Abstract Three-dimensional finite-difference modeling studies conducted over subsalt structures in the deepwater Gulf of Mexico confirm the deficiencies of narrow-azimuth towed-streamer surveys and predict significant improvement in image quality with wide-azimuth methods. Finite-difference modeling has provided important design parameters for two separate approaches for wide-azimuth surveys: ocean-bottom receivers distributed in a sparse grid on the ocean floor coupled with a dense grid of source points on the surface, and a wide-azimuth towedstreamer method using multiple seismic vessels in a novel configuration. These two methods complement each other. Oceanbottom receivers may be used effectively where field development has resulted in many obstacles that might interfere with towed-streamer methods, where the required size of the 3D survey is not too extensive, or where very long offsets are required for all azimuths. Towed-streamer methods are more efficient for large surveys, and key parameters in the wide-azimuth towed-streamer method can be varied to provide a wide range of cost versus data-quality options to make the method suitable for application in scenarios ranging from exploration to field development.