t | | 
is ie Nat [ Ahi + { Se + Shy, + Sear, 
where Tp = temperature of freezing in © C., 
T = temperature of the water in ° C., 
k; ® heat conductivity of sea ice, 
k, = heat conductivity of snow, 
/; = density of sea ice, 
K = latent heat of fusion, 
1, = thickness of the ice in cme, 
1, = thickness of the snow in cme, and 
Qr = amount of sensible heat loss in kg. cal. 
Using this method, ice~growth curves, figures 21 and 22, are plotted 
against degree days of frost (° F.) for various snow depths that may be 
covering the ice. The growth curves are based on oceanographic data 
taken at site A on 29 September and 12 October. Ice grows more rapidly 
during the early stages and/or with no snow cover. Greater ice thickness 
and/or snow cover offer more insulation therefore resulting in 4a slower 
rate of ice growth as compared to degree days. It will be noted here that 
these degree days are based on 26.6° f compared to the previous 29.0° F. 
However, differences are very small, less than 1 percent. The computed ice 
growth curve (in inches) versus degree days of frost (° F.), using the actual 
measured snow depths, is shown in figures 23 and 2h. Figure 23 and figure 
2h are for the new and old ice, respectively. For comparison purposes, 
the actual growth curves are plotted. In the case of the new ice, these 
curves parallel very closely. she old ice curve does not verify as well. 
Nevertheless, after 150 degree days (° F.) with 13 inches of ice, the 
computed thickness of the old ice is only 1.5 inches less than the actual 
depth. It is possible that this discrepancy may be the result of measur= 
ing the ice thickness in a comparatively shallow area, whereas the oceano=- 
graphic data is based on information from site A, in deeper water. The 
same was not true in the case of the new ice. Im this latter case the 
ice growth was the same over the entire areas; i.eo, from the pier to site 
A, which is approximately 6 nautical miles to the southwest. Neverthe- 
less, the actual growth curves parallel the computed values very closely, 
especially in the case of the new ice, thus reflecting the accuracy of the 
method in predicting ice growth. 
Naturally, if one were to predict the air temperature and snow depth 
values, a forecast of this natwre probably would not be so accurate as in 
the case pointed out here is figures 23 and 2h, where the observed values 
for snow cover and temparatures are used. Difficulty in making accurate 
long-range predictions of snow depths and temperatures will vary for dif- 
ferent areas. In general, the larger the monthly variation of these par= 
ticular parameters, the more difficult the forecast will be. Comparatively 
the variation for these two parameters is small in the area of Thule. 
