DAUBIN: STATISTICAL ANALYSIS OF SHIP GENERATED NOISE 



data taken near critical depth, about 4400 meters. Axis depth is 

 about 1100 meters, and the bottom is at about 4900 meters. The depth 

 effect is rather strong where the variance is highest, near critical 

 depth and the axis. These observations seem to be consistent with 

 the observations of Marshall and others. 



How do we interpret such data and explain the observed means, 

 variances, and other statistics? One of the assumptions we make — 

 perhaps erroneously — is that a major factor in the generation of 

 noise is surface shipping. 



Figure 6 shows two examples of ship distributions which might be 

 encountered in an ocean area radius of 450 miles. In one, the overall 

 ship density is about 0.1 of a ship per degree square, while in the 

 other it is about 0.5 of a ship per degree square. 



To investigate the relation between shipping and ambient noise, 

 we perform aircraft surveys of surface traffic as part of the field 

 experiment. We use radar or visual methods to obtain position, 

 course, speed of ships; if we are lucky, we get their size and even 

 their identification. We then attempt to relate these observations 

 to the ambient noise data. We have no good measure of the source 

 level of each ship, however. Even if we did know the source levels 

 and the coupling mechanisms to the ocean, and were able to relate 

 them to the noise level at one time, we would have only one point on 

 a ourve which is varying with time. We need the statistics of the 

 noise for a long period of time. Hence, this represents an attempt 

 to solve deterministically a problem that is essentially statistical. 

 Therefore, we will treat the problem in its parts, namely, in the 

 statistics of source level and the statistics of ship distributions. 

 From this we should be able to remove from the observed data those 

 components which result from the ship distribution. 



851 



