No abstract

No abstract

Reverse-time migration (RTM), being based on the two-way wave equation, can generate superior subsurface images in areas where other migration methods based on the Kirchhoff integral or the one-way wave equation cannot correctly image reflectors below complex overburdens such as salt or basalt. In addition, because of its ability to image reflection events that cannot be properly imaged by other imaging techniques (e.g., prism waves), RTM can be used for refining a velocity model. Thus, RTM is a tool of choice when exploring targets below salt canopies or other complexly shaped salt bodies.Despite considerable advances in computer technology, however, the cost of running RTM is still very high. The amount of computation far exceeds that of conventional one-way wave equation migration (WEM). In addition, it requires a large amount of core memory for computation. Thus, it is critical to speed up the computation and reduce the amount of memory for production usage of RTM.By dividing a velocity model into a few regions in depth, we are able to substantially reduce the computation time and the memory required to run RTM, thereby making RTM more cost-effective for large-scale production applications. The computation time can be reduced by using a larger grid size for deeper regions. The use of a larger grid size is possible because subsurface velocities usually increase as the depth increases. Since the top region is much smaller than the entire velocity model, the amount of the core memory can also be substantially reduced.