HR: 16:00h
AN: S34B-01    [Abstracts]
TI: Computation of Travel Time Through 3D Velocity Models for Applications in Real-Time, Global Seismic Event Monitoring
AU: * Ballard, S
EM: sballar@sandia.gov
AF: Sandia National Laboratories, MS0401, Albuquerque, NM 87185-0401, United States
AU: Young, C J
EM: cjyoung@sandia.gov
AF: Sandia National Laboratories, MS0401, Albuquerque, NM 87185-0401, United States
AU: Hipp, J R
EM: jrhipp@sandia.gov
AF: Sandia National Laboratories, MS0401, Albuquerque, NM 87185-0401, United States
AU: Chang, M C
EM: mchang@sandia.gov
AF: Sandia National Laboratories, MS0401, Albuquerque, NM 87185-0401, United States
AU: Barker, G T
EM: gtbarke@sandia.gov
AF: Sandia National Laboratories, MS0401, Albuquerque, NM 87185-0401, United States
AB: Three dimensional velocity models of the Earth have been little used by real-time global monitoring agencies despite the expectation that these models might improve the accuracy and reduce the uncertainty of the seismic event locations they calculate. There are many reasons for this reluctance to adopt 3D models, including 1) uncertainty that adoption of 3D models will in fact significantly improve locations, 2) questions about how to quantify the uncertainty of the travel time predictions, 3) uncertainty as to how to assess the fidelity of computed travel times relative to the input velocity model, 4) questions about the computational architecture most appropriate for calculating predicted travel times, and 5) concern that the computational cost of calculating travel time predictions in a real time monitoring environment will be prohibitive. In this paper we begin to address the implications of using 3D velocity models in real-time global monitoring environments by addressing the last 3 items in the list above, which focus on the computational aspects of using 3D velocity models for travel time prediction. There are three fundamental approaches to computing travel times through 3D velocity models: 1) fix a source location at some position in the Earth model and compute travel times to all nodes in a 3D grid of nodes surrounding the source locations by solving the eikonal equation, 2) fix the locations of a single source and single receiver within the 3D velocity model and find the ray path(s) that honor Snell's Law in between (boundary value problem; ray bending), and 3) fix a single source location within the model and iteratively modify an initial estimate of the ray parameter searching for a ray that arrives at the receiver (initial value problem; ray shooting). To assess the computational issues with the use of these types of travel time calculators we have implemented the Fast Marching Method of de Kool, et al (2006), which is an eikonal solver, and the pseudo-bending algorithm of Um and Thurber (1987). In this paper, we compare the relative merits of these approaches in the context of their use in a real-time global monitoring environment.
DE: 7203 Body waves
DE: 7219 Seismic monitoring and test-ban treaty verification
DE: 7270 Tomography (6982, 8180)
DE: 7290 Computational seismology
SC: Seismology [S]
MN: 2007 Fall Meeting