Decimation of OBN Shot and Receiver Locations for Evaluating Impact on Velocity Model Estimation by FWI

Velocity model estimation is a key factor in generating optimal imaging of subsurface structures. We study how the velocity model is affected when changing the density of ocean bottom receivers and seismic sources, in order to establish an understanding of how to reduce costs for such surveys or possibly conducting hybrid ocean bottom and streamer acquisitions. A set of tests are specified by increasing the acquisition separations, also simulating reduction of azimuthal distribution between sources and receivers. The results show that for the chosen velocity model inversion method, the receiver spacing can be increased, whereas for source spacing an increase will lead to higher degrees of model noise. The results representing reduction in azimuthal coverage show a tendency to induce false structuring at target, whereas if shot lines were decimated to sparse axial or orthogonal patterns and maintaining azimuthal distribution, the estimated model was fairly in compliance with the non-decimated model case. In this work, the impact and sensitivity of the initial model has not been studied. Thus, results do not uniquely define a threshold for how sparse the acquisition geometry can be specified while still maintaining the model accuracy that a dense acquisition geometry would offer.