Anatoly Baumstein1*, Mohamed T. Hadidi2, Ramesh Neelamani1, and Warren Ross1
We present a comprehensive discussion of the main steps involved in applying data-driven surface-related multiple attenuation (SMA) and inter-bed multiple attenuation (IMA) algorithms, illustrated with field data examples from the Middle East. Despite recent advances (e.g. Kelamis et al., 2002), inter-bed multiple attenuation continues to present a formidable problem due to sparsity of data acquisition, noise, statics and the difficulty in distinguishing between primaries and multiples in many geologic settings. We demonstrate that accurate identification of key multiple generators is an important first step in applying IMA and describe a vertical seismic profile (VSP) deconvolution technique that may be helpful in analyzing multiple-generation mechanisms. Based on this analysis, we apply both post-stack (1-D) and pre-stack (1.5-D) multiple attenuation. In the case of pre-stack multiple attenuation, we address the data regularization procedures necessary to reconstruct the regularly and densely sampled common mid-point (CMP) gathers that are required for successful application of IMA. We also show that surface-related multiple attenuation is an important step in the pre-stack case and discuss how to overcome statics-related problems that arise during the application of SMA. Finally, we describe the trade-offs involved in choosing adaptive subtraction parameters in areas with overlapping primaries and multiples. We conclude that in the target reservoir zone, the application of pre-stack CMP-domain, data-driven, multiple-attenuation methods resulted in significant structural interpretability improvements. The signal-to-noise ratio was further enhanced through an application of a novel post-processing curvelet-based coherent noise attenuation technique, described in detail in the companion presentation by Neelamani et al.
AAPG Search and Discovery Article #90077©2008 GEO 2008 Middle East Conference and Exhibition, Manama, Bahrain