Clark and Yarnall 



Using the RBR model with "sliding profile" it is now possible 

 to perform an interesting simulation of actual data taken at H43 . The 

 data to be simulated are cw signal phase and amplitude over a span of 

 12 hours, taken from an interval of exceptionally stable transmission 

 during 27-28 November, 1966. The simulation is accomplished by sliding 

 the profile, as a function of time, sinusoidally over a small selected 

 range of sound speeds not far removed from those specified for the 

 propagation model in Figure 6. 



Specifically we choose: 



Cg = 1540 + 0.34 sin (2it t) 



T 



C = 1482 + 0.34 sin (2rt t) 



T 

 The acoustic data and the results of the simulation are shown in 

 Figure 9. The period, T, of the simulated phase change is about 

 26 hours. 



This simulation is surprisingly consistent over the full 

 12 hours illustrated, but it should by no means be mistaken for a 

 direct comparison of theory and experiment. It is, rather, a strong 

 suggestion that the physical concepts used in constructing the pro- 

 pagation model are of valid application; but, at this stage, a one- 

 sided story has been presented. The data of Figure 9 have been selec- 

 ted from an interval of exceptionally stable transmission. An object- 

 ive approach demands representation of the opposite situation, i.e . , 

 data from an interval of unstable transmission. This will be provided, 

 but it is instructive to first present preliminary results from the 

 model study with swinging profile. Calculations of this kind promise 

 eventually to be considerably more informative than the sliding pro- 

 file case. Results available at this time are limited to a detailed 

 study of a multipath effect, that was not demonstrated by the sliding 

 profile calculations. 



This effect is illustrated in Figure 10a. These results were 

 obtained by swinging the profile of the propagation model (Fig. 6) 

 linearly through the values of the sound speed gradient designated on 

 the horizontal axis. Again, we are dealing with the resultant phase 

 and amplitude of 10 RBR cw ray arrivals. The amplitudes of the 

 individual arrivals change negligibly throughout the calculation, 

 identifying the events illustrated in Figure 10a as deep interfer- 

 ence nulls accompanied by very nearly discontinuous "jumps" in phase. 

 This phenomena can be understood by recourse to a very simple phasor 

 diagram (Fig. 10b) . We picture the received multipath signal as being 

 resolved into two phasors, approximately equal in amplitude, which are 

 rotating either in opposite sense or at different angular rates. Since 

 superposition applies, the existence of an interference null is a 

 sufficient condition to justify this interpretation. If we fix our- 

 selves in the reference frame in which one of the phasors is station- 

 ary, the stage is set for the sequence illustrated in Figure 10b. 



317 



