3D Modeling and Migration of a Wide-Azimuth Towed Streamer Survey

Abstract

A 3D synthetic wide-azimuth towed-streamer (WATS) dataset is modeled and migrated. The modeling was performed on MareNostrum at the Barcelona Supercomputer Center using a maximum of 2000 cpus. This dataset is then migrated with three imaging techniques suited for complex overburden. First, a wave-equation method based on a one-way propagator is used. Second, a two-way method that utilizes one-way propagators for the wavefield extrapolation downward and upward is tested. Finally, a method based on the solution of the twoway acoustic wave equation, also known as Reverse Time Migration (RTM) is selected. We compare the migration results in 2D and 3D and show that the best results are obtained when more information is incorporated in the imaging process, e.g., turning and/or prismatic waves. In practice, the selection of a migration algorithm is based on computational and geophysical considerations. For instance, the complexity of the subsurface indicates if turning waves are needed or not. Our ability to estimate an accurate velocity model influences which method will produce the best results. Finally, computing resources could present challenges when large datasets need to be migrated, especially for advanced imaging techniques such as RTM. Here, our ability to use MareNostrum, the 5th largest computer in the world, proved extremely valuable.