The importance of non-hyperbolic and stretch effects in far-offset P and PS NMO processing

Abstract

The general aim of seismic surveying is to map the subsurface in as much detail as possible, subject to economic constraints. As technology has developed so has the ability to acquire and process larger amounts of data. This often translates to acquisition of larger offsets, potentially increasing the fold and signal-to-noise ratio of the stack. Far-offset traces are subject to non-hyperbolic NMO, which may be handled by incorporating higher-order anisotropic terms. However, even if far-offset nonhyperbolic events can be flattened, they are likely to suffer from NMO stretch. This can result in a serious reduction in dominant frequency, and hence in vertical resolution. Several techniques have been published which apply NMO to P-wave data without introducing stretch. We have focused on extending one of these techniques through analysis of modelled and production data. We have also extended the analysis to include converted-wave (PS) data, where NMO stretch can have even greater impact. For PS surveys, reflections on the near offsets have lower amplitude and are often swamped by noise. Therefore, most of the data contributing to the stack are from the mid to far offsets, particularly at the shallower coal scale. The dominant frequency of PS data can be significantly reduced by NMO stretch. This may be one of the factors that contribute to the poorer than expected resolution observed on some PS imagery. Both the P and PS non-stretch NMO algorithms developed in this investigation successfully incorporate anisotropic parameters to accommodate non-hyperbolic NMO effects.