Attenuation of multiple diffractions using a cascaded noise removal sequence

John Brittan and Andy Wrench, PGS Geophysical, UK, describe a seismic processing sequence to suppress diffracted multiple energy. Multiple diffractions are a significant noise problem on seismic data from many of the world’s most important hydrocarbon-rich provinces. These noises are most prevalent in the case of deep water and a complex near-surface reflection sequence. In such regimes the multiples of diffracted energy by the near surface geology are often coincident in time with primary reflections within the subsurface. The energy composing the primary reflections will have a considerably reduced amplitude and frequency content due to the long travel-paths through the Earth. By contrast, the multiples from the near-surface reflectors have travelled most of their propagation path through the very weakly absorbing sea-layer. Thus the deeper parts of the seismic section are dominated, particularly at high frequencies, by the incoherent, high amplitude diffracted multiple arrivals. Due to their aliased, non-hyperbolic nature these arrivals are difficult to suppress using standard demultiple methods, e.g. 2D SRME (Surface Related Multiple Elimination), parabolic Radon demultiple. It has been shown that the use of high-fold, multiazimuth data acquisition can considerably increase the effectiveness of CMP stacking in removing diffracted multiple energy (Widmaier et al., 2002). However, for standard marine acquisition, a rigorous noise attenuation methodology must be adopted. Indeed, while techniques such as 3D SRME hold great promise for multiple diffraction attenuatio (van Borselen et al., 2004), we discuss in this paper how for typical field data a cascaded sequence of (up to) four different processes can be used to attenuate the problem events (Figure 1).