Paper No. 5
Presentation Time: 9:00 AM


SHEN, Weisen, Department of Physics, University of Colorado at Boulder, 2000 Colorado Avenue, Department of Physics, University of Colorado at Boulder, Boulder, CO 80309 and RITZWOLLER, Michael H., Department of Physics, Univ of Colorado, Campus Box 390, Boulder, CO 80303-0390,

Recently, large seismic arrays with over 1,000 seismic stations have been deployed in various locations around the world (e.g. CEArray, VEBSN, USArray transportable array (TA)). These arrays provide unprecedented high quality seismic data that enable the application of methodological innovations including: 1) the observation of surface wave dispersion maps extending to uncommonly short periods based on ambient noise, 2) the development of more accurate surface wave tomography methods (e.g. eikonal/Helmholtz tomography) for both ambient noise and earthquake data, and 3) the application of Bayesian Monte Carlo inversion that assimilates multiple seismic data sets in order to construct a 3-D shear wave speed model with uncertainties. By processing more than six years of seismic data recorded at over 1,000 stations from the TA, 8-80 sec phase velocity maps are generated for the western and central US based on both ambient noise cross-correlations and teleseismic earthquake data. Azimuthally independent receiver functions are produced by a method referred to as harmonic stripping. These data sets are interpreted jointly via a Bayesian Markov Chain Monte Carlo (MCMC) inversion algorithm to produce a 3-D Vs model for the crust and uppermost mantle. The application of receiver function data reduces the traditional velocity-depth trade-off that besets surface wave inversions at discontinuities such as the Moho. As a result of the MCMC method, uncertainties are traced through the inversion from the input data to the final model. The seismic model that results provides new information about diverse geological/tectonic features and processes beneath USArray (e.g., hypothesized mantle plume, lithospheric delamination, active/failed continent rifts). For the model to be useful to seismologists and non-seismologists alike, realistic uncertainty estimates are needed and are produced by the Bayesian MCMC method. We discuss the utility of the model to improve the location and characterization of small intercontinental earthquakes and to provide information about the thermal/density structure of the lithosphere.