Deghosting and Free-Surface Multiple Attentuation of Multi-Component OBC Data

Abstract

Summary Inverse scattering multiple attenuation (ISMA) is a method for removing free-surface multiple energy while preserving primary energy. The other key feature of ISMA is that no knowledge of the subsurface is required in its application. I have adapted this method to multicomponent OBC data (i.e., to arrays of sea floor geophones and hydrophones) by selecting a subseries made of even terms of the current scattering series used in the free-surface multiple attenuation of conventional marine surface seismic data (streamer data). This subseries a p proach has allowed me to remove receiver ghosts and free-surface multiples in multi-component synthetic and real OBC data. I have processed each component separately. As for the streamer case, my OBC version of ISMA preserves primary energy and does not require any knowledge of the subsurface. Moreover, the preprocessing steps of muting for the direct wave and interpolating for missing near offsets are no longer needed. However, the knowledge of source signature is still required. The existing ways of satisfying this requirement for streamer data can be used for OBC data without modification. This method differs from the present dual-field deghosting method used in OBC data processing in that it does not assume a horizontally flat sea floor nor does it require the knowledge of the acoustic impedance below the sea floor. Furthermore, it attenuates receiver ghosts (receiver-side reverberations) as well as free-surface multiples (source-side reverberations). Introduction The inverse scattering multiple attenuation (ISMA) series allows us to remove free-surface multiples of first and higher orders from seismic data and to preserve primary reflections and diffractions. The series is made up of an infinite sum of terms, each term corresponds to an order of multiple which may then be scaled and substracted from the field data. ISMA is an attractive method for attenuating freesurface multiples because it is multidimensional, it does not require any knowledge of the subsurface and more importantly it preserves primary energy [e.g., Carvalho et al. (1991); Matson and Weglein (1996); Ikelle et al. (1997)). I am here going to present an adaptation of this method to the OBC (ocean bottom cable) data. As significant amount of reverberations in water are categorized as receiver ghost (Figure 1) in the ODC experiment, I will aim to produce an OBC version of ISMA which also attenuate receiver ghosts. The main difficulty arising in the multiple attenuation of ocean bottom cable (OBC) data is that the receivers are located at the sea floor which can have a very heterogeneous structure. The bathymetric mapping of the Gulf of Mexico performed by Hilde et al. (1991) provides a good illustation of how heterogeneous the sea floor can be. It is therefore important to try to develop multiple attenuation methods which do not require a knowledge of the sea floor and when possible, any knowledge at all of the subsurface. The current multiple attenuation methods (more precisely deghosting) of OBC data are based on up/down separation at the sea bottom. As shown in Figure 1, seismic events in OBC data can be grouped into downgoing and upgoing wavefields. The algorithms for performing this wavefiled separation can be found in Barr and Sanders (1989), Amundsen and Reitan (1995) and Osen et al. (1996).