Over more than a decade beginning in 1996, we conducted a series of time reversal mirror (TRM) experiments in collaboration with the NATO Undersea Research Centre in coastal waters. The method leverages spatial diversity to achieve both spatial and temporal focusing in complex environments, offering a promising alternative to conventional multichannel equalizers (M-DFE) for underwater communications. Temporal focusing enables self-equalization and mitigation of intersymbol interference (ISI), which can then be followed by an adaptive decision feedback equalizer (DFE), known as TR-DFE. Additionally, spatial focusing facilitates multiuser communications, with adaptive TR further reducing crosstalk among users. Our findings reveal that TR-DFE outperforms M-DFE, particularly when high-order constellations, such as 32-QAM, are employed, significantly enhancing data throughput and spectral efficiency. This presentation will showcase examples inspired by Bill’s invaluable contributions and insights.

During the Office of Naval Research (ONR) New England Seamounts Acoustics (NESMA) experiment, low- to mid-frequency acoustic measurements were made of scattering from steep and irregular seamount flanks. These measurements included bistatic scattering from a stationary source (operated by Scripps Institution of Oceanography) to a towed horizontal line array (the Penn State Three Octave Research Array). The use of broadband waveforms and horizontal beamforming allows the separation of seabed/surface interaction and the azimuthal extent of scattering from the bathymetrically complex and acoustically rough seamount flank. The impact of these features is seen in the form of anisometric, out-of-plane scattering with an angular extent greater than 30 deg. Acoustic receptions appear to have significant incoherent energy due to seamount roughness, but also contain deterministic components from specific regions of the seamount flank. This talk will discuss acoustic measurements and analysis, the spatial distribution, and statistical characteristics of scattered returns.

In shallow-water downward-refracting ocean environments, hydrophone measurements of shipping noise encode information about the seabed. In this study, neural networks are trained on synthetic data to predict seabed classes from multichannel hydrophone spectrograms of shipping noise. Specifically, ResNet-18 networks are trained on different combinations of synthetic inputs from one, two, four, and eight channels. The trained networks are then applied to measured ship spectrograms from the Seabed Characterization Experiment 2017 (SBCEX 2017) to obtain an effective seabed class for the area. Data preprocessing techniques and ensemble modeling are leveraged to improve performance over previous studies. The results showcase the predictive capability of the trained networks; the seabed predictions from the measured ship spectrograms tend towards two seabed classes that share similarities in the upper few meters of sediment and are consistent with geoacoustic inversion results from SBCEX 2017. This work also demonstrates how ensemble modeling yields a measure of precision and confidence in the predicted results. Furthermore, the impact of using data from multiple hydrophone channels is quantified. While the water sound speed in this experiment was only slightly upward refracting, we anticipate increased advantages of using multiple channels to train neural networks for more varied sound speed profiles.

The Seabed Characterization Experiment 2022 (SBCEX22) was carried out in the spring and summer of 2022 with the goal of studying the fine grain sediment off the coast of New England and evaluating different methodologies for the estimation of the sediment geoacoustic properties. Towards this goal, tonal data measured during the experiment at a vertical line array are employed for source localization and geoacoustic inversion via traditional matched-field inversion (MFI) and Gaussian process (GP) based MFI. The latter approach relies on the generation of virtual arrays with functions that capture the coherence of the acoustic field at different depths in the ocean. The predicted data at densely spaced virtual sensors, resulting from interpolation of the original array, are used for inversion in place of raw measurements. A Gaussian kernel is integrated in the prediction process and different spacings between virtual sensors are considered for array interpolation. Genetic algorithms are used for optimization of the inversion for both methodologies, which are compared through an analysis of their estimates and the ensuing uncertainty. The GP-based technique is found superior, with the results in good agreement with ground truth information and with reduced uncertainty in comparison to the traditional approach.

During the 2023 New England Seamounts Acoustics (NESMA) pilot experiment, continuous measurements were made of underwater acoustic propagation and scattering near a seamount from close range through the first deep-water acoustic convergence zone. Measurements were made by positioning a stationary mid-frequency source at a shallow depth over the plateau of the Atlantis II seamount and then towing Penn State’s Three Octave Research Array (THORA) to a range of 85 km from the source. Low- to mid-frequency acoustic data were beamformed to improve signal-to-noise ratio and reduce own-ship noise, and transmission loss was estimated using narrowband processing techniques. Comparison of the data with ray and parabolic equation-based models helped to develop an understanding of the influence of the seamount bathymetry, roughness, and sound speed profile. This talk will discuss details of these unique measurements, modeling, and data analysis confirming a frequency-dependent change in acoustic enhancement within the first deep-water shadow zone.

Since the ocean covers more than two-thirds of the planet and mud covers most of the seabed, understanding the acoustic propagation of mud is essential for acoustic performance prediction. We aim to determine the acoustic properties and their spatial variability on the New England Mud Patch by using data collected during the SBCEX17 experiment. In particular, we use tonal signals between 53 and 953 Hz emitted by sources towed on circular tracks and recorded by vertical line arrays. For Bayesian estimation, we use a new implementation of Metropolis-Hastings Markov chain Monte Carlo (MCMC) sampling that combines adaptive covariance estimation, sequential sampling in eigenvector space, and parallel tempering. Acoustic sound propagation is modeled by an adiabatic modes model that can provide a good tradeoff between inversion speed and modeling accuracy in the moderately range-dependent SBCEX17 environment. Our description of the seafloor includes a mud layer in which the sound speed increases moderately with depth and a thinner mud-sand transition layer where sound speed increases strongly due to a sand content that increases with depth. Preliminary results match well with established results based on broadband reflection-coefficient data. In addition to the spatial variability, we investigate the effects on uncertainty quantification of geoacoustic parameters when switching from a range-independent propagation model to the considered adiabatic modes model.

During the New England Seamount Acoustics (NESMA) 2023 experiment, the majority of temperature and salinity measurements from fixed moorings did not reach the surface due to a mooring blowdown caused by the Gulf Stream. To provide a statistical model of upper ocean properties, NESMA 2023 mooring measurements are extended to the surface using the altimetry-informed gravest empirical mode. The altimetry-informed gravest empirical mode is a tool that allows for the determination of interior, three-dimensional water column profiles from surface measurements of sea surface height, as well as the depth-dependent uncertainties of these properties. Sound propagation models based on the parabolic equation method will be implemented using the augmented sound speed profiles. Outputs from the acoustic models will be compared with sound propagation measurements to evaluate the accuracy of the sound speed profile augmentation. [Work supported by the Office of Naval Research.]

Passive acoustics is a versatile tool for maritime situational awareness, enabling applications such as source detection and localization, marine mammal tracking, and geoacoustic inversion. This study focuses on estimating the range between an acoustic receiver and a transiting ship in an acoustically range-independent shallow water environment. Here, acoustic propagation can be modeled by a set of modes that are determined by the shallow water waveguide and seabed characteristics. These modes are dispersive, with phase and group velocities varying with frequency, and their interference produces striation patterns that depend on range and frequency in single-hydrophone spectrograms. These striation patterns can often be characterized by the waveguide invariant (WI), a single parameter describing the waveguide's properties. This paper presents a statistical model and corresponding WI-based range estimation approach using a single hydrophone, leveraging broadband and tonal sounds from a transiting ship. Using data from the Seabed Characterization Experiment 2017 (SBCEX17), the method was evaluated on two commercial ships under different environmental conditions and frequency bands. Range estimation errors remained below ±4% up to 62 km in the best case, with robust performance demonstrated in the 40-60 Hz band.

Large surface wave breaking events in deep water are acoustically detectable by beamforming at 5–6 kHz with a mid-frequency planar array located 130 m below the surface. Due to the array's depth and modest 1 m horizontal aperture, wave breaking events cannot be tracked accurately by beamforming alone. Their trajectories are estimated instead by splitting the array into sub-arrays, beamforming each sub-array toward the source, and computing the temporal cross-correlation of the sub-array beams. Source tracks estimated from sub-array cross-correlations match the trajectories of breaking waves that are visible in aerial images of the ocean surface above the array.

The very low-frequency noise from merchant ships provides a good broadband sound source to study the deep layers of the seabed. The nested striations that characterize ship time-frequency spectrograms contain unique acoustic features corresponding to where the waveguide invariant beta becomes infinite. In this dataset, these features occur at frequencies between 20 and 80 Hz, where pairs of modal group velocities become equal. The goal of this study is to identify these beta = infinity frequencies in ship noise spectrograms and use them to perform statistical inference for the deep layer sound speeds and thicknesses in the New England Mudpatch for a larger number of ships and acoustic arrays over a larger geographical region than previously studied. Marginal probability distributions of the data indicate that using singular points for a feature-based inversion yields an estimate of the sound speed and a limiting value for the thickness of the first deep layer. Heterogeneity is examined by correlating spatial variability of the deep layer sound speeds with ship tracks.

The coherent recombination of a direct and seabed reflected path is sensitive to the geophysical properties of the seabed. The concept of feature-based inversion is used in the analysis of acoustic data collected on a vertical line array (VLA) on the New England continental shelf break in about 200 m of water. The analysis approach for the measurements is based on a ray approach in which a direct and bottom reflected path is recombined, resulting in constructive and destructive interference of the acoustic amplitudes with frequency. The acoustic features have the form of prominent nulls of the measured received levels as a function of frequency as a broadband (500–4500 Hz) source passes the closest point of approach to the VLA. The viscous grain shearing (VGS) model is employed to parameterize a two-layer seabed model. The most likely seabed is a sand sediment with a porosity of about 0.42. There is a possibility of a thin (less than 0.5 m) surface layer having a slightly higher porosity between 0.45 and 0.50. Using the estimates for the VGS parameters inferred from the short-range frequency features, a normal mode model is used to predict the received acoustic levels over larger range scales.

Convergence zone propagation has a cycle distance of order 50 km with most of the path spent below the thermocline. Calculations using Garrett-Munk statistics and rays computed with a smoothed background sound speed profile over this single cycle predict propagation in the partially saturated regime and an integration time of 500 seconds (time spent in a single output bin of a discrete Fourier transform). Parabolic equation simulations with synthetic fine structure and ocean dynamics validate the ray-based calculations. Data of narrowband transmissions from 1.5 to 7.5 kHz from a shallow towed source (150 m) to a shallow drifting array (150 m) depth at a range of 59 km demonstrate smoothly varying phase in a beam permitting coherent integrations of 3.5 minutes. The shorter integration time in data is consistent with non-uniform motion estimated from GPS.

Passive sonar is a useful tool for underwater acoustics that can be used to detect ships or other sound sources in the ocean. The input signals into the sonar system in turn can also be used to make inferences about the ocean environment. In recent work, ship noise has been used to infer seabed information from 15–20 minute spectrograms, with CPA in the center of the time window. Waiting for 15–20 minutes to record the full time window slows down real-time inference efforts. In this work, our goal is to complete ship-of-opportunity (SOO) spectrograms given the first few minutes. Our approach is to use a self-supervised attention-based transformer, which has been found to be effective for machine learning problems, particularly for natural language processing. The transformer is trained on simulated SOO spectrograms and tested on spectrograms from the Seabed Characterization Experiment in 2017 in the New England Mud Patch. The resulting predicted spectrograms contain the key features of the measured spectrograms over the full time window. If successful, this work may allow for real-time applications of ship detections and seabed inferencing methods. [Work supported by the Office of Naval Research, Grant N00014-22-12402.]

The very low-frequency noise from merchant ships provides a good wideband source to study the deep layers of the seabed. The nested striations which characterize ship spectrograms contain unique acoustic features where the waveguide invariant (β) becomes infinite. This occurs at frequencies between 20 and 80 Hz where pairs of modal group velocities are equal. The goal of this project was to identify the β = ∞ frequencies in ship noise spectrograms and use these frequencies to perform statistical inference for the deep layer sound speeds and thicknesses. The Seabed Characterization Experiment of 2022 on the New England continental shelf had three vertical line arrays strategically placed between two shipping lanes. The average water depth was 75 meters with less than one meter bathymetry change between the arrays. The results of this study are based primarily on five ships. There was a gradual shift in the β = ∞ frequencies between the three arrays, suggesting a gradual change in the deep sediment layers. [Work supported by Office of Naval Research.]

The shallow water environment is challenging for geoacoustic inversions due to site-specific complex bathymetry and sub-bottom characteristics as well as time-evolving oceanography. Shallow water mid-frequency (0.5–10 kHz) ambient noise and source tow transmission observations are presented that were collected with a 2D array (512 elements arranged in 8 vertical staves of 64-elements each all half-wavelength spaced at 6 kHz). The data were collected in August-September 2021 WSW of the Ports of Los Angeles and Long Beach and NW of Santa Catalina Island. The water depth at the array site was 311 m with the bathymetry gradually becoming deeper to the north and was ∼600–900 m deep in the east-west region for ships transiting in/out of LA/Long Beach. The mid-frequency observations included tonal source tow transmissions for characterizing propagation in the region, the radiated signatures of ships transiting the area, and the relatively quiet periods between. The 2D array geometry facilitates decomposing the acoustic field in azimuth as well as elevation. The long-term intent is to implement azimuthally and range-dependent geoacoustic inversions for the seafloor properties in this region as well as to compare shipping as sources of opportunity to the tonal transmissions.

An ocean acoustics experiment in 2017 near a shipping lane on the New England continental shelf in about 75 m of water provided an opportunity to evaluate a methodology to extract source signatures of merchant ships in a bottom-limited environment. The data of interest are the received acoustic levels during approximately 20 min time intervals centered at the closest position of approach (CPA) time for each channel on two 16-element vertical line arrays. At the CPA ranges, the received levels exhibit a frequency-dependent peak and null structure, which possesses information about the geophysical properties of the seabed, such as the porosity and sediment thickness, and the characterization of the source, such as an effective source depth. The modeled seabed is represented by two sediment layers, parameterized with the viscous grain shearing (VGS) model, which satisfies causality, over a fixed deep layered structure. Inferred estimates of the implicit source levels require averaging an error function over the full 20 min time intervals. Within the 200-700 Hz band, the Wales-Heitmeyer model captures the inferred frequency dependence of the source levels.

To support the modeling of reverberation data collected during the Target and Reverberation Experiment in 2013 (TREX13), transmission loss was measured in the 1.5-4.0 kHz band using a towed source and two moored vertical line arrays. The experiment site was located off the coast of Panama City Beach, FL, and the transmission loss measurements took place along a 7-km-long isobath, which ran parallel to the shore with a water depth of approximately 19 m. The seafloor at the TREX13 site consists of sand ridges, which run perpendicular to the track of the experiment, with narrow bands of softer sediments on the western sides of the ridges and in the ridge swales. Using data from a multibeam echosounder survey and direct measurements of the seafloor properties, a geoacoustic description of the seafloor is developed and used to model the transmission loss at the site. Although the soft-sediment bands only occur in 27% of the seafloor, they are found to have a significant impact on the transmission loss, increasing it by roughly 5 dB at 4 km over what would be expected from an entirely sand sediment. This is consistent with the previous work by Holland who showed that lossiest sediments play the largest role in propagation over range-dependent seabeds. Simulations also show that the exact locations of the soft sediments are less important for controlling propagation in the TREX13 environment than the proportions of the sediments. This suggests that a range-independent, effective media description of the sediment could be used to model propagation at the site. The limits of the use of an effective medium in describing both propagation and reverberation measurements made during TREX13 are considered.

This paper presents inversion results for three datasets collected on three spatially separated mud depocenters (hereafter called mud ponds) during the 2022 Seabed Characterization Experiment (SBCEX). The data considered here represent modal time-frequency (TF) dispersion as estimated from a single hydrophone. Inversion is performed using a trans-dimensional (trans-D) Bayesian inference method that jointly estimates water-column and seabed properties along with associated uncertainties. This enables successful estimation of the seafloor properties, consistent with in situ acoustic core measurements, even when the water column is dynamical and mostly unknown. A quantitative analysis is performed to (1) compare results with previous modal TF trans-D studies for one mud pond but under different oceanographic condition, and (2) inter-compare the new SBCEX22 results for the three mud ponds. Overall, the estimated mud geoacoustic properties show no significant temporal variability. Further, no significant spatial variability is found between two of the mud ponds while the estimated geoacoustic properties of the third are different. Two hypotheses, considered to be equally likely, are explored to explain this apparent spatial variability: it may be the result of actual differences in the mud properties, or the mud properties may be similar but the inversion results are driven by difference in data information content.

Deep learning can assist in characterizing seabeds using sources of opportunity such as shipping noise. While previous work focused on seabed classification, this study uses a residual convolutional neural network to find individual seabed properties. The training data were labeled with sound speed, density, attenuation, and thickness of the layer values of the top sediment layer. A comparison was made between predictive capabilities of ResNet-18 networks when trained to learn a single parameter and those trained to simultaneously learn multiple parameters. For stiff parameters—those with high information content in the data—learning an individual parameter performed better. These single parameter predictions are fundamentally different from a geoacoustic inversion for one parameter. In geoacoustic inversion, all other parameters are held at a fixed value. In deep learning, variability in all other parameters is contained in the training data, but the network focuses on features in the data related to a single property. The trained networks are applied to ship noise measured during the 2017 Seabed Characterization Experiment. [Work supported by the Office of Naval Research and the National Science Foundation’s REU program.]

Remote sensing using passive sonar in the ocean is a challenging problem due to variations in the geoacoustic structure of the seabed and unknown source location and strength. One way to improve remote sensing is to perform an optimization for geoacoustic and source parameters. We use a Bayesian maximum entropy (BME) approach with a viscous-grain shearing parameterization for two sediment layers. The statistical optimization provides probability distributions for porosity and thickness of the sediment layers as well as ship speed, closest point of approach, and the source strength for the Wales-Heitmeyer empirical source level spectrum. We use this approach on spectrograms of transiting ships collected on a vertical line array during the 2017 seabed characterization experiment. We compare the resulting parameter distributions from distinct ships as well as previous estimates of geoacoustic values and source properties. This research shows that the BME approach obtains estimates for porosity and source strength that have narrow posterior probability distributions. [Work supported by the Office of Naval Research.]