Surface Waves Suppression Using Interferometric Prediction and Curvelet Domain Hybrid L 1 / L 2 Norm Subtraction

Yibo Wang,Shuqian Dong,Yanwei Xue
DOI: https://doi.org/10.1190/1.3255543
2009-01-01
Abstract:Surface waves are coherent seismic noises which obscure valuable reflection signals and therefore it is crucial to remove them. We present a prediction‐subtraction surface waves suppression method. There are two steps for implementation. First, surface waves are predicted from raw data by cross‐correlating recorded traces in a shot gather followed by summing correlated trace pairs over all shot positions. Second, the predicted surface waves are subtracted from raw data by minimizing a curvelet domain hybrid and norm objective function. Synthetic and field seismic data contaminated by high amplitude and dispersive surface waves are used for testing. Results show that most dispersive surface waves can be well suppressed; but because of the approximations introduced by a limited recording aperture, not all dispersive surface waves can be thoroughly predicted, so there are still some residual surface waves after subtraction. For comparison, a time‐distance domain least squares subtraction method is also used in second step. The comparisons show that curvelet domain hybrid and norm subtraction is more suitable for integration with interferometric prediction in some cases.
What problem does this paper attempt to address?