218 



TRANSMISSION OF EXPLOSIVE SOUND IN THE SEA 



ness of the interval between any pulse and its reflec- 

 tion from the surface near the hydrophone. In 

 Figure 21A a similar grouping occurs for the bottom- 

 reflected pulses, which continue after the last sound 

 channel arrival, but the range for this case is so short 

 that the sound channel arrivals have not yet had time 

 to form into well-defined groups. 



In the records obtained when both source and re- 

 ceiver were shallow, each theoretical group of four is 

 of course entirely unresolved and appears as a single 

 pulse. 



When source and hydrophone are both deep the 

 total duration of the sound channel arrivals, that is, 

 the time interval between the first arrival and the 

 last arrival via the sound channel is found to agree 

 nicely with the duration predicted by Figure 18. At 

 the longer ranges the last sound channel arrival is 

 easily spotted by the conspicuous "piling-up" effect 

 which occurs just before it, as shown in Figure 21B. 

 At shorter ranges, however, as in Figure 21 A, the 

 last sound channel arrival can only be identified by 

 its high intensity and the fact that, like all the ar- 

 rivals which do not involve reflection from the sur- 

 face, it is absent from the record of the shallow hydro- 

 phone. For short ranges the number of sound channel 

 arrivals is small because of the short range and the 

 fact that neither source nor receiver is on the axis 

 of the sound channel, and the bottom reflections, 

 which come in both before and after the last sound 

 channel arrival, may mask the abrupt termination of 

 the sound channel arrivals even when a piling up 

 occurs. At ranges beyond about 300 miles, on the 

 other hand, the bottom reflections become so weak 

 that they no longer appear on the records, and the 

 piling up of the sound channel rays is followed by 

 sudden silence. This disappearance of the reflected 

 pulses is of course due to the fact that according to 

 Figure 18 any ray traveling by bottom reflections 

 cannot go more than about 70 kyd between successive 

 reflections; and since an appreciable amount of 

 energy is lost at each reflection, pulses traveling 

 along such rays are much more rapidly attenuated 

 than those which travel in the water alone. 



According to Table 5 at the time of the first series 

 of experiments there was a shallow sound channel 

 with its axis at a depth of about 75 ft. That trans- 

 mission to considerable distances near the surface 

 was possible at this time is shown by a comparison 

 of the velocity of propagation of the first arrival with 

 the velocity for the bottom-reflected rays. The ratio 

 of these velocities is found to agree nicely with the 



ratio of the velocity of sound at the surface to the 

 velocity given by Figure 18 for the particular bottom 

 reflection studied, showing that the first arrival does 

 indeed come by a path lying entirely in the region 

 near the surface. 



The most significant results obtained in these ex- 

 periments have to do with attenuation and with the 

 reflection coefficients of bottom and surface. To study 

 quantitatively the variation of intensity with dis- 

 tance and with number of bottom reflections, some 

 sort of measurements must be carried out on the os- 

 cillograms. The most obvious thing to measure would 

 be the peak pressures or momentums of individual 

 arrivals. However, in many cases the individual ar- 

 rivals of a group were not completely resolved, and in 

 all cases the pressure-time curves may have been 

 distorted by small-angle scattering or off-specular re- 

 flection. For these reasons it was concluded that. the 

 most suitable characteristic of the records from which 

 to estimate attenuations and reflection coefficients is 

 the energy of a poke, or of a group of two pokes, 

 rather than the peak deflection, this energy being as- 

 sumed to be a constant times the integral of the 

 square of the deflection. One may hope that this 

 quantity will represent a suitably weighted average 

 of the spectrum level of the pressure pulse in the 

 water in the region of frequencies covered by the re- 

 cording channel being used. It is of course not strictly 

 true that the "energy" measured in this way on an 

 oscillogram represents this weighted average, since, 

 for example, the phase of the transient disturbance 

 produced by the first arrival at the time of the second 

 arrival will determine whether the second arrival 

 increases or decreases the amplitude. However, we 

 may expect that the desired correspondence will be 

 valid for an average over many pokes. 



Because of the very large ranges covered by the 

 second series of experiments, it was possible to meas- 

 ure the very small attenuations suffered by sound at 

 the comparatively low frequencies to which the re- 

 ceiving channels responded, frequencies at which no 

 other measurements of attenuation have been ob- 

 tained. The results, based on the total energy of all 

 the sound channel arrivals taken together, are sum- 

 marized in Table 6. The data beyond about 200 miles 

 are fairly consistent, as the sample plot given in 

 Figure 22 shows. At shorter ranges the measured 

 energies vary erratically, perhaps because the number 

 of sound channel rays is too small to give a uniform 

 spatial distribution of energy. 



The interpretation which should be given to these 



