HR: 1330h
AN: S32B-0847    [PDF]
TI: Bayesian Approach to Short Period Slowness Tomography in Central Asia
AU: * Maceira, M
EM: mmaceira@geosc.psu.edu
AF: The Pennsylvania State University, Universty Park, State College, PA 16802 United States
AU: Taylor, S R
EM: taylor@lanl.gov
AF: Los Alamos National Laboratory, Los Alamos, Los Alamos, NM 87545 United States
AU: Ammon, C J
EM: cammon@geosc.psu.edu
AF: The Pennsylvania State University, Universty Park, State College, PA 16802 United States
AB: We are investigating the utility of short-period, high-resolution surface-wave slowness maps for lowering detection thresholds and allowing for discrimination of explosions and earthquakes at lower magnitudes (3.0-4.8). We focus on the region of central Asia between 69 and 108 degrees east longitude and 29 and 54 degrees north latitude. We retrieved broadband waveform data from more than 1,100 events that occurred between January 1997 and May 2002. Waveforms were obtained from four different networks comprising 13 individual stations. Using multiple-filter and phase-matched filter techniques, we measured the dispersion characteristics of the signals between 6 and 30 seconds. These Rayleigh-wave group velocity dispersion curves were used to compute high-resolution, half a degree cell size, slowness tomographic maps for each period. We adopted a Bayesian tomography method to solve the equation that relates travel-time data with the slowness structure. We used a declustering technique to minimize the effect of event clusters on the results. Our model is similar, to some extent, to the a priori model, but contains more structure and the cell size is smaller (0.5 compared to 1 degree cell size for the a priori model). The tomographic patterns correlate well with known geologic and tectonic features in the area. Accumulations of relatively young sediments across Eurasia are greater than on any other continent due to the continuing rapid uplift across much of central Asia. The short periods are primarily sensitive to upper crustal structures. The tomographic images display low velocities associated with the known sedimentary basins - Tarim, Junggar and Quaidam basins - in the area of central Asia under study. High velocities are associated with mountainous tectonic features such as the Tien Shan. We validated our maps using Rayleigh group-velocity dispersion curves for 640 events that occurred on the region of interest between January 1993 and December 1996. We analyzed group velocity residuals between the new model and the a priori model using these independent measurements. The overall refined model predictions are good. More over, they show a significant improvement at short and intermediate periods (6 to 15 s) with respect to the a priori model. The refined model also shows a 15 percent improvement in the detection of surfaces waves with respect to previous 1D models. Therefore, these improved high-resolution, short-period tomography maps can be used to lower the Ms detection thresholds to be used for construction of regional mb:Ms discriminants.
DE: 7219 Nuclear explosion seismology
DE: 7255 Surface waves and free oscillations
SC: Seismology [S]
MN: 2003 Fall Meeting