Feature-based maximum entropy for geophysical properties of the seabeda).

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.

Inference of source signatures of merchant ships in shallow ocean environmentsa).

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.

Maximum entropy inference of seabed properties using waveguide invariant features from surface ships.

Acoustic data were recorded on two vertical line arrays (VLAs) deployed in the New England Mud Patch during the Seabed Characterization Experiment 2017 in about 75 m of water. The sound recorded during the passage of merchant ships permits identification of singular points for the waveguide invariant ß for mode pairs (1,n):ß1,n,for n=2,3,4,5, in the 15-80 Hz band. Using prior geophysical information and an acoustic data sample from the merchant ship KALAMATA, a geoacoustic model M of the seabed was developed. Then, using data samples from other merchant ships, a feature-ensemble maximum entropy method is employed to infer the statistical properties of geoacoustic parameter values for the sound speeds in a surface mud layer and a deep sand layer. Technical challenges include a sparsity of observed singular points, the unique identification of mode pairs for an observed singular point, and the deviation of the waveguide from horizontal stratification. A geoacoustic model M is developed that reproduced the observed ß≈-1 for f < 20 Hz and mode cutoff features at about 15 Hz. The statistical low-frequency inference of the singular point structure from multiple ships provides evidence of an angle of intromission at the water sediment interface with an average sound speed ratio of about 0.986 and an average sound speed for the deeper sand layer of about 1775 m/s.

Impact of data augmentation on supervised learning for a moving mid-frequency source.

Two residual networks are implemented to perform regression for the source localization and environment classification using a moving mid-frequency source, recorded during the Seabed Characterization Experiment in 2017. The first model implements only the classification for inferring the seabed type, and the second model uses regression to estimate the source localization parameters. The training is performed using synthetic data generated by the ORCA normal mode model. The architectures are tested on both the measured field and simulated data with variations in the sound speed profile and seabed mismatch. Additionally, nine data augmentation techniques are implemented to study their effect on the network predictions. The metrics used to quantify the network performance are the root mean square error for regression and accuracy for seabed classification. The models report consistent results for the source localization estimation and accuracy above 65% in the worst-case scenario for the seabed classification. From the data augmentation study, the results show that the more complex transformations, such as time warping, time masking, frequency masking, and a combination of these techniques, yield significant improvement of the results using both the simulated and measured data.

Influence of seabed on very low frequency sound recorded during passage of merchant ships on the New England shelf.

An examination of the received spectrogram levels of about twenty merchant ship recordings on two vertical line arrays deployed on the New England continental shelf during the Seabed Characterization Experiment 2017 has identified an acoustic feature that can be attributed to the group velocities of modes 1 and 2 being equal at a frequency f=F. The observation of such a feature is a result of ßnm(2πF)=∞, where ßnm is the waveguide invariant for modes n and m. For the New England Mudpatch, the average value of F is about 24.5 Hz. An effective seabed model is inferred from a feature inversion method that has a deep sediment layer which lies between 190 m and 290 m beneath the seafloor with sound speeds on the order of 1810 m/s. This effective sediment model appears to be consistent with a previous seismic survey on the New England shelf that identified a deep low speed layer about 250 m beneath the water sediment interface.

Learning location and seabed type from a moving mid-frequency source.

While source localization and seabed classification are often approached separately, the convolutional neural networks (CNNs) in this paper simultaneously predict seabed type, source depth and speed, and the closest point of approach. Different CNN architectures are applied to mid-frequency tonal levels from a moving source recorded on a 16-channel vertical line array (VLA). After training each CNN on synthetic data, a statistical representation of predictions on test cases is presented. The performance of a single regression-based CNN is compared to a multitask CNN in which regression is used for the source parameters and classification for the seabed type. The impact of water sound speed profile and seabed variations on the predictions is evaluated using simulated test cases. Environmental mismatch between the training and testing data has a negative impact on source depth estimates, while the remaining labels are estimated tolerably well but with a bias towards shorter ranges. Similar results are found for data measured on two VLAs during Seabed Characterization Experiment 2017. This work shows the superiority of multitask learning and the potential for using a CNN to localize an acoustic source and detect the surficial seabed properties from mid-frequency sounds.

Maximum entropy inference of seabed attenuation parameters using ship radiated broadband noise.

The received acoustic field generated by a single passage of a research vessel on the New Jersey continental shelf is employed to infer probability distributions for the parameter values representing the frequency dependence of the seabed attenuation and the source levels of the ship. The statistical inference approach employed in the analysis is a maximum entropy methodology. The average value of the error function, needed to uniquely specify a conditional posterior probability distribution, is estimated with data samples from time periods in which the ship-receiver geometry is dominated by either the stern or bow aspect. The existence of ambiguities between the source levels and the environmental parameter values motivates an attempt to partially decouple these parameter values. The main result is the demonstration that parameter values for the attenuation (α and the frequency exponent), the sediment sound speed, and the source levels can be resolved through a model space reduction technique. The results of this multi-step statistical inference developed for ship radiated noise is then tested by processing towed source data over the same bandwidth and source track to estimate continuous wave source levels that were measured independently with a reference hydrophone on the tow body.

Maximum entropy approach to statistical inference for an ocean acoustic waveguide.

A conditional probability distribution suitable for estimating the statistical properties of ocean seabed parameter values inferred from acoustic measurements is derived from a maximum entropy principle. The specification of the expectation value for an error function constrains the maximization of an entropy functional. This constraint determines the sensitivity factor (ß) to the error function of the resulting probability distribution, which is a canonical form that provides a conservative estimate of the uncertainty of the parameter values. From the conditional distribution, marginal distributions for individual parameters can be determined from integration over the other parameters. The approach is an alternative to obtaining the posterior probability distribution without an intermediary determination of the likelihood function followed by an application of Bayes' rule. In this paper the expectation value that specifies the constraint is determined from the values of the error function for the model solutions obtained from a sparse number of data samples. The method is applied to ocean acoustic measurements taken on the New Jersey continental shelf. The marginal probability distribution for the values of the sound speed ratio at the surface of the seabed and the source levels of a towed source are examined for different geoacoustic model representations.

A nonlocal effective operator for coupling forward and backward propagating modes in inhomogeneous media.

In an acoustic waveguide spatial inhomogeneities couple the forward and backward propagating modal amplitudes. To address the nature of such coupling the integral equation for the range-dependent modal amplitudes is decomposed into components that satisfy the asymptotic boundary conditions of the free Green's function operator. An equivalent set of equations is obtained by eliminating the components that become the asymptotically backward propagating channels to leave a set of integral equations that describe only the components that become asymptotically the forward propagating channels. The elimination of the components that become asymptotically the backward propagating channels is done at the expense of introducing a nonlocal effective coupling operator. The nonlocal operator contains all the effects of the asymptotically backward propagating field on the asymptotically forward propagating field. An expansion of the effective coupling operator allows an investigation of the importance of the coupling and provides a systematic approach to add correction terms to the forward only equation. Idealistic underwater waveguides with various degrees of inhomogeneities are used to illustrate the main features of the convergence characteristics for the expansion.

Low-frequency geoacoustic model for the effective properties of sandy seabottoms.

The debate on the sound speed dispersion and the frequency dependence of sound attenuation in seabottoms has persisted for decades, mainly due to the lack of sufficient experimental data in the low-frequency (LF) to high-frequency speed/attenuation transition band. This paper analyzes and summarizes a set of LF measurements in shallow water that have resulted in the identification of nonlinear frequency dependence of sound attenuation in the effective media of sandy seabottoms. The long-range acoustic measurements were conducted at 20 locations in different coastal zones around the world. The seabed attenuations, inverted from different acoustic field measurements and characteristics, exhibit similar magnitude and nonlinear frequency dependence below 1000 Hz. The resulting effective sound attenuation can be expressed by alpha(dB/m)=(0.37+/-0.01)(f/1000)((1.80+/-0.02)) for 50-1000 Hz. The corresponding average sound speed ratio at the bottom-water interface in the 50-600 Hz range is 1.061+/-0.009. Both the LF-field-derived sound speed and attenuation can be well described by the Biot-Stoll model with parameters that are consistent with either theoretical considerations or experimental measurements. A combination of the LF-field-inverted data with the SAX99, SAX04, and other high-frequency measurements offers a reference broadband data set in the 50-400 000 Hz range for sonar prediction and sediment acoustics modeling.

Seabed acoustics of a sand ridge on the New Jersey continental shelf.

Acoustic measurements were made on a sand ridge on the New Jersey continental shelf. Data collected on two L arrays separated by 20 km from a single multi-frequency tow suggest small horizontal environmental variability. Values for the sound speed structure of the seabed are extracted by first applying a geo-acoustic inversion method to broadband and narrowband acoustic data from short-range sources. Then, a parabolic equation algorithm is used to properly include the bathymetry and sub-bottom layering. Finally, the frequency dependence of the seabed attenuation is inferred by optimizing the model fit to long-range transmission loss data in the 50-3000 Hz band.

Analysis of wind-driven ambient noise in a shallow water environment with a sandy seabed.

On the New Jersey continental shelf ambient sound levels were recorded during tropical storm Ernesto that produced wind speeds up to 40 knots in early September 2006. The seabed at the position of the acoustic measurements can be approximately described as coarse sand. Differences between the ambient noise levels for the New Jersey shelf measurements and deep water reference measurements are modeled using both normal mode and ray methods. The analysis is consistent with a nonlinear frequency dependent seabed attenuation for the New Jersey site.

Broadband modeling of downslope propagation in a penetrable wedge.

Sound propagation in a wedge-shaped environment with a penetrable bottom is simulated with broadband adiabatic mode, coupled mode, and parabolic equation model computations. Simulated results are compared to measured data taken in a tank experiment by Tindle et al. The coupled mode formalism is shown to predict, in agreement with that experiment, that modal wave fronts in penetrable wedges are approximately circular arcs centered at the apex of the wedge for a source near the apex. It is also shown that for wedge angles up to 6 degrees, the received waveforms are well approximated by the adiabatic waveforms time-shifted by a depth-dependent interval to account for the curvature of the modal wave fronts. A small deviation from circularity in the modal wave fronts is possibly observed in the 6 degrees case.

Geoacoustic inversion of short range source data using a plane wave reflection coefficient approach.

Acoustic time series data were collected in a shallow, hard bottom lake environment located in central Texas using both short range (2 m) implosive data, obtained with the source and a single hydrophone located near mid-depth in the waveguide, along with longer range implosive and explosive data from a near surface source to a bottom mounted hydrophone. Matched field inversions using simulated annealing were performed with a ray trace plus complex plane wave reflection coefficient forward propagation model that was validated in previous work. Isolating bottom interacting paths to perform the inversions is shown to be essential to reduce parameter uncertainties in the hard bottom environment and enables a systematic approach to the inversions which establishes the number of layers needed to represent the lake environment. Measured transmission loss data from a towed source were compared through a RMS error analysis to modeled transmission loss, constructed with the parameters from inversions of data from several source types, to further establish the validity of the inversion approach for this environment. Geoacoustic parameters obtained by inversions of short range, low frequency impulsive data are used to predict transmission loss at longer ranges and higher frequencies. The range dependence of the global minimum is discussed.

Low frequency coupled mode sound propagation over a continental shelf.

A two-way integral equation coupled mode method is applied to a continental shelf ocean waveguide proposed for a special session devoted to range-dependent acoustic modeling at the 141st meeting of the Acoustical Society of America. The coupled mode solution includes both sediment trapped and continuum modes. The continuum is approximated by a finite number of leaky modes but neglects the branch cut contribution. Mode coupling matrix elements and the range evolution of the modal amplitudes show the nature of the mode coupling. Transmission loss versus range at 100 Hz predicted by the integral equation approach is compared to the transmission loss predicted by a wide angle parabolic equation method. While there is very good agreement, one observes small differences that can be interpreted as backscattering predicted by the integral equation solution.

Broadband sound propagation in shallow water and geoacoustic inversion.

Part of an experiment to test a measurement package in a shallow water region in the Gulf of Mexico was designed to gather broadband acoustic data suitable for inversion to estimate seabed geoacoustic parameters. Continuous wave tow acoustic signals at multiple frequencies and broadband impulsive source signals were recorded on a horizontal line array in a high-noise environment. Simulated annealing with a normal mode forward propagation model is utilized to invert for a geoacoustic representation of the seabed. Several inversions are made from different data samples of two light bulb implosions, the measured sound speed profiles at the HLA and at the positions of the light bulb deployments, and for two different cost functions. The different cost functions, measured sound speed profiles, and measured time series result in different inverted geoacoustic profiles from which transmission loss is generated for comparison with measurements. On the basis of physical consistency and from the comparison of the transmission loss and time series, a best estimate geoacoustic profile is selected and compared to those obtained from previously reported inversions. Uncertainties in the sound speed profile are shown to affect the uncertainties of the estimated seabed parameters.