Understanding of Crustal Activity with Statistical Approaches

Abstract

We developed a comprehensive statistical validation system of crustal activities with which to easily address spatially and temporally sufficient range of a database of geophysical measures. The system involves carrying out the following four processes: (1) creating the database of geophysical measures with spatially and temporally gridded and other convenient formats, (2) comparing any two geophysical measures, at least one of which is time-variable, (3) classifying the spatiotemporal relationship of these geophysical measures into some types by defining a statistical index, and (4) evaluating and validating the relationships between classification results and the occurrence of target physical events such as large inland earthquakes. With the system, we aim for making a statistical model, or an appropriate rule for monitoring of crustal activities. Formulation of the rule is, in turn, expected to lead to comprehensive understanding of crustal activities. Here, we focused on the relations of seismicity and strain rate to introduce the conception and algorithm of the system. The system requires input of the database and other parameters and leads to output of various spatiotemporal distributions, 2-by-2 contingency tables, and probability gains for prediction and alarm rates for target physical events.