Shallow Water Acoustics Research Articles

Modeling of the out-of-plane pulse propagation in the SWARM95 experiment

The acoustical data collected in the shallow water acoustics in a random medium (SWARM95) experiment on the continental shelf off New Jersey in 1995 have shown that the temporal variations of the time series of the airgun signals (center frequency 75 Hz and bandwidth 50 Hz) are both range and azimuthal dependent. A full-wave broadband PE model is used to simulate such shallow water pulse propagation at different horizontal bearings. The time-dependent shallow water internal wave fluctuations constructed with a broad spectrum of excitation and the solitonlike thermocline displacements based on the in situ observations are included in the numerical simulation as well as the bathmetry. The vertical wavefronts and the 2-h time histories of the repeated transmissions from airgun positions above and below the thermocline are calculated and displayed. The simulated data and the experimental data of the receiver arrays at different locations with an approximately 10-km cross-range distance are compared. The results suggest that the azimuthal dependence of the received signals correlate to both the internal wave fields and the bathmetric variations.

Preliminary results on environmentally robust vertical array design for shallow water mode filtering

Many shallow water acoustics techniques require accurate estimates of the propagating modes from pressure observations at a vertical array. The noise in these estimates is a function of the conditioning of the vertical modes as sampled by the hydrophone array. However, the sound-speed profile and thus vertical mode shapes are usually not known when the array is designed. In addition, economic considerations often dictate the same array be used in a wide range of environments. Consequently, optimizing the array for a specific environment risks a degradation in performance if the experimental environment differs from the environment assumed by the array design. Conventional wisdom holds that in the absence of precise environmental knowledge, the safest approach is to space the hydrophones evenly throughout the water column. This work presents several simplified mode filtering scenarios demonstrating that nonuniform hydrophone spacings exist which are more robust in the minimax sense than uniform spacing, if it is possible to specify a range of operating environments for the array. This means that their worst case performance is better than the worst case performance of the uniformly spaced array over the range of environments.

Acoustic intensity fluctuations induced by internal waves in a shallow-water waveguide

The SWARM (Shallow Water Acoustics in a Random Medium) experiment performed during July–August 1995 made simultaneous acoustic and oceanographic measurements along a 45-km track on the continental shelf off the New Jersey coast. The area is a site of intense internal wave activity, primarily in the form of soliton packets that perturb the sound-speed profile in range, depth, and time. Incorporating environmental parameters estimated from SWARM measurements, computer simulation results are presented for acoustic propagation through this environment at frequencies of 224 and 400 Hz. A wide-angle parabolic equation method was used to compute multiple realizations of the acoustic field. The simulations employed bathymetry and sub-bottom acoustic parameters estimated from chirp sonar surveys. Oceanographic parameters were obtained from thermistor strings and yo-yo CTDs in conjunction with satellite images, which allowed for estimation of internal wave-field parameters that were used to compute sound-speed distributions. Numerical results include estimates of the full-field and modal-intensity fluctuations, as well as the spatial coherence at two vertical arrays located at 33 and 45 km from the moored sources. [Work supported by NRL and ONR. The authors wish to acknowledge the contribution of the SWARM group to this work.]

Three-dimensional chirp sonar surveying of SWARM-95 experimental site

A full-spectrum chirp sonar system [LeBlanc et al., J. Acoust. Soc. Am. 91(1) (1992)] was used to invert three-dimensional density and sound-speed structure of the subbottom at the SWARM-95 (shallow water acoustics in random media) experimental site on the New Jersey Shelf. Biot’s theory was used in a forward model to calculate normal-incident reflection seismograms for layered marine sediments. Efficient inversion algorithms based on global optimization techniques were used to survey large areas for characterization of sediment properties and their spatial distributions. Chirp sonar inversion results were compared with ground-truth core data at the AMCOR-6010 site and good agreement was observed. Three-dimensional images of sediment density and sound-speed structure were constructed from several parallel runs along the SWARM acoustic track 1003 at the outer-shelf sediment wedge area. Corresponding 3-D wave-number spectra of density and sound-speed variability were also calculated.

Observations, characterizations, and acoustic effects of nonlinear internal waves on the Mid-Atlantic Bight continental shelf

During the summer of 1995, a multi-institutional field study called Shallow Water Acoustics in a Random Medium (SWARM) was conducted in the Mid-Atlantic Bight continental shelf region off the coast of New Jersey. Environmental and acoustic sensors were deployed as part of SWARM to measure and characterize the nonlinear internal waves and their impact on the spatial and temporal coherence of the acoustic transmissions. As part of the environmental monitoring network, two bottom-moored, upward-looking acoustic Doppler current profilers (ADCPs) were deployed. An oceanographic, modal, time-series analysis of the ADCP data reveals that: large-amplitude, nonlinear, internal wave packets were generated at multiple sites near the shelfbreak; the generation mechanism was consistent with the lee-wave hypothesis of generation; the propagation characteristics were in good agreement with nonlinear soliton theory; and the power spectral density was spatially varying and changed markedly during the passage of these nonlinear waves. Based on these observations, a canonical model of the induced sound-speed perturbations was developed. Using a coupled normal-mode propagation model, the temporal and vertical structures of the sound field were subsequently calculated for comparison to data obtained by a vertical line array. [Work supported by ONR.]

Internal waves and tides in shallow water

The well-known Garrett and Munk description of the internal wave spectrum synthesizes many observations into a spectral model that reproduces the general features of the mid-latitude, deep-ocean internal wave field. The general applicability of this ‘‘kinematic’’ spectral model is related to the dynamic balance of the internal wave field, but the model itself includes no dynamics and is meant to represent only the average or steady-state spectrum, away from the direct influence of sources, sinks, or boundaries. In shallow water, the kinematic model is typically not appropriate, since some or all of the simplifying assumptions about the wave field may be invalidated. Variable topography, strong tidal currents, fronts, and ice cover (at high latitudes) are the principal reasons for these complications, and result in an internal wave field which varies both spatially and temporally from the steady-state kinematic description. The most straightforward way to address the inadequacy of steady-state models for shallow water acoustics experiments is direct observation of the internal wave field at the site. Profiles sufficient to define the ‘‘mean’’ or background density gradient along with time series of temperature and horizontal velocity are the most common observations.

Hot topics in underwater acoustics

Underwater acoustics is in a period of renaissance with researchers actively pursing studies in many areas from short‐range high‐frequency shallow water acoustics to global low‐frequency acoustics. There has been renewed interest in inhomogeneous sediment acoustics as well as propagation in very shallow water bounded by these sediments. The recent Office of Naval Research ‘‘SRP’’ program studying deep water boundary reverberation has not only been fruitful, but has also laid the foundation for innovative research in the area of boundary scattering in coastal water acoustics. On the other scale, the Heard Island Feasibility Test as described by 18 papers in the October 1994 issue of JASA has re‐awakened interest in the myriad of possibilities of monitoring all scales of the ocean using acoustic methods. In an expanding number of efforts, underwater acoustics is distinguishing itself as developing the most physically based signal processing schemes presently being considered in any field. These and other hot topics will be discussed.

Average acoustic field in shallow water I: Active sonar performance prediction

The seabed dominates the shallow-water acoustics problem, but bottom properties and bottom scattering mechanism are usually poorly known. Detailed interference patterns in shallow water are thus not always physically meaningful for engineering applications. A simple analytical expression averaged over frequency or space can sometimes give better insight into some physical problems. In this presentation, an average angular power spectrum method for calculating shallow-water sound propagation, reverberation, noise, and spatial coherence is briefly introduced. The method is based on normal-mode and ray-mode analogies and originally appeared in Chinese papers which are not available in English [Zhou, Acta Acoust. Sin. 5, 86–99 (1980); Acta Ocean. Sin. 1, 212–217 (1979)]. Taking active sonar performance prediction as an example, analytical expressions for range/depth dependences of the sound propagation, average reverberation intensity, and echo-reverberation ratio will be given for several typical cases, including Pekeris shallow water, wedged continental shelf, and negative gradient waveguide. [Work supported by ONR.]

Time-frequency energy distribution analysis of a shallow water acoustics experiment at the AGS site

The short time Fourier transform (STFT), wavelet transform (WT), and Wigner distribution (WD) are used to characterize the time-frequency distribution of air gun signatures in azimuth and range. The signatures were obtained in a shallow water experiment conducted at the AGS site. The distributions are explained in relation to the acoustic interaction with the seafloor. Accurate descriptions of these distributions are important for better understanding of the acoustic field in shallow water. By means of a combination of the time-frequency distribution and empirical orthogonal functions (EOF), the effects of higher-order modes are studied and localized on the features of the received time series. Numerical simulations using the SNAP model are conducted to interpret some of the observed features. Comparisons between the simulated and measured time series indicate that the higher-order modes are more sensitive to the fine layer structure of the seafloor physical properties. a)Present address: Office of Naval Research, Ocean Acoustics program Code 324OA, 800 North Quincy St., Arlington, VA 22217-5000.

Critical propagation issues in shallow-water acoustics

Several sets of propagation problems in shallow-water acoustics are discussed, covering overlapping frequency ranges. At VLF (say below 300 Hz) the bottom material is important, because energy travels in the bottom or because of attenuation associated with shear-wave coupling. At LF (say above 30 Hz) the presence of gas in surficial sediments can change boundary conditions and bottom losses, and even allow a new type of interface wave. At MF (300 Hz to several kHz) dispersed fish can raise the attenuation, typically by 1 dB/km. Fish have also been observed to dominate the echo structure or reverberation, out to 100 km. At HF (say above several 100 Hz) the sound-speed profile becomes important, partly because of channeling. At VHF (usually above several kHz) the propagation mechanisms tend to change from boundary reflection to boundary scattering. In addition all these effects are greatly influenced by distance and by water depth.

Sonar signal processing and shallow water acoustics

Sonars operating in shallow water usually encounter what signal processors term a reverberation-limited environment; higher signal levels simply lead to higher reverberation and not to improved performance. Consequently, the principal means of improving sonar performance is to use signals appropriate for the task and environment. This implies that robust environmental models are imperative. In both sonar and radar, reverberation, or scattering, is usually modeled to be a random linear system characterized by a scattering function, or some other second moment such as a two-frequency correlation function. Typically, an active sonar resolves a target or an environmental parameter by either range or time gating (high bandwidth signals) or Doppler gating (long duration signals). The resolution properties of a signal are specified by its ambiguity function and the interaction with the scattering function is the single most important factor in determining system performance. This talk will provide a brief overview of some of the signal processing issues for shallow water sonars.

An issue (sediment volume fluctuations) and a nonissue (shear wave propagation) in shallow water acoustics

A two-way full-wave PE reverberation model has been developed that provides predictions of mean reverberation levels in shallow water to long ranges (hundreds of water depths) due to sediment volume fluctuations as well as bottom and basement interface roughnesses. Using T. Yamamoto’s tomographically measured sediment volume fluctuation spectra as one of the inputs, the PE reverberation model has been exercised at several locations on the shelf in the East China Sea over a broad range of frequencies. It has been found that predicted reverberation levels due to both mechanisms are generally comparable in magnitude, and that absolute levels are controlled mainly by total field ensonification of the sediment volume and boundaries, i.e., forward transmission losses. The question whether it is necessary to include shear wave propagation in shallow water acoustic models has been resolved: the answer is definitely negative, and this is no longer an issue. Measured shallow water transmission loss data over a broad range of frequencies with source and receiver in the water column has been successfully modeled using a full-wave range-dependent PE model that includes rough bottom forward scattering (necessary for good agreement), but treats shear wave effects as an equivalent fluid by augmenting the volume attenuation of compressional waves in the sediment layers. The conclusion that it is not necessary to include shear wave propagation when sources and receivers are in the water column has been further confirmed by inter-model and intra-model comparisons using realistic geoacoustic sediment properties as measured by T. Yamamoto’s BSMP system [J. Acoust. Soc. Am. 90, 441–456 (1991)]. [Work supported by ONR, 1125OA.]

Hot topics in underwater acoustics

There have been two major changes that have modulated the temperature of topics in underwater, or ocean, acoustics since the last ‘‘hot topics’’ session. First, close coupling among the acoustics, the ocean environment and its processes, numerical models, and the signal processing continues to lead to new methods for learning about the ocean and to improved performance of acoustic systems. Second, in response to the changing demands upon future sonars, there has been much more attention to both active systems and shallow water environments. Several ‘‘hot’’ (at least warm) topics will be discussed: adaptive focusing systems including phase conjugation, self-cohering arrays, and matched-field processing/tomography; global acoustics and ocean climate monitoring; high resolution bathymetric mapping; reverberation; seismoacoustics of interface waves; shallow water acoustics; ambient noise and acoustic daylight; and high data rate acoustic telemetry. (The order is not an indication of relative temperature.)

Shallow water acoustic channel monitoring with selective mode excitation

The shallow water environment is characterized by multiple processes that cover a wide range of temporal and spatial scales and that have a significant effect on the propagation of acoustic signals [G. V. Frisk, ‘‘Report on the Office of Naval Research Shallow Water Acoustics Workshop,’’ 24–26 April 1991, WHOI-92-06]. The coupling between the modes of acoustic signals is potentially an effective means of monitoring these processes. Traditionally, a signal source is used to excite the acoustic medium. A receiver array is used to sort out the modes of the propagating signal and the coupling between particular modes is inferred from travel time differences. This approach provides relatively little control of the distribution of transmitted energy among the modes and adds uncertainty to the estimation of the coupling between particular modes. An alternative approach is proposed to use an array of sources to selectively excite modes of the medium. There are several advantages and challenges of using such an approach. First, this will improve the ability to estimate the actual coupling between different modes. Second, it is well known that propagation of and coupling between particular modes is most sensitive to environmental perturbations having particular spatial scales. [B. Cornuelle and B. M. Howe, ‘‘High Spatial Resolution in Vertical Slice Ocean Acoustic Tomography,’’ J. Geophys. Res. 92, No. C11 (15 October 1987)]. Therefore, by exciting particular modes and not exciting other modes, the emitted signal can be tuned to be most sensitive to environmental fluctuations on a particular spatial scale. A major challenge in this approach is the tuning of the source array to ensure that only the desired modes are excited.

Some features of ambient noise in shallow tropical waters

This paper discusses the main features of ambient noise in shallow tropical waters near Australia and the implications for shallow water acoustics. Noise from biological sources is usually the dominant component of the ambient, in the absence of high winds and heavy rain. Noise levels fluctuate by more than 30 dB (commonly by about 20 dB) as a result of biological activity. Diurnally varying biological choruses produce the highest noise levels and are evident in the frequency range 40 Hz to 5 kHz. Snapping shrimp noise, extending to frequencies beyond 200 kHz, is the dominant component of ultrasonic noise where habitats are favorable. Biological transients (signals detectable from individual animals, usually fish and whales) are also important. Traffic noise varies from high levels near ports to negligible in areas remote from ports and where propagation is poor.

Approximate methods for shallow water acoustics with solid sea beds

The reflection properties of representative shallow water solid sea beds can be reproduced by several simple approximations. An ‘‘equivalent fluid’’ with real parameters can be found provided the shear wave speed in the bottom is less than about 600 m/s. An equivalent fluid with complex density is useful if the shear wave speed is less than about 1000 m/s. A perfectly reflecting boundary placed at a ‘‘complex effective depth’’ gives a very simple approximation, again for shear speeds up to 1000 m/s. The acoustic field at long ranges is determined by reflection at low grazing angles and is reproduced well by all three approximations. In all cases algebraic expressions are given for the various parameters. If the complex effective depth is matched to each mode by iteration it provides a rapid means of finding exact normal mode eigenvalues for both the trapped and leaky modes. All normal modes are leaky for a soft solid sea bed and have complex parameters even when there is no attenuation.

Forward scatter and signal coherence in shallow water

In shallow water it is difficult to separate the effects of propagation in a waveguide composed of layered media from those of scattering produced by roughness at the layer interfaces. Moreover, to properly account for these combined effects, three-dimensional, fully elastic propagation and scattering models must be used, if a fundamental understanding of shallow water propagation is to be attained. Very few forward propagation studies include scattering effects on forward propagating signal properties such as amplitude and coherence. In shallow water such effects are critical to the condition of the signal when it reaches a target and when it returns to the monostatic receiver. The importance of these effects on the signal properties is magnified for bistatic receivers. Boundary scatter, including out-of-plane scatter, in shallow water significantly affects time and angle (as well as frequency) spreads that lead to degradation of signal coherence. The requirements to deal with scattering in three dimensions and to include full wave field physics in propagation models introduce the need for extensive computer code and computational power. While there are no simple solutions to the problems these requirements impose, the advance of computational technologies and the development of advanced propagation and scattering codes will allow comprehensive and innovative approachs. These issues are addressed and an approach to solving this significant part of the shallow water acoustics problem is discussed.

Shallow water acoustics research issues

Propagation of sound in shallow water is complicated by the proximity of the boundaries and interferring noise sources and by environmental variability. Moreover, geomorphic and geoacoustic properties of the seabed play a major role in the character of the propagating signal. Propagation prediction is complicated by the need to use full wave field models versus simpler ray models. Roughness and inhomogeneities of the interfaces and seafloor are important to both forward and backscatter. Multiple interactions with the interfaces cause signal frequency, angle, and time spreads, coherence loss, and out-of-plane effects, and they exacerbate shear and interface wave problems. The practical effects of all this are complications in the reverberation and noise fields and in signal coherence. On the other hand, match-field processing in shallow water provides some benefits (essentially, by taking advantage of multiple images of the target reflected in the boundaries), but achieving a positive net gain is critically dependent on our basic understanding of the multiplicity of detrimental effects. These issues are discussed in terms of their relevancy to acquiring this understanding to support ongoing technology and advanced development programs in the Navy.

Oceanographic variability in shallow-water acoustics and the dual role of the sea bottom

Acoustic propagation in shallow water is examined. Multipath propagation and extensive boundary interactions, which along with a host of other phenomena produce a highly variable and often unpredictable acoustic field, are discussed. The responsible mechanisms, and hence the acoustic effects, cover a wide range of temporal and spatial scales and are classified as either deterministic or random, although the two types often act in concert. Because of extensive interactions with the sound field, the bottom can severely degrade waterborne propagation, although the sea bottom (and subbottom) can provide a seismic path that not only is relatively stable, but exists even under environmental conditions that preclude an effective waterborne path. Propagation in the bottom is particularly significant at very low frequencies. These various aspects of shallow-water acoustics are illustrated using the results of experiments conducted in diverse geographic areas. >

The maximal range step in parabolic wave propagation in shallow-water acoustics

To quickly obtain the wave field in parabolic wave propagation, this study sweeps the range in the largest possible range steps. Jensen and Krol [Saclanten Memorandum-72, NATO-SACLANTEN, La Spezia, Italy (1975)] obtained the truncation error of the parabolic equation that depends on the range step but did not derive the appropriate range step. The derivation was completed and it was implemented to a parabolic equation code. This study demonstrates that the truncation error of the parabolic wave equation is inadequate for the purpose of determining the maximal range step under acceptable wave-field error. This led to an analysis based on the truncation error of the wave field. It is shown that given a wave-field gain/loss of γ=10 log10‖η‖dB, the maximal range step Δr allowed is given by a formula that relates it to η (that measures the wave-field truncation error), the gradients of the refraction index n (with respect to range and height) and to the wave field and its gradients. The formula had been tested successfully on the parabolic equation code.