Quantifying effect of geological factor on distribution of earthquake occurrences by inhomogeneous Cox processes.

Research on the earthquake occurrences using a statistical methodology based on point processes has mainly focused on the data featuring earthquake catalogs which ignores the effect of environmental variables. In this paper, we introduce inhomogeneous versions of the Cox process models which are able to quantify the effect of geological factors such as subduction zone, fault, and volcano on the major earthquake distribution in Sulawesi and Maluku, Indonesia. In particular, we compare Thomas, Cauchy, variance-Gamma, and log-Gaussian Cox models and consider parametric intensity and pair correlation functions to enhance model interpretability. We perform model selection using the Akaike information criterion (AIC) and envelopes test. We conclude that the nearest distances to the subduction zone and volcano give a significant impact on the risk of earthquake occurrence in Sulawesi and Maluku. Furthermore, the Cauchy and variance-Gamma cluster models fit well the major earthquake distribution in Sulawesi and Maluku.