Particle velocity in geoacoustics: a performance study based on ensemble adjoint inversion

In the last decade or so, methods for the physical characterization of marine sediment have been developed that use waterborne acoustic (scalar) pressure field measured on a hydrophone array and a global optimization scheme based on some form of matched field processing (MFP). In this paper, the benefit of additional information of co-located (vector) particle velocity measurement is investigated. An actual measurement setup of MREA/BP07 experiment is simulated to assess the relative sensitivity of pressure and vertical particle velocity fields to the geoacoustic parameters of a layered bottom and the respective robustness to uncertainty. An ensemble adjoint approach is proposed wherein the inversion process itself is directly and optimally controlled by the physics of acoustic propagation, in contrast to metaheuristic global optimization. In this first comparative study, the inversion is carried out jointly across multiple frequencies by minimizing a Bartlett cost function for the pressure and vertical particle velocity components taken separately. A multidimensionalMonte Carlo technique is applied to the initial conditions and measurement noise realizations. The ensemble results suggest that vertical particle velocity can enhance the estimation of density and attenuation parameters of the layered bottom