Inverse multiple scattering in the downward-continuation approach

Internal multiples have long been recognized as a prob-lem in seismic imaging. Some theoretical developments have been made to attenuate them, but there is no 3D theory as yet that is valid for a wide range of velocity models. The goal of this paper is to present such a theory. Through the development of this theory we show that, unless certain simplifying assumptions are made about the velocity model, estimating multiples is velocity-model-dependent. We propose estimating and subtracting the multiples in the image domain rather than the data domain. In common-image gathers, the redundancy of well-corrected primaries can be exploited to improve adaptive subtraction techniques