320 



SHALLOW-WATER REVERBERATION 



* Angle not calculated. 



absorption loss in decibels, the entries in Tables 4 

 and 5 will be more or less corrected for absorption 

 losses, and the resultant values of m" will be the best 

 estimates which can be made from the data of refer- 

 ence 6. 



Table 6 shows the mean values of m" determined 

 in this way from the data of Tables 4 and 5. The as- 

 sumed attenuation coefficients at 10 kc, 20 kc, and 

 24 kc m db per kyd were 1.3, 3.2, and 4.0 respectively. 

 In Table 6 the mean values for each bottom type were 



Table 6. Mean values of backward scattering coef- 

 ficient at 10-degree grazing angle. 



Bottom type 



10 log to" 



from 

 Table 4 



10 log m" 



from 

 Table 5 



10 log m" 



from 

 Table 2 



determined by averaging the corrected values of m" 

 for all three frequencies 10, 20, and 24 kc, giving each 

 entry in Tables 4 and 5 equal weight. The justifica- 

 tion for averaging m" for different frequencies has 

 been discussed in Section 15.3.2. 



If data for 10 and 20 kc are not averaged with 24-kc 

 values, a large part of the data of reference 3 has to 



be omitted. Another reason for including the 10- and 

 20-kc data is the following. Examination of the cor- 

 rected values of Tables 4 and 5 shows that the as- 

 sumptions which were made concerning the range to 

 the reverberation peak and the value of the attenua- 

 tion coefficient seem to overcorrect m" at 24 kc ; that 

 is, the corrected values of m" at 24 kc frequently 

 tend to be abnormally large, especially in deep water 

 where the range to the peak is long. For example, in 

 the last two MUD entries in Table 5, the range of the 

 peak was estimated to be about 1,400 yd, so that the 

 two-way transmission anomaly correction was 11 db. 

 This made the corrected values of 10 log m" for those 

 entries equal to — 14 db, which is much too high for a 

 MUD bottom. The reason that this overcorrection 

 occurred is not clear. It might indicate that the at- 

 tenuation coefficient for the sound returned as rever- 

 beration is less than the attenuation coefficient de- 

 termined in transmission runs. However, the data of 

 reference 2, which appear quite reasonable, are based 

 on an assumed attenuation coefficient of 4 db per 

 kyd ; in the analysis of volume and surface reverbera- 

 tion in Chapter 14, median reverberation levels were 

 fitted quite well by assuming 4 db per kyd as the value 

 of the 24-kc attenuation coefficient. Whatever the rea- 

 son, the existence of this apparent overcorrection 

 makes it desirable to include values of m" for 10 and 

 20 kc in the averages based on the data of reference 3, 



