Toward a Community Seismic Velocity Model for Alaska

Date: 4/26/2019

Time: 04:15 PM

Room: Cascade II

We have initiated an effort to develop a community seismic velocity model for Alaska. The current goal is to produce a tomographic model based on the best available body-wave and surface-wave data that is useful for obtaining accurate earthquake locations and can be refined via full waveform tomography for purposes of waveform simulations. We are extending the 3D south-central Alaska model of Eberhart-Phillips et al. (2006; EP06) to a region with twice the area, adding more earthquake data, in particular taking advantage of data from the USArray Transportable Array, incorporating Rayleigh wave group velocity data from ambient noise, and carrying out joint inversions of body-wave and surface-wave data. Our current model covers a 1400 km by 100 km region with corners near (53° 20' N, 145° 20'W), (56° N, 161°W), (68° 10'N, 156° W), and (64° 45'N, 135° 20'W). The dataset includes approximately 260,000 P arrivals, 60,000 S arrivals, and Rayleigh wave group velocity maps for frequencies of 6.6 to 15.1 seconds. The surface-wave data used so far only cover the central part of the study region. A progressive inversion strategy was used, starting with body-wave only inversions for Vp and Vp/Vs structure with ~50 km grid spacing, followed by body-wave only inversions with ~25 km grid spacing, and finally the joint body-wave and surface-wave inversion, using the method of Eberhart-Phillips and Fry (2017). The area of good resolution is enlarged compared to EP06 in the overlapping area, especially for Vp/Vs thanks to the addition of the surface-wave data. In addition to providing generally minor refinement of the main features present in the EP06 model, we also confirm the deep horizontal velocity contrast across the Denali fault inferred by Allam et al. (2017) to represent a 10 km offset in the Moho. Our subsequent work will focus on improving the body-wave dataset, incorporating a larger set of surface-wave data, expanding the model area to cover more of continental Alaska, and comparing inversion results to that obtained using the method of Fang et al. (2019).

Presenting Author: Clifford H. Thurber

Authors