SCEC Award Number 19011 View PDF
Proposal Category Individual Proposal (Integration and Theory)
Proposal Title Near real-time earthquake focal mechanism inversion in the Southern California region using the SCEC Community Velocity Models
Investigator(s)
Name Organization
Zhongwen Zhan California Institute of Technology
Other Participants Xin Wang
SCEC Priorities 3a, 1d, 2d SCEC Groups Seismology, CXM, GM
Report Due Date 03/15/2020 Date Report Submitted 04/06/2020
Project Abstract
 Focal mechanism and directivity of earthquakes put primary control on the distribution of ground motion, and also bear on the stress state of the crust. Focal mechanisms of small-to-medium-magnitude earthquakes in Southern California have been routinely processed by the Southern California Earthquake Center (SCEC) using a simple 1D Earth velocity model, yet recent tomographic studies demonstrate strong 3D velocity heterogeneities in the Southern California crust. Here, we adopt the 3D SCEC Community Velocity Models in near real-time focal mechanism inversion. Our study results in a new focal mechanism catalog for the Los Angeles basin region, and provide more accurate focal mechanism solutions. The catalog is made available for further seismological and geological investigations and will contribute to mitigating the seismic hazard and risk in the area.
Intellectual Merit We have developed a highly automated and efficient algorithm to determine the moment tensor solutions for small-to-medium-sized earthquakes using 3D velocity models in the Los Angeles region. Our results show that incorporating the 3D velocity model can refine the existing moment tensor catalogs in the region, resulting in more accurate focal mechanism solutions, focal depth, and moment magnitude. In addition, our highly accurate, efficient, and automatic inversion approach can be expanded in other regions and can be easily implemented in near real-time system.
Broader Impacts We have presented findings from this project at the 2019 SCEC Annual Meeting. In addition, a paper that detailed our findings has been published on the Geophysical Journal International.
Exemplary Figure Figure 2. Comparison of moment tensor inversion results using 1D and 3D velocity models. The obtained focal mechanism solutions are shown as beachballs, where the black lines are optimal results, and the gray lines are uncertainties estimated by a bootstrapping method (95% confidence level). Histograms of the cross-correlation coefficients of waveform fits (including those discarded by our automatic data selection) are shown on the top. The comparison shows that the inversion using the 3D velocity model significantly reduces the moment tensor uncertainties, mainly owing to the accuracy of the 3D velocity model in predicting both the phases and the amplitudes of the observed seismograms.