Info

Figure 6.4 Bathymetric and sound-speed structure in the north Pacific Ocean. The noise from distributed shipping sources at high latitudes can enter the sound channel and propagate with little attenuation to lower latitudes. Relationships between the sound speed structure and the prevailing water masses are also illustrated (Kibblewhite et al., 1976).

Figure 6.4 Bathymetric and sound-speed structure in the north Pacific Ocean. The noise from distributed shipping sources at high latitudes can enter the sound channel and propagate with little attenuation to lower latitudes. Relationships between the sound speed structure and the prevailing water masses are also illustrated (Kibblewhite et al., 1976).

a downward-refracting (negative-gradient) sound-speed profile. Sound waves propagate downslope with small grazing angles until they reach the depth of the sound channel axis in deep water where they then detach from the bottom and propagate within the sound channel. Tappert et al. (2002)

refer to this as the "mudslide" effect, which they characterize as one of a few robust and predictable acoustic propagation effects that occur in range-dependent ocean environments.

6.2.3 Bioacoustic noise

Marine bioacoustic signal sources are typically transient in nature and exhibit diverse temporal, spatial and spectral distributions. The main contributors to bioacoustic signals include certain shellfish, fish and marine mammals. Of the marine mammals, whales are the most notable contributors. Au et al. (1987), Cummings and Holliday (1987) and Watkins et al. (1987) described whale signal characteristics and distributions. A paper by Watkins and Schevill (1977) included a phonograph recording of actual whale codas.

Richardson etal. (1995: chapter 7) provided comprehensive summaries of marine mammal sounds in the form of tabulations that included frequency ranges and associated source levels for each species. Specifically included were the sounds produced by baleen whales (bowhead, right, gray, humpback, fin, blue, Bryde's, sei and minke), toothed whales [Physeteridae (sperm whale), Ziphiidae (bottlenose whale), Monodontidae (beluga and narwhal), Delphinidae (killer whale and dolphin), Phocoenidae (porpoise) and river dolphins], phocid (hair) seals, eared seals (sea lions and fur seals), walruses, sea otters and sirenian (manatees and dugongs).

As discussed by Richardson et al. (1995: chapters 8-11) and the National Research Council (2000), understanding of both the sensitivity of marine mammals' hearing and the reactions of marine mammals to various noise sources has advanced through additional fieldwork. This work provides relevant guidance to the design and operation of high-intensity sources, especially in multistatic and tomographic experiments.

Furthermore, this research aids in the development of meaningful acoustical regulations to ensure the health and safety of marine mammals.

6.2.4 Wind and rain noise

Kerman (1988, 1993) and Buckingham and Potter (1996) provided updated summaries of sea-surface sound. The established relationships between surface weather phenomena and noise levels have been used to advantage by oceanographers. Shaw et al. (1978) demonstrated that surface wind speeds derived from measured noise spectra could be used to calculate the wind stress over the oceans. Synoptic wind stress information is required for the dynamic modeling of wind-driven ocean currents. Scrimger et al. (1987) and Lemon and Duddridge (1987) described the development and operation of WOTAN (weather observation through ambient noise) systems. These systems operate over the frequency range 0.5-30 kHz and have been successfully used to infer wind speeds that were highly correlated with buoy-mounted anemometer measurements. Vagle et al. (1990) made further measurements in support of WOTAN.

Zedel et al. (1999) modified an acoustic Doppler current profiler (ADCP) to record ambient sound in the frequency range 1-75 kHz. The resulting instrument package, called OASIS (ocean ambient sound instrument system), inferred wind speeds and directions from these acoustic measurements that were determined to be in good agreement with direct observations made at Ocean Weather Station Mike in the Norwegian Sea.

Felizardo and Melville (1995) concluded that ambient noise correlated well with wind speed (in the Knudsen range) but correlated poorly with significant wave height. The poor correlation with wave height was attributed to the disproportionate effect of swell on the frequency of breaking waves, which are considered the primary source of wind-dependent noise in the ocean.

The noise attributable to rainfall over the oceans has also been used in an inverse fashion to provide estimates of oceanic precipitation (Nystuen, 1986). The underwater noise spectrum generated by rain has a unique spectral shape that is distinguishable from other noise sources by a broad peak at about 15 kHz. Moreover, the relationship between spectral level and rate of rainfall is quantifiable. Scrimger et al. (1987) made measurements of the underwater noise generated by rain in a lake located on Vancouver Island, British Columbia (Figure 6.5). These data illustrate the peak in noise levels at 15 kHz. Pumphrey and Crum (1990) determined that the major cause of

Frequency (kHz)

Figure 6.5 Rain noise spectra observed at moderate wind speeds for rain rates of 0.5 mm h-1 (A),0.8 mm h-1 (+) and 1.3 mm h (O) compared to the Knudsen curve for sea state 1/2 (Scrimger et al., 1987).

Frequency (kHz)

Figure 6.5 Rain noise spectra observed at moderate wind speeds for rain rates of 0.5 mm h-1 (A),0.8 mm h-1 (+) and 1.3 mm h (O) compared to the Knudsen curve for sea state 1/2 (Scrimger et al., 1987).

rain-generated sound is the production of bubbles upon water-drop impact at the surface. These bubbles then oscillate with small amplitude and radiate as a dipole.

The effects of bubbles and rain as noise-generating mechanisms have been investigated further by Buckingham (1991) and Laville et al. (1991), respectively. As previously noted, there are two sources of rainfall sound: raindrop impacts on the water surface and air bubble resonances. Bubble resonances were found to be responsible for the spectral peak near 13-15 kHz, while raindrop impacts were associated with a broadband spectrum having a negative slope (Laville et al., 1991). The underwater sound due to rainfall can be distinguished further according to raindrop diameters (Medwin et al., 1992): small drops (0.8-1.1 mm) radiate primarily from bubble resonances near 15 kHz; mid-size drops (1.1-2.2 mm) radiate only broadband impact sound; and large drops (>2.2 mm) radiate both impact and resonance sounds. Nystuen and Medwin (1995) proposed a new bubble-entrapment mechanism to account for a missing component in the modeling of underwater sound levels produced by raindrops. In related investigations, Rohr and Detsch (1992) attributed the effect of films on suppressing high-frequency ambient noise to the opposition of bubble-producing events by these films. Crum et al. (1992) and Pumphrey (1994) reviewed the nature of precipitation sounds underwater, while Prosperetti and Oguz (1993) reviewed the physics of drop impacts on liquid surfaces. Collectively, these studies have improved upon earlier investigations such as those reported by Heindsmann et al. (1955), Franz (1959) and Bom (1969).

Indirect measurements of rainfall are important since roughly 80 per cent of Earth's precipitation occurs over the oceans and lakes where only about 10 per cent of the weather stations are located (e.g. islands and buoys). This information allows oceanographers and meteorologists to improve their understanding of the oceanic heat and fresh-water budgets (e.g. Etter et al., 1987) that, in turn, can provide measures of both short-term and long-term fluctuations in the global climate.

Nystuen (1994) developed an acoustic rainfall analysis (ARA) algorithm consisting of: detection of rainfall in the presence of other underwater noise sources, classification of rainfall type based on drop-size distribution and acoustic quantification of rainfall rate. The classification of rainfall type (e.g. according to stratiform or convective) will permit meteorologists to infer the vertical distribution of atmospheric latent-heat release in support of global climate studies.

Oguz (1994) developed a theoretical model to predict the ambient noise levels arising from bubble clouds generated by breaking waves and resultant whitecap formation. Cloud geometries were modeled by inverted hemispherical shapes within which solutions to the wave equation could be obtained analytically. Empirical relationships between wind speed and whitecap occurrence permitted calculation of noise levels as a function of frequency and wind speed.

Sound speed (ms 1)

2,000

Was this article helpful?

0 0

Post a comment