Passive acoustic remote sensing of the coastal ocean using interferometry of diffuse noise fields

Abstract

A two-point correlation function of a perfectly diffuse noise field is known to contain all the information about the environment that can be obtained using transceivers placed at the two points, provided that environmental parameters are time-independent. This theoretical prediction underlies the approach to passive remote sensing that is known as noise (or wave) interferometry. However, acoustic noise in the ocean is never perfectly diffuse, except at very high frequencies, where noise of thermal origin dominates. Moreover, the averaging times necessary for deterministic features to emerge from noise cross-correlations far exceed the time scales of temporal variations of the ocean surface, e.g., due to surface gravity waves and in the water column, e.g., due to internal gravity waves and tides. This paper reviews current theoretical understanding of limitations of noise interferometry, which result from time-dependence of environmental parameters and noise anisotropy in the horizontal and vertical planes. It is demonstrated that, within these limitations, phase-coherent data processing techniques, including back-propagation, waveform matching, and time-warping, can be successfully applied to measured noise cross-correlations to characterize seafloor properties and evaluate current velocity in a coastal ocean. [Work supported by NSF and ONR.]