Regional Seismic Information Entropy for Detecting Precursors of Earthquake Activation

TL;DR

This paper introduces a novel regional entropy measure for seismic data that can detect earthquake precursors, showing it predicts activation earlier than existing methods in Japanese regions.

Contribution

The paper proposes the Regional Entropy of Seismic Information as a new indicator for earthquake precursors, validated on real earthquake data with improved lead time.

Findings

Regional entropy increases before earthquake activation

Method predicts earthquakes 1-2 years earlier than baseline methods

Validated on Japanese earthquake catalog data

Abstract

Here a method is presented for detecting precursors of earthquakes from time series data on earthquakes in a target region. Regional Entropy of Seismic Information, a quantity representing the average influence of an earthquake in the target region to the diversity of clusters to which earthquakes distribute, is introduced. Based on a rough qualitative model of the dynamics of land crust, it is hypothesized that the saturation after the increase in the Regional Entropy of Seismic Information precedes the activation of earthquakes. On the open earthquake catalog, this hypothesis is validated. This temporal change turned out to correlate more with the activation of earthquakes in Japanese regions, by one to two years precedence, than the compared baseline methods.