Use F-test method to discriminate the significance of solution for moment tensor inversions
Wen-Hong Wei
Abstract
I study how to use F-test to determine whether solutions of moment tensor inversion can be significantly different for four different source models: Double-Couple (DC), Deviatoric (DEV), Double-Couple plus Isotropic (DC+ISO), and Full Moment Tensor (FMT).
Dreger and Woods (2002) perform seismic moment tensor inversions on long-period, three-component velocity seismograms from the Caltech TERRA scope network and the Berkeley Digital Seismic Network (BDSN) to determine best-fitting kind of those source mechanisms for the Little Skull Mountain Earthquake and three large Nevada Test Site (NTS) nuclear explosions.
Their results show that strong non-double-couple seismic moment tensors and shallow source depth characterize the nuclear explosions. The full-moment tensor inversions recover a volume increase. An isotropic plus double-couple source model was found to provide the same level of fit as the deviatoric moment tensor inversions. Determination of the nonisotropic source mechanism is not unique and they discuss their results with respect to the proposed source models for NTS.
Templeton and Dreger (2006) use three-component broadband digital waveforms at regional distances to solved for four different source models. And using the F-test as a statistical aid, they determined which of the four models was most appropriate for each event, and then conducted statistic tests to determine the robustness of the focal mechanism solutions and isotropic components. Their results show that Mw 3.5 to 5.6 earthquakes in the Long Valley volcanic region are rarely fluid-influenced earthquakes.
Reference
Dreger Douglas and Woods Bradley, 2002. Regional distance seismic moment tensors of nuclear explosions, Tectonophysics , 356, 139-156.
Templeton Dennise C. and Dreger Douglas , 2006. Non-Double-Couple Earthquakes in the Long Valley Volcanic Region, Bull. Seism. Soc. , 96 (1), 69-79.