When velocity contrasts at the base of the near-surface layer are small, or when it shows some degree of structural complexity, delay times of wavefronts transmitted through this layer may become raypath dependent. Due to the low velocity of S-waves this dependency is translated into significant non-stationary delays. In order to remove this effect we transform the data to a domain in which amplitudes are a function of the raypath angle. Processing the statics in the radial trace (RT), Snell trace (ST) and τ-p domain achieve this goal, but with different degrees of accuracy. Since the first two domains involve only a very simple remapping of the amplitude from the original x-t domain they are free of the numerical problems experienced with the τ-p transform. However, the RT and ST transformations require that some assumptions about the subsurface velocity model must be made. On the other hand, the τ-p transform automatically scans the data in order to capture the rayparameter values that were actually recorded. The surface correction of synthetic data in a depth-varying velocity medium showed that the solutions obtained in the ST and τ-p domain are very similar, while common offset and RT solutions still show some unresolved problems. Further analysis was performed using real data from the Hussar experiment. Results showed that the τ-p solution provided a better stacking power in the deeper part of the section. In the shallow part the RT and ST solution seem to have better resolution. Further work is needed to reduce the artifacts present in the τ-p transformation. Trying a high resolution τ-p transform will be the next step in this research.
View full article as PDF (10.23 Mb)