A Model for Strain Accumulation in Zones of Distributed Faulting

Abstract

Crustal displacements that accumulate across long strike-slip faults during the interseismic period have traditionally been modeled in one-dimension using a buried planar dislocation in an elastic half-space, the so-called “screw dislocation.” Despite its simplicity, this model fits geodetic data during the interseismic period well. Together, the screw dislocation and elastic rebound theory create an idealized framework permitting strain accumulation and earthquake potential to be quantified in terms of accrued slip deficit; however, screw dislocation models make no predictions as to how upper crustal deformation will be distributed during future earthquakes. To integrate upper crustal faulting complexity we replace the traditional screw dislocation with a density-distribution of back-slipping upper crustal dislocations in order to incorporate the finite width of real-world fault zones into the model. We characterize the distribution of plate-boundary wide fault widths from Quaternary fault map data, using the accumulated finite strain implied by mapped faults to approximate the potential distribution of future permanent strain. We applied this model to Southern California, where we utilized variational Bayesian Gaussian Mixture Modeling to objectively determine the distribution of dislocations from fault map data. The model predicts broad areas of potential upper crustal deformation, and thus seismic hazard, between the San Jacinto and the San Andreas faults, and on the northeastern side of the San Andreas fault.