But nevertheless, the formation of depth and bottom ice in a sea, as in fresh water, ordi- 

 narily ceases after the formation of a surface rind of ice, since direct release of heat by the water 

 to the air ceases after this. 



Thus, there is no basic difference between the formation of deep ice in fresh water and in the 

 seas; in both cases, a certain supercooling and mixing of water which guarantee heat release, are 

 necessary. 



LITERATURE: 62, 77. 



Section 39. Ice Formation at Positive Air Temperatures 



Ice formation in the sea ordinarily begins at negative air temperatures and after the tempera- 

 ture of the surface level has decreased to the freezing point. In individual cases, however, ice 

 formation can begin at positive air temperatures as well. 



Two conditions are necessary for this: 1) The surface layer of the water should be very thin 

 and it should differ sharply in density from the lower layers, and 2) effective radiation should be 

 sufficiently strong. For the latter, a very dry transparent atmosphere is necessary. 



I observed a characteristic example of ice fornaation in a very thin, very fresh surface layer 

 on 5 September 1934 in Traurenberg Bay (79° 58' north, 16° 48' east) during the voyage of the 



Perseus . 



It was about noon on a calm day (the sun's elevation was approximately 17°) with a completely 

 cloudless sky, air temperature 2.6° (at a height of about 6 m), atmospheric pressure 1, 030 mb, 

 and relative humidity 65 per cent (the minimum relative humidity, recorded by the automatic re- 

 corder, was 59 per cent for the day). 



The water temperature at the very surface of the sea was 4. 92°. The salinity of a water 

 sample, taken simultaneously, was 37. 73 o/oo. The water, carefully dipped from the surface of 

 the sea by a bucket, was of lower temperature (about 3, 5 per cent). Despite the high temperature 

 of the subsurface layers (bottom temperature at a depth of 64 m was 4. 18°), a very thin film, con- 

 sisting of fresh (to the taste) ice appeared from time to time on the surface of a completely calm 

 sea. But a gust of wind and the appearance of ripples were sufficient to destroy the ice which was 

 melted by the heat of subsurface water particles which rose to the surface of the sea with a new 

 heat supply when ripples formed. 



Such formations of thin surface ice in the presence of relatively high water and air tempera- 

 tures had also been observed earlier. 



Scoresby often observed this phenomenon in 1882, and writes that "during cloudy weather, 

 when the thermometer is at -1.7°, the surface of the sea does not freeze, but during clear, calm 

 weather, when the sun drops towards the horizon, the sea begins to freeze, even though the ther- 

 mometer is at +2. 2° or higher. " 



The same phenomena was observed by Nordenskjold on 31 August 1879 during the voyage of 

 the Vega, off the New Siberian Islands. 



Nordenskjold writes "the sky was clear of clouds at the zenith and in the east . . . despite the 

 fact that the temperature of the air and water was above the freezing point, we had the opportunity 



94 



