Recent advances in earthquake seismology using machine learning

Abstract

Given the recent developments in machine-learning technology, its application has rapidly progressed in various fields of earthquake seismology, achieving great success. Here, we review the recent advances, focusing on catalog development, seismicity analysis, ground-motion prediction, and crustal deformation analysis. First, we explore studies on the development of earthquake catalogs, including their elemental processes such as event detection/classification, arrival time picking, similar waveform searching, focal mechanism analysis, and paleoseismic record analysis. We then introduce studies related to earthquake risk evaluation and seismicity analysis. Additionally, we review studies on ground-motion prediction, which are categorized into four groups depending on whether the output is ground-motion intensity or ground-motion time series and the input is features (individual measurable properties) or time series. We discuss the effect of imbalanced ground-motion data on machine-learning models and the approaches taken to address the problem. Finally, we summarize the analysis of geodetic data related to crustal deformation, focusing on clustering analysis and detection of geodetic signals caused by seismic/aseismic phenomena. Graphical Abstract

Neural Phase Picker Trained on the Japan Meteorological Agency Unified Earthquake Catalog

“…Note that the selection of an appropriate threshold in the actual cataloging process should consider the performance of subsequent processes, especially phase association, as highlighted by Bornstein et al (2024). If the phase-association algorithm proves highly effective, a lower threshold might be advisable to improve recall, thereby increasing the number of detectable events, as suggested by Kubo et al (2024).…”

“…Recently, however, phase picking using deep learning techniques has achieved performance levels comparable to manual picking (Zhu and Beroza, 2019;Mousavi et al, 2020). Various models for this task, or neural phase pickers, trained on manually-read arrival times from regional or global datasets are now publicly available (Kubo et al 2024).…”

“…In this study, we trained the widely recognized neural phase picker, PhaseNet (Zhu and Beroza, 2019), using JMA picks and assessed the performance enhancement of the resulting model on JMA-unified data compared to the original PhaseNet trained with California data. Although numerous publicly accessible neural phase pickers exist besides PhaseNet (Kubo et al, 2024), previous research has consistently shown that PhaseNet is one of the best choices (Münchmeyer et al, 2022;García et al, 2022). The retrained model was integrated into an automatic seismic cataloging process employing the PF method (Tamaribuchi et al, 2016;Tamaribuchi, 2018), which is currently utilized to develop the JMA-unified catalog, to explore the benefits of using our enhanced model.…”