In good shape? The impacts of variable fault geometries on synthetic earthquake catalogues from physics-based earthquake simulators

Abstract

A well-known problem in understanding seismic hazard is the short time period of the historical record relative to the time between large earthquakes. This short record means that not all possible earthquakes have been observed, and that the statistics of earthquake recurrence intervals are poorly constrained. These issues can be mitigated to an extent by paleoseismological investigations, but how these prehistoric data are interpreted is still limited by current understandings of which earthquakes are possible. In particular, recent multi-fault ruptures, such as the 2010 El Mayor-Cucapah and 2016 Kaikōura earthquakes, have demonstrated the potentially complex interactions of faults in single earthquakes, contrasting with the typical assumption of characteristic fault ruptures in seismic hazard assessment. Physics-based earthquake simulators, such as RSQsim offer one approach to expanding our understanding of potential multi-fault earthquakes.

Here we investigate the effects of fault geometry on the outputs of such simulators. We use the Canterbury and North Marlborough regions of the South Island of Aotearoa New Zealand – the epicentral region of the 2016 Kaikōura earthquake – as a case study. Using recently developed fault modelling tools we create 3D fault networks spanning the range of uncertainty of fault geometries in the region, including the potential for missing faults and variable geometries at fault intersections. The different networks we develop are motivated by key observations from the Kaikōura earthquake, such as the high proportion of previously unmapped faults in the rupture, and by explicit uncertainties in the New Zealand Community Fault Model. We generate synthetic earthquake catalogues on these different fault networks and investigate their similarities and differences, both statistically and in terms of the generated multi-fault ruptures. By doing so, we can better understand the effects of uncertainties in fault geometry on earthquake simulator outputs, which is critical before synthetic earthquakes can be used as a starting point for seismic hazard assessment or scenario planning.