Earthquake Nowcasting with Deep Learning

TL;DR

This paper explores deep learning models, including recurrent neural networks and transformers, for earthquake nowcasting, demonstrating promising initial results in Southern California and providing open-source tools and data.

Contribution

It introduces novel deep learning approaches for earthquake nowcasting, utilizing RNNs and transformers, with a focus on regional predictions and open-source implementation.

Findings

Promising initial results in Southern California from 1950-2020

Deep learning models outperform traditional methods in nowcasting accuracy

Open-source software and data are provided for further research

Abstract

We review previous approaches to nowcasting earthquakes and introduce new approaches based on deep learning using three distinct models based on recurrent neural networks and transformers. We discuss different choices for observables and measures presenting promising initial results for a region of Southern California from 1950-2020. Earthquake activity is predicted as a function of 0.1-degree spatial bins for time periods varying from two weeks to four years. The overall quality is measured by the Nash Sutcliffe Efficiency comparing the deviation of nowcast and observation with the variance over time in each spatial region. The software is available as open-source together with the preprocessed data from the USGS.