The error Introduced was found to be small unless the water temperature 

 was unusually high, and in most cases ice will not form under such condi- 

 tions. Since the potential curve 1^(Qt) must always be plotted, it Is 

 readily apparent when the above approximation is invalid; in this situation 

 the formula can be used. In order to facilitate substitution of the 

 actual integral for the approximation, an auxiliary table (Table 13) is 

 included whicn gives the degree-days of frost associated with various 

 values of 1^ and Op. If the potential curve is very Irregular, Table 13 

 is entered, and by comparison with the degree-day figure computed from 

 an average 1^, one can determine the amount of error introduced into the 

 total* Most frequently this will be found to be small, since the potential 

 term is only one of four separate factors contributing to the total $ how- 

 ever, the actual value can be added or subtracted and a more accurate 

 figure secured. 



The form of equation 14 which was used in the tables is 



The degree-days of frost for fixed 1 8 ™ 0, 2*5, 5, 7.5, 10, 15, 20, 30, 

 40, and 50 cm. are given in Tables 3 to 12. The concept of degree-days 

 has an advantage over one of calendar days because of its versatility as 

 far as forecasting is concerned. An ice forecast in days is difficult 

 because It requires a forecast of air temperatures as well as ice growth, 

 while the degree-day forecast permits ice growth forecasts which fit all 

 conceivable temperature variations. 



E. USE OF THE ICE GROWTH TABLES IN FORECASTING 



Forecasting the ice growth in a given area requires the following 

 informations (1) oceanographic data taken after the water mass begins 

 to cool and prior to ice formation, (2) the date of Ice formation, and 

 (3) an expression relating ice growth to degree-days of frost for the 

 given area. With this information the forecaster can. then proceed to 

 make a forecast for the entire season. The procedure is as follows? 

 (1) the ico potential of each oceanographic station is calculated and 

 the ice potential curve (example, Figure 2) is drawn; (2) from available 

 meteorological and oceanographic data, the date of ice formation is 

 determined; and (3) the forecast is made in terms of degree-days of 

 frost. 



Since the calculation of the ice potential, the date of first ice 

 formation, and the ice potential curve have been previously explained, 

 the remaining discussion will describe the preparation of an actual 

 forecast. As before., Station #37 is used as an example. The ice poten- 

 tial curve for this station is given in Figure 2. In order to draw the 

 ice forecast curves for this station it is necessary to read the ice 

 thickness figures corresponding to integral values of Qp from the Ice 

 potential curve. The degree-days of frost corresponding to each combi- 



10 



