Interferometric application of static corrections

Abstract

Correcting reflection seismic data for the effects of near-surface irregularities is a persistent problem usually addressed at least partly by static corrections applied to traces. However, there are areas where static corrections are ineffective because basic assumptions are violated. The assumptions which fail most often are surface consistency and stationarity, which are central to the concept of static corrections. To address this failure, I mapped raw seismic traces into the radial trace domain and gathered the radial traces by common surface angle. Then I imposed a more general constraint, raypath consistency, which simultaneously introduces nonstationarity. Conventional static correction also assumes implicitly that reflection events consist of single discrete arrivals. This is not true, however, in regions where near-surface multipathing and scattering complicate reflection event waveforms. Borrowing from recent work in seismic inferometry, I removed the single-arrival assumption by using trace crosscorrelations to estimate and deconvolve surface functions from traces, rather than applying time shifts. The entire crosscorrelation function is used in every case, so both timing and waveform variations are removed by the deconvolution. The operation is applied in the common-angle domain, so it is raypath consistent and nonstationary. The method, dubbed “raypath interferometry,” was applied successfully to a set of 2D Arctic field data with serious surface consistency and multipath problems, and to a set of 3C 2D land data with very large S-wave receiver statics. Although intended primarily for use on seismic data for which conventional statics corrections fail, raypath interferometry can be used on any seismic data; its assumptions include single-arrival events and surface consistency as special cases.