Recently, a number of seismic processing techniques, collectively termed `interferometry' have been devised to help image various types of seismic data, or to remove unwanted effects from raw data attributable to source/receiver coupling, surface layer irregularities, and scattering. Although varying in their objectives and formulation, these techniques are generally based on cross- correlations of raw seismic traces. Previously, we demonstrated success in applying near-surface corrections to two sets of arctic data using two different implementations of interferometry, `virtual trace gathers' and `static deconvolution'. Additionally, for one of these data sets, surface- consistency was generalized to `raypath-consistency', and stationarity was abandoned. Interferometry was then applied successfully in the radial trace domain.
Since interferometry can handle very large `statics', as demonstrated with the arctic data, we made some initial attempts to apply similar techniques to converted wave data, but with limited success. We demonstrate here a more recent attempt to remove surface effects from converted wave data, using a `hybrid' approach, in which we correlate traces from two different vector components at the same receiver positions in order to detect and remove the receiver `static differences'. The results show greatly improved coherence; but have yet to properly account for the `structural' term and can exhibit spurious coherent events due to fortuitous correlation of unrelated PP and PS events.
View full article as PDF (6.87 Mb)