A Renewal Process Model for Use in Seismic Risk Analysis

Abstract

A model of random occurrence of earthquakes using a renewal process is proposed on the basis of the catalogue of historical earthquakes for Kyoto, Japan. The model accounts for the variation of annual occurrence rate depending on the intervals between successive earthquakes. It can incorporate widely recognized nonstationary features of earthquake occurrences that can not be explained with the conventional simple Poisson process models. By using the renewal process model, called herein the double Poisson process model, one can estimate the probability of future earthquake occurrences in terms of the conditional probability based on the time since the last earthquake. Numerical results are presented for the Kyoto area.