CONCLUSIONS 



Many of the forms of weather data now available (the data 

 from Fleet Numerical Weather Facility are considered the best 

 in the sense of the overall patterns they describe) do not lend 

 themselves well to this type of study. To use a high coefficient 

 of correlation in a derivation of a regression line prediction 

 model, one must be assured that the accuracy of measurement of 

 the independent variable is significantly greater than that of the 

 dependent variable (ambient noise). Such was not the case with 

 data used in this study. However, even though precise prediction 

 relations cannot be formulated, an indication is provided of the 

 frequency regions in which the noise is affected by wind and wave 

 height at deep-mounted hydrophones in open-ocean locations. 



The results do not support the hypothesis of very-long-range 

 noise/storm dependence. They do suggest that both wind speed 

 and local wave activity have a significant effect on underwater 

 ambient noise greater than about 160 c/s, and that local wave 

 activity may have a small effect on low-frequency noise. By using 

 a 30-hour prediction chart, available at Fleet Numerical Weather 

 Facility, one may be able to predict whether the noise will in- 

 crease or decrease in regions greater than 160 c/s and to provide 

 a gross estimate of the amount of change. The general dip in the 

 correlation coefficient to insignificant values in the midfrequency 

 range (32 to 160 c/s) for most of the stations is probably due to 

 masking by ship and traffic noise. 



23 



