6
Summary and Conclusions
GENERAL APPRAISAL OF THE SSHAC REPORT
The Senior Seismic Hazard Analysis Committee's (SSHAC) report offers substantial contributions to the foundations and practice of probabilistic seismic hazard analysis. But the primary focus of the report is not on how to create an assessment from the inputs; only in Chapter 2, in an introductory fashion in Chapter 6, and in Appendix J is a methodology for calculating the hazard estimates and their uncertainties addressed.
Instead, the central theme of SSHAC is guidance on the process of eliciting and aggregating expert opinion on seismic sources, seismicity within these sources, and ground motion attenuation, as well as the associated uncertainties and final estimates of the hazard. SSHAC focused on this theme based on its conclusion that the reason for some serious discrepancies in the results of prior studies is differences in ways in which these inputs were derived, even though the work was done by competent specialists working from the same or similar data bases. In the panel's view, SSHAC's most important message is that the quality of a PSHA using multiple experts can be enhanced by careful and wise choice of experts and skillful facilitation of expert discussion and interaction through workshops and other meetings.
The panel believes it very important to emphasize what the SSHAC report is and what it is not. The report presents a procedure for using experts in seismic hazard evaluation and for determining the uncertainties at key stages of the hazard analysis process. Its primary domain of application is to nuclear and other critical facilities. According to SSHAC, if a project sponsor and the analysts choose to do a probabilistic hazard analysis, its procedures will yield stable results. The SSHAC report is not a defense of the probabilistic approach to hazard assessment. In particular, SSHAC explicitly excludes any discussion of the nonprobabilistic methods of seismic hazard assessment. The panel accepts
this decision of SSHAC on the grounds that an evaluation of the relative effectiveness of the two approaches, or their relationship, was not in the committee's charge. The full-blown version of the SSHAC procedure, utilizing the technical facilitator/integrator (TFI) technique where needed, is costly and will almost certainly be used only for major critical facilities. The SSHAC report offers useful guidelines as to the level of effort required for various kinds of problems and for various levels of information already available to analysts. In the view of the panel, simpler methods of probabilistic hazard analysis are appropriate for application to noncritical facilities.
GENERAL SHORTCOMINGS AND LIMITATIONS OF THE SSHAC REPORT
The SSHAC report, with its appendixes, is a lengthy and complex document that requires careful reading. Many important ideas, including clarification of the limitations of the SSHAC procedures, are distributed throughout the text. A casual scanning of the document may leave readers with incorrect impressions as to what SSHAC has recommended, especially with regard to nonnuclear facilities. Most importantly, the report appears to have been written for those already quite familiar with PSHA methods, offering guidance on a preferred way to get stable results from a PSHA.
SSHAC's Executive Summary will be useful to administrators and project sponsors who are not specialists in hazard analysis methodology, but it includes nothing about the excellent earth science materials that are in the report and its appendixes.
SSHAC provides an up-to-date procedure for obtaining stable results from the application of PSHA principles that have been established in past practice. It does provide a consistent and systematic approach to elicitation and aggregation of diverse expert opinion and the uncertainties that arise therefrom, but this is not the same as the calculation of seismic hazard from the information elicited.
The SSHAC report does not make reference to nuclear reactors or other nuclear facilities, thereby lending an air of generality to its final report and the applicability of its recommended procedures. The panel believes, nevertheless, that the flavor of the report is strongly influenced by concern for applications to nuclear facilities and this generality is more
apparent than real. In response to recommendations in the panel's March 1995 letter report (Appendix B) to the U.S. Nuclear Regulatory Commission, SSHAC did attempt to narrow the scope of the applications for which its recommended procedure is intended. Disclaimers are included in several places that are technically adequate to protect a practitioner who chooses not to use the SSHAC prescription against the need to defend that decision in a regulatory situation. Nevertheless, it seems clear that the report was written to support the highest, most sophisticated level of PSHA practice. Because the concept of the TFI is held by SSHAC to be one of its most important contributions to PSHA practice, a great deal of space is devoted to this topic, even though there are repeated comments that it is not needed for many of the issues that arise. The impression is given that this highest level of operation is really the key to success in general.
The panel concludes that the SSHAC contention—namely, that all PSHA projects should share the same basic principles and goals—should be taken as an overarching postulate for project design. But this contention should not be taken as implying or imposing the full elaborate and demanding methodology for application to every PSHA study. That alternate simpler methods may well be adequate for noncritical facilities is acknowledged by SSHAC, but they are not discussed nor is guidance offered as to where readers can learn about them.
In meetings and in its letter report of March 1995 (Appendix B), the panel urged SSHAC to document in adequate detail the manner in which lessons leading to the recommended SSHAC procedures were learned from the study of prior PSHA studies. Although the SSHAC report states that its conclusions are based on a thorough review of a number of such studies, the requested details are not offered and no previous PSHA analyses other than the Lawrence Livermore National Laboratory and Electric Power Research Institute studies are referenced.
The panel's evaluation of SSHAC's treatment of uncertainty is presented in detail in Chapter 3 of this report. The panel acknowledges that recognition of the two kinds of uncertainty is useful in elicitating expert opinion and in making decisions about where additional data gathering and research are likely to lead to reduced uncertainty about hazard estimates. However, as discussed in Chapter 3, the panel has reservations about how this distinction is ultimately helpful to final users, especially because the distinction between uncertainty types is sometimes
ambiguous and the amount of epistemic uncertainty regarding a hazard depends on the type of models used in the analysis.
Moreover, it is the impression of the panel that the statistical analysis and uncertainty separation procedures recommended in the SSHAC report are, at times, more sophisticated than is warranted by the data on which such analysis is based or the purposes for which the results are used.
The problem of integrating the opinions of a group of experts is difficult. It is treated in greatest detail in Appendix J of the SSHAC report. The panel found that this treatment is not easy to follow and that specific aggregation models described are not exhaustive. Therefore, the panel recommends that the quantitative methods of Appendix J be used as examples and not be regarded as prescriptive procedures. Given the current state of the art in formal expert aggregation and the difficulties specific to the earthquake hazard problem, the panel suggests that judgmental combination rules may be at least as valid as quantitative procedures.
SOME CONTRIBUTIONS OF SSHAC TO HAZARD ASSESSMENT
The contributions that the SSHAC report makes to the hazard assessment process are discussed in detail in the preceding sections of this report. A few key items are highlighted here.
The TFI Methodology
SSHAC considers the TFI methodology to be the centerpiece of its work and developed it from lessons it learned from prior hazard analysis studies and from workshops conducted as part of its study. The panel is favorably impressed with the concept and its implementation in the two ground motion workshops (SSHAC's Appendixes A and B). Readers of the SSHAC report should keep in mind that use of a TFI is not recommended or needed for all hazard assessments and should not even be viewed as a rigid prescription for a high-level PSHA. The TFI elicitation procedure is not synonymous with PSHA methodology.
Clear Definition of Experts' Distinct Roles as Proponents, Evaluators, and Integrators
It is important that experts be educated to the significance of their distinct role as proponents of a particular position or as evaluators. The panel is not sure that experts can truly assess the view of the whole informed community on the entire range of relevant issues.
Results of SSHAC-Sponsored Workshops
SSHAC held workshops on seismic source characterization, ground motion estimation, and earthquake magnitudes. The outputs of these workshops (Appendixes A, B, C, H), especially those on ground motion, are a valuable contribution of the SSHAC effort and led to the formulation of many of the recommended procedures in the committee's report.
Considering the broad consensus on ground motion modeling that was reached at the end of Workshop II, the panel believes that a real opportunity exists now to formulate, with further work to fill in necessary details, a ground motion model that can be used as a standard in the eastern United States for PSHA until new data or future theoretical developments warrant a reevaluation. The results of this effort would eliminate the need to elicit again ground motion input for each hazard analysis and could be used as a baseline for more detailed studies as needed for specific problems.