A wavelet-based method for normal moveout correction

ABSTRACT Seismic normal moveout (NMO) correction processes a seismic nonzero-offset trace sample by sample conventionally, resulting in stretching at the NMO-corrected trace. This effect of NMO stretching degrades the time resolution of NMO-corrected seismic traces and compromises the analysis of amplitude variation with offset (AVO). A wavelet-based method for NMO correction (W-NMO) was proposed, which will reduce the NMO stretching effect. The W-NMO method models seismic traces of a common midpoint gather as a linear combination of weighted unit monopole seismic responses, in which the weights correspond to a set of coefficients. These weighted monopole seismic responses are statistically shifted from their nonzero-offset two-way traveltime to the zero-offset two-way traveltime, and as a result, all time samples associated with the same reflection event of a seismic trace share a common NMO correction time. The W-NMO method preserves the integrity of reflection events without introducing stretching, thereby enabling effective subsequent processing and analysis. Synthetic and field data examples demonstrated that W-NMO correction preserves the shape of seismic reflection events when provided with accurate wavelet and NMO correction velocities and enhances time resolution and lateral continuity in seismic data.