Abstract The use of reflection seismic for the exploration of hydrocarbons entered a new epoch with the introduction of the horizontal stacking method in the early 1950s (Mayne, 1962, Sheriff and Geldart, 1982). In his pioneering article Mayne explained the benefit of using shorter receiver patterns for the preservation of reflected signals while delegating some of the groundroll suppression to the stacking process. For a long time the 6-fold stack was the ultimate of the seismic method, due to limitations in acquisition and processing hardware. Not until the late 1970s did reductions in station spacing, combined with a large increase in the number of recorded channels, allow higher stacking multiplicities leading to further drastic improvements in the quality of the final seismic reflection sections. The introduction of multiple-coverage recording called for the use of multichannel processing. Stacking, the multichannel process introduced first, was soon followed by poststack multichannel processes such as velocity filtering (Fail and Grau, 1963; Embree et al., 1963) and migration. Prestack multichannel processes became more feasible and even mandatory when the increase in number of recorded channels led to larger multiplicities and the reduced pattern lengths resulted in lower signal-to-noise (S/N) ratios. Some multichannel processes developed in the 1970s and 1980s are: multiple elimination, surface consistent statics correction, surface consistent deconvolution, slant stacking, dip moveout (DMO) correction, and prestack migration. Most, if not all, of these techniques perform best for high multiplicities and regular spatial sampling.

Abstract The use of reflection seismic for the exploration of hydrocarbons entered a new epoch with the introduction of the horizontals tacking method in the early 1950s (Mayne, 1962, Sheriff and Geldart, 1982). In his pioneering article Mayne explained the benefit of using shorter receiver patterns for the preservation of reflected signals while delegating some of the groundroll suppression to the stacking process. For a long time the 6-fold stack was the ultimate of the seismic method, due to limitations in acquisition and processing hardware. Not until the late 1970s did reductions in station spacing, combined with a large increase in the number of recorded channels, allow higher stacking multiplicities leading to further drastic improvements in the quality of the final seismic reflection sections. The introduction of multiple-coverage recording called for the use of multichannel processing. Stacking, the multichannel process introduced first, was soon followed by post stack multichannel processes such as velocity filtering (Fail and Grau, 1963; Embree et al., 1963) and migration. Prestack multichannel processes became more feasible and even mandatory when the increase in number of recorded channels led to larger multiplicities and the reduced pattern lengths resulted in lower signal- to-noise (S/N) ratios. Some multichannel processes developed in the 1970s and 1980s are: multiple elimination, surface consistent statics correction, surface consistent deconvolution, slant stacking, dip move out (DMO) correction, and prestack migration. Most, if not all, of these techniques perform best for high multiplicities and regular spatial sampling. Though a variety of sophisticated multichannel processing techniques were developed by the best scientists in the industry, not much attention has been paid to the basic theories required for an optimal definition of acquisition parameters. Progress in the field of data acquisition has been driven by technology, experimentation and intuition, rather than by sound theories. A breakthrough in the formulation of seismic data acquisition techniques came with Nigel Anstey’s papers on the stack-array approach (1986a and b). Anstey’s recommendations, which were based on experience and intuition, need to be modified and extended on the basis of a more theoretical approach to the three-dimensional aspects of multiple coverage data. This theoretical background is provided here and the consequences for seismic data acquisition and seismic data processing are elaborated upon. The reader’s basic knowledge of the seismic method and of signal processing is assumed. Sheriff and Geldart (1982) is recommended as a more general introduction to seismic exploration Some aspects