Regional Studies with Broadband Data Array Analysis of Regional Pn and Pg Wavefields from the Nevada Test Site.

摘要

Small-aperture, high-frequency seismic arrays, with dimensions of a few kilometers or less, can improve our ability to seismically monitor compliance with a low-yield Threshold Test Ban Treaty. This work studies the characteristics and effectiveness of array processing of the regional Pn and Pg wavefields generated by underground nuclear explosions at the Nevada Test Site. Waveform data from the explosion HARDIN (mb = 5.5) is recorded at a temporary 12-element, 3-component, 1.5 km-aperture array sited in an area of northern Nevada. The explosions VILLE (mb = 4.4) and SALUT (mb = 5.5) are recorded at two arrays sited in the Mojave Desert, one a 96-element, vertical-component, 7 km-aperture array and the other a 155-element, vertical-component, 4 km-aperture array. Within the aperture of each array, spectral amplitudes vary significantly among sensors, with differences of a factor of 10 over 4 km typical. Among the mean spectra for the mb = 5.5 events, though there are significant differences in low-frequency spectral amplitudes between array sites, reflecting the relative site effects, the spectra become nearly identical beyond about 6 Hz. Spectral ratios are used to briefly examine seismic source properties and the partitioning of energy between Pn and Pg. The principal factor governing array performance is signal correlation, and it is examined here in terms of spatial coherence estimates. The coherence is found to vary between the three sites. The coherence estimates are used to construct spatially-continuous, frequency-dependent models of cross-spectra, which can be used to simulate array processing performance for arbitrary sensor configurations.