With 3D data, multiples are not only function of the inline offset but also of the crossline offset and thus SRME should be applied in 3D. In principle this is possible, since the theory of SRME is not limited to 2D data. In practice, however, the demands of SRME in terms of crossline sampling and crossline aperture make its application challenging. A great deal of research is being carried out on efficient and accurate ways of doing crossline interpolation and extrapolation and in making 3D-SRME practical from the computation point of view Nekut (1998); van Dedem and Verschuur (1998), but this still remains a problem with most real datasets. An alternative, therefore, is needed for those cases where we are unable or unwilling to spend the human and computer resources necessary for 3D-SRME to work properly.
Given the relative simplicity of the Radon method applied in the image space Alvarez and Artman (2005), extending its application to account for the effect of crossline offset seems attractive. The first step in that direction is to understand how the moveout of the multiples will behave on 3D Subsurface-Offset-Domain Common-Image Gathers (SODCIGs) and Angle Domain Common Image Gathers (ADCIGs). To that effect, we will use a very simple 3-D synthetic prestack dataset provided by ExxonMobil.
In the next section we describe the data in some detail to illustrate the difficulties of forming a complete dataset with uniform sampling in all five dimensions (time, inline and crossline position, and inline and crossline offset) small enough to fit in our computers. Then we briefly describe the preprocessing of the data and in the last two sections show the migration results of the primaries and multiples in pseudo and true 3D ADCIGs.