Probabilistic and Alarm-Based Evaluation of a b-Value-Driven Deep Learning Earthquake Forecast

TL;DR

This study assesses a deep learning earthquake forecast model based on b-value evolution, showing it provides limited but consistent predictive information beyond traditional seismicity models, with improved discrimination at low alarm levels.

Contribution

It introduces a novel evaluation of a deep learning model using b-value dynamics for short-term earthquake forecasting, demonstrating its partial sensitivity and incremental predictive value.

Findings

Model shows positive Brier Skill Scores indicating useful information content.

Alarm-based evaluation captures more earthquakes than random models at low alarm fractions.

Model's output correlates weakly with clustering and recent activity, indicating sensitivity to seismic regimes.

Abstract

We evaluate the forecasting performance of a deep learning model, originally introduced as a pattern-extraction framework, that operates on the spatiotemporal evolution of seismic b-values in a short-term forecasting context. Model output is rescaled to account for training on balanced datasets and evaluated relative to a spatial base-rate model using the Brier Skill Score (BSS). Absolute skill values are small, but mean BSS values are consistently positive, including at locations where Mw geq 5 earthquakes occurred during the test period, indicating information content beyond historical seismicity alone. Alarm-based evaluation using Molchan diagrams shows elevated event capture rates at low alarm fractions (5.88 percent of events captured at 1 percent area under alarm), indicating discrimination exceeding random and purely spatial reference models under constrained alarm conditions. Comparison with ETAS-derived triggered probabilities further reveals a weak positive correlation, suggesting partial sensitivity of the model output to seismic regimes characterized by enhanced clustering and recent activity, while remaining distinct from classical aftershock-based descriptions. Together, these results indicate that spatiotemporal variations in b-values contain a persistent, though limited, signal relevant to probabilistic earthquake forecasting, yielding marginal but consistent improvements over baseline models across complementary evaluation frameworks.