HR: 0830h
AN: S41E-0133    [PDF]
TI: Revised Boundary Perturbation Theory for the Direct Solution Method
AU: * Takeuchi, N
EM: takeuchi@seismo.berkeley.edu
AF: Berkeley Seismological Laboratory, 215 McCone Hall, Berkeley, CA 94720 United States
AU: Kobayashi, M
EM: mkoba@nal.go.jp
AF: National Aerospace Laboratory of Japan, 7-44-1 Jindaiji Higashi-machi, Chofu-shi, Tokyo, 182-8522 Japan
AB: The method for computing perturbed synthetic waveforms due to boundary perturbations is formulated by Woodhouse (1980, GJRAS). However, because it is based on the perturbation theory of free oscillations, it is not obvious whether the formulation works when the frequency is not in the vicinity of the reference frequency. It also is not obvious whether the discretization error is small especially when we use basis functions other than the eigenmodes. Because the Direct Solution Method (DSM; e.g., Takeuchi et al. 2000, PEPI) is an efficient computational method based on the optimally accurate discretization method (Geller \& Takeuchi 1995, GJI), if we appropriately re-formulate the previous boundary perturbation theory, we can expect further improvement of the efficiency and accuracy for the forward and inverse approach for 3-D heterogeneous Earth structure. In this study, we express the equation of motion for the perturbed boundary in terms of the solution of the compulsory oscillation with the appropriate boundary conditions (Geller \& Hatori 1995, GJI). Using the similar approach to Woodhouse (1980, GJI), we can derive the matrix elements which are guaranteed to be valid for arbitrary frequencies. Because the basis functions used in DSM are converged to the above basis functions, the obtained results can be applied without any modifications. Error analysis of Geller \& Takeuchi (1995, GJI) shows that the obtained formulation allows the optimally accuracy. We also confirm this fact by numerical experiments.
DE: 7200 SEISMOLOGY
DE: 7203 Body wave propagation
DE: 7255 Surface waves and free oscillations
DE: 7260 Theory and modeling
SC: Seismology [S]
MN: 2003 Fall Meeting