COMPLEX STRESS WHERE EAST MEETS WEST: RESULTS FROM MOMENT TENSOR INVERSIONS ACROSS THE CONTERMINOUS UNITED STATES

Both recent induced seismicity and the expansion of seismic networks such as the Earthscope Transportable Array have provided an unprecedented number of earthquake moment tensors (1,849) across the conterminous United States, allowing appraisal of the state of stress in the North American interior with unprecedented resolution (~20 to ~200 km). We divide the moment tensors into 29 distinct geographic regions and invert each subpopulation for the optimal normalized stress tensor in each region. Contrary to conventional conception, the state of stress varies significantly across the central and eastern United States: more rapidly than across the Basin and Range or the western margin of North America. The most profound heterogeneity exists in the Rocky Mountains and adjacent western Plains. Range-normal extension typifies the Rio Grande Rift and northern Rockies, whereas N-S extension dominates the southern Rockies, and the Wyoming Craton hosts a chaotic mix of strike-slip and thrust faulting. Similar complexity is seen on the Great Plains: E-W extension in the Raton basin is orthogonal to the ~N-S extension of the western Plains and southern Kansas, which then transitions quickly (over ~100 km) to strike-slip deformation in central Oklahoma. To understand the cause of this small-scale variability, we develop a 3-D model of lithospheric density from seismic velocity, gravity, topography, and heat flow and generate a finite-element model of the gravity-derived stress tensor. Gravity-derived stress—with no contribution from the plate edges or base—generally reproduces (independently) observed principal stress directions, and we separately find that modeled deviatoric stress is generally highest where rates of seismicity—both natural and induced—are highest. These results suggest that lithospheric density variations strongly influence the intraplate stress field and intraplate seismicity. As such, they reveal the possibility of quantifying the stress resolved on known or hypothesized intraplate faults. We are currently investigating the feasibility of incorporating such 3-D stress models in the logic tree of the National Seismic Hazard Mapping Project.