Seismic quiescence and b-value decrease before large events in forest-fire model

TL;DR

This study uses a 2D forest-fire model to analyze how seismicity and b-value change before large earthquakes, revealing that cluster dynamics lead to seismic quiescence and b-value decrease.

Contribution

It demonstrates the link between cluster-size distribution changes and seismic precursors using a forest-fire model for earthquake simulation.

Findings

b-value decreases prior to large events

Seismic quiescence correlates with reduced non-percolated stressed sites

Cluster aggregation explains b-value reduction

Abstract

Forest fire models may be interpreted as a simple model for earthquake occurrence by translating trees and fire into stressed segments of a fault and their rupture, respectively. Here we adopt a twodimensional forest-fire model in continuous time, and focus on the temporal changes of seismicity and the b-value. We find the b-value change and seismic quiescence prior to large earthquakes by stacking many sequences towards large earthquakes. As the magnitude-frequency relation in this model is directly related to the cluster-size distribution, decrease of the b-value can be explained in terms of the change in the cluster-size distribution. Decrease of the b-value means that small clusters of stressed sites aggregate into a larger cluster. Seismic quiescence may be attributed to the decrease of stressed sites that do not belong to percolated clusters.