Info

Figure 6.6 Northeast Pacific Ocean ambient noise measurements: (a) sound-speed profile and hydrophone depths and (b) measured noise profiles in one-third-octave bands (Morris, 1978).

6.3 Depth dependence

Measurements of the depth dependence of low-frequency ambient noise were made by Morris (1978) in the northeastern Pacific Ocean. Hydrophones were suspended from the research platform FLIP (floating instrument platform). Figure 6.6 shows the sound-speed profile, hydrophone depths and average noise profiles in one-third-octave bands for this experiment. There is a decrease of noise with increasing depth at low frequencies, with a smaller decrease with depth at 500 Hz as wind noise overcomes the dominance of shipping noise. Below the critical depth, the fall-off with depth is steeper as the bottom is approached. This is the result of the loss of refracted sound energy through the effects of bottom interaction.

Urick (1984: chapter 4) demonstrated that ambient noise at frequencies greater than 10 kHz is rapidly attenuated with increasing depth due to the effects of absorption.

6.4 Directionality

As a first-order approximation, the noise field in the ocean might be considered to be isotropic in nature, that is, uniform in all directions, both horizontal and vertical. Measurements have shown that this is not the case.

Axelrod et al. (1965) made measurements of the vertical directionality of ambient noise at frequencies of 112 and 1,414 Hz. Figure 6.7 presents polar plots of the ambient noise intensity per unit solid angle N(0) arriving at a bottomed hydrophone as a function of the vertical angle 0. At 112 Hz, more noise arrives at the hydrophone from the horizontal than from the vertical. This effect diminishes with increasing wind speed. At 1,414 Hz, the opposite is true in that more noise arrives from overhead than horizontally. This effect increases with increasing wind speed.

112 Hz

1,414 Hz

112 Hz

1,414 Hz

Figure 6.7 Distribution of ambient noise in the vertical plane at a bottomed hydrophone at two frequencies (Axelrod et al., 1965).

This directional behavior is consistent with the view that low-frequency noise originates at great distances and arrives at the measurement hydrophone principally via horizontal paths, suffering little attenuation. Alternatively, high-frequency noise originates locally at the sea-surface overhead (Urick, 1984: chapter 5).

The horizontal (or azimuthal) directionality of ambient noise can be highly variable, particularly at low frequencies. Shipping traffic is the dominant source of noise at low frequencies, and the temporal and spatial variations in shipping densities explain much of the observed azimuthal variation.

6.5 Arctic ambient noise

The noise environment under the Arctic ice is different from that of any other ocean area. Shipping noise is extremely low due to the lack of surface traffic. The ice cover itself affects the ambient noise field significantly. It can decouple the water from the effects of the wind and produce ambient noise conditions that are much quieter than a corresponding sea state zero in the open ocean. The ice itself may produce noises as wind, waves and thermal effects act on it (e.g. Milne, 1967). Other sources of noise in the Arctic include seismic and biological activities.

The character of the ice cover is different in areas of shore-fast pack ice, moving pack ice and the marginal ice zone (MIZ). The under-ice noise levels, directionality, spectrum shape and temporal character are very different in each of these regions. Noise originating within the ice stems principally from its state of stress, which gives rise to fracturing. Noise measurements under the pack ice span the frequency range from 3 Hz to more than 1,000 Hz (Dyer, 1984, 1988).

Pack ice is very dynamic and its characteristics are highly variable in space and time (Makris and Dyer, 1986). Nevertheless, the lack of wind-wave interaction and the absence of local shipping can lead to noise levels 10 dB lower than those encountered in the open ocean.

Noise levels in the MIZ are typically higher than those in either the pack ice regions or the open ocean regions (Diachok and Winokur, 1974; Diachok, 1980). Figure 6.8 illustrates the variations in median ambient-noise levels that occur across a compact ice-water MIZ area. The relative magnitudes of the noise levels generated at the ice-water boundary are a function of the rate of change of ice concentration with distance. Thus, the relative maximum noise level measured at a diffuse ice-water boundary would be smaller than that measured at a compact ice edge. Ambient noise levels in the MIZ also depend on such variables as sea state, water depth and dominant ocean-wave period. The last variable is hypothesized as being related to the efficiency of coupling ocean-wave energy into the ice. Makris and Dyer (1991) demonstrated that surface gravity wave forcing was the primary correlate of ice-edge noise in the MIZ.

Pritchard (1990) developed an ambient noise model to simulate the time history of under-ice noise generated by dynamical ice movement, stress

120 80 40 0 40 80 120

Distance (km)

Figure 6.8 Variation of median-ambient noise sound-pressure spectrum levels with distance from a compact ice edge for frequencies of 100, 315 and 1,000 Hz in sea state 2 (Diachok, 1980).

120 80 40 0 40 80 120

Distance (km)

Figure 6.8 Variation of median-ambient noise sound-pressure spectrum levels with distance from a compact ice edge for frequencies of 100, 315 and 1,000 Hz in sea state 2 (Diachok, 1980).

and deformation. The model included noise contributions from local and distant sources as well as a transmission loss (TL) module appropriate for the Arctic environment. Time series data were simulated for a frequency of 31.5 Hz.

Lewis and Denner (1988) reported observations of higher-frequency (1,000 Hz) ambient noise data in the Arctic Ocean. These noises were attributed to the thermal fracturing of ice. Related aspects of ice-generated noise have been reviewed by Sagen et al. (1990).

6.6 Acoustic daylight

Ambient noise in the ocean can be used to form pictorial images of underwater objects. An analogy can be drawn between the natural optical (daylight) field in the atmosphere and the radiating ambient noise field in the ocean: both fields consist of random, incoherent radiation propagating in all directions. Thus, the concept of imaging with ambient noise has been referred to as "acoustic daylight" (Buckingham et al, 1992). A reasonable operating frequency range for the acoustic daylight system is 5-50 kHz; the lower limit is determined by angular resolution considerations while the upper limit is determined by the onset of thermal noise. A general introduction to imaging underwater objects with ambient noise was presented by Buckingham et al. (1996b).

Buckingham (1993) analyzed the incoherent acoustic imaging of a spherical target in order to quantify the contrast ratio (acoustic contrast) under various degrees of anisotropy. In the case of isotropic noise, the contrast (acoustic visibility) was found to have a maximum value of about 4 dB. This result was consistent with field measurements when the angle of view of the acoustic lens matched the angle subtended by the target at the phase center of the measurement array.

An acoustic daylight ocean noise imaging system (ADONIS) was constructed. The system was designed to operate over the frequency range 15-75 kHz with a corresponding minimum wavelength of 20 mm, formed 126 individual beams of nominal width of 0.76° (at the upper frequency), and operated over the spatial range 10-200 m. Simulations developed to predict the performance of ADONIS (Potter, 1994) demonstrated that near perfectly reflecting objects were unlikely to be imaged in volume isotropic noise, except perhaps in the near field. However, the ocean was expected to exhibit considerable anisotropy, if only in the vertical direction, thus improving performance over the conjectured isotropic baseline. Buckingham et al. (1996a) described the results of an experiment with ADONIS, which in practice operated in the frequency range 8-80 kHz and relied on ambient noise to provide the acoustic contrast between look angles on-target and off-target. Epifanio et al. (1999) described results from the ORB experiments, which were conducted with targets at ranges between 20 and 40 m using ADONIS' 126 receive-only beams spanning the vertical and horizontal. Makris et al. (1994) conducted a careful analysis of this noise-imaging concept and concluded that it pressed the limits of current technology. Furthermore, they traced similar approaches back to 1985 when the possibility of detecting submarines solely by their noise absorbing and scattering properties (acoustic contrast versus acoustic glow) had been investigated by S. Flatté and W. Munk.

Potter and Chitre (1999) extended the concept of acoustic daylight (which uses the mean intensity of backscattered ambient-noise energy to produce images of submerged objects, and is thus analogous to vision) by exploring the information contained in higher moments. Specifically, information embodied in the second temporal and spatial moments of intensity, for which there are no visual analogs like acoustic daylight, was referred to as ambient-noise imaging (ANI), a broader imaging approach.

6.7 Geoacoustic inversion

In shallow water, the spatial structure of the ambient-noise field is strongly influenced by multiple interactions with the sea floor. Consequently, both the vertical directionality and coherence of the shallow-water noise field are determined primarily by the geoacoustic properties of the seabed rather than by any temporal variations in source distributions. Several investigators have deduced geoacoustic parameters through inversion of the ambient noise field in range- and azimuth-independent shallow-water environments in which the noise sources were uniformly distributed. For example, Buckingham and Jones (1987) determined critical angles, Carbone et al. (1998) determined compressional and shear wave speeds and Aredov and Furduev (1994) determined reflection losses.

For non-uniformly distributed noise sources in range-dependent environments, sophisticated vertical hydrophone arrays are required to resolve the arrival structure of the noise field. Furthermore, the experimental data must be analyzed using detailed noise models. For simpler geoacoustic parameters such as reflection loss, however, Harrison and Simons (2001) argued that detailed models are not required in the inversion analysis, although a densely populated vertical array is required to resolve the arrival structure. Harrison and Simons (2002) report additional experimental results.

Was this article helpful?

0 0

Post a comment