applets.

Sound in the Sea

← All apps

About Sound in the Sea

Sound is the only practical way to sense far under water: light and radio fade within metres in sea water, short of very low radio frequencies and blue-green light, which reach tens to a few hundred metres at best. But it does not go in straight lines. Its speed depends on the temperature, the saltiness and the pressure of the water, and wherever the speed changes with depth the sound bends towards the slower water. That is why a sonar's range can halve between morning and afternoon, why a submarine hides "below the layer", and why a ship can hear another 30 miles away but not 15.

PROFILE is a bathythermograph slide: temperature against depth. Drag the points (or pick one and use the sliders); click or tap the slide to add a point. The salinity strip works the same way. The page works out the speed of sound at every depth and inks it on the graph paper, and reads off what a sonar operator looks for: the sonic layer depth, the deep sound channel axis, the critical depth and the depth excess.

RAYS fires a fan of rays from the source and follows each one through the water, reflecting off the surface and the bottom. Green rays have only been bent, blue ones have bounced off the surface, red ones off the bottom. The blue wash is the transmission loss (TL): the darker, the louder. Drag the submarine (or use its sliders) to see how much sound reaches it, and by which paths.

SONAR EQUATION adds up what decides detection. Passive: the target's noise (SL), less the transmission loss (TL), against the background noise (NL) reduced by the array's directivity (DI), with the detection threshold (DT) on top. Active: the sonar's own pulse makes the journey both ways (2TL) and the target sends back an echo set by its target strength (TS). The figure of merit (FOM) is the loss the sonar can afford; the chart shows where along the range, at the submarine's depth, the loss is less than that. The verdict and the list of detected stretches use the same test, one chart column (a 300th of the range) at a time, so they always agree.

THE STORY

After the Titanic, Reginald Fessenden built an electrically driven underwater sound source; in 1914 it picked up an echo from an iceberg about two miles away. During the First World War Paul Langevin in France used quartz crystals to send and receive short, high-pitched pulses, and British, French and American scientists worked on finding submarines by their echoes. The British system was called ASDIC (where the name came from is disputed); the American word sonar came into use during the Second World War.

Before that war American destroyers found that their echo ranges were shorter in the afternoon than in the morning. Scientists from the Woods Hole Oceanographic Institution traced this to the sun warming the surface water, which bends the sound down. To see the temperature layers, Athelstan Spilhaus built the bathythermograph in the late 1930s: a torpedo-shaped instrument lowered from a moving ship that scratched temperature against depth onto a smoked glass slide. During the war it became standard, and submarine commanders learned to hide beneath the layer.

In 1944 Maurice Ewing and J. Lamar Worzel showed that a small charge set off at the depth of slowest sound could be heard hundreds of miles away: the sound is trapped in a channel and never touches the surface or the bottom. They called it SOFAR; Leonid Brekhovskikh found the same channel independently in the Soviet Union. From the 1950s the United States laid lines of hydrophones on the sea floor that listened along this channel for submarines, the Sound Surveillance System, SOSUS, made public in the 1990s.

The speed of sound formula here is K. V. Mackenzie's nine-term equation of 1981. It gives 1550.744 m/s at 25 °C, salinity 35 and 1000 m, as this page does. The absorption is W. H. Thorp's formula of 1967. The sonar equations were set out in this form during the Second World War and became the standard way of working with Robert Urick's book Principles of Underwater Sound (1967).

HOW IT WORKS

Sound speed. Mackenzie's equation adds about 4.6 m/s for every degree warmer (less in warm water), about 1.3 m/s for every unit more salt, and about 1.6 m/s for every 100 m deeper. Near the top the temperature usually wins and the speed falls with depth; deep down the water hardly cools any more and the pressure wins, so the speed rises again. The slowest point in between is the axis of the deep sound channel, around 1000 m in the mid-latitudes and at the surface in polar water.

Rays. Between two points the speed is taken to change in a straight line with depth. In such a layer Snell's law, cos(angle) ÷ speed stays the same all along a ray, makes every ray an arc of a circle of radius speed ÷ (gradient × cos angle), so the page follows each ray exactly from layer to layer instead of in small steps. A ray turns back where it meets water as fast as (speed at the source) ÷ cos(launch angle): that is why rays dive under the layer and come back up at the convergence zone, once the bottom is deeper than the critical depth (where the deep water is as fast as the water at the top of the layer). The convergence zone range is not put in by hand: it comes out of the profile, about 30 nm for the winter profile here.

Transmission loss is worked out from the spacing of neighbouring rays (ray tubes): where rays spread apart the sound is weaker, where they crowd together it is stronger, which gives 20 log(range) in plain water. On top come absorption (Thorp, set by the frequency on the form; it is about 0.07 dB per kiloyard at 1 kHz and 1.2 at 10 kHz) and the loss at each bounce off the bottom (Rayleigh's formula for a fluid sediment: sand and gravel are faster than water, so shallow rays are reflected almost completely; clay is not). Different paths add as powers.

The noise comes from the wind on the sea (Knudsen's curves, by sea state, falling 5 dB per octave) plus the thermal noise of the water molecules that takes over at high frequency.

WHAT IT LEAVES OUT

This is ray theory, which is good when the water is many wavelengths deep and poor when it is not. It knows nothing about wavelength: a real surface duct leaks low frequencies (below a few hundred hertz for a layer about 50 m deep), and real sound leaks into shadow zones by diffraction and scattering, so the page's empty shadows are really "much quieter", not silent. Where rays cross in a fold (a caustic) the true field is finite and the ray tubes are only tamed by the grid cells. Paths are added without interference, so there is no Lloyd's mirror pattern near the surface. The ocean is the same everywhere along the range with a flat bottom, and the earth is flat (over 100 nm this is a small effect). The surface is a perfect mirror; a rough sea or ice scatters, more at higher frequencies. The bottom is a fluid with no shear and no layers. Absorption ignores temperature and acidity, the noise ignores shipping (which usually sets it below about 500 Hz) and rain, and the active sonar is noise-limited only, with no reverberation. The profiles and every sonar figure are representative, made up for teaching: not for planning a real search.

FURTHER WATCHING AND READING

Ocean
Depth shown
Drag a point. Click the slide to add one.Drag a point. Tap the slide to add one.