Systematic comparison of geodetically-derived strain rate accumulation and seismogenic strain release in Western US

Geodetically-derived strain rate states and seismologically-derived stress states in the crust are often compared in the literature. Geodetic data record the accumulation of (largely) elastic strain while earthquake focal mechanism data record the release of stored elastic strain in response to the total stress state on faults. Principal strain rate components derived from geodetic data therefore need not have the same orientation as principal strain or stress directions derived from earthquake focal mechanism data. It has been shown previously in several subduction zone settings that principal directions of strain rate due to interseismic coupling on the plate interface do not match principal stress directions. This is an important observation because we can compute present-day stressing rates accurately and then use the orientation of the principal stresses to place constraints on total stress in the crust. In this study, we systematically examine stress and strain rate states in the western US. We estimate the horizontal strain rate field and uncertainties using geodetic data and physically-constrained interpolation method to derive a continuous velocity field using the solution for body forces in a thin elastic sheet. We also compute the seismogenic strain tensor at the same grid locations as computed geodetic strain rates by summing moment tensors and interpolating. Uncertainties on seismogenic strain tensor components are estimated with bootstrapping. We find systematic differences in principal directions of geodetic strain rates and seismogenic strain in central and southern California along the San Andreas fault system and in the Pacific Northwest above the Cascadia subduction zone. In Cascadia, NE-SW principal shortening rates associated with coupling on the plate interface are detected in a very wide belt up to 500 km from the coastline (700 km from the trench). In contrast, maximum horizontal stress directions in this region are rotated 40-60 degrees to the north relative to the principal shortening rate directions. Along the San Andreas fault, principal stress directions are rotated, relative to shortening rate directions, 25-40 degrees closer to the fault-normal direction. We will show that these inferences allow us to place minimum estimates on total crustal stress magnitudes.