Dilatancy Strengthening As a Mechanism for Earthquake Rupture Barriers and Aseismic Creep Transients on Oceanic Transform Faults

Abstract:

Ocean bottom seismometer deployments along the Gofar, Quebrada and Discovery transform faults on the East Pacific Rise (EPR) have revealed strong along-strike variation in M6 earthquake rupture extents and earthquake swarm activity. An active-source refraction survey along the ~ 100-km-long western segment of Gofar found a ~ 10-km-long zone of ~ 10-20% P wave velocity reduction, which extends from the surface to the Moho and acted as a “barrier” to previous cycles of M6 ruptures [McGuire et al., 2012; Roland et al., 2012]. The low velocity zone is interpreted to result primarily from enhanced fault zone porosity. That this region appears to behave as a rupture barrier is interesting from a fault frictional point of view because it nucleates intense microseismicity and hence has velocity-weakening (unstable slip) characteristics. In this study, we use a 3D strike-slip fault model with rate-state friction to investigate how the presence of a high-porosity, strong dilatancy zone embedded in a velocity-weakening transform fault could lead to a persistent earthquake rupture barrier. Rate-state frictional parameters are based on experimental results on gabbro gouge under hydrothermal conditions, and constrained by the tomoDD relocation of seismicity on Gofar [Froment et al., 2014]. Our modeling results reproduce the ~ 5 year recurrence interval of M6 earthquakes on two ~ 20-km-long fault segments separated by a ~ 10 km zone with effective dilatancy strengthening. A stronger dilatancy effect leads to a lower seismic coupling coefficient in the barrier zone. The release of energy in the barrier zone is manifested in various forms of aseismic deformation, including postseismic slip and interseismic slow slip events. The modeled slow slip migration speed and equivalent stress drop are comparable to those estimated from earthquake swarms on transform faults [Roland and McGuire, 2009], and suggests that such swarm activity is primarily driven by aseismic transient slip events.