We apply Pn tomography beneath the entire USArray footprint to image uppermost mantle velocity structure and anisotropy, as well as crustal thickness constraints, beneath the United States. The sparse source distribution in the eastern United States and the resulting longer raypaths provide new challenges and justify the inclusion of additional parameters that account for the velocity gradient in the mantle lid. At large scale, Pn velocities are higher in the eastern United States compared to the west, but we observe patches of lower velocities around the New Madrid seismic zone and below the eastern Appalachians. For much of the mantle lid below the central and eastern United States we find a moderate positive velocity gradient. In the western United States, we observe a moderate gradient in the region of the Juan de Fuca subduction zone, but no significant gradient to the south and east of this region. In terms of anisotropy, we find that the Pn fast axes generally do not agree with SKS splitting orientations, suggesting significant vertical changes in anisotropy in the upper mantle. In particular the circular pattern of the fast polarization direction of SKS in the western United States is much less pronounced in the Pn results, and in the eastern US the dominant Pn fast direction is approximately north‐south, whereas the SKS fast polarizations are oriented roughly parallel to the absolute plate motion direction.

We use measurements of mantle P-wave velocity from the Moho refracted phase, Pn, to estimate temperature within the uppermost few km of the western U.S. mantle. Relative to other approaches to modeling the deep geotherm, using Pn velocities requires few assumptions and provides a less uncertain temperature at a tightly constrained depth. Assuming a homogeneous mantle composition, Moho temperatures are lowest in an arc that extends from the High Lava Plains through western Montana and the high-plains region of Wyoming and western Kansas/Nebraska. Highest temperatures are observed under recent (<10 Ma) volcanic provinces and are consistent with melting. Estimates of lower crustal viscosity suggest that the western U.S. west of the Laramide deformation front likely has regions of mobile lower crust that decouple upper crustal and upper mantle tractions.

[1] USArray has now provided several years of high-quality seismic data and improved ray coverage for much of the western United States, which will enable increased resolution for studies of the lithospheric and deeper structure of the North American continent. Here we analyze Pn arrival times from the transportable stations of USArray to resolve crustal thickness and uppermost mantle structure. We use 123,008 Pn picks from April 2004 to October 2009 as measured by the Array Network Facility at epicentral distances from 180 to 1450 km. These picks are derived from 778 stations at ∼70 km spacing and 7903 earthquakes and quarry blasts. Applying the classic time-term method, we use a regularized least squares inversion to estimate crustal thickness variations and image velocity perturbations in the uppermost mantle just below the Moho. We also consider upper mantle anisotropy and describe the velocity perturbations with a cos 2 azimuthal variation. Our crustal thickness map generally agrees with receiver function results from other researchers but differs in some details. We obtain an average upper mantle velocity of 7.93 km/s, with higher velocities beneath eastern Washington and northern Idaho, and lower velocities near the California-Mexico border, the Sierra Nevada, the northern coastal California region, and the greater Yellowstone area. We observe large anisotropic anomalies in southern California as well as in the Snake River Plain area. These results should complement other seismic studies (e.g., body and surface wave tomography and shear wave splitting) to provide information about composition, temperature, and tectonic processes in the western United States.