FITZGERALD: CONVERGENCE ZONE DEPENDENCE ON FREQUENCY 



high-frequency convergence zone, the low-frequency peaks have moved 

 sufficiently far apart to overlay the peaks of other convergence zones. 



These calculations will be compared with some experimental data 

 in the next three Figures. Figure 5 shows the location of an experi- 

 ment in which two cw sources were towed in a northerly direction away 

 from a hydrophone located near the West Indies. Source frequencies 

 and depths correspond to those of the preceding calculation. The 

 receiver was located on the SOFAR axis. The sound-velocity profiles 

 measured at ranges less than 1085 km from the receiver are super- 

 imposed in the leftmost box in Figure 6. The average of these pro- 

 files corresponds roughly to the bilinear profile of the calculation. 



Figure 7 shows the experimental transmission-loss data. Curves 

 (a) and (c) are the transmission- loss curves for the 13.'89-Hz and 

 111.1-Hz signals, respectively. Curves (b) and (d) are 7-km running 

 averages of intensity for curves (a) and (c) . In curve (b) we see 

 a double peak in the low- frequency convergence zone which opens with 

 increasing range until the lefthand peak of one zone has moved over 

 to the righthand peak of the previous zone. Figure 8 shows the 

 detail of Figure 7 for ranges less than 1000 km. 



Profile 1 in Figure 9 was used by Gordon to calculate the 

 modal parameters of the next Figure. The profile continues into the 

 bottom depths as a steep negative gradient. As a result, higher 

 modes (BRSR) are very leaky and do not influence the field at long 

 ranges. 



Figure 10 is a plot by Gordon of interference wavelength, 

 X , , versus average phase velocity, X ^ , for frequencies 

 of 31 Hz and 200 Hz. The plotted points for the two frequencies 

 follow each other closely except in the region of the higher SOFAR 



674 



