Comparison of seismic wavefield simulation between the curvilinear-grid finite difference method and ray method in undulated layered medium

Commonly it is not possible to verify and test the accuracy and validity of the seismic wavefield modeling method, especially when the model surface (or subsurface) is undulated. For this reason, this paper tries an alternative way to mutually verify and test the computational accuracy and the solution correctness of both the ray theory (the multistage triangular shortest-path method) and the wave-equation simulation method (the curvilinear-grid finite difference method of the DRP/opt MacCormack format) in layered isotropic medium. Through the analysis and comparison of wave field snapshot, common source gather profile and synthetic seismogram, it is not only able to verify the accuracy and correctness of each of the methods at least for kinematic features, but also to thoroughly understand the kinematic and dynamic features of the wave propagation in isotropic media. The results show that the curvilinear-grid finite difference method is able to yield the correct results even for complex isotropic media; the multistage triangular shortest-path method is capable of predicting similar kinematic features as the wave-equation simulation method does, which can be used to mutually test each other for methodology accuracy and solution correctness. In addition, with the aid of the ray tracing results, it is easy to identify the multi-phases (or multiples) in the wave field snapshot, common source point gather seismic section and synthetic seismogram predicted by the wave-equation simulation method, which is a key issue for later seismic application.