SONIC TRANSMISSION 



155 



harmonics present because of overloading, while 

 UCDWR used a noise source of the type employed 

 for acoustic minesweeping. The receiving equipment 

 consisted of hydrophones, whose output was ampli- 

 fied and sometimes put through band filters to be 

 recorded by means of power level recorders. Trans- 

 mission was always continuous throughout the run. 



6.3.1 Long Island Area Survey 



Long Island Sound is mostly shallow, less than 

 15 fathoms deep, and the bottom is predominantly 

 sandy, although some runs were made over MUD, 

 SAND-AND-MUD, and STONY bottoms. All runs 

 were made with a single-frequency source. Fre- 

 quencies used were 0.6, 2, 8, and 20 kc. Geographi- 

 cally, the survey was divided into three areas; the 

 Fisher's Island area, the New York Harbor ap- 

 proaches, and Block Island Sound. In all three 

 areas, hard bottoms were predominant. Depths 

 varied from about 50 ft to 200 ft. During the New 

 York Harbor and Fisher's Island area surveys, re- 

 fraction was mostly upward, owing in part to salinity 

 gradients. Off Block Island, some negative gradients 

 were found. Sea states were low with the exception 

 of the New York Harbor runs where sea states up to 

 4 were encountered. Table 6 summarizes the results 

 obtained. To obtain this table, the investigators at- 

 tempted to fit each run by a formula of the form 



H = n-10 logr -f- ar 



(4) 



in which n is the power of spreading and a represents 

 the attenuation in decibels per kiloyard. Since it is 

 difficult to determine both n and a simultaneously by 

 a best fit calculation, n was chosen arbitrarily to as- 

 sume the values 1, 1.5, and 2. The best value of a was 

 then determined by inspection for each of the three 

 assumed values of n. 



The three fits for n = 1, 1.5, and 2 were classified 

 in order of decreasing preference as I, II, III. In addi- 

 tion, the individual fits were graded on an absolute 

 standard as "good" (g), "fair" (/), "poor" (p). Table 7 

 is a summary of the goodness of the fit obtained for 

 these runs. Despite the equal standard deviation 

 values of Table 6, the value 1 for n seems to be most 

 frequently the best fit to the data, although 1.5 is 

 probably the best average, especially at the higher 

 frequencies. 



This survey resulted in the following general con- 

 clusions. Higher frequencies were attenuated more 

 than the lower frequencies. At high frequencies the 

 transmission loss increased with increased disturbance 



e 



13 





c 



"=5 .-I 



CD 



