NO. 4 TEMPERATURE VARIATIONS IN THE NORTH ATLANTIC 12/ 



or for Torungen, The curves for air temperature at Torungen and 

 for all Norway show genefally a similarity with the curve of the 

 air pressure gradients for Torungen for the month of March, but 

 there are nevertheless man}^ disagreements, as shown in figure i on 

 plate 48. 



For August, the air temperature curve for Ona and all Norway 

 shows agreement with the air pressure curves of August and the 

 mean of July and August for Stad. 



AGREEMENT BETWEEN THE TEMPERATURE VARIATIONS IN THE COAST 



WATER AND IN THE AIR OVER SCANDINAVIA, BOTH 



DETERMINED BY AIR PRESSURE DISTRIBUTION 



From what has gone before, we can conclude with certainty that 

 the variations in the air pressure distribution control not only the 

 surface temperature on the Norwegian coast, as at Ona and To- 

 rungen, but also the air temperature in Norway during the coldest 

 and warmest part of the year. Since the air pressure distribution 

 (wind) has a simultaneous action upon the temperature of the coast 

 water and the temperature of the land, we can expect that both 

 these temperatures would follow in such a sequence that the char- 

 acteristic variations would be a little earlier in the air than in the 

 water. Such an agreement cannot be merely local but must be 

 shown over great regions, because the air pressure distribution has 

 a very extended sphere of influence. If, for example, the temperature 

 variations for all Norway are compared with those in Stockholm, 

 we find a complete agreement. We shall later return to the con- 

 sideration of such comparisons (see fig. 75). 



In order to examine these correlations more carefully we have 

 made a comparative investigation of the temperature variations 

 of the Norwegian lighthouse stations, Torungen, Utsire, Heliso, 

 and Ona, and the temperature variations at so far removed a 

 locality as Stockholm. The results are given in figure 53. The 

 pairs of curves A show the variations of the temperature devia- 

 tions from month to month for 37 years, 1874 to 1910, From the 

 monthly values we have computed consecutive 12-month means 

 as shown in curves B, and from the latter also 24-month means, as 

 shown in curves C, in order to bring out periodic phenomena. The 

 sun spot curve, S, is introduced lowest in the figure. The tempera- 

 ture scale is twice as great for the coast water, which has the scale 

 on the right hand, as for the air with the scale on the left, because 

 the variations of the air temperature are greater than those of the 



