- Copyright © 2000 Society of Exploration Geophysicists
The problem of vertical resolution is to determine the wavelet at the target. Mis-ties at wells of ±20 ms are normal in the best 3-D seismic data, and they present no simple pattern. The seismic wavelet in poststack 3-D data clearly varies both laterally and vertically. The lateral variations are caused by two effects: shot-to-shot and line-to-line variations in acquisition parameters (especially source-signature variations) and prestack, trace-dependent filtering processes (such as predictive deconvolution). The vertical variations are caused by attenuation in the earth. The approach described in this paper is to eliminate the lateral variations in acquisition using independent measurements of the source wavefield and to ensure that any trace-dependent filtering has no effect on the wavelet. Vertical variations are then tackled as a separate problem. This approach provides a firm framework within which to interpret subtle stratigraphy.
During the last 30 years, resolution of lateral acoustic discontinuities in 3-D seismic data has improved very rapidly, and it is now substantially better than vertical resolution. This is a consequence of many developments in acquisition/processing (particularly migration) of 3-D seismic data for the determination of structure. Modern data permit routine resolution and mapping of faults with throws on the order of 10 m and the mapping of meandering river channels over large distances using horizontal time slices.
By comparison, vertical resolution has been somewhat neglected. The problem is to determine the wavelet present in the data at the zone of interest. This problem is intimately related to the well-tie problem and the determination of fine stratigraphy at the theoretical limit of vertical resolution, which is usually reckoned to be some fraction of the shortest wavelength in the data. As Jon Claerbout has explained, the difference between theory and practice is much less in theory than in practice. In this case, the …