De-coupling Residual Statics and Velocity Picking

Land surveys suffer from rapid variations in velocity in the near-surface weathering layer that degrade seismic images when the velocity variations are not accounted for in imaging. While elevation information or refraction tomography can compensate for long and mid-wavelength variations, residual static algorithms are used to correct for the shorter wavelengths. Residual statics are commonly estimated by maximising lateral coherency in gathers after Normal Move-Out corrections, either by measuring trace-to-trace correlation or overall stack power. This work introduces an alternative method able to estimate residual statics using a cost function based on sparseness in gathers without a-priori knowledge about velocities needed for NMO correction. By maximizing sparsity in the tau-v and tau-p domains the method enhances coherency assuming, respectively, hyperbolic reflections in the case of mid-wavelength residual statics and local linearity in the case of shorter wavelength residual statics. Results from 2D and 3D data show overall improvement in stack power and focus, as well as reflector positioning, when the sparseness method is used before picking velocities for imaging. To complete the flow, stack optimisation methods can be run as a post-process to further add constructively to the stack.