5. SUMMARY OF RESULTS AND CONCLUSIONS 



1 . Nine shallow water areas were examined for the presence of positive sound 

 speed gradients. Ninety percent of the fall-winter sound speed profiles and 3 1 percent of 

 the spring-summer profiles were of the positive gradient type. 



2. In three areas 60-76 percent of the spring-summer profiles were positive 

 gradient. These were the low latitude areas, East of Singapore, the North Coast of Brazil 

 and the Bass Strait, the region between Australia and Tasmania. 



3. Positive gradient (winter) profiles generally resulted in at least 10 dB less loss 

 at the 50 km range than did the non-positive gradient (summer) profiles. 



4. Three of the areas in which propagation losses were calculated, North Sea, 

 Strait of Juan de Fuca and Lands End, had coarse grained sediments resulting in small 

 bottom reflection losses and relatively good propagation by bottom reflected paths. In 

 these cases optimum propagation is in the 500 Hz frequency range. 



5. Optimum frequencies of propagation for positive gradient surface ducts varied 

 from 200 to 1000 Hz. 



6. When a surface duct exists in shallow water it can dominate the propagation 

 and lead to a completely different optimum frequency of propagation. However, to 

 demonstrate this effect, high bottom loss is required. This is illustrated at the East of 

 Singapore site. 



7. The 500 m depth. Straits of Sicily results support the generally accepted con- 

 clusion that the optimum frequency of propagation occurs at a relatively low frequency 

 for the case of downward refraction. However, high bottom loss is dominant in this case. 



8. A change in the magnitude of the compressional wave attenuation in the 

 sediment model produces a proportional change in the bottom reflection loss and can lead 

 to large changes in the calculated propagation loss. This was evident in the Straits of 

 Sicily calculations. 



9. Propagation by bottom reflected paths is a very sensitive function of sediment 

 types. Sediment models used here produced propagation ranging from very good to very 

 poor. Many details of this propagation in the frequency domain were unexpected. 



6. RECOMMENDATIONS 



1 . The prevalence of surface ducts in shallow water needs to be determined in 

 many more areas. 



2. The relative dominance of a surface duct when present in shallow water propa- 

 gation needs further clarification. 



3. Since bottom reflection loss is a large factor in shallow water propagation, the 

 dependence of bottom loss on sediment properties, particularly attenuation, should be 

 determined more precisely. 



4. Include surface scattering in an assessment of surface duct versus bottom 

 reflected paths. 



66 



