Multiple suppression for 2D marine seismic data using surface-related multiple elimination and parabolic radon filtering

Abstract

Marine seismic reflection method is the most efficiency method and widely used in petroleum industry to map and interpret the potential of petroleum reservoirs. Multiple reflections are the particular problem in marine seismic reflection investigation, as they often obscure the target reflectors in seismic profiles. Multiple reflections can be categorized by considering the shallowest interface on which the bounces take place into two types: internal multiples and surface-related multiples. Besides, the multiples can be categorized on the interfaces where the bounces take place, a difference between long-period and short-period multiples can be considered. The long-period surface- related multiples on 2D marine synthetic and real marine seismic data were focused on this research. The 2D marine synthetic seismic data was generated by Hess Corporation, and the 2D real marine seismic data of the East Coast of the United States-Southern Atlantic Margin was conducted by the U.S. Geological Survey between 1974 and 1978 in order to address hydrocarbon resource potential and stratigraphic history. The seismic profile demonstrates the effectiveness of the results from predictive deconvolution and the combination of surface-related multiple elimination (SRME) and parabolic Radon filtering. First, predictive deconvolution applied on conventional processing is the method of multiple suppression. The other, SRME is a model-based and data-driven surface-related multiple elimination method which does not need any assumptions. And the last, parabolic Radon filtering is a moveout-based method for residual multipleMarine seismic reflection method is the most efficiency method and widely used in petroleum industry to map and interpret the potential of petroleum reservoirs. Multiple reflections are the particular problem in marine seismic reflection investigation, as they often obscure the target reflectors in seismic profiles. Multiple reflections can be categorized by considering the shallowest interface on which the bounces take place into two types: internal multiples and surface-related multiples. Besides, the multiples can be categorized on the interfaces where the bounces take place, a difference between long-period and short-period multiples can be considered. The long-period surface- related multiples on 2D marine synthetic and real marine seismic data were focused on this research. The 2D marine synthetic seismic data was generated by Hess Corporation, and the 2D real marine seismic data of the East Coast of the United States-Southern Atlantic Margin was conducted by the U.S. Geological Survey between 1974 and 1978 in order to address hydrocarbon resource potential and stratigraphic history. The seismic profile demonstrates the effectiveness of the results from predictive deconvolution and the combination of surface-related multiple elimination (SRME) and parabolic Radon filtering. First, predictive deconvolution applied on conventional processing is the method of multiple suppression. The other, SRME is a model-based and data-driven surface-related multiple elimination method which does not need any assumptions. And the last, parabolic Radon filtering is a moveout-based method for residual multiple reflections based on velocity discrimination between primary and multiple reflections, thus velocity model and normal-moveout correction are required for this method. The predictive deconvolution is ineffective for long-period surface-related multiple removals. However, the combination of SRME and parabolic Radon filtering can attenuate almost long-period surface-related multiple reflections and provide high- quality seismic image of marine seismic data.