CONCLUSIONS

I have rewritten the TTI pure P-wave equation in a form which reduces the number of FFTs per time step. Also, a new hybrid pseudospectral/finite-difference is proposed to solve this equation, where wavenumber operators are replaced by inexpensive finite-difference spatial operators. The computational costs for 2D and 3D RTM using the hybrid scheme are reduced by $31\%$ for the 2D model and $29\%$ for the 3D model. Therefore, the hybrid pseudospectral/finite-difference scheme makes the TTI pure P-wave equation more practical and realistic for industrial-scale 3D migration problems.