Regional Moment Tensor Source-Type Discrimination Analysis
Technical Report,23 Aug 2013,22 Nov 2015
University of California, Berkeley Berkeley United States
Pagination or Media Count:
In this study we developed a new iterative inversion approach to find the maximum fit surface in moment tensor source-type space that has the advantage that it does not rely on a distribution of a large number of random moment tensors, is not dependent on the particular source-type mapping used, and since it inverts for the best fit solution for specified source-types it returns the true maximum fit surface. We have also investigated the effects of shallow depth of burial on seismic moment tensor recovery, bias in solutions, and the estimation of yield. The results indicate that while shallow depth of burial can affect moment tensor recovery due to the effects of vanishing traction on computed Greens functions the combination of long-period waveforms and first-motions results in unique discrimination of explosive events, and reasonable estimates of yield. The primary focus of the research was the investigation of 3D velocity structure effects on the recovery of seismic moment tensors at regional distances. A series of synthetic sensitivity tests and applications to NTS explosions utilizing 3D Greens functions obtained by invoking source-receiver reciprocity using the LLNL SW4 finite-difference code were carried out. The results indicate that minor improvement is afforded by the consideration of 3D velocity model, path average 1D structures perform as well at less computational cost. Additional work is needed to investigate more 3D velocity structures to assess whether a different model can improve performance and to assess the uncertainties resulting from variation in 3D velocity structure.