A Model of Spontaneous Complex Tremor Migration Patterns and Background Slow-Slip Events via Interaction of Brittle Asperities and a Ductile Matrix

Abstract:

Slow-slip events (SSE) and tectonic tremors offer a unique window into a broad spectrum of earthquake behavior and fault mechanics at the bottom of seismogenic zones. Our previous heterogeneous fault models composed of brittle asperities embedded in a matrix with velocity weakening-to-strengthening transition reproduced tremor migration patterns observed in the Cascadia subduction zone including forward tremor propagation and Rapid Tremor Reversals (RTRs). Tremors were driven by a background SSE whose existence did not require tremors. Here we assess the importance of feedback between tremors and SSE by considering a ductile (pure velocity-strengthening) background matrix. In this new model complex tremor swarms and SSE result solely from the interaction between brittle asperity failures via intervening transient creep. In particular, in the absence of tremor asperities the model produces no SSE. The new model is less complicated than our previous one and matches some observations better. It reproduces the scale and recurrence interval of large tremor episodes and the migration speed and distance of RTR observed in Cascadia, and the ratio of RTR to forward migration speeds spans a wider range. Observations of tremor-genic and tremor-less slow slip occurring in a same segment of the Cascadia subduction zone suggest that natural faults are in between these two end-member models.

We study the tremor-driven SSE model through numerical simulations of heterogeneous rate-and-state faults with mixed velocity-weakening (VW) and velocity-strengthening (VS) materials. We find that the (b-a)*sigma value (effectively the frictional properties and pore pressure) of the VW fault portions determine the criticalness of the whole fault (i.e. its ability to produce spontaneous transient events). A combined analytical and numerical study of fault criticalness as a function of density of VW material and contrast of (b-a)*sigma values reveals subcritical-to-supercritical transitions both with increasing criticalness of the individual VW patches and, less trivially, by increasing the (b-a)*sigma value of the VW patches. Higher effective normal stress of the VW portions can bring the whole fault to a supercritical state even if the VW patches are individually subcritical.