PISAS v1.0: Physics-Informed Neural Surrogate for Earthquake Productivity and Aftershock-Sequence Modeling

PISAS v1.0 is a physics-informed neural surrogate for predicting 30-day earthquake aftershock productivity and the modified Omori–Utsu decay exponent from information available at mainshock origin time. The archived release contains the complete source code, frozen input and derived datasets, trained neural-network checkpoints, benchmark and validation results, figure-source data, and the scripts required to reproduce the analyses and Figs. 1–8 of the associated manuscript. The analytical decoder enforces positive productivity, a bounded decay exponent, monotonic temporal decay, and exact consistency between predicted productivity and the integrated aftershock-rate history.