Efficient finite difference calculation of partial derivative seismic wavefield using reciprocity and convolution

Sheen, Dong-Hoon
Tuncay, Kağan
Baag, Chang-Eob
Ortoleva, Peter J.
Seismic waveform inversion is commonly studied by the least‐squares method which transforms the inverse problem to an iterative minization problem of residuals between the synthetic model response and observed data. One can solve this problem using the gradient method, the Gauss‐Newton method or the full Newton method. The performance of the inverse method depends on the algorithm for calculating the partial derivative seismic wavefield due to perturbations in the subsurface. In this work, 2‐D elastic staggered‐grid finite difference method is used to explicitly calculate the partial derivative wavefields by utilizing the source‐receiver reciprocity and the convolution