On Separation of Reflections and Diffractions Using a CRS/MUSIC Approach

Seismic diffracted waves carry high-resolution information about important subsurface structures associated with potential hydrocarbon traps like faults and pinch outs. However, the amplitudes of these waves are in general much weaker than those of the reflected waves and diffractions often fall within the noise level. As a consequence, classical seismic processing schemes mostly regard these waves as noise. Recently, however, different approaches have been proposed to separate diffractions from reflections and to further make use of these contributions within velocity analysis and local imaging of the subsurface (Fomel et al., 2007; Moser and Howard, 2008). In this paper we propose to employ the Common Reflection Surface (CRS) technique (Jager et al., 2001)to enhance the diffractions through optimized stacking. To further ensure a high-resolution separation of diffracted energy we also replace the classical coherency measure of semblance with that of MUSIC (MUltiple SIgnal Classification; Schmidt (1986)). The potential resolving power of this combined technique of CRS and MUSIC has been demonstrated employing a synthetic data set involving two closely separated point diffractors placed in the vicinity of a dipping reflector.