Overburden-borne internal demultiple formula

Abstract

Imaging targets buried under highly reflective overburdens might require removal of a substantial amount of internal multiples. However, existing data-driven internal demultiple methods are either not accurate, not very user friendly, or both, as they (1) (only) kinematically predict only some of the multiples, (2) allow for false predictions (multiples leakage), (3) require primary-harming adaptive filtering, (4) can be computationally expensive or require user intervention, (5) require other than reflection data nontrivial inputs, and/or (6) the output cannot be readily quality controlled. We have developed a unifying surface-related multiple estimation/elimination (SRME)-like macrolayer internal demultiple algorithm, which addresses these issues and accurately eliminates all overburden-borne internal multiples. We find that the leading order terms of our formula correspond to some well-known approaches. This strongly suggests that the main benefit of using the Marchenko-equation-based methods comes from higher order terms, which provide transmission amplitude correction and suppress unphysical events predicted by lower order terms. We illustrate the efficacy of our formula with a numerical example, use it to identify outstanding limitations, and suggest some practical solutions to them. We also show that this approach is sufficiently general to be relatively easily extended to include, e.g., viscoelastic effects.