Clark and Yarnall 



ray arrivals are also negligibly affected in amplitude, and suffer 

 progressively smaller phase changes. A consistent trend was esta- 

 blished in this calculation, and it may be stated that the cw resultant 

 of any number of RBR ray arrivals suffers a phase change, due to tidal 

 depth variation, of less than 260°. 



Two additional simulations will be of interest. The first, 

 given the descriptive designation, "sliding the profile", repeats with 

 the propagation model the uniform sinusoidal change in the medium 

 considered in the ideally simple experiment discussed previously. In 

 the earlier discussion acoustic phase provided a linear measure of the 

 change in the medium. Referring to Figure 6, we can see that a verti- 

 cally uniform change in sound speed (temperature) in the acoustic path 

 does not change the value of the sound speed gradient, g, but does 

 slide the position of the profile along the sound speed axis. This 

 simulates an external mode of variation in the medium, in which the 

 stratification is undisturbed. 



The second simulation of interest " swing ^' rather than slides 

 the profile. This is accomplished by fixing the surface sound speed 

 (temperature) and varying the sound speed at the bottom. Since the 

 sound speed gradient, g = (Cq - Cs)/H (where C^ = bottom sound speed, 

 Cg = surface sound speed, H = water depth), the gradient varies 

 linearly with the change in sound speed at the bottom. The swinging 

 profile is as close as the propagation model can come to simulating 

 an internal mode of variation in the medium in which the stratifica- 

 tion is disturbed. 



Returning to the sliding profile, the calculation that has 

 been performed slides the profile sinusoidally over a peak to peak 

 change in sound speed of 4m/sec, approximately equivalent to a 1.0°C 

 peak to peak change in the temperature of the entire propagation path. 

 The variation in cw resultant phase and amplitude resulting from this 

 change in" the medium" is illustrated in Figure 8. Note that the peak 

 to peak variation in resultant phase is about 60 cycles. 



The following conclusions have been derived from the sliding 

 profile model calculations: 



1) Within the accuracy of the calculation, acoustic 

 phase provides a linear measure of the sound speed 

 (temperature) variation in the medium. This is 

 identical to the result obtained in the "ideal" 

 experiment. 



2) There is negligible change in the amplitudes of 

 the individual ray arrivals over the full varia- 

 tion in the medium. This implies that the cal- 

 culated variation in the resultant amplitude 

 (Fig. 8) can be interpreted strictly as a multi- 

 path interference pattern displayed in time. 



316 



