Indonesia Conference Directory

Surface Related Multiple Elimination (SRME) technique is known to have ineffectiveness to attenuate shallow water multiples, primarily because the water-bottom reflection required by SRME for predicting the multiples is not recorded due to lack of near offset data. The conventional technique using predictive deconvolution in either x-t or tau-p domain is often used to suppress these multiples, however, it also attenuates primary events that have a periodicity close to that of water layer. These limitations cause unrealistic seismic images in final output. In this paper, we propose an enhanced technique by modelling the water-bottom reflection and then adding it to the recorded seismic data. The modified data can then be used to predict first order multiple. We present multi stages processing workflow for removing water-bottom 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 F-K filter for isolating signals from noise in frequency-wavenumber (f-k) domain. Lastly, SRME technique for suppressing other long-period surface multiples generated by sub-surfaces underneath the water-bottom. Through real-data examples from Nias Waters, North Sumatra, we demonstrate that our workflow provides an optimal multiple attenuation solution in shallow water environment in comparison with conventional methods such as predictive deconvolution or SRME alone. Tests result show that the multi stages technique can produce better result than using predictive deconvolution or SRME alone. Furthermore, we also show SRME approach can give optimum seismic migration images after we combined these stages. We successfully applied these techniques in 2D seismic data of Nias Waters.