HR: 0830h
AN: S31E-0797    [PDF]
TI: Travel Times in the Regional Distance Range: Results of the 3-D Ray Tracing.
AU: * Kustowski, B
EM: kustowsk@eps.harvard.edu
AF: Harvard University, 20 Oxford St., Cambridge, MA 02138 United States
AU: Ekstr\"{o}m, G
EM: ekstrom@eps.harvard.edu
AF: Harvard University, 20 Oxford St., Cambridge, MA 02138 United States
AU: Dziewo\'{n}ski, A M
EM: dziewons@eps.harvard.edu
AF: Harvard University, 20 Oxford St., Cambridge, MA 02138 United States
AB: Travel times of regional body waves are primarily sensitive to seismic velocities in the upper mantle, and they are therefore useful for constraining the heterogeneous structure in tomographic inversions. Traditionally, tomographic inversions seek velocity variations with respect to a 1-D reference model, and travel time perturbations are calculated for 1-D reference rays. Strong lateral heterogeneity in the upper mantle significantly deviate rays from their reference paths, which may considerably affect the predicted travel times. The effect of misidentification of multiple arrivals observed at regional distances may be even bigger. Owing to these complications, measurements at regional distances are often neglected in tomographic studies. Travel times can be calculated more precisely by using regional 1-D velocity models, but only if the region of interest is sufficiently small and uniform to be represented by a 1-D model. To remedy this limitation, one may predict travel times using a global 3-D model at a cost of implementing a 3-D ray tracing technique. We employ a pseudo-bending algorithm ({\it Koketsu and Sekine}, 1998) to trace rays in global 3-D velocity models expanded over spherical and radial spline basis functions. To distinguish between multiple arrivals we trace rays bottoming in several depth ranges. We show that the first arriving phases recorded at a given distance may travel along significantly different paths in distinct tectonic settings. To improve the accuracy of the travel time calculations we incorporate a detailed global crustal model with the 3-D mantle models. We use the ray tracer to model an extensive data set of regional travel times detected at 130 stations from 192 underground nuclear explosions in the Former Soviet Union. To investigate the importance of the 3-D ray tracing we compare travel times calculated along the rays in 3-D velocity models versus those integrated along the paths predicted by a 1-D model. The comparison of the resulting residuals shows a systematic small bias of the 1-D rays towards longer travel times. Higher resolution models, constrained by surface waves and regional body waves, are needed to predict the travel times more accurately.
DE: 7200 SEISMOLOGY
DE: 7203 Body wave propagation
DE: 7218 Lithosphere and upper mantle
DE: 7219 Nuclear explosion seismology
SC: Seismology [S]
MN: 2003 Fall Meeting