Clark and Yarnall 



qualitative results have been obtained. Only a summary of the results 

 of the model study will be given. 



The propagation model assumes that a vertical gradient in the 

 sound speed is the single most important factor in determining the geo. 

 metry of the ray paths, and that useful results are obtained by linear 

 approximations to profiles such as those illustrated in Figure lb. 

 The linear profile illustrated in Figure 6 is the starting point for 

 the computations discussed below. 



An alternative theoretical description stresses the role of 

 turbulent processes in the oceans and pictures the medium as an iso- 

 tropic random distribution of thermal "patchiness" . Available in- 

 formation indicates that this description is inappropiate for use in 

 the present application. 7 This point is emphasized because the inter- 

 pretation of some of the more interesting data to be presented is 

 based on phenomena which occur only in a stratified fluid. Attenua- 

 tion due to absorptive processes in the fluid volume has been taken 

 to be negligible." 



Under the assumption of a smooth bottom, reflection loss and 

 phase shift at bottom reflections have been simulated by means of the 

 modified - Rayleigh plane wave reflection coefficient." Details of 

 the use of this reflection coefficient in a similar application have 

 been provided in a previous publication. ^0 Phase shifts of 180° at 

 surface reflections and 90° at RBR ray turning points have also been 

 incorporated.* Scattering effects within the fluid and at irregular 

 boundaries cannot be handled by the model. It is probable that this 

 is the most serious deficiency. There is perhaps sufficient theore- 

 tical and experimental background available to include coefficients 

 of scattering loss for reflections at the boundaries. The work of 

 H.W, Marsh, for example, might be applied to SRBR rays at the 

 surface. This has not yet been attempted. 



The ray model will be discussed for the Fowey Rocks to 

 Bimini (source to H43) path only. The calculations, which include 

 spreading loss for each ray, are based on techniques recently devel- 

 oped by M.J. Jacobson. 12,13 These methods account analytically for 

 all surface-reflected bottom reflected (SRBR) and all refracted-bottom 

 reflected (RBR) rays between fixed source and receiving points 

 (Figs. 7a and 7b) . A tool is thus provided which describes analyti- 

 cally how the medium is "sampled" in space by the acoustic energy 



* Phase shift at RBR turning points was neglected in the previous 

 publication by Jacobson and Clark (Ref. 10). The essential 

 theoretical conclusions of the publication appear to remain 

 unchanged . 



313 



