Prestack Elimination of Complex Multiples: a Gulf of Mexico Subsalt Example

A. Guitton,G. Cambois,C. Houston

DOI: https://doi.org/10.1190/1.1820146

Abstract:Summary Since the surface-related multiple elimination technique popularized by the DELPHI consortium is independent of subsurface structure, it can theoretically be used in any complex environment. In practice, however, the limited offset range of field data introduces time-, dip- and offsetdependent amplitude artifacts into the multiple model. This, plus the fact that sources and receivers are located below the water surface, make it impossible to extract a relevant wavelet, which is a critical step in the method. Consequently, this technique is not adaptive enough to provide adequate multiple attenuation. A new method has been developed that eliminates the need for wavelet extraction in the DELPHI technique. It utilizes the kinematically accurate multiple model and applies a pattern-recognition algorithm to remove the actual multiples. This adaptive technique offers optimum multiple attenuation without damaging the primaries. However, its effectiveness is still somewhat limited by the amplitude inaccuracies of the multiple model. In areas where first-order multiples can be isolated, it is possible to extract a posteriori the seismic wavelet using a simple matching filter between estimated and modeled multiples. Studies of these extracted wavelets have shown that they are indeed offset-dependent, confirming the limitations of the single wavelet approach. Thus, the pattern-recognition technique always gives better results than the costly non-linear wavelet inversion associated with the DELPHI technique.