Advances in methodology and field data examples for attenuating internal multiples using the inverse scattering series

Abstract

The leading-order term in the inverse scattering series can predict internal multiples without assumptions about the subsurface. However, there are challenges related to data conditioning and computational costs in the inverse scattering series method for 2D or 3D media. To address these challenges, we enhance the method from several aspects for its practical applications. We introduce an efficient data regularization strategy by applying nearest neighbor search and differential moveout to accommodate various acquisition situations. We apply the angle constraints over the dip and opening angles to reduce the computational cost without compromising the model's quality. We propose a new method to define the optimized dip and opening angles before performing the prediction. We also propose segmented prediction by limiting the range of the multiple's generator, which can benefit the subsequent adaptive subtraction. We illustrate the proposed workflow with examples from a 2D field data acquired in a shallow marine environment; the results demonstrate the effectiveness of the new workflow.