52 



OBSERVATIONS AND RESULTS IN PHYSICAL OCEANOGRAPHY 



Table 1. Data used for graphs to determine correction 



factors to basic velocity at any salinity and 



temperature at atmospheric pressure 



correction factors applicable to the various depths was 

 from this point on a more or less obvious one of taking 

 means. A specimen set of computations of correction 

 factors is reproduced in table 3. 



Criticism may be made of the arbitrary selection of 

 the basic velocity of 1450 meters per second for the 

 compilation of calibration tables. Consideration was 

 given to the selection of some velocity which would have 

 some significance other than merely being the base for 

 a set of tables. For instance, such as the velocity of 

 sound in water of 35.00 per mille salinity, 0° C temper- 

 ature, and atmospheric pressure. So far as could be 

 learned, however, practice had not crystallized to the 

 point of selecting such a velocity which could be consid- 

 ered as standard, and as any velocity might be used 

 equally as well as any other velocity, it was considered 

 best for the purpose to select a figure which was approx- 

 imately a round number, was somewhere near the true 

 velocity, and would give corrections which would be ad- 

 ditive in nearly all cases. It was for these reasons that 

 1450 meters per second was the velocity selected. 



An estimate of the accuracy of each sounding was 

 made and recorded at the time of the measurement. The 

 method of arriving at these estimates may be of interest. 

 The rotary converter and its controlling tuning fork were 

 of 60-cycle frequency. As long as synchronism was 

 maintained, the two had to maintain a phase relation 

 which was constant within a quarter-cycle, and it is 



probable that the successful synchronizing range was 

 about one-eighth of a cycle. This meant that relative to 

 the tuning fork, the rotating parts were varying in phase 

 by a maximum of 1/480 second or about 3 meters in 

 distance. As the distance traveled was twice the depth, 

 the uncertainty in depth due to this cause was about 1.5 

 meters. As there was no temperature control or com- 

 pensation on the tuning fork, and as there was about 

 10° C range on either side of the mean, and as the tuning- 

 fork rate had a temperature coefficient of about 0.007 

 per cent per degree centigrade, it was considered that 

 the time intervals indicated were subject to an error of 

 0.1 per cent. Further, there was an uncertainty of the 

 dial setting within which the outgoing and returning sig- 

 nals sounded as one to the operator. This uncertainty 

 was converted into depth and if greater than 0.1 per cent 

 and greater than 1.5 meters, it was recorded as the un- 

 certainty of the measurement. If it was less than 0.1 

 per cent but greater than 1.5 meters, 0.1 per cent of the 

 sounding was recorded as the uncertainty. And if the 

 distance 1.5 meters was greater than both the uncertain- 

 ty of setting and 0.1 per cent, then it was recorded as 

 the uncertainty of measurement. It was thought that 

 this was a reasonable procedure of estimating the accu- 

 racy of soundings. This is assuming, however, that the 

 frequency of the tuning fork was accurately adjusted to 

 60 cycles per second, that the sounding velocity used 

 was accurate, that the sounding distance was vertical, 

 and that no gross errors were involved. The conditions 

 of temperature and salinity at nearby oceanographic 

 stations are on record, and if in the future it is found 

 that the velocities used were inaccurate, corrections 

 may be made. Very often echoes would be reflected 

 from more than one surface. In such cases the first 

 echo to return was selected as being from that surface 

 which was most nearly vertically beneath the ship. Be- 

 cause of the comparatively gentle slopes of the ocean 

 bottom, such a procedure is probably not greatly in er- 

 ror in soundings at sea, although it is recc^nized that in 

 steep gradients, such as are encountered in certain ap- 

 proaches to land, the error may be considerable. Gross 

 errors are possible when the returned echoes are 

 matched with second or third succeeding signals instead 

 of with the immediately succeeding signal, thus giving 

 one-half, or one-third, the actual depth. Such errors 

 are easily avoidable by sending single signals in order 

 to determine the order of magnitude of the depth. As 

 the single signal was usually used to determine the num- 

 ber of reflecting surfaces and the number of echoes, 

 there was little possibility of gross errors entering the 

 Carnegie results from this cause. Actually, the fre- 

 quency of the tuning fork was not accurately adjusted to 

 60 cycles par second and corrections, which will be 

 dealt with below, have been applied to the soundings taken 

 with the sonic depth finder. 



A program of sounding every four hours was attempt- 

 ed. During such times as the ship was becalmed or 

 making little headway, soundings were taken about every 

 ten miles. This program was, in general, followed but 

 in areas of rapidly changing depth more frequent sound- 

 ings were made. Other deviations from this schedule 

 sometimes occurred to avoid interference with pilot- 

 balloon ascensions, or radio schedules, and occasionally 

 because of the press of other work. Short interruptions 

 to the sounding program were sometimes caused by the 

 necessity of making repairs to the depth finder or to the 

 gasoline engine which drove the main generator. The 



