Poster #152, Fault and Rupture Mechanics (FARM)

Numerical simulations of dynamic earthquake rupture with different failure laws

Ekaterina Bolotskaya, & Bradford H. Hager
Poster Image: 

Poster Presentation

2020 SCEC Annual Meeting, Poster #152, SCEC Contribution #10320 VIEW PDF
Geodetic and seismic observations indicate that most earthquakes result from unstable shear on quasi-planar faults, usually represented as shear cracks. Multiple failure laws have been proposed to describe fracture- and friction- related processes governing earthquake rupture, with the common feature that they relate the evolution of shear stress along the fault to slip. However, there is currently little discussion in the literature on the implications of each failure law for earthquake cycle characteristics or for near field rupture observables. We propose that the proper physical formulation of the problem of propagation of a seismic rupture should consider different failure laws (or a co...mbination of several of them), depending on the context of the earthquake (tectonic or induced), observation scale, and specifics of the problem.

To compare the implications of different failure laws, we test them in simplified 1D spring-slider models to determine the characteristics of the resulting earthquake cycle. To study rupture nucleation and propagation in more detail we perform dynamic finite element simulations, applying different failure laws along the fault: slip-weakening friction, double slip-weakening friction, rate-and-state friction, exponential and trapezoidal cohesive zone, etc., and compare the observables including rupture velocity, traction, slip, and slip rate distributions along the fault throughout the simulation. We obtain very similar results for different failure laws in both approaches, provided we set the same fracture energy and only consider planar ruptures. There are minor differences in slip and stress distributions resulting from the different shapes of the failure curves. However, it is challenging to detect them in the simplified modeling results, suggesting that the failure laws studied will be almost indistinguishable for the case of real seismic data. Distinguishing among which failure laws are applicable seems to require independent constraints on the friction law parameters, e.g., from laboratory experiments.