Reproducing Magnitude-Invariant Stress Drops in Fault Models with Thermal Pressurization

Abstract:

Stress drops, observed to be magnitude invariant, are a key characteristic used to describe natural earthquakes. Theoretical studies and lab experiments indicate that dynamic weakening, such as thermal pressurization, may be present on natural faults. At first glance, these two observations seem incompatible, since larger events may experience greater weakening and should thus have lower final stresses. We hypothesize that dynamic weakening can be reconciled with magnitude-invariant stress drops due to larger events having lower average prestress when compared to smaller events. The additional weakening would allow the final stresses to also be lower, but the stress drops may be similar.

To explore this hypothesis, we study long-term earthquake sequences on a rate-and-state fault segment with enhanced dynamic weakening due to thermal pressurization using a fully dynamic simulation approach. Our seismogenic segment has uniform friction properties. Our results show that such models can explain both the stress drop magnitude invariance and observationally inferred breakdown energy increase for a range of event sizes. Smaller events indeed have larger initial stresses than medium-sized events, and we get roughly constant stress drops for events spanning up to five orders of magnitude in moment. These events also have increases in breakdown energy consistent with observations. We are working toward quantifying the robustness of these findings by exploring a range of fault properties, including the efficiency of dynamic weakening. Note, however, that our largest, segment-spanning events tend to have much larger stress drops, possibly due to being contained by velocity-strengthening regions of the models rather than arresting due to insufficient prestress. An important next step is to consider models of heterogeneous faults that eliminate model-spanning events and may extend the magnitude-invariant stress drops to the largest events simulated.