SECT. 4] SOUND IN THE SEA 489 



3. Noise 



At any point of the ocean there are, in addition to sound waves due to the 

 signal under observation, changes of pressure covering a very wide frequency 

 range. This "noise", dealt with at length by Horton (1957), is due in part to 

 movements of water, in part to marine fauna, in part to man-made devices, 

 and in part to vibrations of the ground. The quasi-periodic pressure variations 

 due to tides and to swell can be left out of consideration because their fre- 

 quencies are lower than what we have agreed to call sound. 



Much of the noise in the deep open sea seems to be due to impact of wavelets 

 and collapse of cavities trapped just below the surface when wave crests break 

 to form "white horses". As soon as the temporary reduction of pressure forming 

 the cavity is over, it collapses producing a pressure change of a type such as to 

 give the continuous spectrum proper to "cavitation", of level decreasing by 

 6 dB per octave. 



The number of cavities is very large since the whole surface of the sea 

 contributes. The intensity is nevertheless finite under the reasonable assump- 

 tion that the contribution from any area is proportional to the cosine of the 

 angle between the normal and the line joining it to the point of observation. 

 If z is the depth and a the pressure absorption coefficient, the intensity is 

 proportional to 



At low frequencies a is negligible and this integral is equal to unity, which means 

 that the intensity is independent of depth. That is in fact what is found in 

 practice. Even at high frequencies the decrease is small for moderate depths : 

 at 32 kc/s, for instance, a is approximately 8 dB per kiloyard and at a depth 

 of 300 ft the intensity would be only about 1 dB less than near the surface. This 

 result thus agrees with the statement by Horton (1957) that, for depths down to 

 300 ft, neither the magnitude nor the frequency characteristic of water noise 

 depends on the depth at which the observation is made. "The quality of the 

 noise does, however, vary somewhat with depth. As the hydrophone is lowered 

 the sounds of individual waves, which can be separately identified near the 

 surface, merge into a more nearly continuous sound. At the same time there is 

 a noticeable decrease in short-time variations from the average noise level" 

 (Horton, 1957). 



Another aspect of water noise is more difficult to understand : measurement 

 has established that the spectrum of water noise is pretty well a straight line 

 from 100 c/s up, decreasing with frequency by 5 dB per octave down to the 

 point where it merges with the thermal noise of the equivalent electric resistance 

 of the detector. That it is due to the motion of the surface of the sea is well 

 demonstrated by its dependence on sea state (Fig. 5) ; but experiment shows 

 that cavitation noise from any other source whatever decreases at the rate of 

 6 dB per octave, and this result is consistent with the assumption that the 

 pressure near the time of collapse of a cavity varies exponentially. It is 



