III. COMMENTS ON EXPERIMENTAL RESULTS 



3.1 Transmission Loss Curves . Theoretical transmission losses as functions 

 of range and frequency were computed by Code 3185, USNEL on the basis of 

 Pedersen's normal mode theory, Reference (2). A single set of surface channel 

 constants, as listed in Table B-l, was selected for each area from the data of 

 Appendix B . These two sets of constants were used for all theoretical curves 

 of propagation loss in this report. The theoretical predictions are presented 

 as solid lines in Curves 1-65. 



Generally agreement between theory and experiment is good, particularly when 

 the theoretical curve is simple. On Channels 2 and 3, the experimental data 

 agree with the theoretical curves within 5 dbs, in approximately 67% of the 

 cases, the exception being the following: (An index of experimental Curve num- 

 bers is provided in Table 2 .) 



Curves 4-6 

 Curves 9 



Curves 10 



Curves 22 - 24 

 Curves 25 - 27 

 Curves 32 



Experimental transmission loss too low 



Experimental transmission loss too high 



Experimental transmission loss too low 



Inadequate data (theoretical curves below expected noise level) 



Experimental transmission loss too high 



Experimental transmission loss too high 



There is some evidence in the bathythermograph data (Figures B-5, B-6) of the 

 existence of a secondary duct below the thermocline in the Boundary Region. If 

 this is the case, then it might explain the low (compared with theory) transmis- 

 sion loss observed in Curves 4-6 which involve deep source and receivers. 

 There are no obvious reasons for the discrepancies in the other cases . For the 

 reasons noted earlier the data obtained on the shallow receiving hydrophone 

 (Channel 1) are omitted from further discussion. 



A detailed comparison between averaged experimental data and theory is diffi- 

 cult if in the theoretical curve, the transmission loss varies rapidly with range. 

 Accordingly, several curves were analyzed point by point. They are presented 

 as Curves 55-65. In one case (Curve 55A) an expanded range scale was employed 

 to make clearer the behavior of the received signal. The point by point analysis 

 confirms the agreement between the average transmission loss curve and the 

 average theoretical prediction; but in general the signal fluctuations mask any 

 close agreement with the details of the theoretical curves . In a few cases, e.g., 

 Curves 57 and 61, there is, in fact, very good agreement. 



On reciprocity arguments, transmission loss curves between two transducers 

 are independent of which is source and which is receiver. Accordingly, we 

 have compared the transmission loss curves between sources and receivers 



10 



arthur ZD.ltttleJnr. 



