1. Project Goals/Objectives:

Probabilistic seismic hazard analysis (PSHA) is a commonly used approach for assessing earthquake ground shaking hazard. Allin Cornell of Stanford University first developed the approach in 1968 and this was followed by a publicly available code developed by Robin McGuire. Since then, numerous individuals have developed PSHA codes.

As seismic source characterization has advanced and source models have become more complex in terms of source geometry and earthquake recurrence, PSHA codes have become more sophisticated. In most cases, analytical checks cannot be conducted for the hazard codes. Preliminary comparisons of the hazard computed using various computer programs found large differences in the results for some cases.

In this project, a working group on PSHA code validation has been formed to develop standard validation exercises, and test existing hazard codes. Both publicly available codes as well as proprietary codes that have been used extensively in hazard evaluation in California will be tested. The focus of this evaluation will be numerical verification of the codes and analysis and comparison of their various features.

PG&E, Caltrans, and the CEC use PSHA results, which may be computed by consultants using different computer programs. Thus there is a need to develop standard validation tests to qualify seismic hazard codes.

2. Role of this project in supporting PEER’s mission (vision):

In addition to providing a means for Caltrans, PG&E, and the CEC to qualify hazard codes, the results of this project will also be useful to the PEER program as it begins to implement the PEER framing equation, which will require new hazard calculations for a vector of parameters. PEER will need to adopt a hazard code or codes that it will use.

3. Methodology Employed:

The Working Group consists of 15 persons involved in developing or using PSHA computer programs. The Working Group will develop about 20 to 25 test cases, which will range from the simplest to more sophisticated. The simplest cases have analytical solutions, but the more complex cases will not. Test Case Set #1 was provided to each of the participants on 12 March 2002. Most of the participants provided their results to URS by 26 April 2002. Two publicly available codes are being tested by URS.

The results of the initial test case set are being provided to each participant without identifying the names of the codes (except for the participants own code). This initial feedback is intended to allow the code developers to identify numerical mistakes and to correct them. For the publicly available codes, the results will be sent to the authors. The test cases will then be re-run with the revised code and submitted to URS.

A second set of more complex test will be sent out in early June. The results will again be compiled by URS and distributed to all of the participants with the names of the programs identified. Differences in the codes will be examined and a workshop will be held to go over the detailed differences in program features. A final workshop will be held in October for final evaluations.

4. Brief Description of past year’s accomplishments (Year 6) & more detail on expected Year 7 accomplishments:

Almost all milestones completed in Quarters 1 and 2.

5. Other Similar Work Being Conducted Within and Outside PEER and How This Project Differs:

None.

6. Plans for Year 8 if project is expected to be continued:

None.

7. Describe any actual instances where you are aware your results have been used in industry:

8. Expected Milestones & Deliverables:

The project is scheduled to be completed in 12 months from the date of authorization. Expected milestones include:

Quarter 1:

Quarter 2:

Quarter 3:

Quarter 4:

A report describing the test cases, the identification of errors, and the differences in the features of the codes will be prepared. We anticipate that consensus hazard results can be identified for the test cases that can serve as benchmark for other codes. We expect that this verification exercise will identify errors in some of the existing hazard codes and it will clearly identify the differences in features and assumptions in the hazard codes. The report will recommend minimum standards for meeting the benchmark results (e.g., 20% in probability level) that can be used to qualify hazard codes.

The test cases and their results established as part of this project will be entered on the PEER website to allow other interested parties to evaluate their codes. This will be of value to the other Engineering Centers and to other countries in their development or use of probabilistic hazard codes.