Inversion of Deep-Water Velocity Using the Munk Formula and the Seabed Reflection Traveltime: An Inversion Scheme That Takes the Complex Seabed Topography Into Account

The sound speed of seawater is not constant. It varies with season, daytime, depth, and range. This variation was not considered in marine seismic data processing and imaging before deep-water seismic acquisition became a routine activity. As a result, nonignorable errors may be contained in the final migrated image of deep-water seismic data. To eliminate such errors, we propose here a scheme for inverting the seawater sound velocity under the condition that the seabed has a complex topography. In this scheme, the seabed topography is represented by step grids constrained by measured water depth data, and the sound profile is assumed to be the one satisfying the Munk formula. Thus, only the parameters appearing in the Munk formula need to be inverted. This is quite different from the conventional inversion schemes that take the velocity as the inversion target. We use the conjugate gradient method to minimize the objective function given as the sum of squares of traveltime differences. Tests on synthetic common-shot datasets show that our scheme presented here works as expected. Applications to real marine datasets further validate the reasonability and feasibility of our scheme under complex seabed topography. To investigate the reliability of our inversion scheme, we migrate the real 2-D dataset using the Kirchhoff migration with the inverted Munk profile as the velocity model. The corresponding results show that the seawater sound velocity obtained by our inversion scheme improves the imaging quality of underwater structures.