Mantle deformation and tectonics: constraints from seismic anisotropy in the western United States

We have examined shear-wave splitting in teleseismic shear waves (SKS, SKKS, S) from 15 stations in the western United States, based on analysis of 123 records from 67 events. The varied past and present tectonic styles in this region (subduction, transform faulting, extension, and stable domains) are expected to cause a wide variety of anisotropic behavior and therefore make it an excellent natural laboratory. Fast polarization azimuths (φ) vary from E-W to nearly N-S and time delays ( δt) range from being barely detectable (less than 0.6 s) to 1.6 s. Most stations yielded consistent measurements independent of station-event geometry. The exceptions were stations situated very close to the San Andreas Fault, which yielded well-constrained but inconsistent splitting parameters. These have been successfully modeled by two anisotropic layers with different horizontal symmetry axes. The upper layer has a fast direction parallel to the fault and δt of about 1 s. The lower layer, with δt from 0.6 to 0.9 s, is oriented E-W near the San Francisco Bay Area and NE-SW in the Mojave Desert. Other measurements of E-W fast φ are observed as far east as western Nevada, with large delay times of 1.3-1.5 s, but disappear to the north. Stations in the northwestern Basin and Range have values of φ oriented at about + 70° with delay times ranging from 0.7 to 1.2 s. φ varies in other regions of the Basin and Range, from - 70° in eastern Nevada to + 20° in the transition zone between the Colorado Plateau and the Basin and Range, with δt of 1.0 s. Measurements are geographically coherent, and in many cases changes correspond to geologic or tectonic boundaries. The E-W fast feature extends across several tectonic provinces, but disappears near the southern boundary of the Gorda plate, and may therefore be related to the dynamics of plate interactions. Regional variations in the fast directions of anisotropy within the Basin and Range suggest that the present extension has not been strong enough to align the fast directions over a great depth extent. Fast polarization directions within the transition zone between the Basin and Range and Colorado Plateau are nearly parallel to the geologic and geophysical boundaries between the two regions and to other geophysical boundaries. The size of the delay times suggests that in many of the areas studied, either the mantle is almost twice as anisotropic as we estimate, or the anisotropy, and therefore mantle deformation, extends into the sublithospheric mantle. We used the two-layer method to examine the effect of a postulated lower layer with fast direction parallel to the absolute motion of the North American Plate, and found the best estimate of contribution from a uniform layer is 0.1 s of δt, although a contribution of up to 0.7 s cannot be ruled out. A value as low as 0.1 s suggests either that olivine is not strongly aligned in a horizontal direction in this layer, or that the layer thickness is of the order of 10 km.