ANNUAL VARIATION OF CLOUD. 



Table 85. 



153 



Monthly variation of percentage frequency of Clear Skies (b in 1902 and 1903 and — 1 in 1911 



and 1912) from mean for each two years. 



Table 86. 



Monthly variation of percentage frequency of Overcast Skies (0 (overcast) in 1902 and 1903 and 

 9 — 10 in 1911 and 1912) //-oH! wiertw o/ each turn years. 



These values have been plotted on fignre 50, the curve for clear skies being inverted 

 so that its variations may be compared directly with the other two. 



It will be seen that the annual variation of cloud amount is well marked, while its 

 variations are followed very closely by the frequency of clear and overcast skies. Also an 

 examination of the values for the individual periods given in the tables clearly indicates 

 that the main variations are repeated in each. The curve of variation is a simple one : cloud 

 is at a maximum near the equinoxes and at a minimum in both summer and winter. 



There is little difficulty in explaining the summer minimum. November, December and 

 January are the fine weather months of the Antarctic. There are then relatively few blizzards and 

 in their absence the Antarctic anticyclone produces clear unclouded skies. Ihe frequency of 

 blizzards in February and March, to which attention has already been drawn in the section 

 on wind, is directly responsible for the large increase then in cloud above the summer 

 value. Winds of more than 30 miles an hour from the south are from three to four times 

 as frequent in February and March as in November and December, and as these winds have 

 on the average a cloud amount of 8'4, it is quite understandable that the average cloud 

 for these months is high. Coming now to the winter months we at once find difficulties. 

 During May, June and July blizzards are somewhat more frequent than during February 

 and March, and they are accompanied by nearly as much cloud — the mean cloud during 

 winds > 30 miles an hour from the south was 8-9 in February and March and 8-0 in 

 May, June and July — but the average cloud amount is less than during the three summer 

 months. We therefore have the paradox that the months with the maximum blizzards have 

 the minimum cloud amount. 



It is clear that we must seek for some action which clears away the clouds as soon as 

 the blizzards have ceased to produce them. Such an action was described by Sir Napier Shaw 



3Q 



