10137-1 .0-00 101 39-2.0-00 Tor0 (1 998) Tor0 (1 998) Yucca Mountain is within the Southern Great Basin of the Basin and Range tectonic province. Tectonically, the Basin and Range is experiencing extensional strain at a low to moderate rate, with !ow to moderate historical seismicity (Figure l), and it has a thin crust. Yucca Mountain is located on the south flank of a large Miocene caldera complex. It is considered to be an erosional remnant of a 11.4 to 14.0 million-year-old volcanic apron (Fridrich et al. 1999). Structurally, the mountain is dominated by subparallel north-trending and east-dipping fault blocks. The blocks of ash-flow tuffs are bounded by typical Basin and Range, high-angle, generally west-dipping, normal faults formed by rapid east-west extension during the waning phases of Miocene volcanism. Secondary intrablock faults are common. Although Basin and Range tectonic structure defines the structural pattern of Yucca Mountain blccks, the whole proposed repository area lies within the Walker Lane, a 100-km-wide , structural belt along the western edge of the Basin and Range province. The Walker Lane is characterized by long, northwest-striking and shorter, north-to-northeast-striking, strike-slip faults that accommodate much of the early extension in this region. The peak-phase of tectonism took place 12.7 to 11.6 million years ago (middle Miocene); and the region has since experienced declining strain rates. The current pattern of Quaternary deformation mimics the middle Miocene activity, however, at substantially lower rates (Fridrich et al. 1999). Within a 100-krn radius of Yucca Mountain, more than 100 Quaternary faults were identified as potential seismic sources (Figure 6). With the exception of the Death Valley-Furnace Creek-Fish Lake Valley fault system, these faults are interpreted to have low slip rates (SRs) (generally less than 1 rnrntyr). I The faults closest to Yucca Mountain are the most important to vibratory ground motion and fault displacement hazards (Figure 6). Within 10 krn of the proposed repository 8 of 14 mapped faults show evidence of multiple surface-rupturing earthquakes during the Quaternary. These faults are characterized by trace lengths shorter than 26 km, SRs of 0.001 to 0.05 mmlyr, and average recurrence intervals of 20,000 to 100,000 years (e-g. Whitney 1996). Several faults are spaced only a few krn apart and may merge at depth. Pdeoevent data modeled from all trench studies suggest that distributed faulting may have been common at Yucca Mountain (Pezzopane et al. 1994). 6.3.2 Seismic Source Characterization for Ground Motion Assessment The objective of seismic source characterization for the ground-shaking PSHA was to identify and characterize the seismic sources capable of producing earthquakes significant to groundshaking hazard at the site. Evaluations were conducted following the structured elicitation process that was adopted for the project, which included information assimilation and interpretation workshops and individual team elicitations. The process was facilitated by the SSFD Facilitation Team. The elicitation process included a total of six workshops and a 1-day elicitation meeting with each team (Figure 2). Each SSFD expert team evaluated seismic sources for ground motion and fault displacement hazard computation. The evaluations included alternative interpretations, each weighted to express the teams' uncertainty. Two basic types of seismic sources were evaluated and characterized by the SSFD experts: fault specific sources and areal seismic source zones. Both local faults (defined here as .within about ANL-CRW-GS-000003 REV 00 21 of 89 \ ~ A K ~ ~ ~ ~ ~ ~ ~ ~ ~ \ ~ - W C ~ W R O I E C ~ Y U O C A M T M A M R ~ - ~ C ~ ( ~ ~ \ ~ ~ ~ ~ ~ ~ I C L W C 04/18/00 April 2000 10 km) and regional faults (to a distance of about 100 km) were evaluated. Areal source zones were defined to represent zones of distributed seismicity not apparently associated with known specific faults, and were also used by some teams to characterize known structures that were not explicitly included as fault-specific seismic sources. Detailed descriptions of each expert team's seismic source characterization for the ground motion hazard assessment can be found in Appendix E of CRWMS M&O (1998a), and summaries of the evaluations can be found in Section 4.3 of CRWMS M&O (1998a). Table 5 is from Section 4.3 of CRWMS M&O (1998a) and it summarizes the key components of each team's seismic source characterization model, including issues regarding tectonic models and potential seismic sources (areal seismic source zones, regional faults, and local faults). Although tectonic models are not seismic sources per se in the PSHA, they are included in Table 5 because their evaluation was integral to development of seismic source characterization models. Tectonic models provide the framework that can help define or constrain some seismic source parameters, such as maximum seismogenic depths, fault dips, rupture models, and probabilities of activity. Thus, uncertainty in tectonic models was an integral part of seismic source characterization and this uncertainty is fully captured in the hazard results. The following sections discuss the overall treatment of the three main types of seismic sources. 6.3.2.1 Imal Fault Specific Seismic Sources An important part of the evaluations by the SSFD experts focused on characterizing the local fault-specific seismic sources due to their proximity to the site (figure 6). The local Yucca Mountain faults can be subdivided into three categories: (1) north-striking block-bo~mding faults, (2) northwest-striking faults, and (3) intrablock faults. The close spacing between faults, their anastornosing pattern, and their relatively short lengths suggest that the local faults may be structurally interconnected either along strike or at depth and, thus, may rupture either partially or fu!ly together. The local faults were characterized in terms of their rupture behavior, probability of activity, locations, rupture lengths, sense of slip, fault dips, maximum seismogenic depths, maximum magnitude (M,,), and rate of activity. Their geometric characterization depended on the experts' evaluations of the tectonic models. Parameters for the most significant local faults are shown in Table 6, and include the block-bounding Paintbrush Canyon, Bow Ridge, Stagecoach Road, and Solitario Canyon faults (Figure 6). The parameters in Table 6 (also Table 7) generally show the range of preferred values interpreted by the expert teams. Note that because the expert teams varied considerably in their fault characterizations, particularly with regard to interpretations of independent seismic sources and linked or combined fault sources, not all teams characterized the faults individually as listed in Table 6. Approaches used to evaluate the M, for faults were based on empirical relationships between magnitude and surface-rupture length (SRL), rupture area (RA), and/or maximumdisplacement (MD) and average displacement (AD). M,, values ranged from Mw 5.7 to 6.8 for some of the linked systems. Earthquake recurrence rates for the faults were described using either recurrence intervals and/or SRs with most teams using the latter due to the lack of ANL-CRW-GS-D00003 REV 00 22 of 89 \ \ O A K ! ~ ~ ~ ~ ~ Y ~ ~ ~ - ~ ~ M P R ~ ~ E C T S \ W C C A M T M A M R \ A ~ ~ - ~ C ~ - ~ ~ ~ . W C W181W) April 2000 Table 5. Summary of Seismic Source Characterizations for Ground Motion Hazard Assessment a able 4-1 of CRWMS M&O 1998a) Issue TECTONIC MODE1 Overall Approach Planar Block- Faulting Models AAR Team I ASM Team 1 DFS Team I I SBK Team 1 RYA Team SM) Team Viable models based on observations and inferred processes for the Crater Flat structural domain, with simple shear model given full weight (1 .O). Superposed NW-SE dextral shear manifested as specific structures (tectonic models A, 6, & C) (0.5) or not (tectonic model D) (0.5). Regional faults are modeled as independent and linked (for selected faults) planar faults to maximum seismogenic depth. Local faults include linked and coalesced models; planar faults to maximum seismgenic - depth, to depth of local detachment, or in some cases to a depth constrained by allowable aspect ratio or by intersection with a higher-order fault. The source model incorporates various aspds of planar block fault (preferred), detachment, iateral shear, and volcanictectonic models. Regional faults are modeled as independent planar faults to maximum seismogenic depth. Local faults-the preferred model Is that the faults are planar b a depth controlled by the brittleductile transition and the Bare Mountain fault: treated as independent and coalescing faults that merge at dep?. structural models are presented provides a unified characterization of local seismic, geologic, and faults: geophysical data. domino model (0.8) Alternative tectonic and Iolanar fault): structural models are detachment'(0.2) I considered primarily in the nncludes hv~othetical characterization of local hdden stril&lip fault of either local or regional extent ). faults. A coalescing fault model best fits the Yucca Mountain area. to maximum seismogenlc depth. I Local faults-include models of independent (0.95) and distributed (0.05) fault behavior; alternative structural models (dominoplanar and detachmentlistric) used to constrain downdip geometry and extent. Regional faults are modeled as lnde~endent ~lanar faults Independent planar faults to maximum seismogenic depth. Bare Mountain and regional faults are modeled as Local faults--planar to listric (1 to 3 coalescing systems). Preferred model: oblique rift-planar faults. Threedimensional strain accommodated on planar, strike-slip, normal, and oblique-slip faults. Rock Valley and Highway 95 faults act as accommodation zones in the rift. Regional faults are modeled as independent planar faults to maximum seismogenic depth. Local faults-Yucca Mountain faults are part of a half-graben, with Bare Mountain as the master fault, predominantly normal slip with a left-lateral component. Alternative tectonic and structural models are considered in the characterization of local faults. Preferred model for Crater Flat - Yucca Mountain k a half-graben formed within a larger rift that opens and deepens to the north. Deformation history and structure are associated with carapace effect, clockwise vertical axis rotation, basaltic volcanism, age and behavior of Bare Mountain fault. Regional faults are modeled as independent planar faults to maximum seismogenic depth. Local faults: half-graben model (1) end member-all Yucca Mountain faults are seismogenic, continuous planar faults to maximum seismogenic depth. (2) carapace effect--only major Mock-bounding faults are throuah-the-crust seismogckic faults; other intrablock faults are confined to the carapace (i.e., are aseismic) or link to faults having different attitudes and aspect ratios below the unconformity. ANLCRW-GS-000003 REV 00 X:\r~s\PROJECTS\YUCCAMTMAMR\Am-mvOO\TABLESR00.DOC Ul&PO - April 2000 Table 5 (Continued) Issue Shear Models (buried strike-slip faults or fault systems) Detachment Models AAR Team Included three alternatives: Model A - Throughgoing regional dextral shear zone (0.05); Model B - right-stepping dextral shear zone that produces a pull-apart basin WITHOUT an underlying cross-basin fault (0.6); and Model C - right-stepping dextral shear zone that produces a pull-apart basin WITH an underlying cross-basin fault (0.35). Regional detachment not viable (0.0), but hypothesized local detachments included, with weights dependent on the type of dextral shear structures assumed to be present. Local detachments not included as specific seismic sources; detachments affect only downdip fault extent for local fault sources. Depths induded for local detachments range from 3 km to the maximum thickness of the seismogenk crust, with 3 to 10 km preferred. ASM Team Model 1 -Continuous, long (240-kml strikesb fault zone as proposed'by Schweikert considered. Regional (60-km-long) sbike-slip fault given low weight. Model 2 - Shorter (25-km), more complex or segmented zone. Assessment of existence of buried strikeslip fault condiiional on whether or not detachment exists; assessment of the seismogenic potential of the buried strike-slip fault is conditional on the depth of the detachment (shallow- 0.8, moderate-0.6, deep 0.0). Detachment Model (0.1 5): Hypothesized detachment affects down-dip geometry and extent of local fault sources; seismogenic detachment Is included as possible fault source with very low probability (see below). DFS Team Model allows for component of northwest-directed rightlateral strikeslip strain. HypoWtical hidden strikeslip fault source (probability of activity [P J = 0.05) is included in detachment model. Two postulated strikeslip fault sources are lncluded: regional strike-slip fault (0.5) local strikeslip fault (0.5) Detachment Model (0.2): Hypothesized detachment chiefly affects down-dip aeometn, and extent of local {auk sokes; seismogenic detachment Is included as possible fault source with very low probability (see below). ANLCRW-GS-000003 REV 00 X\rrc(s\PAOJECTS\WCCAMTWR\Amr-revOO\TABLE5RW00M>C 4/18/00 Page 24 of 89 buried source covered by background source). Detachments are not explicitly modeled. Possibility that local faults truncate down dip in a detachment or zone of decoupling Is included in coalescing fault model. SBK Team I SDO Team A buried reaional shear I Three sources of dextral zone modeiis given low weight (0.01); no evidence for a buried strike-slip fault trending northwest across Crater Flat that would result in a earthquake larger than the maximum assigned to the host source zone. detachment affects only I (modeled as an Independent shear were evaluated to account for vertical axis rotation at Yucca Mountain: (1) distributed shear (restricted to Crater Flat basin; basin Is a discrete domain controlled by local bounding faults); (2) extemal transcurrent strike-slip fault (passes through the basin, totally hidden); and (3) extemal strike-slip fault enters basin from southeast (manifested at Yucca Mountain by the N25"W striking %ingeline")nd terminates In Crater Flat. Only (1) and (3) are credible modifications to the basic model. . Hypothesized the down-dip extent of - source) was thoroughly local fault sources. I considered but could not be substantiated bv the available A seismogenic detachment April 2000 Table 5 (Continued) Issue I AAR Team Volcanic-Tectonic I The ~ossibilitv of Models ("ash event") simultaneou~rupture on subparallel Yucca Mountain faults as postulated for the "ash evenr is included in coalesced fault models for local faults. -- ASM ~e&n The ~ssibility that some surface rupturing eatthquakes in Crater Fiat are accompanied by dike injection (e.g., the 70-ka 'ash evenr) is included in simultaneous rupture models for local faults. DFS Team I RYA Team The possibility of I The coalescing fault model simultaneous-rupture on used to model-local faults subparallel Yucca Mountain (see below) would explain faults as postulated for the the apparent sychroneity of 'ash event' is included in faulting on Yucca Mountain the distributed faulting faults (is., the 70 ka 'ash model for local faults. event"). Thickness of Dmaxl 12 (0.1) 12 (0.6) 12 km 15 14 15 km (0.2) Seismogenic Crust 11 km (0.185) (0.6) 16 (0.3) 17 20 km (0.7) 15 km (0.63) (0.3) (0.1 1 (0.1) Zones-- Recurrence Four scenarios: Scenario I w13 zones (0.3), Scenario II w/2 zones (0.3), Scenario Ill w/3 zones (0.3), and Scenario IV wll zone (0.1). For all scenarios, a host zone (within 20-km radius) is defined only for assigning a lower Meat for separate recurrence estimate. Truncated exponential recurrence rnodel (1 .O) 300-km radius catalog Version 7 (1 .O) Adjustment made for underground nuclear explosions in relevant source zones. Two source zones within 100-km radius of site. A local zone (within 50-km radius) is induded that is defined solely for assigning a lower M-. Truncated exponential recurrence rnodel (1 .O) 300-km radius catalog Version 7 10.7) Version 5 (0.3j Adjustment made for underground nuclear expiosions. Model A (0.2) One zone Model B (0.8) Three zones Both models include a local zone that Is defined for constraining Mm, in the area of the detailed site characterization studies. Truncated exponential recurrence model (1 .O) 300-km radius catalog Version 7 10.5) Version 5 io.5j Three primary source zones within 100 km of site; two alternative configurations to model Zone A (local Yucca Mountain region) and Zone B (the zone surrounding Zone A). SBK Team I SDO Team Explicitly models a ( Distributed fault models simultaneous rupture involve simultaneous rupture event (triggered by I of local faults that are parallel volcanic event: see to each other. Such models Local Fault ~odei) would account for volcanism and tectonic faulting as a coupled process. Model A (0.7) 3 zones Model B (0.3) 4 zones Both models include a local zone that is defined solely for assigning a lower M,. Truncated exponential Truncated exponential recurrence model (1 .O) I recurrence model (1 .O) 100-km radius catalog 100-krn radius catalog Version 5 (0.5) Version 7 10.3 to 0.6) Version 7 (0.5j version 5 (0.4 to 0.7j Weights vary depending on source zone. In relevant zones, adjustments made for underground nuclear explosions weighted (0.4) versus no adjustment (0.6). Eight source zones within a 300-km radius of the site were considered initially, but only 3 remained given a filter of radius 400 km. Truncated exponential recurrence model (1 .O) 300-km radius catalog Version 5 (0.6) Version 7 (0.2j Version 8 (0.2) ANLCRW-GS-000003 REV 00 Page 25 of 89 April 2000 Table 5 (Continued) AAR Team For Scenarios I - 111: . - - - .~ . - Uniform (1 -0). For Scenario IV: h-5 km (0.25) h = 10 km (0.5) h = 20 km (0.25) Excluding Host Zone 6.6 10.3) Host Zone (within 20 km) 6.0 (0.3) 6.3 6.6 (0.4) (0.3) 19 regional fault sources; includes faults with Pa of <1 .O; includes two possibly linked fault systems: Death Valley with Furnace Creek (0.8), and Arnargosa River with Pahrump (0.1); also indudes five faults considered as segmented (ma. rupture length