RESTRICTED 



by its insulating power. This is particularly true of loosely packed 

 snow. 



A common assumption in the North is that heavy snow in the fall 

 means a rapid break-up in the spring. AVith the subsequent de- 

 creasing I'ate of growth, ice which has grown steadily throughout the 

 winter is seldom more than 4 to 5 feet in thickness by the following 

 summer. 



Perennial sea ice may grow in thickness during the summer by re- 

 freezing of thaw water. Snow on the surface melts, and the water 

 runs down through cracks and holes to form a layer of fresh water 

 under the ice. Since the temperature of the underlying salt water is 

 usually lower than the freezing point of fresh water, a layer of fresh 

 water ice is formed on the bottom of the sea ice. In summer, there- 

 fore, a floe melts away on top, but at the same time may be growing 

 slowly on its undersurface. By this process, mud, stones, seaweed, 

 or shells originally frozen to the under side of grounded floes may 

 work right up to the surface. Diatoms frozen to the under side will 

 similarly rise. An autumn period follows, with lower temperature 

 but without ice formation, the supply of fresh water being no longer 

 renewed and the sea temperature not being low enough for the freezing 

 of salt water to begin again. In the second winter, growth continues 

 by salt water freezing. If the ice is unbroken through the second 

 winter, its thickness may reach 7 to 8 feet at the most. Ice in the 

 Arctic polar basin is seldom less than 31/2 to ii/2 feet thick, and Nansen 

 reports a maximum thickness of 13 feet 10 inches produced by about 

 4 years of normal growth. 



The action of blocks and floes being forced over each other or turned 

 on end by some form of pressure is called rafting. Ice of much greater 

 thickness than ordinary floes can be formed by rafting, tidal over- 

 flow, or other types of flooding such as spray and splashing, but such 

 areas will be of limited extent. 



The approximate thickness of ice may be predicted from figure 3 

 if the temperatures at a specified locality are known. Even if exact 

 temperatures are not available, estimates can probably be made from 

 a general knowledge of weather conditions in the region. The only 

 complication in using this graph lies in calculating the "degree days 

 of frost." First it must be remembered that a temperature of 0° F., 

 for example, is equal to 32° of frost. Secondly, the mean number of 

 degrees of frost for each day, or group of days with the same mean 

 degrees of frost, is to be used, not the mean degrees of frost for the 

 entire period. Days on which the temperature was below freezing 

 for only a part of the 24 hours can be ignored unless exceptionally 



RESTRICTED 5 



