An essential component in the detection and characterization of underwater munitions is knowledge of the acoustic response of the environment as well as the environment's effect on the acoustic response of munitions. Simulation tools and technologies have been developed under SERDP-funded research initiatives, such as the Personal Computer Shallow Water Acoustic Toolset (PC SWAT), to model the acoustics of both the environment and the munitions. The evaluation of these technologies, as well as their future use in unexploded ordnance (UXO) remediation, rely on knowledge of the underwater environment, particularly the properties of the seafloor. Conventional methods for determining relevant seabed properties employ time-consuming point measurements. Commercially available high-frequency multi-beam echo sounders (MBES) offer a solution to this problem by making faster measurements over larger areas. While these systems are primarily intended for high-resolution bathymetry, acoustic inversion techniques can be used to estimate seafloor parameters relevant to the UXO problem. The objective of this project was to develop and rigorously test a physics-based algorithm that can invert high-frequency acoustic data for sediment parameters. The result is a collection of new algorithms for high-frequency acoustic data inversions and system-independent environmental assessment in terms of measurable seabed parameters, which can then be used as inputs to acoustic and electromagnetic systems.

Technical Approach

The inversion algorithms developed by this project incorporate the results of recent research in high-frequency sediment acoustics, utilizing both models and technologies developed through Office of Naval Research (ONR)-funded experiments such as the Sediment Acoustics Experiments 1999 (SAX99) and 2004 (SAX04). To develop and test the algorithms, acoustic data was collected during three field experiments:

  1. Gulf of Mexico Experiment in 2011 (GulfEx11)
  2. Target and Reverberation Experiment in 2013 (TREX13)
  3. St. Andrew’s Bay Experiment in 2014 (BayEx14)

These field efforts were sponsored by both SERDP and ONR under several different programs, and this project utilized the infrastructure and logistics of those efforts to collect data. In each of these efforts, the acoustic data was collected using Seabat 7125 multibeam echo sounders that were provided along with technical support by Teledyne-RESON. The acoustic data collection in each of these experiments was accompanied by extensive environmental characterization in order to evaluate the accuracy of the inversion algorithms.


While the data collected during GulfEx11 provided the basis for developing the inversion algorithm and learning what was needed to collect a quality dataset, equipment problems, the abundance of fish, and saturation of the MBES data prevented the use of that dataset for inversion testing. The lessons learned were applied to TREX13 and BayEx14 and helped to insure that the data was of sufficient quality to evaluate the inversion. The environments in both of these experiments were complicated, but the inversion performed well producing stable results that could be compared to ground truth measurements at the sites. These ground truth measurements were either compared directly to the inversion outputs or used in scattering models to compare to the inversion.

For TREX13, the inversion compared favorably to the ground truth measurements for a majority of the data products. Discrepancies between the scattering models and the inversion are most likely due to the inability of the scattering models to account for the multiple scattering from shell pieces at the site. The roughness spectrum obtained from the inversion compared well with the measurements across the entire test area for high wavenumbers indicating that the results would be applicable to modeling shallow grazing angle scattering for frequencies at least as low as 50 kHz.

The environment at the BayEx14 site was significantly more complicated than the TREX13 site. The presence of the mud layer and the transition layer between the mud and sand sediments was not accounted for in the inversion, which treated the seafloor as a half-space. Despite this approximation, the inversion results are stable and give effective parameters for the sediment that differ from the ground truth measurements. The scattering models, which incorporate the ground truth measurements, also do not account for the complexity of the environment and could potentially be modified to more accurately represent the scattering physics.


Over the course of this project, the MBES inversion has reached a level of maturity that puts it at the cutting edge of remote seabed characterization particularly for the UXO detection problem. Taking a physics-based approach insures that the output of the inversion addresses the needs of modeling codes and simulations that incorporate similar physical models. This inversion algorithm and the data collection approach needed to apply it have the potential of being implemented into sonars without modifying existing hardware or software.