Preprints
https://doi.org/10.48550/arXiv.2608.18791
https://doi.org/10.48550/arXiv.2608.18791
20 Nov 2026
20 Nov 2026
Status: this preprint is open for discussion and under review for Natural Hazards and Earth System Sciences (NHESS).

Scaling-law-informed neural point processes for earthquake sequence forecasting

Tianlu Xiong, Zaibo Zhao, Yunrui Li, Wenqi Liu, Yosef Ashkenazy, and Yongwen Zhang

Abstract. Earthquake sequence forecasting requires models that can learn nonlinear history dependence while retaining robust statistical structure. We develop a scaling-law-informed neural marked point process, termed Fusion, that combines neural representations of catalog history with temporal features derived from the Epidemic-Type Aftershock Sequence (ETAS) model and magnitude information derived from the Gutenberg–Richter law. The model separates the magnitude cutoff applied to the input catalog from the fixed target-event threshold, allowing lower-magnitude earthquakes to inform forecasts without changing the target-event set. For the 2016–2017 Amatrice–Visso–Norcia sequence, Fusion achieves the highest target-event temporal likelihood when lower-magnitude events are retained, outperforming both the ETAS model and a purely neural point-process baseline. Event-wise and cumulative analyses show sustained timing gains through substantial portions of the Visso and Norcia sequences. Across five benchmark catalogs, catalog-specific neural training with a fixed ETAS prior yields the highest temporal likelihood at the minimum evaluated magnitude cutoff. Magnitude likelihood shows no consistent predictive gain beyond the Gutenberg--Richter-based ETAS reference, indicating that the additional information captured by Fusion is primarily temporal. These results show that lower-magnitude catalog histories and empirical scaling-law information complement neural sequence learning for target-event timing.

Share
Tianlu Xiong, Zaibo Zhao, Yunrui Li, Wenqi Liu, Yosef Ashkenazy, and Yongwen Zhang

Status: open (until 01 Jan 2027)

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse
Tianlu Xiong, Zaibo Zhao, Yunrui Li, Wenqi Liu, Yosef Ashkenazy, and Yongwen Zhang
Tianlu Xiong, Zaibo Zhao, Yunrui Li, Wenqi Liu, Yosef Ashkenazy, and Yongwen Zhang
Metrics will be available soon.
Latest update: 20 Nov 2026
Download
Short summary
Fusion combines machine learning with established patterns of earthquake activity. Tests on records from Italy and California show that, when smaller events are retained, Fusion generally forecasts earthquake timing better than traditional statistical and purely neural models. The gains are mainly in timing, with no consistent improvement over established models of earthquake size. These results support using earthquake laws as inputs to neural models for better timing forecasts.
Share