Diffraction stacking with stacking velocity analysis in a surface seismic survey

摘要

A new type of prestack time migration method whose data processing scheme is analogous to that of conventional CMP method is presented. This proposed data processing method does not require multiples iterations for achieving the velocity structure in the case of prestack migration. The optimum constant stacking velocity can he determined at each image point from a stacking velocity analysis based on primary diffraction patterns.Both the conventional CMP stacking method and the conventional stacking velocity analysis are modified as follows.1. In the case of the CMP stacking method, the amplitudes of observed data are stacked along the reflection pattern in a CMP gather. On the other hand, in the case of diffraction stacking method with stacking velocity analysis, the amplitudes of observed data are stacked along the diffraction pattern in a common source gather or common receiver gather.2. Stacking velocity analysis can be a useful tool for detecting characteristic patterns. Conventional stacking velocity analysis is modified to detect patterns caused by diffraction events. Final stacked records can be obtained based on maximum coherency peaks on stacking velocity analysis panels.Firstly, proposed data processing procedure is described in detail. Secondly, the characteristic analysis of the proposed data processing is presented. We studied the S/N ratio of reflection images by numerical experiments which especially considered the effect of stacking aperture, defined as an angle range of collecting seismic traces. Our numerical experiments revealed that diffraction stacking generally has the ability to produce a seismic reflection image with higher S/N ratio than CMP stacking for a horizontal and dipping reflector. This can be explained by Huygens' principle. In case of dipping reflector, however, source/receiver geometrical arrangements are important. And also, stacking coverage, defined as an angle range of collecting seismic traces, is an important factor in obtaining the highest S/N ratio. This can be explained in terms of Fresnel zones. We also found that S/N ratio and horizontal resolution are in inverse proportion to each other.