18 



BJ0RN HELLAND-HANSEN 



[REP. OF THE "MICHAEL SARS" NORTH 



The following examples may be given. At Stat. 25 B 

 a temperature-observation was made at 800 m., but no 

 determination of salinity. The curves will be found on 

 p. 66*. The density-curve and the temperature-curve are 

 probably correct. The salinity-curve was constructed in 

 correspondence with these curves, in the way shown by 

 the continuous line in the figure, showing two secondary 

 maxima, at about 800 m. and at a little less than 1 200 

 m. The broken curve is drawn as it would have been 

 without regard to the other two curves. The difference 

 between the two salinity-curves is quite significant, the 

 greatest difference corresponding to about 0-08 "/"o- — At 

 Stat. 68 (p. 71-) we have temperature observations but 

 no salinities at 457 and 549 m. The temperature-curve 

 seems to be well defined and does not leave room for 

 any doubt. This station should especially be compared 

 with Stat. 65. A comparison between the two stations, 

 and the fact that in these regions — as in most parts of 

 the ocean — a temperature-curve and a density-curve 

 generally show similar variations, renders it highly 

 probable that the density-curve should have a form 

 such as that shown in the figure. It follows that the 

 salinity-curve for depths between 300 and 600 m. must 

 be drawn in quite a different manner than what it 

 would be if drawn solely from the few observations of 

 salinity, the maximum difference representing about 0-20°/oo. 



All the serial observations from the "Michael Sars" 

 Expedition have been treated in this way by the inter- 

 polation for standard depths, and it is not probable that 

 any considerable errors exist in the values of temperature 

 and salinity published in Table III. The errors are cer- 

 tainly quite minute as far as densities are concerned. This 

 is due to two facts: the densities (without compression) are 

 practically always increasing, or constant, everywhere from 

 the surface downward to the bottom of the sea, even 

 though there may be considerable variations in temperature 

 and salinity; — and the density-curve usually has a com- 

 paratively regular form with only small differences in 

 "constructive" details within the same region of the ocean. 

 This holds good also for the other data in Table 111: 

 stability, pressure, etc., as they depend essentially upon 

 the densities. 



18. Correlation between Temperature and 

 Salinity (Salinity-Anomaly). 



Mention was made above (sect. 7) of the fact that, 

 upon the whole, a certain positive correlation exists between 

 temperature and salinity in the surface waters of the North 

 Atlantic. High salinity is generally connected with a high 

 temperature, and low salinity with a low temperature. 

 The salinity variations are chiefly due to variations in the 



difference between evaporation and direct or indirect 

 precipitation. The variations in temperature are essentially 

 due to the difference between heat absorption and radiation. 

 Even if heat is consumed by evaporation, the correlation 

 mentioned indicates that in the North Atlantic the heat 

 absorption is generally in excess where the evaporation 

 is predominant. There are, of course, many exceptions 

 to the general rule as far as the surface waters are con- 

 cerned. Thus we often find that the diluted surface water 

 near the coasts is comparatively warm in summer, so that 

 much the same temperature may be met with in water 

 of widely different salinity. The seasonal variations are 

 generally greater in temperature than in salinity, so that 

 the correlation will alter with the season as far as the 

 surface is concerned. 



In spite, however, of the manifold variations in the 

 climatological conditions we shall find the above- 

 mentioned correlation rather distinctly shown when we 

 plot out on millimeter-paper the corresponding values of 

 temperature and salinity. This has been done for the 

 surface-observations of the "Michael Sars" Expedition in 

 the North Atlantic proper, with the result shown in the 

 figure on p. 73*. In this figure the following marks have 

 been used: a recumbent cross (X) for the observations 

 between the English Channel and Gibraltar, a ring for 

 those between Spain and the Canary Islands, a triangle 

 for observations along the route from the Canary Islands 

 to the Azores, a standing cross (+) for those between the 

 Azores and Newfoundland, and, finally, a black dot 

 indicates that the observation has been made on the route 

 from Newfoundland to the British Isles. It will be seen 

 that the marks have quite a distinct distribution, which is 

 especially the case within the separate groups for limited 

 locality and time. 



The variations in temperature and concentration of 

 the water are predominantly started from the sea surface. 

 Of especial importance are the vertical convection ''currents" 

 in winter, which may proceed far downwards in homohaline 

 waters. The effect on the temperature will upon the whole 

 be inversely proportional to the depth reached by the 

 vertical convection "current". The conduction of heat 

 from above in summer is very slow. At a depth of some 

 few hundred metres the seasonal variations are, therefore, 

 quite insignificant. In general, they may be ignored at 

 200 metres and deeper (cf. chapter VI). 



The intermediate and the deep waters are very 

 conservative as to their physical properties. Practically 

 speaking they are only altered by mixing processes. Where 

 very different currents meet and mix the alterations 

 may, of course, be considerable, as e. g. in the sea 

 W. of the Gibraltar Strait where deep water from the 

 Mediterranean is poured into the water-masses of the 



