One problem encountered in treating the noise data was 

 contamination by ship and traffic noise. Traffic noise, the contri- 

 bution of distant ships, is frequently not recognizable in the noise 

 records. Ship noise, however, displays a pronounced transient 

 effect on a continuous display in both broadband records and third- 

 octave -band records, particularly in the 50-to-63-c/s region. 

 For times during which there were obvious transient effects in the 

 broadband records, all noise data were deleted from the data from 

 station C, which were heavily contaminated. Figure 4 shows some 

 of the characteristic transients of these broadband records. In 

 this example, all of record C and all the third-octave -band data 

 corresponding to the times designated by arrows on records A and 

 B were deleted. 



The method of recording sea-state conditions was to take 

 the hydrophone location as the center for the generation of a series 

 of ranges at equal radial increments as depicted: 



WAVE-HEIGHT CONTOURS 

 SHORE 



RANGE, N. M. 





200 



400 



600 



800 



MAX WAVE HEIGHT, FT 



2 



3 



5 



9 



8 



HYDROPHONE 

 400 N. M. RANGE 



In most instances, the storm configurations were such that it was 

 convenient and practical, as a first approximation, to use linear 

 interpolation. Data were recorded according to wave height in the 

 immediate vicinity of the hydrophone and the maximum values of 

 wave height observed anywhere on arcs at radial ranges of 200, 

 400, 600, and 800 nautical miles from the hydrophone. 



These data were compared with ambient-noise levels re- 

 corded at the center of the system, which were reduced by the 

 third-octave digital analyzer 6 ' 7 to third-octave-band data starting 

 at 10 c/s and ending at 400-c/s midband frequency. Coefficients 

 of correlation between the third-octave -band noise levels and the 



