Computation Of Traveltimes And Amplitudes Based On Wavefront Construction Ray Tracing

Nowadays, Seismic imaging is a most important part of exploration seismology, which faces a major challenge: to realize forward modeling wave propagation and 3-D imaging through large complex 3-D geological models. Numerous geophysicists have concentrated on the analysis and processing of seismic data in order to obtain perfect 3-D geological images of the subsurface. As for forward modeling, several papers have shown the importance of multivalued traveltime tables for quality of migrated images (Geoltrain and Brac, 1993; Operto et al., 2000). A variety of fast traveltime and amplitude tools have been developed to overcome the difficulty during the last years. The standard way is to perform dynamic ray tracing (Cerveny, 1985) for each point under consideration. But it is time consuming. So much effort has gone into finding simplified, more economic algorithms for traveltime calculation. An efficient method of solving kinematic ray tracing equations by Runge-Kutta method was proposed (Cerveny, 1985). Also, a method of solving eikonal equation by finite difference was proposed (Vidale, 1988) and was improved (Podvin 1991). Shortest path (Moser, 1991) method based on Fermat principle actually used eikonal equation from another perspective. Huygens wavefront tracing (Sava, Formel, 1998) was proposed as a robust algorithm, though its accuracy needed improving. Several other methods for traveltime calculation have also been proposed to calculate multi-arrival, e.g., Slowness matching (Symes, 1998) and wavefront construction method (Vinje, 1993, Ettrich, Gajewski, 1996). And these methods were extended to application in 3-D complex media (Vidale, 1990, Vinje, 1996, 1999, Schneider, 1998, Cerveny, 2001, Coman 2003). Efforts to realize accurate and efficient forward modeling never stop. For complex 3-D structures imaging a true 3-D pre-stack imaging method, such as pre-stack depth migration, is essential for data processing and interpretation. 3-D pre-stack migration has been studied and developed in exploration geophysics at a small-scale experimental stage for many years. However, due to the formidable amount of computation time and storage required by the procedure, its practical application has been limited. There are several papers and books published summarizing various methods and stating current problems on migration (Gray, 1997, 1998, 2001, Bancroft, 1999, Biondi, 2003). The two most popular methods are the acoustic Kirchhoff-ray algorithm (Sun, 1999, 2000, 2004) and the acoustic reverse-time finite difference algorithm. The former uses 3-D ray-tracing, and the latter uses 3-D finite difference, as the back-propagator, rendering the procedures time-consuming and prohibitive for large-size problems. Wave back-propagation, involving a huge amount of computation, becomes the bottle-neck for pre-stack depth migration or other 3-D imaging/inversion procedures. In this thesis, I concentrated on doing research related to seismic modeling and imaging. Under my supervisor, I assisted in the Laboratory of Wave Theory and Imaging Technology, and managed to code wavefront--oriented ray tracing (WFRT) program. I accomplished calculation of acoustic traveltime and amplitude in...