3-D Traveltime Modeling With Application To Seismic Imaging And Tomography

Abstract

Fast algorithms exist for performing traveltime modeling, even in three dimensions. These algorithms have the nice property that the computational time and memory requirements scale linearly with the number of grid points used represent subsurface velocities in discrete form. While traveltime modeling is typically used to predict first arrival times, later arrivals can also be simulated through the incorporation of a priori reflector information. For two-dimensional seismic imaging and tomography applications, the traveltime modeling algorithms presented here greatly expedite solution and can be readily deployed on distributed-memory parallel computers. Three-dimensional applications present a greater challenge, but by coupling an understanding of algorithm complexity with the promise of faster computers having greater quantities of physical memory, one can begin to predict future capabilities.