IMPLICATIONS OF INTERSEISMIC DEFORMATION IN THE WESTERN U.S. FOR THE MECHANICS OF STRAIN LOCALIZATION

A viscoelastic-cycle model is applied to a large GPS dataset of crustal velocities throughout the western U.S., including measurements from the Plate Boundary Observatory. Using a viscoelastic stratification appropriate for the western US, the crustal velocity field is explained with a `blockless' model consisting of: (1) major faults that contribute time-dependent viscoelastic relaxation from large past earthquakes, (2) minor faults with uncertain rupture history that contribute time-independent velocity fields, (3) other (unspecified) faults that represent distributed slip, and (4) lateral variations in vertically-averaged rigidity of an elastic upper crust. While the first two contributors account for episodic movements on faults, the third and fourth contributors represent a combination of internal deformation of crustal domains between faults and structurally-controlled strain localization, each potentially broadly distributed. The fourth contributor in particular can represent either real variations in the elastic parameters over a reference elastic plate thickness or variations in elastic plate thickness, e.g. laterally-variable depth to the ductile lower crust. Model parameters include long-term slip rates and creep rates on known faults, rates of distributed moment release, and the apparent lateral variations in vertically-averaged rigidity; long-term slip and creep rates may be either fixed or variable, while all other parameters are variable, to be determined in an inversion of the interseismic velocities. Models which fit the western US velocity field well require a significant component of apparent lateral variations in vertically-averaged rigidity. This can range from an inferred rigidity contrast across the San Andreas fault (west side more rigid along most of the SAF) to reduced elastic plate thickness along zones concentrated near the traces of the Eastern California Shear Zone, Walker Lane, and the Intermountain Seismic Belt. These results are consistent with thermal perturbations inferred from seismic tomography. They suggest that different mechanisms of strain localization may apply to different fault zones depending on the elastic structure of the surrounding crust and mantle, which may also control how fault zones evolve once they are initiated.