Modeling Complex Evolution of Slow to Fast Earthquakes on Simple and Heterogeneous Faults

Abstract

The goal of my PhD research is to understand the behavior of earthquakes and slow slip events on fault systems ranging from simple planar geometries to complex fault networks with structural and material heterogeneity. To achieve this, I developed both analytical and numerical models to investigate the processes of nucleation, propagation, and termination. I investigate the occurrence of back-propagating earthquake fronts—secondary rupture fronts that travel in the reverse direction of the main rupture. Through dynamic rupture simulations using rate-and-state friction, I demonstrate that such phenomena can emerge even on frictionally homogeneous faults, provided that the rupture is unilateral and driven by stress gradients. An analytical model shows that velocity-dependent friction destabilizes steady crack-like propagation, leading to transitions between pulse and crack modes and generating back-propagating fronts. These findings suggest that back propagation may be more common than previously recognized, and does not require structural heterogeneity. I examine how fault roughness affects the dynamics of slow slip events (SSEs). Using quasi-dynamic simulations of rough faults with rate-and-state friction, I show that variations in normal stress, serving as a proxy for geometric roughness, induce irregular rupture patterns. SSEs exhibit temporal clustering, diverse rupture speeds, and back-propagating fronts. These arise from stress oscillations caused by roughness-modulated frictional properties, offering a new mechanism for the complex SSE evolution seen in geophysical observations. I perform quasi-dynamic and fully-dynamic earthquake simulations using rate-and-state friction to reproduce laboratory experiments on 760 mm-long faults with three velocity-weakening patches separated by velocity-strengthening barriers composed of PMMA and Teflon. Simulations show that multi-patch ruptures become more frequent under higher normal stress and narrower barriers, consistent with laboratory observations. Fully-dynamic simulations, which incorporate elastic wave propagation, allow ruptures to more easily overcome barriers and better reproduce the observed rupture size and speed compared to quasi-dynamic simulations. I develop novel 2.5D earthquake cycle simulations using boundary element methods to study the evolution of foreshocks and aftershocks in fault systems. This approach offers the computational efficiency of 2D modeling while incorporating more realistic 3D stress transfer effects. Collectively, these works provide new insights into the fundamental physical mechanisms driving both fast and slow fault slip with implications for seismic hazard assessment and earthquake forecasting.