Normal Moveout through partial corrections

In seismic data processing, Normal Moveout (NMO) correction is applied to common mid point (CMP) gathers to simulate a zero-offset trace from traces recorded at variable source to receiver offset x. Then, properly NMO corrected CMP gathers can be the input for a number of different tasks: AVO analysis, residual statics computation, and most commonly, stacking to produce a seismic image with higher S/N ratio and with attenuated multiple reflections. The NMO correction in exploration seismology is a common practice. The wavelet stretching and the event duplication resulting from this procedure are removed from the seismic data by the application of mute functions, computed on the basis of the maximum stretch threshold allowed. Usually, enough data are still available after the mute, because stretching and event superposition have a limited influence at depths of interest in exploration seismic. This is not the case in near surface seismic, where the NMO stretch and the corresponding mute constitute a major problem especially for shallow reflections. Indeed, not only the mute limits the stack potential in enhancing the S/N ratio, but sometime this effect is so severe that all the recorded seismic data at shallow times are zeroed-out. Various methods have been proposed in literature to overcome the problem of the NMO stretching. Trickett (2003) made a synthetic and clear review of the different methodologies that have been proposed since 1975, with an accurate description of the main drawbacks each of the proposed algorithms is affected by. The stretch-free stacking proposed by Trickett (2003) replaces the NMO correction and stacking with a single inversion step to zero-offset. As pointed out by the author, no NMO corrected CMP gather is formed by this procedure, limiting the method’s applicability. Our procedure tries to overcome this limitation suggesting, at the same time, an alternative way to estimate the zero-offset stack trace. This new algorithm was applied to synthetic and several field data-sets with satisfactory results