Integrating Long and Short-Term Time Dependencies in Simulation-Based Seismic Hazard Assessments

Abstract

Conventional probabilistic seismic hazard analysis (PSHA) only considers mainshock events and uses a time-independent earthquake rupture forecast to describe the occurrence of mainshocks. This approach neglects the long-term time-dependency of mainshocks on specific fault segments, the interaction of adjacent faults, and the spatial and temporal clustering of aftershocks. This study integrates a recently proposed advanced fault-based PSHA framework with an aftershock simulator based on the Epidemic-Type Aftershock Sequence (ETAS) model, to overcome the aforementioned limitations. Central Italy is used as a case study to illustrate the proposed framework. The case-study hazard curves show that including state-of-the-art advances in fault-based PSHA may increase the ground-motion amplitudes for low return periods, and decrease the ground-motion amplitudes for high return periods, relative to a conventional PSHA approach. The magnitude of ground-motion estimate increases due to aftershock inclusion is dependent on the selected site and return period. Sensitivity analyses of the results demonstrate that: (a) close to modeled faults, high-return-period hazard estimates are more affected by fault-related inputs than the aftershock inclusion; (b) close to modeled faults, the influence of aftershock inclusion is higher for low return periods than high ones; (c) away from modeled faults, aftershocks inclusion greatly affects the variance of the hazard results; and (d) fault interaction has a limited effect on the variance of the hazard estimates compared to other analysis inputs. This study could help with the future implementation of complex fault modeling approaches and aftershock hazard in seismic risk models.