LONG-RANGE SOUND CHANNEL PROPAGATION 



211 



greatly influenced by scattering, whereas sound scat- 

 tered through any sizable angle would arrive too late 

 to be recorded on oscillograms like those of Figure 7. 



9.2.5 



Variations 



Ideally it shoidd be possible to determine the magni- 

 tude and time scale of the fluctuations and variations 

 in transmission by firing a number of caps in rapid 

 succession from the same place. Unfortunately, no 

 systematic experiments of this sort have been carried 

 out. In the UCDWR work,''' a few repeat shots 

 were made at intervals of a few minutes ; however, the 

 number of such repeat shots was curtailed by the 

 need for obtaining data at different ranges and depths 

 in a time short enough so that oceanographic condi- 

 tions could be assumed constant. Most of the ma- 

 terial in the following paragraphs represents infer- 

 ences obtained when some of the oscillograms for 

 these experiments were restudied in the course of 

 preparing material for the present report; because of 

 the paucity of the data, these inferences must be re- 

 garded with caution. 



In isothermal water, peak pressures from succes- 

 sive shots seem to vary but little out to ranges of 

 over 1,000 yd. Most of the shots studied were con- 

 sistent to within 1 or 2 per cent, though occa- 

 sional shots deviated by 5 or 10 per cent. These 

 variations are of the same order as those which are 

 found at short ranges and attributed to nonuni- 

 formity of the caps themselves. The fact that they 

 are so small is evidence of the uniformity of adjust- 

 ment of the hydrophone and recording system. 



When the cap is in the thermocline and the hydro- 

 phone in an isothermal layer above it, or when there 

 is a negative temperature gradient at all depths, the 

 fluctuations in peak pressure seem to be distinctly 

 greater; for these cases successive shots at a few 

 hundred yards range often differ by 20 per cent or 

 more. In the experiments with single-frequency 24-kc 

 soiuid, which were reported in Section 7.1.1, the 

 fluctuation was found to decrease somewhat if the 

 hydrophone was placed beneath the thermocline. 

 Apart from the fact that neither the evidence on 

 explosive sound nor that on single-frequency sound 

 is based on an adequate number of samples, the ap- 

 parent contradiction may be readily explained by the 

 fact that in 24-kc single-frequency work the surface- 

 reflected signal usually cannot be resolved from the 

 direct signal, while in the experiments reported here, 

 these two signals are generally received one after the 



other. Both in the present case and in the preceding 

 the surface-reflected pulse seems to be a little more 

 variable than the direct pulse, although the evidence 

 for this is not very conclusive because of the irregu- 

 larities, real and instrumental, which are present in 

 the tail of the direct pulse on which the reflection is 

 superposed. 



Beyond a shadow boundary successive shots are 

 often surprisingly consistent. The difference between 

 the first pressure peak and the first trough, for ex- 

 ample, has been observed in several cases of repeat 

 shots to be reproducible to within 20 per cent or so 

 although at least one case of a much larger fluctua- 

 tion has been observed. 



Figure 16 shows some typical oscillograms of shots 

 made a few minutes apart, for three types of trans- 

 mission conditions. One must be cautious in attrib- 

 uting physical reality to all the differences in detail 

 which appear in successive oscillograms of this sort. 

 For example, it has been demonstrated that slight 

 changes in the orientation of a hydrophone from one 

 shot to the next can sometimes produce considerable 

 changes in the recorded pressure-time curve.' Other 

 variable factors mentioned in reference 1 which can 

 have an appreciable effect include scattering of sound 

 by supports and other bodies near the hydrophone, 

 and the possible presence of bubbles or other foreign 

 matter on the hydrophone. Thus differences between 

 successive records may or may not be real. On the 

 other hand, any feature which is consistently repro- 

 duced in all records made under a given set of condi- 

 tions, and for which an instrumental origin can be 

 ruled out by virtue of its nonappearance under most 

 other conditions, is probably a reproducible char- 

 acteristic of the true pressure-time curves under the 

 given conditions. A feature of this type is the shape 

 of the first cycle or so of the oscillatory pressure-time 

 curve in a shadow zone, as shown in Figure 16C. 



9.3 LONG-RANGE SOUND CHANNEL 

 PROPAGATION IN DEEP WATER 



It has long been known that explosive sound from 

 a charge of moderate size can be detected at ranges 

 of the order of tens of miles, and this fact has re- 

 ceived practical application in acoustic position find- 

 jj^g 18,19 ^g ^jji }jg shown later, ranges of thousands 

 of miles can be achieved by proper arrangement of 

 source and receiver. It is to be expected that many 

 of the characteristics of the signals received at long 

 ranges will be determined by refraction and by re- 



