In spite of a plethora of novel innovative algorithms for multiple identification and elimination, the fact is that their applicability and success is in most cases questionable and limited to marine environments especially deep waters. In land seismic datasets however, a clear understanding of internal multiples and their generation is not readily available and thus efforts to attenuate them become cumbersome and quite often ineffective. Particularly in desert terrains with a complex near surface and low-relief structures, surface multiples are rare. In most cases, we are dealing with so called "near surface related multiples". These are basically internal multiples generated within the complex structure of the near surface. They exhibit little moveout discrimination with respect to primary reflections and thus cannot be removed with conventional technologies commonly based on Radon transforms. Their proper estimation and subsequent elimination requires a rigorous approach firmly rooted to wave equation concepts.
In this paper we employ theoretical principles of Inverse Scattering Series (ISS) and develop distinct, cost-effective, practical methodologies for internal multiple elimination applicable for land and shallow water seismic data. These fully-automated, data-driven algorithms require no picking of events and identification of multiple generators, assume no subsurface velocity information and can be applied pre and/or post-stack. They can accommodate the broadest set of internal multiples of all orders and can be extended to multi-dimensional earth models. Thus, multiples are surgically removed by predicting their amplitudes and phases and do not harm primary reflections, even if they are proximal and overlapping. A series of realistic synthetics and field land datasets from the Arabian Peninsula are used in order to demonstrate the effectiveness of ISS-based algorithms towards the attenuation of internal multiples.
