KUTSCHALE: LOW-FREQUENCY PROPAGATION IN THE ICE-COVERED ARCTIC OCEAN 



A comparison of the FFP computations with those computations by 

 normal-mode theory from the corresponding integral solution are 

 identical, but the FFP technique is far more convenient and often 

 faster, since the computations are done directly from the integral 

 solution without first computing the roots of the dispersion equation 

 and then summing the noirmal modes. 



Figure 18 shows an example of propagation loss computed by 

 the FFP. Loss at several thousand range points is computed simul- 

 taneously giving the fine detail of the fluctuating loss with range 

 from the 10 Hz CW source. When comparing computed loss with loss 

 measured from explosive signals, it is necessary to average intensity 

 with range by taking a running average in range and converting to 

 propagation loss. 



Figure 19 shows the absolute value of the integrand for the 

 computations corresponding to Figure 18. The peaks in the integrand 

 correspond to the normal-mode poles at phase velocities less than the 

 speed of sound in the lower half space. For higher phase velocities, 

 peaks correspond to "leaky modes" computed from the branch -line 

 integral contribution (Kutschale, 1970) . 



Figure 20 shows a comparison between data, the FFP and ray 

 theory. The data curves of Mellen and Marsh (1965) and Buck (1968) 

 are based on measurements made from several hundred shots at various 

 depths over paths in various areas. Nevertheless, the composite 

 curves are in reasonably good agreement with computed curves by FFP 

 and ray theory. The computations by ray theory were made by Chow 

 (personal communication) of Aerophysics Research Corporation, Bellevue, 

 Washington. 



Figure 21 is a sample of comparison of experiment and theory for 

 a constant shot and hydrophone depth. These data were measured in 



709 



