TX 75083-3836, U.S.A., fax +1-972-952-9435.
AbstractMultiple is a long-standing problem in petroleum seismology. Despite significant achievements in developing advanced techniques to attenuate multiples, there is currently no single multiple-attenuation technique that can be applied to attenuate all the various types of multiples. The performance of each technique depends on the particular dataset. Selecting an optimal technique (or an optimal combination of schemes) lies in determining whether the Earth models implicit in a particular technique match the Earth model from which the seismic data were physically acquired. In this paper we present an integrated workflow using seismic modelling and SRME (Surface-Related Multiple Elimination) to attenuate multiples for the GlyVeST (Glyvursnes -Vestmanna Seismic Tie) seismic data, which were acquired in the Faore Islands. First, we carried out the GlyVeST seismic modelling, which has shown, among other things, that surface-related multiples are much stronger and more of a problem than internal multiples. Subbasalt seismic processing is well known to face challenges from severe scattering losses by impedance contrasts and rugose surfaces, geometrical spreading, velocity heterogeneity and strong multiples. As there can be great uncertainty in going from real data to interpretation, we gain insight by starting with an assumed or 'known' geology and generating synthetic data from it. Then, processing these data, knowing the correct result, guides us to an optimal processing strategy. The synthetic data, produced by an elastic finite-difference modelling scheme, offer excellent possibilities for detailed studies of how seismic waves interact with heterogeneous basalt flows, as well as better understanding of the multiple problem. Instead of resorting to trial-and-error, i.e. testing many different multiple-attenuation methods and picking the most pleasing result -wasteful of time and computer resources -the GlyVeST seismic modelling has helped us to determine a multiple-attenuation strategy for the Earth model from which the real GlyVeST seismic data were acquired. Based on the characteristics of our multiples (i.e. surface-related multiples much stronger than internal multiples), we have applied SRME, a technique that attenuates surface-related multiples as much as possible, followed by predictive deconvolution on the GlyVeST data. Finally, having conditioned the data for velocity analysis, we apply a high-resolution Radon demultiple routine in a series of iterative passes (velocity analysis -Radon demultiple -velocity analysis), progressively harsher as confidence in the velocity trend increases. As a result, multiples have been effectively attenuated without harming primary events in the data.