Workflow For Surface Multiple Attenuation In Shallow Water

Abstract

Abstract Surface related multiple elimination (SRME) method is known to have difficulties in attenuating shallow water multiples. Particularly, for water-layer related multiples in shallow water environment such as the Gulf of Thailand because the primary water-bottom reflection that is required by SRME for predicting the multiples is not recorded due to the near offset gap between the source and the streamers. Consequently, these multiples often obscure the interpretation of hydrocarbon reservoirs and significantly reduce the effectiveness of seismic inversion. Conventionally, predictive deconvolution in either x-t or tau-p domain is often used in processing workflow to suppress this kind of multiples. However, deconvolution also attenuates primary events that have a periodicity close to that of the water-layer. In this paper, we present a two-step processing workflow for removing free-surface multiples in shallow water situations. Firstly, we use a multi-channel prediction filter estimated from the multiples for attenuating short-period water-layer related multiples. Secondly, we apply SRME for suppressing other long-period surface multiples generated by sub-surfaces underneath the water-bottom. Our workflow is applicable to any marine data that is plagued with surface multiples, particularly the short-period multiples caused shallow water-layer. It can also be used for reconstructing the water-bottom reflection that is not recorded in the data. Through real-data examples from China and Australia, we demonstrate that our workflow provides an optimal multiple attenuation solution in shallow water environment in comparison with conventional methods such as t-p deconvolution or SRME alone. This work adds to the understanding of multiple attenuation in shallow water by highlighting the shortcomings of the existing methods that are routinely used in seismic processing. Further, it presents how the issue can be tackled by separating the multiples into short-period and long-period ones so they can be handled by our proposed two-step workflow.