Exploration of the Moho Structure and Density Profile in Western United States

Abstract:

The base of the continental crust is typically marked by a strong seismic discontinuity due to a sharp contrast in seismic velocities and densities between crust and underlying mantle. Tectonic and magmatic processes over time can strongly modify the lower crust and upper mantle and the structure of the Moho. Moho depth and lower crustal character is, therefore, a reflection of the tectonic amalgamation and magmatic evolution of the continents. We examine Moho topography and lower crustal character in the western U.S. (lower 48) using receiver functions (RFs) and common conversion point stacking (CCP). Data derive from the Earthscope Transportable Array and other suitable seismic stations, with 2198 stations in total. P waves from 3764 suitable earthquakes recorded at these stations were downloaded from the IRIS DMC. We estimate Moho depths directly from the CCP stacks. These results compare well with extant models of Moho depth in the western U.S. (Crust 1.0; Buehler and Shearer, 2017). In addition to Moho depths, we calculate lower crustal densities in the study area from estimates of the P-to-S conversion transmission coefficients, which depend on source-receiver geometry and seismic velocities and densities above and below the Moho. We use crustal seismic velocities from Shen and Ritzwoller (2016) and uppermost mantle seismic velocities from Buehler and Shearer (2017) to determine transmission coefficients of P-to-S conversions at the Moho for three crustal models, Crust 1.0, Buehler and Shearer, and our CCP stacking Moho model. We used both constant (3.35 g/cm3) and regionally varying uppermost mantle density estimates (Mooney & Chulick, 2002) in these calculations. We also estimated transmission coefficients directly from the RF amplitudes. Lower crustal densities are most sensitive to seismic velocities - both P and S - in the lowermost crust. In order to estimate uncertainties in estimated lower crustal densities, we perturbed values of the seismic velocity models using a random number generator with Gaussian distribution, thereby introducing 'noise' in the resulting RFs. Transimission coefficients and densities where then recalculated, yielding mean values and standard deviations of the lower crustal densities. We find thinner crustal thicknesses, in general, along the Rocky Mountain Front than those reported in Crust 1.0 or Buehler & Shearer. Apparently thin crust from RFs also correlates with loci of double-peaked anisotropic RFs , as reported by Schulte-Pelkum et al. (2017). Thus, apparently thin crust along the Rocky Mountain Front is likely due to presence of a lower crustal 7.X layer; seismic velocity contrast at the top of this layer generates the strongest mode conversion in the CCP stacks. We find low density lower crustal values beneath the Basin and Range, and relatively higher density crust beneath the Colorado Plateau.