Surface waves, earth structure, and seismic discrimination.

摘要

I compute short-period, high-resolution surface-wave slowness maps for central Asia using Bayesian tomography. I focus on the region between 69 and 108 degrees east and 29 and 54 degrees north and used seismograms from more than 1100 events. Using multiple-filter and phase-matched filter techniques, I measured the dispersion characteristics of the signals between 6 and 30 seconds period. These Rayleigh-wave group velocity dispersion curves were used to compute high-resolution, half-degree cell size, slowness tomographic maps. Because short periods are primarily sensitive to upper crustal structures, the images display low velocities associated with the Tarim, Junggar, and Qaidam basins. Relatively high velocities are associated with mountainous tectonic features such as the Tian Shan. I validated these maps using dispersion curves from 640 events that were not used to construct the tomographic model. The model predictions show a significant variance reduction at short and intermediate periods (6 to 15 s) with respect to the prior model. My model also shows 15% improvement in surface-wave detection capability with respect to previous 1D models. These high-resolution, short-period tomography maps can help improve regional magnitude estimations for construction of m b : Ms discriminants. Moreover, the short-period surface-wave tomographic results show unprecedented resolution that reveals greater geologic detail than has previously been achieved using surface waves, and which give us insight into the shear-velocity structure of the crust underlying this part of Asia.; I also validate an alternative method to the traditional m b - Ms seismic discriminant. This new technique uses a probability of detection model (PXD) to estimate the probability that a surface wave detection came from an underground explosion. I measured noise and signal levels of over 1,300 earthquakes and 26 underground explosions recorded at the broadband, digital station WMQ in western China. The PXD analysis shows a reduction of the false alarm rate from 22% to 17% for this dataset. Using path-dependent attenuation and the previously computed group velocity values, the false alarm rate decreases even more to 15%. I also present a new procedure to accurately make Rayleigh wave amplitude measurements and I compute new surface wave magnitudes, Ms, following the methodology of Marshall and Basham and the methodology of Russell. The new Ms computed following Russell´s formulation proved to be an effective mb - M s discriminant and do not require farther analysis with the probability of detection model. This finding is of great importance to seismologists working on seismic discrimination problems.; To end, I implement and apply a method to jointly invert surface-wave group velocities and free-air gravity observations. (Abstract shortened by UMI.)