Summary
The simulation of a zero-offset section leads to a first interpretable time image and is still one of the key processing steps in seismic imaging. While recent works have indicated that common-offset stacking results in improved resolution and illumination in complex settings, the zero-offset approximations are still reasonably accurate, especially when lateral heterogeneity is moderate. Due to the increased dimensionality of the problem, the common-offset stack is computationally expensive, though. The partial common-reflection-surface stack uses local subsets of globally defined zero-offset operators to perform common-offset stacks. In this work, we suggest an extension of this scheme, in which not only travel times but also slope information is extracted from the zero-offset surfaces. We show, with simple synthetics and for a complex field data example from the eastern Mediterranean, that the presented method allows for efficient full prestack slope analysis and refinement, which can help to further automate the picking-intensive process of stereotomography.
Introduction
The 2D common-reflection-surface (CRS) stack is a multiparameter extension of the classical CMP method (Mayne, 1962). It was formulated for a zero-offset (ZO, Jäger et al., 2001) and for an arbitrary common-offset (CO) central ray (Zhang et al., 2001; Höcht et al., 2009). While the ZO approximation is fast and reasonably accurate for moderate lateral heterogeneity, the CO counterpart shows its strengths in complex settings, providing improved resolution and illumination at the cost of higher computational expenses (see, e.g., Spinner et al., 2012).
In most implementations of the CRS stack, events are described by a hyperbolic operator (Schleicher et al., 1993; Jäger et al., 2001), which, for moderate reflector curvature and comparably weak lateral heterogeneity can lead to high accuracy over a wide range of offsets and midpoints. Based on this assumption of globality, the partial CRS stack introduced by Baykulov and Gajewski (2009) utilizes CO subsets of ZO CRS operators for efficient prestack data enhancement, interpolation and regularization (e. g., Eisenberg-Klein et al., 2008).
In this work, we seek to extend the approach of partial CRS by not only extracting local traveltimes but also slope information from the estimated ZO operators. Based on the work by Lavaud et al. (2004) we suggest a simple scheme, in which the extracted slopes, i. e., first-order coefficients are used to perform a CO stack. We show, with simple synthetics and a complex field data example, that the presented method, due to its formulation in terms of CO attributes, allows for efficient coherence-based local refinement, whose output may directly be used in prestack slope-based stereotomography (Billette and Lambaré, 1998). In the following section, we briefly review the basics of the CRS method and formulate the theoretical framework of the suggested approach.
