Statistical modelling of seismicity in stable continental regions

Various statistical models have been developed in the past to model the seismicity of a region and to evaluate its seismic hazard. Gutenberg and Richter model based on magnitude frequency relationship in combination with Poisson process assumption is a simple and widely used model to evaluate the seismic hazard. The quadratic magnitude frequency model, doubly truncated exponential model and the extreme value models are some advances in modelling the seismicity. Such models are either based on the type of catalogue data (i.e. its completeness) or on the mechanics of the earthquake occurrence process. In this paper, various statistical models have been briefly discussed and a case study to estimate the seismic hazard for Peninsular India has been presented. The paper provides an insight into the various models available for seismic hazard estimation, and their applicability to the particular region. Based on the study, it has been found that the extreme value model is well suited for modeling the magnitude uncertainty in the region with medium to low seismicity and results in statistical estimation of maximum magnitude earthquake expected in the region corresponding to design life of various structures. Such estimates provide very important information for design and maintenance of various critical facilities like dams and nuclear power plants. It has been concluded that the seismicity and hazard to critical facilities in Peninsular India should be estimated using multiple hypotheses. (author)