A Bayesian approach to modal decomposition in ocean acoustics

Abstract

A Bayesian approach is developed for modal decomposition from time-frequency representations of broadband acoustic signals propagating in underwater media. The goal is to obtain accurate estimates and posterior probability distributions of modal frequencies arriving at a specific time and their corresponding amplitudes, which can be employed for geoacoustic inversion. The proposed approach, optimized via Gibbs sampling, provides uncertainty information on modal characteristics via the posterior distributions, typically unavailable from traditional methods.