Testing the Basic Assumption for Probabilistic Seismic‐Hazard Assessment: 11 Failures

Abstract

In the probabilistic seismic‐hazard assessment (PSHA), one calculates the peak ground acceleration (PGA) (alternatively velocity, displacement, or intensity) at a given location that is expected not to be exceeded by a probability chosen by expert judgment (differing in different cases) and during a selected period. This is usually done by considering the influence of several seismogenic sources in the vicinity of a site. Here we consider the simple case in which one source capable of a large earthquake dominates the PGA calculation, a situation often encountered in practice. The first parameter one needs to know for estimating the PGA is the maximum magnitude of earthquakes M max that is possible in each seismogenic source. When one of the M max events clearly dominates the PGA at the site in question, as in this case, one speaks of the maximum credible earthquake (MCE) affecting the site. Estimating the PGA due to the MCE is specified as one of the requirements for assessing the seismic hazard for large dams as detailed by the International Commission on Large Dams (ICOLD, 2010). Once the MCE is identified, one needs to know the annual probability of the MCE, which is proportional to its recurrence time. The PSHA method assumes the occurrence rate (or probability of occurrence) of the MCE can be estimated by extrapolating the occurrence rate and magnitude distribution of small earthquakes in the volume in question. In cases of large MCEs, the extrapolation is performed over one to five orders of magnitude. This basic assumption of the standard PSHA method that the occurrence probability of MCEs can be derived by extrapolation from the occurrence rate of small earthquakes is not supported by a physical theory. I am showing here that this hypothesis fails 11 out of 11 tests. The tests consist of …