Constraints on the Geometry of the Farallon Slab from the Joint Interpretation of All Available Imaging Results from the Earthscope USArray Deployment in the Lower 48 States

We assembled all available seismic tomography models distributed through the IRIS DMC and other sources. We combined these images with our own results using 3D plane wave migration of P to S conversion data derived from the USArray data set and other broadband seismic stations in the lower 48 states. All the tomography models were converted into SEGY format and interpolated onto a regular grid in a UTM reference frame. That innovation makes joint interpretation feasible using a seismic interpretation software (Petrel) because we treat both the tomography models and scattered wave image results as if they were 3D seismic reflection data. The careful designed interface of a modern exploration package makes exploring a range of interpretation packages much faster and allowed us to produce a more comprehensive interpretation of all available data. The tomography models are nearly an order of magnitude smoother than the scattered wave images, so we use the tomography models as a cross-validation in interpretation unless the scattered wave images are ambiguous. The focus of this study is testing a conjecture in an earlier paper (Pavlis, 2011) for the presence of a single continuous horizon interpreted as the top of the Farallon Slab. As in the previous paper we constrained the western edge of this surface with the location of Cascadia trench as well as a virtual edge from a back projection of the Mendocino triple junction using Pacific-North America motion over the past 30 Ma. We also simulated crustal multiple effects on the plane wave migration results using crustal geometry estimates produced by the Earthscope Automated Receiver Survey (EARS). This confirmed the scattered wave images were not reliable in the upper mantle at depths shallower than 200 km due to contamination by crustal multiples. Most tomography models show a steep dip in the slab immediately east of the volcanic arc and our surface follows the average geometry defined by a visual comparison of all the models. In eastern Oregon and northern Nevada the tomography models consistently show a general flattening of the slab over the 410 km discontinuity. A consistent horizon is observed in the most recent plane wave imaging and at we use that horizon to define the top of slab there. Our interpretations also confirmed a sharp increase in dip of the slab in eastern Wyoming and Montana.