Scaling Laws in Earthquake Memory for Interevent Times and Distances

TL;DR

This paper investigates the long-term memory and scaling behaviors of earthquake interevent times and distances, revealing differences between real data and the ETAS model, which impacts earthquake forecasting strategies.

Contribution

It demonstrates the presence of long-term memory and scaling laws in earthquake interevent data and highlights discrepancies with the ETAS model, advancing understanding of earthquake dynamics.

Findings

Long-term memory exists in earthquake interevent times and distances.

Scaling functions in real data differ from those predicted by the ETAS model.

Memory functions decay slowly and crossover to fast decay at characteristic times.

Abstract

Over the past decades much effort has been devoted towards understanding and forecasting natural hazards. However, earthquake forecasting skill is still very limited and remains a great scientific challenge. The limited earthquake predictability is partly due to the erratic nature of earthquakes and partly to the lack of understanding the underlying mechanisms of earthquakes. To improve our understanding and potential forecasting, here we study the spatial and temporal long-term memory of interevent earthquakes above a certain magnitude using lagged conditional probabilities. We find, in real data, that the lagged conditional probabilities show long-term memory for both the interevent times and interevent distances and that the memory functions obey scaling and decay slowly with time, while, at a characteristic time, the decay crossesover to a fast decay. We also show that the ETAS model, which is often used to forecast earthquake events, yields scaling functions of the temporal and spatial interevent intervals which are not consistent with those of real data.