Preprocessing issues associated with multiple attenuation in water depths of less than 150 meters: ISMA and predictive deconvolution

Abstract

Over the past three decades, marine exploration has been essentially limited to water depths within the 200-500 m range. With most reservoirs in this range depleting, exploration is moving towards very shallow water, less than 100 m, as well as deeper water, greater than a kilometer in depth. These two frontiers pose new challenges for seismic acquisition and processing. Our interest in this thesis focuses on the challenges associated with acquisition and processing in shallow water regions, in particular the attenuation of free-surface multiples. Free-surface multiple elimination in shallow-water must address removal of the direct wave, interpolation of missing near-offsets, and the presence of guided waves and strong refracted wave energy. In order to study these challenges associated with acquisition and processing of shallow-water data, synthetic seismic data for 150, 75 and 25 m water depths were generated using a fully elastic, finite difference algorithm. The direct wave was removed with a "model then subtract" method due to the interference with the sea-floor primary event. The synthetic data contained offsets from 0-2000 m, requiring no interpolation of missing near-offsets. The guided wave energy was removed with an f-k filter, and the refracted energy was muted after the data was corrected for normal moveout. Predictive deconvolution (Preddcon) and Inverse Scattering Multiple Attenuation (ISMA) were two algorithms compared for their effectiveness in eliminating free-surface multiple energy from the data. For the 150 m case, Preddcon removed most of the multiple energy in the very near-offsets, where theoretical assumptions remained valid. Preddcon removed multiple energy over a larger offset range for the 75 m case, within the design window. For larger offsets, where the assumptions cannot be satisfied, multiples were not attenuated and primaries were altered. For both water depths, the two-pass approach adopted for ISMA performed far superior across all offsets for removing free-surface multiples, while preserving primary amplitudes, because the primaries and multiples were well-separated. However, for the 25 m case, where there is no separation between primaries and multiples and moveout difference is almost negligible, Preddcon had superior results because the multiple reverberation pattern remained similar across offsets.