Shelfbreak PRIMER Research Articles

Geoacoustic parameter estimation using machine learning techniques

The Airy phase region corresponding the minimum group velocity of the dispersive acoustic normal modes in shallow water are extremely sensitive to bottom properties. The group speed minima and the associated frequency data for two lower order modes were used to train a neural network to predict the bottom parameters in a previous study (Potty et al., 2019). The geoacoustic model for the bottom consisted of a sediment layer over acoustic basement. Synthetic data generated by varying the values of the sound speeds in the sediment layer and basement were used to train and evaluate the technique. The output of the neural network was used as the background model for a linear perturbation inversion. Data collected from the Shelf break Primer experiment were used to test the algorithm and the preliminary results were promising. This study will continue the previous work by incorporating higher order modes in the training data, testing with data from other experiments and using other machine learning techniques. The performance of this hybrid approach (neural network combined with linearized inversion) will be compared with fully non-linear inversion, both in terms of accuracy and computation time. [Work supported by the Office of Naval Research.]

Effect of Shear on Modal Arrival Times

The sea bottom is generally modeled as a fluid for many of the shallow water acoustic propagation modeling applications. The inherent assumption is that the bottom does not support any shear wave propagation. This study explores the impact of this assumption on the dispersion behavior of acoustic normal modes. A sensitivity study is performed to investigate the effect of density, shear, and compressional wave speeds on modal dispersion. This study focuses on low-frequency (less than 100 Hz) and lower order modes (mode four and lower) in shallow water (70–90-m water depth). These modes and frequency bands are characterized by deeper penetration and larger depth averaging scales. This enabled the use of a simple half-space elastic model of the sea bottom for most of the analyses in this study. A depth-dependent bottom model was used to investigate the effect of a mud layer on modal dispersion. Sediment shear wave speeds were also estimated using broadband data from two shallow water experiments: Shelf Break Primer (1996) and Seabed Characterization Experiment (2017). The inversion results from these two experiments are compared and they are also compared with deep core data and prior inversions.

Shallow water acoustics and oceanography at the Woods Hole Oceanographic Institution over the last quarter century

Shallow water (coastal) acoustics has been a topic of great interest over quarter century, both from the point of view of learning coastal oceanography and of learning coastal acoustics. Indeed, the two are difficult to separate. In this talk, the research done at the Woods Hole Oceanographic Institution (very often in collaboration with other institutions) over the past 25 years will be discussed, with an emphasis on the major experimental efforts that have been fielded, but also including theoretical and computational efforts. The Barents Sea Polar Front experiment, the Shallow Water Acoustic Random Medium experiment, the Shelfbreak PRIMER experiment, the Asian Seas International Acoustics Experiment, the Shallow Water 2006 experiment, and the Quantifying, Predicting and Exploiting Uncertainty program are some of the at-sea efforts to be discussed. Theoretical advances in shallow water acoustics, as well as computational efforts, such as the Integrated Ocean Dynamics and Acoustics project, will also be ...

Acoustic and sediment data in the southern New England Bight

This study provides a review of the acoustic and ocean bottom sediment data collected during the Shelf Break Primer experiment conducted in 1996. The location of the experiment was in the southern New England Continental Shelf called the “New England Mud Patch.” The mud patch is a 13,000 square kilometer area covered by fine-grained sediment. Previous surveys in this area have estimated the thickness of this fine grained sediment deposit to be as much as 13 m. This layer of sediment rests on a reflector that is geomorphically similar to and continuous with the Holocene transgressive sand sheet, which is exposed on the shelf to the west of this area. This fine-grained sediment layer, which is oriented in an east-west direction seaward of the 55–65 m isobath, contains more than 30% silt and clay. During the Primer experiment, broadband acoustic sources were deployed in the western side of the “mud patch” along and across the continental shelf. The acoustic data collected on a vertical line array will be analyzed. Sediment data from gravity cores from the area will also be presented. [Work sponsored by Office of Naval Research, code 322 OA.]

Experimental design for sediment characterization in the New England Mud Patch

This study focuses on the design of an experiment to estimate the shear wave properties of ocean bottom sediments at a location in the southern New England Continental Shelf called the “New England Mud Patch.” The mud patch is a 13,000 square kilometer area covered by fine-grained sediment. The inversion technique is based on collecting interface wave (Scholte wave) data using geophones on the sea bottom. The data for the inversion consist of Scholte wave phase velocity dispersion calculated from the geophone array data. The present study aims at applying this Scholte wave based shear wave inversion technique to the “mud patch” area. Two different interface wave measurement systems will be presented. Appropriate source to receiver ranges will be explored based on simulations. The engineering challenges associated with deploying the system in a soft seabed will be investigated and design modifications will be investigated. Acoustic and sediment data from the 1996 Shelf Break Primer experiment, which was conducted in the western side of the mud patch, will be reviewed. The simulations will be based on historic sediment and acoustic data. [Work sponsored by Office of Naval Research, code 322 OA.]

Inversion of Internal Wave-Perturbed Sound-Speed Field Via Acoustic Data Assimilation

Model-based signal processing methods, such as matched-field processing, are concerned with exploiting the combination of the sound propagation physics in an ocean waveguide and the observation data. As such, they require accurate knowledge of the ocean acoustic environment. This paper presents an approach of environmental inversion via acoustic data assimilation (ADA), intended specifically to reconstruct the sound-speed field perturbed by linear internal waves. It exploits a variety of information sources, including direct local sound-speed measurements, an internal wave model, a full field acoustic propagation model, and acoustic pressure measurements. Compared to the previous work on ADA, four important improvements have been made: 1) a linear internal wave model is included; 2) a more realistic shallow-water environment from the Shelf Break PRIMER 1996 experiment is considered in the numerical simulations; and 3) effect of the sediment on acoustic propagation is considered in the inversion algorithm. The results further demonstrate the validity of the ADA approach.

Time frequency analysis techniques for long range sediment tomography

Long range sediment tomography inversion technique requires accurate estimation of the arrival times of acoustic normal modes. The inversion technique is based on minimizing the difference between the forward model predictions and the data using a global optimization scheme. Predictions are computed using a trial parameter set which is iteratively modified until the algorithm converges. The modal arrival times are calculated from the time-frequency analysis of broadband acoustic data collected on a single hydrophone. During the initial stages of the development of the inversion scheme, Fourier-based spectrograms and wavelet-based scalograms were used. Taking advantage of some of the recent developments, dispersion based short time Fourier transform (DSTFT) and warping transform techniques were used for the time-frequency analysis, in recent times. This paper will summarize and compare the performance of these time frequency techniques in the context of long range sediment tomography. Data from some of the recent field tests (Shelfbreak Primer and Shallow Water-06 experiments) will be analyzed for this study. Finally, time-frequency analysis performance of another new technique, Modified S transform, will be examined using the field data. [Work supported by Office of Naval Research].

Nonlinear optimization for beamforming a geometrically deficient vertical line array: Application to sediment tomography

The 1996 Shelfbreak PRIMER experiment included broadband acoustic transmissions in the vicinity of a shelfbreak front collected using a geometrically deficient vertical line array (VLA). The broadband signals were used to tomographically image sediment sound speed and attenuation. The VLA spanned the lower half of the water column. Hence, the orthonormal properties of the underlying normal modes cannot be guaranteed for any replica vector based on the local modal structure. This produces modal sidelobes in the output of the geometrically deficient VLA. Various methods of extracting the normal modes from VLA data have been developed including a generalized least squares method shown to improve the performance of a geometrically deficient VLA processing phase encoded tomography signals [Chiu et al., IEEE J. Ocean. Eng. 22, 522–533 (2005)]. The present study seeks to enhance the modal resolution for the reception of high-amplitude, impulsive signals transmitted over a wide variety of source receiver paths in shallow water where ensemble averaging is not an option. The method uses nonlinear optimization techniques to compute a set of array element coefficients that enhance suppression of modal sidelobes, while maintaining adequate spatial matching with the desired mode. [Work supported by Office of Naval Research.]

Propagation of impulsive broadband signals in a coastal ocean setting during the 1996 Shelfbreak Primer experiment

The 1996 Shelfbreak PRIMER experiment provided an opportunity to study the impulse response of a coastal ocean waveguide using broadband acoustic signals from explosive sources. Acoustic transmission paths in the vicinity of the shelfbreak included various orientations ranging from up‐slope to along‐slope, including interactions with a shelfbreak front. Acoustic receptions were acquired with a 16‐element vertical line array (VLA). In addition, thermistor string data acquired at the receiving array suggest that a group of high‐amplitude, internal waves traversed the test site during this portion of the experiment. Modal decomposition of broadband signal arrivals was performed to investigate variation of the impulse response during source deployments. Variations in the received spectra of individual normal modes were compared with numerical predictions for impulsive broadband signals propagating in range‐dependent environments constructed using oceanographic data available for the test site. The potential for mode coupling and three‐dimensional effects to shape the modal and spectral distribution of received energy is being investigated. Numerical and experimental results will be discussed. [Work supported by Office of Naval Research.]

Vertical directionality of noise at two shallow water sites

Noise recorded with 16-element vertical line arrays moored at two sites near 80-m water depth has been analyzed. One site is south of New England (Shelfbreak PRIMER, June 1996), the other south of China in the South China Sea (ASIAEX, May 2001). Both sites exhibited downward refracting conditions, so that a noise notch is realizable, and both exhibit internal gravity waves. The waves were more plentiful and of larger amplitude at the South China Sea site. Sound in a 40-Hz band centered at 135 Hz was extracted and separated into vertical beams using delay and sum beamforming. At each location, noise lower than the median level exhibited a peak near zero degrees for each data set (near-horizontal arrival). Noise higher than the median level in the New England data set exhibited a 5-dB noise notch, with peak levels at about ±15 deg. Noise higher than the median level in the South China Sea data set did not show the notch and was essentially flat between 10 and −10 deg. The results are consistent with observed fluctuating propagation between fixed sources and re- ceivers at the sites, and with differences in coupled-mode propagation through the two internal-wave fields.

Climatological Estimation of Environmental Uncertainty Over the Middle Atlantic Bight Shelf and Slope

Historical hydrographic data are used to determine the spatial and seasonal patterns of uncertainty in thermohaline and sound-speed fields in a well-sampled region, the continental shelf and slope in the Middle Atlantic Bight (MAB). Several different historical databases are combined to produce two-dimensional (2-D) plan view and cross-shelf fields of temperature, salinity, and sound speed in two separate regions, the New England shelf and the shelf off Delaware and Maryland. In addition, spatial maps of the sound-speed fields reveal that the maximum variance of the sound speed occurs at the edge of the continental shelf, in the vicinity of the shelfbreak front. The standard deviation of the sound speed was largest during the spring and summer, with magnitudes as large as 14 m/s in a narrow band coinciding with the mean position of the shelfbreak front. During spring the peak in variance was located near the surface outcrop of the front, but during summer the maximum variance was centered at a depth of 30 m, immediately beneath the seasonal thermocline. Comparisons with both synoptic measurements from the Shelfbreak PRIMER experiment as well as moored time series from the Nantucket Shoals Flux Experiment confirm that the shelfbreak front is a “hotspot” of uncertainty (maximum variance), and that the vertical structure of the peak variance is dependent on the presence or absence of the seasonal thermocline.

Oceanographic and Sound Speed Fields for the ESME Workbench

The authors describe the effort to provide three-dimensional global thermohaline and sound speed fields for use in the effects of sound in the marine environment (ESME) workbench suite of programs. The primary fields used are from the modular ocean data assimilation system (MODAS), developed by Fox et al. The system provides global thermohaline and sound speed fields on a daily basis using environmental inputs from the U.S. Navy as well as remote sensing of sea surface temperature and sea surface height. To examine the MODAS fields, the authors also used data from the Southern California Bight collected by the California Cooperative Fisheries Investigations as well as high-resolution hydrographic data collected over the continental shelf south of New England as part of the shelfbreak PRIMER experiment. MODAS performs well for features such as large-scale boundary currents and eddies but is more limited in resolving features such as shelfbreak and coastal fronts, which have small spatial and temporal correlation scales. Because of the considerable computational needs of other ESME modules and its use as a planning tool, the authors present a pragmatic approach for future applications

Analysis of multipath scintillations from long range acoustic transmissions on the New England continental slope and shelf

Observations of intensity fluctuations for broadband, 400-Hz multipath arrivals observed during the 1996 New England shelfbreak PRIMER study are described. Acoustic signals were generated by two bottom-mounted sources located on the continental slope in roughly 290-m water depth and were received on a 52-m-long vertical-line-array (VLA) located in 93-m water depth. Propagation ranges were 42.2 and 59.6 km. The bathymetry, oceanography, and bottom geology of the PRIMER site are described. Acoustic observables of point intensity, peak intensity, and integrated energy over the VLA are treated in terms of the scintillation index, log-intensity variance, and intensity probability density functions (PDFs). Variability of the observables are decomposed into high and low frequency components with time scales less than and greater than 2 h, to facilitate correlation to ocean processes at different timescales. Parabolic equation numerical simulations using a quasi-random undular tidal bore model are able to reproduce many of the observed intensity fluctuation to within a factor of 2, and they allow investigation of scintillation behavior as a function of range.

A large‐amplitude meander of the shelfbreak front during summer south of New England: Observations from the Shelfbreak PRIMER experiment

In order to examine spatial and temporal variability of the shelfbreak front during peak stratification, repeated surveys using a towed undulating vehicle (SeaSoar) are used to describe the evolution of shelfbreak frontal structure during 26 July to 1 August 1996 south of New England. Spatial correlation (e‐folding) scales for the upper 60 m of the water column were generally between 8 and 15 km for temperature, salinity, and velocity. Temporal correlation scales were about 1 day. The frontal variability was dominated by the passage of a westward propagating meander that had a wavelength of 40 km, a propagation speed of 0.11 m s−1, and an amplitude of 15 km (30 km from crest to trough). Along‐front geostrophic velocities (referenced to a shipboard acoustic Doppler current profilers) were as large as 0.45 m s−1, although subject to significant along‐front variations. The relative vorticity within the jet was large, with a maximum 0.6 of the local value of the Coriolis parameter. Seaward of the front, a small detached eddy consisting of shelf water was present with a diameter of approximately 15 km. Ageostrophic contributions to the velocity field are estimated to be as large as 0.3 m s−1 in regions of sharp curvature within the meander. These observations strongly suggest that during at least some time periods, shelfbreak exchange is nonlinear (large Rossby number) and dominated by features on a horizontal scale of order 10 km.

Inversion for sediment geoacoustic properties at the New England Bight

This article discusses inversions for bottom geoacoustic properties using broadband acoustic signals obtained from explosive sources. Two different inversion schemes for estimating the compressional wave speeds and attenuation are presented in this paper. In addition to these sediment parameters, source-receiver range is also estimated using the arrival time data. The experimental data used for the inversions are SUS charge explosions acquired on a vertical hydrophone array during the Shelf Break Primer Experiment conducted south of New England in the Middle Atlantic Bight in August 1996. The modal arrival times are extracted using a wavelet analysis. In the first inversion scheme, arrival times corresponding to various modes and frequencies from 10 to 200 Hz are used for the inversion of compressional wave speeds. A hybrid inversion scheme based on a genetic algorithm (GA) is used for the inversion. In an earlier study, Potty et al. [J. Acoust. Soc. Am. 108(3), 973-986 (2000)] have used this hybrid scheme in a range-independent environment. In the present study results of range-dependent inversions are presented. The sound speeds in the water column and bathymetry are assumed range dependent, whereas the sediment compressional wave speeds are assumed range independent. The variations in the sound speeds in the water column are represented using empirical orthogonal functions (EOFs). The replica fields corresponding to the unknown parameters were constructed using adiabatic theory. In the second inversion scheme, modal attenuation coefficients are calculated using modal amplitude ratios. The ratios of the modal amplitudes are also calculated using time-frequency diagrams. A GA-based inversion scheme is used for this search. Finally, as a cross check, the computed compressional wave speeds along with the modal arrival times were used to estimate the source-receiver range. The inverted sediment properties and ranges are seen to compare well with in situ measurements and historical data.

Intensity fluctuations from deterministic and random azimuthal variability

Shallow-water sites for shelfbreak PRIMER and ASIAEX show substantial ocean horizontal variability arising from nonlinear internal wave trains. These relatively coherent structures strongly influence acoustic intensity observations, and deterministic propagation modeling has demonstrated their azimuthal effects. Moreover, random variations in the nonlinear internal waves, as well as processes such as linear internal wave fields, contribute to observed intensity fluctuations. Stochastic modeling is performed to show the azimuthal changes induced in the intensity fluctuations. The principal metric of interest is the scintillation index (SI), the ratio of standard deviation to mean intensities. Limiting cases of the SI are examined for examples including both nonlinear internal waves and random ocean variations. Simulations using a three-dimensional parabolic equation model [Collins, J. Acoust. Soc. Am. 94, 2269–2278 (1993)] are discussed, along with an extension of a random modal fluctuation approach [Dozier and Tappert, J. Acoust. Soc. Am. 63, 353–365 (1978)]. [Work supported by ONR.]

Characteristics of acoustic propagation to the eastern vertical line array receiver during the summer 1996 New England shelfbreak PRIMER experiment

During July and August of 1996, the summer component of the New England shelfbreak front PRIMER experiment was fielded in the Mid-Atlantic Bight, at a site due south of Martha's Vineyard, MA. This study produced acoustic transmission data from a network of moored sources and receivers in conjunction with very-high-resolution oceanography measurements. This paper analyzes receptions at the northeast array receiver from two 400 Hz acoustic tomography sources, with the transmission paths going from the continental slope onto the continental shelf. These data, along with forward acoustic-propagation modeling based on moored oceanographic data, SeaSoar hydrography measurements, and bottom measurements, reveal many new and interesting aspects of acoustic propagation in a complicated slope-shelf environment. For example, one sees that both the shelfbreak front and tidally generated soliton internal wave packets produce stronger mode coupling than previously expected, leading to an interesting time-and-range-variable population of the acoustic normal modes. Additionally, the arrival time wander and the signal spread of acoustic pulses show variability that can be attributed to the presence of a frontal meander and variability in the soliton field. These and other effects are discussed in this paper, with an emphasis on creating a strong connection between the environmental measurements and the acoustic field characteristics.

Directional noise and correlation of signal and noise variability in the ASIAEX South China Sea Experiment

As part of the 2001 ASIAEX South China Sea experiment, known acoustic signals and noise were recorded for 16 days with co-located horizontal and vertical receiving arrays in 120-m depth water at the edge of the continental shelf. The location was south of China and southwest of Taiwan. Vertical and horizontal directionality of both signals and noise can be estimated using the two arrays. Directionality can be linked to vertical mode bandwidth. Furthermore, because of mode stripping effects, it can indicate the amount of mode coupling in the vicinity of the receiver. Previously, arrival time and total energy of the signals have been shown to respond to the strong nonlinear internal waves and internal tides of the area. Here, we investigate the response of noise energy and directionality to the waves, and compare with the response of signal energy and directionality. The results can be compared to the Shelfbreak PRIMER experiment in the Mid-Atlantic Bight, where signal and noise parameters recorded with a VLA were correlated. Correlation of signal and noise violates assumptions used in many signal processing algorithms.

Optimal search strategies for nonlinear, multi-parameter geoacoustic inverse problems

A mode travel time based technique for geoacoustic inversions has been developed using broadband explosive sources [Potty et al., J. Acoust. Soc. Am. 108, 973–986 (2000)]. In this inversion scheme we search for various parameters including water depth, sound speed in the water column represented by Empirical Orthogonal Function (EOFs), compressional wave speed in the sediment and sub-base, thickness of the sediment and range. The sensitivity of the objective function to perturbations in these parameters is different, i.e., different parameters have different sensitivities. The most sensitive parameters can be estimated with higher accuracy compared to parameters with low sensitivities. In this study we aim to link the search process to parameter sensitivities by matching the fineness of search to sensitivities. The highly sensitive parameters will be searched with a fine scale sampling of the model space with a coarser sampling adopted for less sensitive model parameters. This condition will be incorporated into the search tool (Genetic Algorithm) used for the inversion scheme. The objective of this approach is to utilize the computational resources in an optimum manner by giving more emphasis to parameters which are more sensitive as opposed to a uniform weighting for all parameters. This approach will be applied to data from the ASIAEX-2001 and Shelf Break Primer data sets. [Work supported by ONR.]

Spatial and temporal variations in acoustic propagation characteristics at the new england shelfbreak front

The spatial and temporal variability of the acoustic field in the region of a strong coastal shelfbreak front are examined, using the high-resolution environmental data from the 1996-1997 New England shelfbreak PRIMER experiment to provide input to acoustic propagation models. Specifically, the "SeaSoar" undulating conductivity-temperature-depth (CTD) probe across-shelf transects provide 1-km along-track resolution, including the front, during the spring, summer, and winter seasons. These data allow one to study the diurnal and seasonal temporal variation of the acoustic field, as well as the varying spatial structure of the field. Using the RAM parabolic equation code, across-shelf acoustic field structure at 200, 400, and 1000 Hz is studied for various source depths. A number of interesting propagation effects are noted, the most interesting of which are the inhibition of strong acoustic-bottom interaction by the warm shelf water beneath the shelfbreak front and the existence of small-scale ducts near the front, due to offshore transport. Data from the vertical line arrays deployed as part of PRIMER offer partial validation of the predictions made. Specifically, it is seen that the mean received levels are in reasonable accord with propagation calculations made using locally measured bottom properties and the SeaSoar water-column measurements.