ATLANT. DEEP-SEA EXPED. 1910. VOL. i] PHYSICAL OCEANOGRAPHY AND METEOROLOGY 



29 



and at 500 metres at the Scottish station, without any 

 very definite accordance with the vertical temperature- 



gradient. The quantity 



gives, of course, higher 



values still. 



We may in different ways find an expression of 

 the part played by tiie tidal variations in the total varia- 

 tions. If for each lunar liour we subtract the deviation 

 of temperature due to the calculated tidal effects from 

 the deviation (Ar) of the observed (or interpolated) tem- 

 perature from the average temperature we obtain a re- 

 mainder, R. The sum for the whole series of 24 lunar 

 hours of the numerical values of A? and the correspond- 



V I /^ I 



ing sum of R have been found, and the ratio — ^, .' '.- 



calculated. The results are recorded in column 9 of the 

 table above. The figures show that in most cases the resi- 

 dual variations are on an average considerably smaller than 

 the total variations. Especially at 100 and 400 metres at 

 Stat. 115 the total variations are nearly covered by the 

 tidal variations. At the Scottish station the closest coin, 

 cidence between the total and the tidal variations is found 

 in the deep water-stratum at 600 metres, but it is also 

 fairly good at the other depths except at 300 metres. 



V I /^ I 



The course of the vertical variations in the value of -p [ 



- I At] 



between 100 and 400 metres at Stat. 115 corresponds to 

 that between 200 and 500 metres at the Scottish station, 

 in accordance with the above mentioned difference of tem- 

 perature between the two stations. 



The partial temperature-"tides" are illustrated in Fig. 2, 

 where the time (in lunar hours, L. H.) is plotted along 

 the abscissa at the top of the figure, and the deviations 

 from the average temperature for each single depth along 

 the ordinate (scale to the right in the figure). Most of 

 the tidal variations are very marked, especially at 300 

 and 400 metres below the surface, where the vertical 

 temperature-gradient is great. 



The amplitude of the temperature variations must not, 

 of course, be confounded with the amplitude of the un- 

 derlying motions of the water-layers. In a "discontinuity- 

 layer" even a fairly large temperature-variation may 

 correspond to a small displacement only. When we as- 

 sume that the variations in temperature are chiefly caused 

 by vertical or horizontal movements it is important for a 

 study of the kinematical conditions to have the temperature 

 observations converted into values of height or length. A 

 vertical displacement of the water-layers will generally 

 be connected with a horizontal one, and we should not one- 

 sidedly ascribe the variations either to purely vertical or 

 to purely horizontal movements. We may, however, in 

 any case estimate the variations in depth of the isothermal 



surfaces of /„ along the station-vertical. From curves 

 representing the mean vertical distribution of tempera- 

 ture (cf. Fig. 1) we may find approximately the vertical 

 distance corresponding to a variation of OOP C, or 



10 - -V- metres (the reciprocal value of the vertical 

 at 



temperature-gradient). A valuation of this kind has been 

 made for the stations in question and the results are re- 

 corded in column 1 1 of the table on p. 27. By means of the 

 numbers recorded in columns 4-6 of the table we may 

 then find approximately the assumed diurnal and semi- 

 diurnal, as well as the combined, variations in depth of the 

 isothermal surfaces in question. The results are recorded 

 in columns 12 — 14. 



The lower figures on pp. 84* and 85* illustrate the 

 total variations in height — "tidal" and otherwise — at 

 Stat. 115, calculated in the manner just mentioned.') 



24. Diurnal Variations. 



The daily variations in surface temperature are quite 

 considerable at Stats. 115 and Sc. A daily variation may 

 be expected on account of the heating by day and cool- 

 ing by night. The harmonic analysis gives at Stat. 115 a 

 maximum temperature in the daily variation at A^= 19-6 lu- 

 nar hours, /. e. August 13th, 13'' 43™ local mean time. The 

 total amplitude in the daily variation is 0-11° C. It is 

 probably due to the heating during the day and cooling 

 during the night. At the Scottish station the correspond- 

 ing maximum in the daily variation occurs early in the 

 forenoon, at about 10 o'clock, the total amplitude being 

 0-30° C. It may be due to a combination of the daily 

 variations just mentioned and real tidal variations. But 

 beside these daily variations we find half-daily variations 

 with nearly the same amplitudes (0-11 and 0-32° C). A, 

 has practically the same value at both stations, correspond- 

 ing to temperature maxima shortly after noon and after 

 midnight on August 13th — 14th. A closer examination 

 of the observations of temperature and salinity shows that 

 the surface currents have evidently oscillated in a horizontal 

 direction, which was also to be expected, rotary tidal currents 

 being distinctly discernible in the Faeroe-Shetland Channel 

 after elimination of the rest current. The greater amplitude 

 of the semi-diurnal temperature variation at Stat. Sc. com- 

 pared with Stat. 115 is probably explained by greater hori- 

 zontal variations in temperature at the former station than 

 at the latter. If a semi-diurnal oscillation occurs at the 



')Tlie scales of lunar lioiirs at the bottoin of ttie figures are un- 

 fortunately incorrect because of a misinterpretation of some nautical 

 tables casually used. 



