Quantifying uncertainty in geoacoustic inversion. II. Application to broadband, shallow-water data.

Area of Science:

• Ocean acoustics
• Geophysical inversion
• Bayesian statistics

Background:

• Seabed geoacoustic parameter uncertainty estimation is crucial for interpreting acoustic survey data.
• Existing methods often lack robust validation for experimental conditions.
• Fast Gibbs sampling (FGS) offers a novel Bayesian approach to address these limitations.

Purpose of the Study:

• To apply and validate the fast Gibbs sampling (FGS) method for estimating uncertainties in seabed geoacoustic parameters.
• To interpret results from a broadband, shallow-water acoustic survey using FGS.
• To assess the reliability and stability of FGS uncertainty estimates against experimental data and ground truth.

Main Methods:

• Application of fast Gibbs sampling (FGS), a Bayesian inversion technique.
• Estimation of marginal probability distributions and parameter covariances via posterior probability density sampling.
• Validation of FGS by comparing results with independent data sets, synthetic test cases, and geophysical ground truth.

Main Results:

• FGS provides reliable and stable uncertainty estimates for geoacoustic inverse problems.
• Validation confirms the accuracy of FGS-derived uncertainties across various experimental conditions (bandwidth, range, prior information).
• The method successfully interprets experimental data from a site off the west coast of Italy.

Conclusions:

• Fast Gibbs sampling (FGS) is a validated and effective tool for quantifying uncertainties in seabed geoacoustic parameters.
• The study confirms the practical applicability of FGS for real-world acoustic survey data.
• FGS enhances the interpretation and reliability of shallow-water geoacoustic inversion.