Common reflection surface stack, new method in seismic reflection data processing: A synthetic data example

Abstract

The NMO/DMO/stack method is a traditional and well-known method in the oil industry that needs an accurate macro velocity model to image the subsurface structures. Making such a macro velocity model is a time consuming process that is error prone. New introduced common-reflection-surface (CRS) stack is a data driven method which is independent of velocity information apart from the surface velocity. It comes from common reflection point (CRP) trajectory concept for finite offset. In NMO/DMO/stack, the stacked data are obtained from a summation over a curve along the offset coordinate but the principal of CRS stack is to sum along a surface of specular contributions from a segment of a reflector instead of reflection point. The summation over a segment of a coherent reflector drastically improves the signal/noise (S/N) ratio as the stacked data will show. An important aspect of the method is that the estimated parameters provide us with significant information on the subsurface structure. These are three new parameters called kinematic wavefield attributes; a, RN and RNIP. The parameter a is the emergence angle of normal-ray which will be used later for a normal-ray map migration. RN is the curvature of the exploding reflector wavefield measured at the surface and RNIP is the curvature of normal-incidence-point wave which could be used later to yield information on the propagation velocity and inversion NIP tomography. Here we processed a synthetic data to derive zero offset or CRS stacked section with high S/N ratio and better continuity on the reflection events.