Estimating the uncertainty of geoacoustic parameters of a range‐dependent environment

In this study, data from a range‐dependent environment are inverted to obtain estimates of the geoacoustic properties and their uncertainties. The technique consists of combining the results of a series of range‐independent inversions to produce a model of the range‐dependent environment. A rigorous uncertainty analysis provides a way of discerning whether variations in the inversion results are due to range‐dependent features of the environment or simply due to uncertainty/variability in the results. Broadband acoustic data from a track off the island of Sicily were analyzed. Ground‐truth information in the form of core measurements and a high‐resolution seismic profile were also collected and used for verification of the results. The method of fast Gibbs sampling (FGS) was used to estimate the uncertainties of the geoacoustic properties. FGS is based on a Bayesian approach to inversion which samples the posterior probability distribution to estimate marginal probability distributions and parameter corre...