Rapid estimation of empirical Green’s functions for passive acoustic characterization of dynamic shallow-water environments

Applications of acoustic noise interferometry to passive remote sensing of the ocean and improvement of sonar performance predictions rely on retrieval of empirical Green’s functions (EGFs) from synchronized measurements of ambient sound at spatially separated points. Usually, long noise averaging times of many hours and even days are employed to achieve precise passive measurements of the ray or normal mode travel times that are necessary for evaluation of the environmental parameters with oceanographically relevant accuracy. However, long averaging times have limited utility when nonlinear internal waves or other oceanographic processes cause significant, time-dependent variations in sound propagation conditions. Analysis of the noise records acquired during the Shallow Water 2006 experiment reveals that EGFs can be reliably retrieved from cross-correlations of noise recorded over periods as short as 10 minutes by moored receivers horizontally separated by 40–50 ocean depths. The measured EGFs evolve as the sound speed field changes. Suitability of the EGFs for acoustic remote sensing is demonstrated by geoacoustic inversion of the passively measured normal mode dispersion curves and comparison to previously reported inversion results. Rapid EGF retrieval has far-reaching implications for operational use of the underwater acoustic noise interferometry. [Work supported by NSF.]