Fast least-squares imaging with surface-related multiples: application to a North-Sea data set

In marine seismic acquisition, surface-related multiples constitute a significant portion of the acquired data. Typically, multiples are removed during early-stage data-processing as they can lead to phantom reflectors during migration that may result in erroneous geological interpretations. However, if properly dealt with, multiples can provide valuable extra information and complement primaries in illuminating the subsurface. In this article, we demonstrate the limitation of the reverse-time migration in imaging these multiples, and present an alternative inversion procedure that is computationally efficient, that jointly maps both primaries and multiples to the true reflectors, and where the source function is estimated on the fly. As a result, we obtain high-quality, mostly artifact-free, broad-band images where the imprint of the source-function are partly removed at a computationally affordable expense compared to the combined costs of the wave-equation based surface-related multiple elimination and the reverse-time migration. We achieve all this by including the total downgoing wavefields as areal sources in the least-squares migration in combination with curvelet-domain sparsity promotion. We apply the proposed method to a shallow-water marine data set from the North sea, which contains abundant short-period surface-related multiples, and show its efficacy in eliminating coherent imaging artifacts associated with these multiples. We also demonstrate the benefits of joint imaging of primaries and multiples compared to imaging these signal components separately.