Robert L. Field

Measurements of ambient noise and sperm whale vocalizations in the northern Gulf of Mexico using near bottom hydrophones

The Littoral Acoustic Demonstration Center (LADC) consisting of the University of Southern Mississippi (USM), the University of New Orleans (UNO), and the Naval Research Laboratory at Stennis Space Center (NRL-SSC), with guidance and technical assistance from the Naval Oceanographic Office (NAVOCEANO), was formed to do ambient noise and marine mammal acoustic measurement and analysis. Three Environmental Acoustic Recording System (EARS) buoys, designed and produced by NAVOCEANO, were deployed by LADC in the northern Gulf of Mexico (GoM) in the summer of 2001. These bottom-moored omni-directional hydrophone recording systems were modified by NAVOCEANO to sample almost 12 kHz, so that the vocalizations of sperm whales could be recorded. The Sperm Whale Acoustic Monitoring Program (SWAMP) was conducted during that summer by the Minerals Management Service and its collaborators. The EARS buoys recorded during the entire 36 days of SWAMP from 17 July through 21 August of 2001. The EARS buoy hydrophones, 50m above the bottom, were placed on a downslope line, ending at the largest concentration of sperm whale sightings in the northern GoM, in 600m, 800m, and 1000m water depths. The moorings were instrumented with self-recording environmental sensors to obtain time series data of temperature, conductivity, and pressure at specified depths spanning the water column. Four cruises were made to deploy and recover the buoys and to collect a suite of environmental measurements, including CTD and XBT casts and a chirp sonar survey for bottom properties to support propagation modeling. In between the first and second cruises, Tropical Storm Barry moved through the area and changes in the oceanographic properties were observed. Each EARS buoy recorded a bandwidth of 5859 Hz for 36 days. These data clearly reveals sperm whale vocalizations, passing ships, and seismic airguns. Marine mammal vocalizations and airgun signatures have been isolated and are being analyzed. Spectral levels for ten minute averages of ambient noise on four different days show moderate shipping levels except during the passage of the tropical storm. A plateau in the noise spectrum from 200 to 1000Hz on one day is due to the presence of sperm whales. Spectrograms show sperm whale clicks and creaks and the seismic airgun signal is very clear.

Comparison of double and triple cross correlation for arrival time identification of amplitude‐ and frequency‐modulated acoustic transient signals

The triple cross correlation of three signals is a simultaneous function of two lags. It is an alternative to cross correlations taken two at a time for determining the lags for a given source at three distributed sensors. It should offer improvement in arrival time identification only when the statistics of the signal have significant third moment components. In this study, amplitude‐ and frequency‐modulated snythetic transient signals are propagated over several possible paths to three sensors, and the triple correlation of the received pulses computed, as well as the cross correlations of the same three signals two at a time. The efficacy of these two approaches is compared for a variety of amplitude‐ and frequency‐modulated transient signals and multipath interference conditions

Near‐bottom hydrophone measurements of ambient noise and sperm whale vocalizations in the northern Gulf of Mexico

Three bottom‐moored hydrophones, 50 m above the bottom, were placed on a downslope line, ending at the largest concentration of sperm whale sightings in the northern Gulf of Mexico, in 600 m, 800 m, and 1000 m water depths. These depths were chosen after upslope propagation modeling, using historical databases, showed transmission losses greater than 110 dB at hydrophones near the bottom in water shallower than 600 m for a 500 m deep source at the 1000 m contour. These autonomously recording hydrophones were environmental acoustic recording system (EARS) buoys obtained from the Naval Oceanographic Office. They were capable of recording signals up to 5500 Hz continuously for 36 days and were deployed from July 17 through August 21. During this period a major marine mammal exercise was being conducted at the surface by the Minerals Management Service and the National Marine Fisheries Service, with other government and university scientists, in which temporary acoustic recording devices were attached to the ...

The effects of a dynamic shallow water front on acoustic propagation

This paper investigates the effects of a dynamic shallow water front on narrowband (1 kHz) and broadband (50 Hz bandwidth centered at 1 kHz) acoustic propagation. A generic shallow water front is simulated with the Princeton Ocean Model. The sound velocity field is calculated as a function of range, depth and time. A parabolic equation acoustic model is used to simulate acoustic propagation from the warm to the cold side of the front. When the front is in geostrophic balance, acoustic mode coupling is from the higher order, bottom-interacting modes to lower order modes that become trapped in the upper 50 meters of the water column. In the narrowband case, acoustic propagation across the front is enhanced by 10 to 15 dB in the upper 50 meters due to mode coupling. In the broadband case, the coupling from higher order to lower order modes mitigates the multipath, causing the pulse to emerge from the cold side of the front almost undistorted in the upper 50 meters. However when the front becomes unstable and begins to move, acoustic propagation becomes highly variable and the above effects may be observed only intermittently. The study shows that acoustic propagation across a shallow water front can change significantly over a period of hours and hence the need to consider the dynamics of fronts when trying to draw general conclusions about their effects on acoustic propagation.

Acoustic Propagation in Turbulent Layers

The objective of this work was to determine the extent to which acoustic propagation varies in the vicinity of topographic features where the flow oscillates between laminar and turbulent states. Since these topographic features are ubiquitous in coastal areas, the results will impact ACOMMS performance in these areas. In a recent experiment by Mourn and Nash, oceanographic measurements were made around a small bank off the western continental shelf (Fig. 1). Temperature, salinity, and turbulent dissipation rate measurements were obtained from this experiment and broadband (9-llkHz) acoustic simulations done to determine the impact of the turbulent sound speed field on acoustic propagation. Acoustic simulations show an overall increase in transmission loss of about 10-15 dB within the 10-11 KHz band. This corresponds to the times where intense turbulence occurs. The transmission loss correlates well with the turbulent dissipation rate.

A comparison of normal mode and parabolic equation range‐dependent propagation models as tools for acoustic communication in shallow water

Accurate prediction of the effects of a shallow‐water propagation channel on acoustic signal distortion would aid in the optimal design of shallow‐water communication systems. It has been shown that the correlation between modeled (based on a parabolic equation approach) and measured ocean responses decreases significantly with increasing propagation range because of the mismatch in relative amplitudes and time delays between the modeled and measured bottom‐interacting arrivals. This raises the question whether the discrepancies are related to the particular modeling algorithm or to insufficient information about the propagation channel. A new normal mode range‐dependent algorithm is used to simulate the acoustic response for the source–receiver configuration used in an experiment in the Atlantic Ocean at the Southern end of the Blake Plateau in which a 25–150 Hz linear, frequency‐modulated source signal is received on a 15‐element vertical array. The normalized correlation coefficients between source and...

The effects of dispersion, attenuation, range, and other factors on the duration of transients

In this paper, the results of a theoretical investigation into the effects of various factors including dispersion, attenuation, and range on the duration of transients in shallow water ocean environments are presented. An acoustic source with the time history s(t;ω) = 0 for t 0 is placed in the ocean several kilometers from a receiver. The duration of the transient is the amount of time between the first arrival and the achievement of a time harmonic state at the receiver. Using. a range‐dependent time domain model developed at NORDA, the Klein‐Gordon equation is solved to obtain the received signal P(t;v) due to the Gaussian source S(t;v) = exp[ − (vt)2]. Then, P(t;v) is convolved to obtain the reponse p(t;ω) due to s(t;ω). Qualitatively, one would expect transient duration to increase with dispersion and range and to decrease with attenuation. A presentation of quantitative results to illustrate these relationships and discuss the possibility of the use of transient duration to...

Least‐squares and single‐filter always‐convergent iterative deconvolution of transient signals for correlation processing

Correlation processing for distributed sensors is most accurate for short pulses and those whose autocorrelation is sharply spiked. For longer transient signals, multipath arrivals at each sensor have significant interference with each other, and it is difficult to identify individual arrival times. Deconvolution of the received signal to sharpen the transients is one method to decrease the overlap and increase the accuracy with which travel times can be identified. Deconvolution can also be applied after cross correlation to sharpen the autocorrelation of the transients. Least‐squares deconvolution is the most commonly used approach for acoustic signals. It has the disadvantage of being computer intensive when filters for long transients are needed. An alternative approach, the single‐filter application of the always‐convergent iterative technique, is faster and provides variable control for noise. The two techniques are compared for actual underwater acoustic multipath transient signals. Single filter a...

Spatial coherence computed from a plane‐wave model

Spatial coherence of bottom‐interacting sound is dependent on sediment properties of the bottom and the experimental geometry. The influence of two of these sediment properties, namely, compressional sound‐speed gradient and attenuation on the coherence is investigated and compared with the influence of the experimental geometry. Complex plane‐wave reflection coefficients, generated from a multilayered bottom model for various grazing angles, were used to synthesize frequency spectra from which spatial coherences were calculated. The coherences were computed both by ensemble averaging over a set of lossy fluid bottom and by frequency averaging over a single bottom with properties equal to the average of the ensemble. The two methods are in close agreement for those grazing angles where the cross‐power spectra are sufficiently smooth. It is concluded that the general pattern of the coherence, as a function of frequency and receiver separation, is more sensitive to experimental geometry and attenuation, whi...