Paper Summary
The abstract describes a methodology to compute time-varying, depth-dependent water column interval velocity profiles from Ocean bottom seismic data. The objective is to use these profiles to correct for cold water statics. They are inverted over time slots using direct arrival picks. The inversion is an integrated part of a flow that jointly inverts also for the nodes positions and their clock drift. The main innovation in this work is the use of a second-order polynomial approximation to model the velocity profile as function of depth. The rationale is to relax the need for an accurate initial velocity profile which is required by the standard methods and to improve the fitting for a more accurate velocity estimation. The parameters of the polynomial model are constrained to give a physically sensible velocity profile. The method is tested on synthetic and real data in comparison with a standard method. It performs quite well in terms of fitting the direct arrival picks and gives an overall better velocity estimation in terms of accuracy and time resolution. For the real data, the most visible uplift was for the far offsets (outer lines), where the standard method is known to produce biased velocity values.