Spatio-Temporal Analysis of Eearthquake Occurrences Using a Multiresolution Approach

Abstract

Catalogs of earthquake occurrences are conveniently modeled as spatial-temporal marked point processes. The widely used Epidemic-Type Aftershock Sequence (ETAS) models has proven to be extremely useful in the description and modeling of earthquake occurrence times and locations. The basic idea of ETAS models is that the conditional intensity function is composed by a long-term background component and a short-term, time-dependent clustering component, which represents the aftershock activity. Many extensions have been proposed in order to incorporate some geophysical information (e.g. orientation of fault-rupture). In this work we propose a multiresolution approach based on directional wavelet transforms for mapping the background seismicity, or to estimate the moment rate in a seismic area. The aim is to identify anisotropic spatial patterns and estimate the background earthquake rate in ETAS models. Then we apply the proposed methodology to the earthquake catalogue of Chile and discuss the effects of spatial clustering for the past earthquake events. Finally, we produce hazard maps in order to identify the areas with highest seismic risk, which are in turn critical for many purposes, including civil engineering and insurances.