Resolving simulated sequences of earthquakes and fault interactions: implications for physics-based seismic hazard assessment

Abstract

Long-term interactions of fault segments qualitatively differ among fully versus quasi-dynamic simulations, and when using oversized cells.• Reproducing frequency-magnitude distributions and static stress drops is not sufficient to constrain the rate of multi-segment ruptures.• Simulated earthquake sequences can differ due to compounded effects of numerical errors, even when individual ruptures are well-resolved.