3D surface-related multiple elimination: Data reconstruction and application to field data

Abstract

The wide success of 2D surface-related multiple elimination (SRME) in attenuating complex multiples in many cases has spurred efforts to apply the method in three dimensions. However, application of 3D SRME to conventional marine data is often impeded by severe crossline aliasing characteristic of marine acquisition geometries. We propose to overcome this limitation using a dip-moveout (DMO)-based procedure consisting of the following steps: resorting the data into common offsets to improve crossline sampling, performing DMO to eliminate azimuth variations in the common-offset domain, and efficiently implementing inverse shot-record DMO to reconstruct densely sampled shot records required for 3D SRME to predict multiples correctly. We use a field data example to demonstrate that the proposed shot reconstruction procedure leads to kinematically accurate reconstruction of primaries but may not be able to simultaneously position multiples correctly. The mispositioning of multiples becomes a problem when second- and higher-order multiples must be predicted. We propose to resolve this difficulty by using a layer-stripping approach to multiple prediction. Alternatively, an approximate algorithm that relies on adaptive subtraction to compensate for inaccurate positioning of predicted multiples can be used. Application of the latter approach is illustrated with a field data example, and its performance is evaluated quantitatively through a measurement of S/N ratio improvement. We demonstrate that a DMO-based implementation of 3D SRME outperforms conventional 2D SRME and can accurately predict and attenuate complex 3D multiples.

Surface‐related multiple prediction for ocean‐bottom node data based on demigration using downgoing wave imaging data

“…With the development of marine threedimensional seismic exploration, scholars have carried out research on the surface-related multiple elimination (SRME) method for three-dimensional conventional streamer seismic data, which solved some problems for multiple prediction, such as sparse source spacing (i.e. navigation line spacing) in the direction of the connecting line, less receiver points and limited offset range, and developed a variety of practical methods finally (Baumstein & Hadidi, 2006;Hokstad & Sollie, 2006;Lin et al, 2004;Pica et al, 2005;Van Dedem & Verschuur, 2005;Zhang & Verschuur, 2021).…”