18 
to the melting point was 102 cal cm? (Table 2). 
After raising its temperature to O°C, another 
1,000 cal cm’? would be required to melt it. 
However, the actual melt process is complicated 
by localized percolation and refreezing of melt- 
water in the snow after slight melting occurs on 
the snow surface. This percolating water re- 
freezes to form a complex network of ice glands 
(nearly vertical, pipelike structures) and ice 
lenses throughout the snow and at the tundra- 
snow interface. The process is analogous to the 
formation of superimposed ice on glaciers 
(Trabant, Fahl, and Benson 1975). These masses 
of ice within the tundra snow may remain for 
several weeks and may significantly modify the 
snowpack as a habitat for small animals such as 
lemmings, which live under the snow, or large 
animals such as caribou, which feed by breaking 
through it. The 1972 spring provided an excel- 
lent example of the process involved in forming 
these ice masses in the snow. The snow was 
subjected to slight surface melting on 6-7 May 
when the maximum air temperature was above 
freezing. After 7 May the maximum air tempera- 
ture remained below freezing until 27 May, as 
summarized in Table | and Fig. 8. 
Figs. 9a, 9b, and 9c show snow profiles* 
measured along a traverse line east from P-2 
(Fig. 5) on 14 May. The pit study plotted in Fig. 
9a, made only 8m east of the road, is in the 
drifted snow alongside the roadbank. The pit 
studies in Figs. 9b and Qc are far enough from 
the road to be essentially unaffected by it; the 
same is true of the profiles in Figs. 9d and Ye, 
which were measured on 16 May. A deposit of 
fresh new snow appears at the top of each of 
these profiles. It was deposited between 11-13 
May and overlies the surface melt crust pro- 
duced on 6-7 May. In addition to the melt crust 
which formed on the snow surface of 6-7 May, 
there were ice glands, lenses, and layers in the 
snow. These are shown in black in the strati- 
graphic columns of Figs. 7 and 9. The location 
of these ice lenses at the base of a fine-grained 
layer is common. They also form in such strati- 
graphic locations in the deep snow of glaciers. In 
Temp.°C Density, g cm? 
-20 -10 ie) 0.25 0.20 0.30 0.40 
Depth, cm 
Ap AR 
0 " 
Fig. 7. Combined photo and data plot, 200m 
east of pingo, 14 April 1972. 
the examples shown here, they are most com- 
monly at the top of the depth hoar layer. Later 
in the melt season ice layers form at the base of 
the depth hoar layer; at Prudhoe Bay in 1972 
this did not occur until the end of May and 
during the first week of June. Similar timing for 
the formation of ice masses at the base of the 
snow was observed at Barrow and Prudhoe Bay 
during the spring of 1973. 
“Unfortunately, a Rammsonde penetrometer was not available during these measurements. 
Typical Rammsonde profiles in tundra snow are available (Benson 1969), and a detailed study of 
the snow at Barrow using Ram hardness profiles was carried out during the 1972-1973 winter by 
Melchior and Benson. 
