A Competitive Comparison of Multi-parameter Stacking Approaches

In literature four multi-parameter stacking approaches can be found that stack along midpoint and offset direction: multifocusing, common-reflection-surface (CRS), implicit CRS and non-hyperbolic CRS. In recent years some comparisons were published. However, in these works some methods were missing and conflicting dips, mostly caused by diffractions, were not considered. Diffractions are of higher-order and therefore benefit the most from these methods. Schwarz et al. introduced a new parametrization for CRS-type stacking operators in terms of time and slowness shifts. They found, that multifocusing utilizes a time shift to account for overburden heterogeneity, while the common-reflection-surface based methods adopt slowness shifts. For a fair comparison we transform all operators in both representations and discuss their performance on an industrial field data set. The focus is on diffractions since they benefit the most from non hyperbolic operators. The required attributes are estimated via differential evolution, a global optimization technique. The non-hyperbolic CRS needs almost the same computation time as the standard CRS while achieving the same quality as multifocusing and implicit CRS which need more computational time. The improvements of non-hyperbolic CRS with respect to CRS are mainly visible for diffractions. A diffraction separation clearly shows improved results for the non-hyperbolic CRS.