82 



EXPERIMENTAL PROCEDURES 



The upper strip was obtained at a range of 1 10 yd in 

 the direct sound field, while the lower strip is a typical 

 record of sound received at a long range, 1,700 yd, in 

 the shadow zone. This fluctuation of received sound 

 field intensity has become the subject of special in- 

 vestigations, which are summarized in Chapter 7. 

 The principal purpose of transmission runs, however, 

 is to obtain the average transmission properties of the 

 ocean with a given set of oceanographic conditions. 



To obtain a representative average, it is necessary 

 to select a sample of signals, assign to each signal an 

 individual sound field amplitude, and then to strike 

 an average. The final result of these steps, the average 

 sound field intensity or sound field level, will depend 

 not only on the record obtained, but also on the 

 details of the sampling and averaging procedure 

 employed. UCDWR has standardized these pro- 

 cedures to insure intercomparability of results ob- 

 tained at different times and by different research 

 groups. The procedure is described in a report by 

 UCDWR^ and will be briefly recapitulated in the 

 following paragraphs. 



In the selection of a sample several requirements 

 must be satisfied. The sample of individual signals 

 must be large enough so that the standard deviation 

 of the average is not much larger than of the order 

 of 1 db. Moreover, the benefit of averaging will be 

 obtained only if the sample covers a period of time in 

 which the transmission passes through a number of 

 maxima and minima, for otherwise the average would 

 be an average of individual signals most of which 

 may be relatively high or relatively low. On the other 

 hand, the period of time covered by the sample must 

 be short enough so that it corresponds to a negligible 

 change of range between the two ships and a negligible 

 change in the large-scale temperature structure. 



The standard procedure for supersonic work, de- 

 signed to strike a compromise between these require- 

 ments, has been to select five signals, equally spaced 

 during a period of 20 sec. Since the standard devia- 

 tion of an individual signal from average intensity is 

 between 2 and 4 db in most samples, the standard 

 deviation of the arithmetical average of five signals 

 from the average of a very large number of signals is 

 between 1 and 2 db, (1/ V n — 2 times the standard 

 deviation of the individual signals). 



At WHOI, the rule has been to use as a sample ten 

 consecutive signals. Since signals are transmitted 

 about 1.2 sec apart, a sample extends over a period of 

 12 sec. This method, although slightly different from 

 that employed at UCDWR, leads to averaged ampli- 



o 

 a 



03 

 C 

 bC 



P 



