Date of Award
Doctor of Philosophy (PhD)
Michael H. Ritzwoller
Craig H. Jones
The recent deployment of the USArray/Transportable Array (USArray/TA) has stimulated new methods (e.g., ambient noise eikonal tomography; teleseismic Helmholtz tomography) to produce high resolution surface wave dispersion maps. These dispersion maps, combined with other geophysical data sets derived from the array such as receiver functions, present the opportunity to image the crust and uppermost mantle for the continental US at unprecedented resolution. However, new methods are needed to overcome the limitations of traditional methods that may generate unstable models and do not estimate model uncertainties.
In this thesis, I present a new approach that jointly interprets new surface wave observations with other geophysical observables using a Bayesian Monte Carlo framework. In this approach, prior constraints and assumptions are explicitly expressed as prior distributions, and data uncertainties are rigorously interpreted in the resulting models by the Monte Carlo sampling of the posterior distributions. Thus, a 3D model with attendant uncertainties at all depths and for all discontinuities is estimated.
I show that with this approach it is feasible to interpret both surface wave data and other geophysical data observed at most stations from the USArray/TA with simple models. In addition, the vertical resolution of the model is enhanced by improvements to estimates of Moho depth and upper crustal structures using both receiver functions and Rayleigh wave H/V ratio. By applying the new method to multiple data sets, a set of 3-D models is constructed for the crust and uppermost mantle beneath the contiguous US. These models reveal many geological features.
Shen, Weisen, "Bayesian Monte Carlo Inversion of Multiple Data Sets to Image the Crust and Uppermost Mantle Beneath the Continental United States" (2014). Physics Graduate Theses & Dissertations. 104.