SCEC Award Number 20058 View PDF
Proposal Category Collaborative Proposal (Integration and Theory)
Proposal Title Improving near surface crustal model across southern California by jointly inverting Rayleigh wave phase velocity/ellipticity and receiver functions
Name Organization
Fan-Chi Lin University of Utah Vera Schulte-Pelkum University of Colorado, Boulder
Other Participants Elizabeth Berg
SCEC Priorities 4a, 4c, 3a SCEC Groups Seismology, CXM, GM
Report Due Date 03/15/2021 Date Report Submitted 03/20/2021
Project Abstract
Near-surface seismic velocity structure plays a critical role in the amplification of ground motion during large earthquakes. In particular, the local Vp/Vs ratio strongly influences the amplitude of Rayleigh waves. Previous studies have separately imaged 3D seismic velocity and Vp/Vs ratio at seismogenic depth, but lack regional coverage and/or fail to constrain the shallowest structure. Here, we combine three datasets with complementary sensitivity in a Bayesian joint inversion for shallow crustal shear velocity and near-surface Vp/Vs ratio across Southern California. Receiver functions – including with an apparent delayed initial peak in sedimentary basins, and long considered a nuisance in receiver function imaging studies – highly correlate with short-period Rayleigh wave ellipticity measurements and require the inclusion of a Vp/Vs parameter. The updated model includes near-surface low shear velocity more in line with geotechnical layer estimates, and generally lower than expected Vp/Vs outside the basins suggesting widespread shallow fracturing and/or groundwater undersaturation.
Intellectual Merit In order to provide a model with resolution of Vs and Vp/Vs in the upper few km, we combine the complementary sensitivities of Rayleigh-wave phase velocities (upper crust), ellipticity (upper few km), and the initial pulse of teleseismic receiver functions (shallow Vp/Vs ratio and shallow interfaces) to create a self-consistent model at the regional scale across southern California. The idea to combine receiver functions and surface wave data in a Bayesian joint inversion to determine Vs and Vp/Vs is relatively new, and only recently shown to be promising in resolving near-surface Vs and Vp/Vs in sediments. By including Vp/Vs as a parameter we are able to fit receiver functions on a regional scale for the first time across 231 Southern California stations, including in basins where receiver functions have long been discarded as nuisance signals or “corrected” with ad-hoc models, as reverberations overprint Moho and other crustal signatures. The results, presented in the technical report, include a map of Vp/Vs across the region and 3D shear-velocity (Vs) model with very low near-surface velocities in basins more in line with previous measurements of shallow, local Vs.
Broader Impacts The funded SCEC project helped to support several graduate students at University of Utah. Elizabeth Berg, in particular, had been supported by multiple SCEC projects in the past and successfully defended her PhD thesis in early 2021. A new PhD student, Konstantinos Gkogkas, is now involved to continue working on refining the 3D model of the southern California.
Exemplary Figure Figure 2. Vs results at each station, with Gaussian-smoothed underlying map, at (a) the surface and (b) 1 km depths, and (c) depth to 3 km/s. (d) Cross-section A-A’ for Vp/Vs ratio in the top linear layer (top) and Vs to 10 km depth (bottom), including white dashed line at 1.5 km/s and black dashed line at 3 km/s. (credit: Berg et al., submitted to GRL).