Full-wave inversion (FWI) to a 2D-ocean-bottom node (OBN) dataset after offset-continuation-trajectory (OCT) data regularization

Abstract

The estimation of geophysical attributes for a better understanding of subsurface geological structures is essential in many stages of seismic exploration and production. As a powerful seismic inversion tool, the full waveform inversion (FWI) has been widely used to estimate many of such attributes (e.g., seismic velocities) with high resolution. In many cases, however, real acquisitions are unable to provide the quantity and density of traces required for optimal inversion results. As a consequence, schemes to produce (simulate) new traces by interpolation/extrapolation of neighboring available traces, play a crucial role. In this paper, we use the offset-continuation-trajectory (OCT) stacking method to generate new (non-recorded) traces and include them in an ocean-bottom node (OBN) acquisition. We next test the FWI inversion on the enhanced dataset and analyze the results. As proof-of-concept, our approach is applied to the synthetic dataset obtained for an OBN acquisition carried out on the 2D Marmousi model for which several additional traces, simulated by the OCT procedure, have been included. In spite of the recognized complexities of the Marmousi model, we have obtained very encouraging results.