Spatiotemporal Seismic Hazard Assessment Using VQ-VAE and Seismic Statistical Features

2026-06-08

AI Summary

This study aims to improve the accuracy of forecasting M≥5.0 earthquakes within a 15-day horizon in localized regions (within a 24-km radius). The authors propose a novel approach that, for the first time, incorporates a two-dimensional spatial feature derived from reconstruction errors of a Vector Quantized Variational Autoencoder (VQ-VAE) to characterize crustal stress accumulation. This feature is fused with conventional one-dimensional seismicity statistics and fed into an XGBoost model for spatiotemporal prediction at the local scale, complemented by SHAP analysis for interpretability. The VQ-VAE–based feature substantially outperforms the traditional b-value and emerges as the most important predictor in the model. Its inclusion not only matches but surpasses the performance of prior global-scale forecasts in terms of AUC, demonstrating its innovative potential and effectiveness for seismic hazard assessment.

Abstract

In this paper we build upon a previous study in which we demonstrated, using XGBoost and earthquake catalogue data from Japan and Chile, that a set of 60 seismic statistical features (SSFs) had much greater predictive value than a set of 428 generic time series features from the tsfresh package. We here extend this previous work in two key ways, focusing on data from Japan as a large dataset is necessary in order to allow for the training of a deep learning (autoencoder) model. First, we move from whole-region prediction (considering, for each candidate event, the likelihood of an event M $\geq$ 5.0 anywhere in the region in the next 15 days) to localised predictions in which both the region of feature computation and the region of prediction are restricted to a circle of radius 24 km around the candidate event, and we show that performance remains excellent, similar to our previous whole-region study for the same area. Second, we here couple this proven set of SSFs, based on one-dimensional (catalogue) data, with a novel feature based on two-dimensional seismic maps, obtained by training a VQ-VAE model to reproduce such maps as output and identifying a measure of its error in doing so with a localised build-up of crustal stress. We show that while localised prediction based on SSFs can be effective alone, with test AUC values as high as those obtained in the case of Japan in our previous whole-region study, the inclusion of the new natively-spatial VQ-VAE-derived feature, top-ranked by SHAP analysis, can enhance performance and additionally appears to near-wholly replace the traditionally-computed $b$-value in terms of feature usage.

Problem

• seismic hazard assessment
• spatiotemporal prediction
• localised forecasting
• seismic statistical features
• VQ-VAE

Innovation

• VQ-VAE
• seismic statistical features
• localized spatiotemporal prediction
• reconstruction error as stress proxy
• SHAP feature importance