92 



BJ0RN HELLAND-HANSEN 



[rep. of the "MICHAEL SARS" NORTH 



waters streaming from lower latitudes return as an under- 

 flow after having been cooled at higher latitudes, preferably 

 in the cold season. 



A difference of 1 ° C. in the North Atlantic corre- 

 sponds, on an average, to a difference of 006" oo at 5^ 

 0-107oo at 10° and 0-147oo at 15° C. A variation in 

 temperature from 10° to 15° C. corresponds to a variation 

 in salinity of from 35-40 to SS-QZ^/oo, or a difference of 

 less than 0-67oo. From the standpoint of biology the 

 differences of temperature are evidently much more im- 

 portant than the differences of salinity in the open ocean 

 where diluted coastal (or Polar) water does not occur. 

 The vertical variations in salinity are generally quite small. 

 In the central parts of the North Atlantic, where the salinity 

 at the surface may be as high as 377oo while it is 357oo 

 at about 2000 metres, the mean vertical variation per 100 

 metres is only — O-P/oo. A variation of — 0-57oo per 100 

 metres would be e.xtraordinary. In coastal and Arctic waters 

 with increasing salinities downwards, the variations may 

 be very considerable, as is well known. At Stat. 70 A, 

 for instance, the difference in salinity between 37 and 

 46 metres was as much as 0-547oo. 



On account of the agreement between temperatures 

 and salinities and the law of parallel solenoids, some of 

 the chief results mentioned in section 36 may be directly 

 transferred and applied to the horizontal distribution of 

 salinity. In particular it may be emphasized that great 

 horizontal variations of salinity usually indicate a strong 

 current (or great vertical differences of current velocity), 

 and that salinity charts for different levels resemble each 

 other when only the course of the isohalines is considered, 

 and not the absolute values of salinity represented by 

 them. The temperature charts on p. 96* may be almost 

 directly 'translated' into salinity charts. Figs. 28 and 29 

 show the horizontal distribution of both temperature and 

 salinity at 400 and 1000 metres' depth below the surface 

 in the eastern North Atlantic, where the comparatively 

 large number of observations have made it possible to 

 give many details. 



39. The Distribution of Salinity-Anomalies. 



In the preceding sections we have repeatedly men- 

 tioned the agreement between salinities and temperatures. 

 A deviation from the average conditions as regards the 

 correlation between these two elements indicates that the 

 water has either an abnormal temperature in relation to 

 the salinity or an abnormal salinity in relation to the 

 temperature. In section 32 we have started from the 

 salinity as an independent variable quantity and used the 

 anomalies of temperature for examining the seasonal va- 



riations of temperature in the upper water-strata. We shall 

 now briefly examine the salinity-anomalies, starting from 

 the temperature as an independent variable (section 18). 

 This may preferably be done in the case of depths below 

 200 metres where the seasonal variations of temperature 

 are comparatively small. It is self-evident that a negative 

 anomaly of temperature corresponds to a positive anomaly 

 of salinity, and vice versa. 



On p. 97* we reproduce some schematic charts illu- 

 strating the distribution of salinity-anomalies at 100, 400, 

 1 000 and 2 000 metres below the surface in the North 

 Atlantic Ocean. The chart for 100 metres cannot give 

 a true picture of the distribution, because the seasonal 

 variations of temperature are considerable at this depth. 

 In spite of the uncertainty it seems, however, to be an 

 established fact that the salinities at 100 metres are rela- 

 tively high as compared with the temperatures in the 

 north-east trade region, and relatively low in the north- 

 zvesteni part of the Atlantic. The high positive salinity- 

 anomalies in the former region should probably be attri- 

 buted to cooling, caused by the northerly winds prevailing 

 there. The low salinity-anomalies in the north-western 

 part of the ocean at 100 metres' depth show the influence 

 of Arctic (or Polar) water of low salinities. The curves 

 representing the distribution of salinity-anomalies at 100 

 metres in the region S. and E. of New Foundland show 

 almost the same shape as the isotherms reproduced in 

 the charts on p. 96*. The curves lie very close together 

 in a belt corresponding to the site of the Great Atlantic 

 Current. If we have 'Gulf Stream' water with a salinity 

 of 36-37oo and a temperature of 17° C. mixed with equal 

 quantities of Polar water with a salinity of 337oo and a 

 temperature of — 1°, we obtain water of 34-657oo and 

 8°C. At the latter temperature the salinity should nor- 

 mally be about 35-2°/oo in the North Atlantic, according 

 to our curve on p. 74*. Hence the salinity-anomaly of 

 the mixed water should be about — 55. If more Polar 

 water is mixed with the Atlantic water, the negative ano- 

 malies of salinity become greater. A mixture of one part 

 of Atlantic water with two parts of Polar water, both 

 constituents having the salinities and temperatures men- 

 tioned above, would give a salinity-anomaly of about 

 — 90. The chart for 100 metres on p. 97* seems to 

 show that the negative values of the salinity-anomalies 

 are greatest just in the regions where the Polar water 

 comes nearest to the Great Atlantic Current. 



The chart for 400 metres on p. 97* shows very small 

 variations in the horizontal distribution of salinity-ano- 

 malies in spite of the fact that the horizontal variations 

 of temperature are very great (cf. the chart on p. 96*). 



At / 000 metres very high positive values of salinity- 

 anomalies appear from Southern Europe and north-western 



