Connecting depth limits of interseismic locking, microseismicity, and large earthquakes in models of long-term fault slip

Thickness of the seismogenic zone is commonly determined based on the depth of microseismicity or the fault locking depth inferred from geodetic observations. The relation between the two estimates and their connection to the depth limit of large earthquakes remain elusive. Here we explore the seismic and geodetic observables in models of faults governed by laboratory-based friction laws that combine quasi-static rate-and-state friction and enhanced dynamic weakening. Our models suggest that the transition between the locked and fully creeping regions can occur over a broad depth range. The effective locking depth, D-elock, associated with concentrated loading and promoting microseismicity, is located at the top of this transition zone; the geodetic locking depth, D-glock, inverted from surface geodetic observations, corresponds to the depth of fault creeping with approximately half of the long-term rate. Following large earthquakes, D-elock either stays unchanged or becomes shallower due to creep penetrating into the shallower locked areas, whereas D-glock deepens as the slip deficit region expands, compensating for the afterslip. As the result, the two locking depths diverge in the late interseismic period, consistent with available seismic and geodetic observations from several major fault segments in Southern California. We find that D-glock provides a bound on the depth limit of large earthquakes in our models. However, the assumed layered distribution of fault friction and simple depth estimates are insufficient to characterize more heterogeneous faults, e.g., ones with significant along-strike variations. Improved observations and models are needed to illuminate physical properties and seismic potential of fault zones.