ANDERSON: VERTICAL NOISE DISTRIBUTION 



RESULTS 



Figure 21 presents the final interpretation of noise direction- 

 ality as a function of depth (assuming no significant frequency 

 dependence) . The basic technique was to start with the 100 Hertz 

 deep data and, moving shallower, correlate the noise shoulders 

 with the 36 Hertz data. Still shallower, the 36 and 23 Hertz data 

 were correlated to continue the trends. The out-of-band noise 

 associated with the filter characteristics prevented estimates of 

 the noi^e in the steeper "indeterminate" areas. 



These data show a basic increase in noise level and broadening 

 in aperture corresponding to the broader aperture available to long- 

 range paths. For the two deepest receivers, these curves are some- 

 what tentative since they are pressing the resolution of the system. 



883 



