10 



BJ0RN HELLAND-HANSEN 



(REP. OF THE "MICHAEL SARS" NORTH 



amplitude, as well as the variations in wind are, however, 

 very often apt to cause displacements in the opposite 

 direction. 



Along the same zero-lines and with the same scales 

 as the B-curves for air temperature some broken curves 

 for the variations in the difference between sea temperature 

 and air temperature have been drawn. The difference, 

 sea temperature minus air temperature li — i ^= A/) 



has been found from the values in Table I b for every 

 second hour, and the 24-hour averages referred to the 

 hours of observation have been calculated | A/ .„ r) The 



differences between the individual values and these averages 

 (Ar — ■^'^yoo) have been smoothed by taking means of 



3 values continuously /(A? — ^'/^^l V ^'"^ ^^^ values 



found in this way are represented in the broken curves 

 pp. 76*, 78*, 80*, and 82*. Positive values (drawn upwards 

 from the abscissa) mean that the surface temperature was 

 higher than the air temperature, negative values (downwards) 

 that the air was warmer than the sea. 



The curves show that the variations in the difference, 

 sea temperature minus air temperature, in the great majority 

 of cases occurred inversely to the variations in air 

 temperature. This means either that the variations in the 

 temperature of sea and air occurred in inverse relation to 

 each other, or that the variations in the air temperature 

 were more prominent than in the sea temperature. Both 

 cases occur, and will mostly be related to a predominant 

 daily period of the air temperature. The air was as a 

 rule warmer than the sea by day and colder by night, 

 the broken curves showing in most cases a minimum by 

 day and a maximum by night. 



9. Humidity and Fog. 



The general features exhibited by the A-curves for 

 the North Atlantic show the same regional distribution 

 of the quantity of water-vapour as of the temperature 

 of the air, and consequently (sections 7 and 8) of the 

 surface temperature and salinity as well. Those regions 

 where the surface salinity and temperature were high 

 had also a high air temperature and a high absolute 

 humidity. Only in the case of coastal wateis in summer 

 time may this general correspondence fail, because the 

 temperatures may be high in relation to the salinity (cf. 

 sect. 7) or the air may be dry when it has come from a 

 continent. 



The variations in the general distribution of air 

 temperature and absolute humidity also corresponded very 

 closely to each other, as will clearly be seen from the 

 curves (see especially the figure representing the third series 



of observations, p. 80*). For these variations the corre- 

 spondence with the surface temperature is less pronoun- 

 ced, or even absent. We have seen (sect. 8) that the air 

 temperature sank or rose in comparison with the surface 

 temperature according to the direction of the wind. For 

 the sake of brevity we shall assume that southerly winds 

 are winds which bring the air from warmer to colder 

 regions, and that northerly winds are those which bring 

 the air from colder to warmer regions. In this sense of 

 the words we find that southerly winds gave higher 

 temperature and absolute humidity, and northerly winds 

 made both these elements decrease. The correspondence 

 is especially pronounced when not merely the air tem- 

 perature but also the difference between the surface tem- 

 perature and the air temperature is taken into consideration. 

 The absolute humidity was comparatively high when the 

 air was warmer than the sea and low when it was colder, 

 and the variations in humidity corresponded to the varia- 

 tions in temperature-difference between water and air. 

 The exceptions to this rule are very rare (in the first and 

 last series of observations only; there are none in the 

 second and third series). 



Now it is rather surprising that the variations in relative 

 humidity demonstrated in our A-curves generally coincided 

 almost e.xactly with the variations in absolute humidity. 

 We can disregard those instances when the air was for 

 some considerable time saturated with humidity, and fog 

 occurred. Otherwise there is almost perfect agreement 

 even in minor details as regards the absolute and relative 

 moisture in the air above the sea. Divergencies in the 

 rate of rise or fall (the derivate) of the average quantities 

 are easily explained by an examination of the wind 

 conditions. 



Northerly winds in the sense mentioned above would, 

 when increasing, generally cause an extra decrease in 

 absolute as well as relative humidity (cf. June 7th- 10th, 

 19th, July nth, 23rd, 25th). On June 28th we had a 

 relatively low air temperature and a very low relative 

 humidity in spite of a southerly wind veering westerly at 

 the time. We have already shown, however, that the same 

 air had in all probability come directly from cold northern 

 regions only a few hours before. The air from northerly 

 regions in the North Atlantic contains a comparatively 

 small amount of moisture, and such "Polar" air appears 

 to be comparatively dry when heated. Similarly, we 

 generally find a comparatively low humidity in con- 

 tinental air moving seawards. On August 12th we e.xpe- 

 rienced a sudden fall of humidity because the wind on 

 that day brought air from Northern Europe. The next 

 day we had an increase of absolute and relative humidity 

 because the wind changed and brought air from a rela- 

 tively warm part of the Norwegian Sea. 



