Geoacoustic Inversion Research Articles

Frequency dependent effects of environmental parameters on the broadband acoustic wave propagation in shallow water waveguides

Acoustic wave intensity plays a crucial role in underwater acoustics, particularly in shallow water regions in which the signal suffers substantial energy fluctuations. The physical properties of the environment significantly impact the intensity of broadband acoustic waves in shallow water waveguides. The accuracy of the intensity is vital for the reliability of any inverse algorithm used to obtain the physical properties of the waveguide boundaries. In this paper, we explore the effects of sound speed fluctuations in the water column on broadband acoustic propagation across different frequencies. Our observations reveal that at higher frequencies the sound intensity is more sensitive to oceanographic variability. This poses challenges for geo-acoustic inversions when parameters in the water column are not well-known, particularly at higher frequencies. To validate our findings, we present simulations using normal modes and parabolic equation with examples from recent experimental observations conducted on the New England Mud Patch Area and the continental shelf region of the United States during 2017, 2021, and 2022. These experiments provide valuable insight into the intricate behavior of acoustic waves in shallow water environments at different frequency bands. [Work supported by ONR (Grant No. N00014-21-1-2760).]

A general model for sediment-column structure on the New England Mud Patch from Bayesian geoacoustic inversion of seabed reflection data

Muddy sediments cover significant portions of continental shelves, but their physical properties remain poorly understood compared to sandy sediments. To explore the spatial and frequency dependencies of mud properties, wide-angle seabed reflection coefficients versus grazing angle and frequency were measured on the New England Mud Patch (NEMP) during the 2017 Seabed Characterization Experiment. This paper presents trans-dimensional Bayesian inversion of reflection coefficients within a frequency band of 1–3 kHz and an angular range of ∼15–25° to obtain geoacoustic profiles and associated uncertainties, as well as frequency dependencies of sound speed and attenuation. The estimated geoacoustic profiles are similar to those from previous inversions of reflection-coefficient data at lower frequencies (0.4–1.3 kHz) collected at two different sites on the NEMP. Based on the inversion results at all three sites, a general interpretive model for sediment-column structure and variability is synthesized for the NEMP. This model includes an upper mud layer in which sediment properties change slightly with depth due to near-surface processes, an intermediate mud layer with uniform properties, and a transition layer where properties change rapidly with depth due to increasing sand content in the mud above a sand layer. [Work supported by the Office of Naval Research.]

Application of trans-dimensional particle filtering for geoacustic inversion

The acquiring geoacoustic model and associated geoacoustic parameters is vital for sound propagation modelling in a range-dependent environment. Conventional sequential methods, e.g., particle filtering and Kalman filtering have been widely used for geoacoustic inversion in range-dependent environment, which encounter challenges when confronted with unknown geoacoustic models that intrinsically varying with range. As an attempt to estimate geoacoustic model and associated geoacoustic parameters as well, a trans-dimensional particle filtering method is presented here. This method integrates birth-death rules into the filtering process, enabling automatic selection of appropriate geoacoustic models and estimating geoacoustic parameters in the same time. Numerical simulations were conducted based on an environment model of the South China Sea, where a sea trial was conducted in 2022. Numerical results demonstrate the efficiency in accurately estimating the geoacoustic model variations and associated parameters. Preliminary results of sea trial data processing are also presented here.

Geoacoustic inversion of wide-angle reflection-coefficient data to estimate sediment shear properties

This paper considers Bayesian geoacoustic inversion of broadband, wide-angle reflection-coefficient data including shear parameters in the seabed model to investigate the ability to estimate these parameters as well as effects on the estimation of other geoacoustic parameters, particularly compressional-wave attenuation. The seabed parameterization is based on the viscous grain-shearing (VGS) sediment-acoustic model, including the grain-to-grain shear modulus as an unknown parameter. VGS sediment parameters are transformed to density and frequency-dependent compressional- and shear-wave speeds and attenuations. Data prediction involves spherical-wave reflection-coefficient calculations. Trans-dimensional inversion, which samples probabilistically over the number of layers in the seabed model, is applied to combine quantitative model selection with parameter/uncertainty estimation. The inversion is applied to reflection-coefficient data sets collected on the New England Mud Patch, and inversion results are compared to those obtained under the common assumption of negligible shear effects (i.e., inverting for a fluid sediment model).

Passive acoustic glider for marine acoustic environment characterization

Acoustic payload-equipped underwater gliders have great potential for assessing the marine soundscape (oceanic acoustic environment), including characterizing properties of the sound field, of acoustic sources, and of the marine environment (water column and seabed). This talk presents observations and analysis of passive acoustic recordings carried out by gliders during the 2017 Seabed Characterization Experiment (SBCEX17). Two Teledyne Webb Research Slocum gliders equipped with omni-directional hydrophones were deployed as virtual acoustic moorings in 72 m deep water on the New England Mud Patch for time periods of ∼6 days for one glider and ∼3 days for the other. Acoustic recordings collected by the gliders captured the marine soundscape, including natural ambient noise, marine-life sounds, ship noise, and acoustic signals transmitted by the SBCEX17 participants. The acoustic environment in the water column is characterized in terms of the acoustic power spectral density as a function of time, depth, and frequency. As an example of the use of passive acoustic gliders for environmental characterization, Bayesian geoacoustic inversion for seabed properties is presented, making use of recordings of controlled sources associated with SBCEX17. [Work supported by the Office of Naval Research.]

An Inversion Method for Geoacoustic Parameters in Shallow Water Based on Bottom Reflection Signals

The inversion method based on the reflection loss-grazing angle curve is an effective tool to obtain local underwater acoustic parameters. Because geoacoustic parameters vary in sensitivity to grazing angle, it is difficult to get accurate results in geoacoustic parameter inversion based on small-grazing-angle data in shallow water. In addition, the normal-mode model commonly used in geoacoustic parameter inversion fails to meet the needs of accurate local sound field simulation as the influence of the secant integral is ignored. To solve these problems, an acoustic data acquisition scheme was rationally designed based on a sparker source, a fixed vertical array, and ship drifting with the swell, which could balance the trade-off among signal transmission efficiency and signal stability, and the actual local acoustic data at low-to-mid frequencies were acquired at wide grazing angles in the South Yellow Sea area. Furthermore, the bottom reflection coefficients (bottom reflection losses) corresponding to different grazing angles were calculated based on the wavenumber integration method. The local seafloor sediment parameters were then estimated using the genetic algorithm and the bottom reflection loss curve with wide grazing angles, obtaining more accurate local acoustic information. The seafloor acoustic velocity inverted is cp=1659 m/s and the sound attenuation is αp=0.656 dB/λ in the South Yellow Sea. Relevant experimental results indicate that the method described in this study is feasible for local inversion of geoacoustic parameters for seafloor sediments. Compared with conventional large-scale inversion methods, in areas where there are significant changes in the seabed sediment level, this method can obtain more accurate local acoustic features within small-scale areas.

Geoacoustic Inversion with a Single Vector Sensor and Multichannel Dispersion Curves

This paper discusses the value added by using a single vector sensor over a conventional pressure-only hydrophone for geoacoustic inversions. Inversion methods based on genetic algorithms are used to estimate the seabed properties. Synthetic signals of impulsive arrivals first are modeled using KRAKEN and RAM propagation models, each being modified to predict components of the vector field. While KRAKEN is utilized to directly compute dispersion curves, RAM provides full-field results that require the application of time warping to separate the modal arrivals. Combinations of dispersion curves utilizing all vector sensor channels are compared to curves estimated with the pressure-only channel. Within the time warping analysis, both binary masking and band-pass filter masking methods are applied to compare stability of results. The environment modeled for the synthetic analysis and inversion method utilize sound speed profiles measured during the Monterey Bay 2019 at-sea experiment and assume a sediment layer of constant thickness overlying a deeper sub-bottom type. White noise is added to the synthetic data at different signal-to-noise ratios to evaluate the impact of signal excess on the results. A hybrid optimization approach is used to improve the results of the genetic algorithm method. The analysis with synthetic data is consistent with the analysis of broadband, impulsive data collected from the experiment, indicating that the additional information from the vertical velocity channel further improves the geoacoustic parameter estimates.

Geoacoustic inversion using very-low-frequency modal interference characteristics.

This paper presents a modal-based geoacoustic inversion method adapted for a very-low-frequency leaky waveguide. It is applied to air gun data collected by a seismic streamer during the multi-channel seismic exploration experiment in the South Yellow Sea. The inversion is carried out by filtering the waterborne and bottom-trapped mode pairs from the received signal and comparing the modal interference features (waveguide invariant) to replica fields. The effective seabed models are inferred at two positions, and the two-way-travel time of basement interface reflected waves calculated using these models exhibit good agreement with geological exploration results.

GIT: A Transformer-Based Deep Learning Model for Geoacoustic Inversion

Geoacoustic inversion is a challenging task in marine research due to the complex environment and acoustic propagation mechanisms. With the rapid development of deep learning, various designs of neural networks have been proposed to solve this issue with satisfactory results. As a data-driven method, deep learning networks aim to approximate the inverse function of acoustic propagation by extracting knowledge from multiple replicas, outperforming conventional inversion methods. However, existing deep learning networks, mainly incorporating stacked convolution and fully connected neural networks, are simple and may neglect some meaningful information. To extend the network backbone for geoacoustic inversion, this paper proposes a transformer-based geoacoustic inversion model with additional frequency and sensor 2-D positional embedding to perceive more information from the acoustic input. The simulation experimental results indicate that our proposed model achieves comparable inversion results with the existing inversion networks, demonstrating its effectiveness in marine research.

Geoacoustic Inversion Using Modal Dispersion Curves

This paper presents a broadband geoacoustic inversion method based on matching the estimated dispersion curves extracted from low frequency pulse signal by warping transform and the simulated dispersion curves. The single receiver broadband inversion of the geoacoustic parameters is implemented using the first three orders of the dispersion curves, and the results are assessed using the parameter distribution and propagation loss deviation. The simulation outcomes demonstrate the method’s broad applicability for inverting geoacoustic parameters and the warping transform’s relative accuracy for dispersion curves estimation in waveguide environments with significant vertical sound speed modification.

Bayesian geoacoustic inversion in shallow water with vertical coherence of ship-radiated noise

Spatial structure of underwater acoustic field depends strongly on the geoacoustic properties of the seabed. The geoacoustic parameters can be estimated using the spatial coherence dominated by a single broadband source. In this paper, geoacoustic inversion of a shallow water environment is carried out using the vertical coherence of ship-radiated noise at a close range. Data were measured in an experiment in the South China Sea, and the vertical coherence is extracted from a single hydrophone pair in the vertical line array. The closest point of the passing research ship approach is determined first, and then the broadband coherence curves are extracted as the input for the inversion algorithm. Due to the lack of an appropriate parametrization, the trans-dimensional inversion method combined with the auto-regressive error model is adopted in this paper. The inversion results show that the sound speed is the most sensitive, followed by the density, and then the attenuation coefficient is the least sensitive, which is consistent with previous works. The data error hypothesis is verified via a posterior examinations.

Geoacoustic Inversion Using Simple Hand-Deployable Acoustic Systems

This article proposes the use of a simple, low-cost, hand-deployable pair of experimental assets to conduct geoacoustic inversion at sea. The system consists of an expendable, fully mechanical acoustic source called a rupture induced underwater sound source (RIUSS) and a new ropeless passive acoustic mooring called a TOSSIT (not an acronym). Used together, RIUSS and TOSSIT enable the collection of acoustic data suitable to perform single-hydrophone geoacoustic inversion. The method is illustrated using data collected on the New England Mud Patch in May 2021 from a relatively small (22 m) and inexpensive chartered fishing vessel. Modal time-frequency dispersion from 15 to 385 Hz is extracted from the TOSSIT/RIUSS data using warping, and used as input for Bayesian transdimensional geoacoustic inversion. The inversion results compare favorably to results obtained with data collected on the same track with traditional assets (e.g., a vertical line array) during the 2017 Seabed Characterization Experiment, even when jointly inverting for the water-column sound speed profile and seabed geoacoustic parameters. This further demonstrates inversion repeatability in a given location using data sets collected years apart, and under different (and potentially unknown) oceanographic conditions.

Broadband Geoacoustic Inversions for Seabed Characterization of the New England Mud Patch

Geoacoustic inversions using broadband acoustic data acquired during the Seabed Characterization Experiment 2017 conducted in the New England Mud Patch area in March 2017 are presented in this article. The primary goal of the data inversions is to estimate the compressional wave speed and density profiles of the sediment layers. Both linear perturbative and nonlinear Bayesian inversion methods are utilized. These two inversion methods share the same principle of solution schemes, i.e., to minimize the difference between the signal acquired (or derived quantities) at the receiver and the prediction by the optimized bottom model. The differences between the two approaches are in the type of data utilized for inversion and on the inversion procedure used to determine the bottom geoacoustic properties. The linear inversion method uses the time difference of modal arrival, while the Bayesian inversion uses the bandpass-filtered sound pressure waveform. One of the study objectives is for the Bayesian inversion to provide a reference solution to the inverse problem and determine the ability of the linear inversion method to provide comparable results as compared with a more exhaustive search used in the Bayesian method. Another objective is to use the verified linear inversion method that is computationally faster to explore different models of sediment layering structure and to estimate the most appropriate bottom model for the experimental area.

Correlating shipping noise on multiple beams

There has been much interest in using transiting surface ships as sources of opportunity for tomography and geo acoustic inversion. In tomography, it would be valuable to increase the ranges at which ship signatures can be processed by measuring times of arrival at the output of beamforming processes, which provide spatial processing gain. This additional gain may enable fainter multipath arrivals to be identified and exploited as additional paths to sample the ocean water column. This seems readily achievable when cross-correlating multiple vertical beams formed on a single vertical line array for example. However, cross-correlating beams from multiple arrays, vertical or horizontal, raises questions. In both cases, different beams may have different Doppler, which is manageable with a correlation process that includes Doppler compensation. But obeserving a ship from different vantage points may be problematic, if the ship signature is due to horizontally displaced noise sources distributed around the ship. We will present results of processing surface ships observed on single and multiple arrays and assess the use of such processing in ambient noise tomography.

Estimating seabed properties with a residual convolutional network

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.]

Waveform modeling with Gaussian Processes for inversion in ocean acoustics

It has been shown that, in ocean acoustics, Gaussian processes can predict a densely sampled field on a receiving array, when only sparse samples of the sound in the water column are initially available. In a similar manner, a Gaussian process approach can be designed to densely sample signals in the time domain, resulting from the transmission and propagation of broadband waveforms. A mean waveform can be obtained that allows the high-resolution estimation of multipath arrival times. These can then be used for source localization and geoacoustic inversion. Uncertainty quantification in the time-series characterization, readily available from the Gaussian process modeling, facilitates uncertainty quantification in inversion, obviating the need for onerous computations. We investigate the potential of different kernels in aiding the inversion process. [Work supported by ONR.]

Joint trans-dimensional inversion for seabed and water-column properties in the seabed characterization experiments

This paper considers the information content to resolve both a geoacoustic model of the seabed and a model of the water-column sound-speed profile (SSP) in the inversion of modal-dispersion data. A Bayesian formulation of the inverse problem allows prior information representing varying levels of independent knowledge to be applied separately to the seabed and water-column models. Issues of interest include the extent to which knowledge (or lack thereof) of the SSP affects geoacoustic inversion, and the ability to remotely estimate the SSP from acoustic data in cases where the seabed is either well known or poorly known. The joint inversion for seabed and water-column properties is formulated in terms of a separate trans-dimensional model for each, providing an automated approach to quantitative model selection. The seabed model is formulated in terms of an unknown number of uniform sediment layers, while the SSP is formulated as an unknown number of depth/sound-speed nodes, with the reciprocal of sound speed squared varying linearly between nodes, as per normal-mode propagation models. The approach is applied to modal-dispersion data measured during the seabed characterization experiments at the New England Mud Patch. [Work supported by the Office of Naval Research]

Effects of weak shear rigidity of the seabed on sound propagation in a range-dependent ocean

The shear wave speed is often small compared to the compressional wave speed in unconsolidated marine sediments. Sediment stratification and especially mass density variation in the seabed enhance the effects of weak shear on acoustic normal modes in range-independent ocean and produce significant shear wave contributions to mode attenuation in shallow water [O. A. Godin, J. Acoust. Soc. Am., 149, 3586–3598 (2021)]. A distinctive feature of the shear-induced perturbations in the mode phase speed and attenuation is their rapid variation with sound frequency due to shear wave interference within a sediment layer. This paper investigates theoretically the shear wave-induced perturbations in normal mode phase, travel time, and attenuation when water depth and/or sediment layer thickness vary gradually with range. Frequency dependencies of the adiabatic normal mode travel time and attenuation are found to be highly sensitive to range dependence of the seabed layering. Gradual changes in the sediment layer thickness have an effect similar to frequency averaging or artificially increased shear wave attenuation. Implications of these findings for geoacoustic inversions will be discussed. Contributions of shear rigidity and cross-range gradients of the sediment layer thickness to horizontal refraction of normal modes will be also examined. [Work supported by ONR.]

Bayesian geoacoustic parameter inversion based on dispersion characteristics of normal mode water wave and ground wave

Most of shallow water geoacoustic inversions based on modal dispersion cannot reliably estimate the deep geoacoustic parameters. Because these studies focused on the dispersions of water waves but ignored the dispersions of ground waves. Therefore, in this paper a Bayesian geoacoustic inversion is studied based on wideband modal dispersions of water waves and ground waves. Firstly, the modal dispersion curves with Airy phase components are discussed. Secondly, the Bayesian inversion theory and a novel sample-efficient inference algorithm, namely Variational Bayesian Monte Carlo, are introduced briefly. In the Bayesian inversion, the posterior probability densities of unknown parameters are inferred, which can provide the prediction closest to the observation data and the uncertainty of the prediction. Considering that the forward acoustic model is computationally intensive, the posterior analysis is carried out by using the Variational Bayesian Monte Carlo method. It is performed by finding the variational distribution closest to the target distribution and requires less computation time than the Markov chain Monto Carlo method. In the simulation study, a range-independent two-layer seabed, including the sediment layer and basement layer, is modeled, on the assumption that the water column is homogeneous. The function of shear wave in waveguide is ignored. The compressional sound speed of the sediment layer varies linearly from <i>c</i><sub>1U</sub> to <i>c</i><sub>1L</sub> between 0 and <i>h</i><sub>1</sub>, while other geoacoustic parameters are constant. By comparing the inversion results with and without the information of ground waves for different signal-to-noise ratios, it can be concluded that the deep geoacoustic parameters are more sensitive to the dispersions of ground waves. And then, a shallow-water experimental study is carried out in the Bohai Sea of China. The average water depth is about 20 m. The wideband pulse signals are recorded by a hydrophone with a sensitivity of –170 dB re 1 V/μPa. The received signals include well-defined Airy phase components, and the modal dispersion curves of water waves and ground waves are extracted accurately. The experimental results indicate that the Bayesian inversion combining water and ground wave dispersions can not only estimate the deep geoacoustic parameters reliably, but also reduce the inversion uncertainties of other model parameters, such as the shallow geoacoustic parameters, water depth, etc. The estimated source-receiver range and water sound speed are close to their measured values. The modal dispersion curves predicted by the posterior mean samples are in good consistence with those extracted from the experimental data in different ranges. In addition, the good forecast of transmission loss also demonstrates the reliability of the joint inversion.

The effect of seafloor roughness on passive estimates of the seabed reflection coefficient.

In this work, a model is developed for the effect of seafloor interface roughness on passive estimates of the reflection coefficient. The main result is an expression for the total intensity reflection coefficient, with separate coherent and incoherent contributions. Assumptions of this model include constant sound speed in the ocean, stationary and Gaussian seafloor roughness, and ambient noise. Numerical examples for the coherent, incoherent, and total contributions to the intensity reflection coefficient are presented for halfspace and layered environments-all using the small slope approximation. To illustrate the potential parameter errors that results from using a flat interface wave model when roughness is present, a geoacoustic inversion is performed using the proposed model as input data. A joint roughness-geoacoustic inversion of simulated data using the proposed model was also performed. It was found that the true roughness and geoacoustic parameters can be inverted using this model, but the sensitivity to the outer scale of the rough surface has the highest error compared to the other parameters.