In the case of the iceberg supply factor it is difficult to find 

 environmental data that is both intuitively reasonable and available. 

 Direct iceberg counts in Baffin Bay have been made only a few times 

 since 1880, and are thus not available. Smith (1931) suggested a 

 relationship between summer air temperatures at Upernavik, Greenland 

 and iceberg supply. It is not unreasonable to suspect that increased 

 air temperatures at Upernavik might be related to a decrease in ice- 

 berg supply through melting. 



In the case of efficiency of iceberg transport, the Belle Isle/ 

 Ivigtut pressure difference as a measure of the strength of north- 

 westerly winds has already been described. This pressure difference 

 is probably also associated with the strength of the Labrador Current. 



In the case of iceberg mortality during the drift south, Groissmayr 

 (1939) found that the December-February air temperature at St. Johns, 

 Newfoundland correlated with iceberg intensity. It is reasonable to 

 infer that this correlation was related to increased melting caused 

 by either the higher air temperatures or higher water temperatures 

 associated with the higher air temperatures. It should be mentioned 

 that to some degree this correlation must also be attributed to the 

 fact that warmer air temperatures are related to a decrease in north- 

 westerly winds which carry icebergs south. 



To investigate the possible iceberg supply, efficiency of transport, 

 and mortality since 1880, the 19 year running average for each of the 

 following was computed and is shown in Figure 4: June-September 

 Upernavik air temperature, December-March Belle Isle/Ivigtut pressure 

 difference, and December-March Torbay, Newfoundland air temperature. 

 The record at Torbay is actually a combination of Torbay and nearby 

 St. Johns. The record at Torbay began in 1946, and the record at 

 St. Johns ended in 1956. A correction which was constant within 

 l0.1°C in 75% of the cases during the period of overlapping record 

 was applied to the St. Johns data to convert it to Torbay data. 



Considering the overall trend of the 19 year running average of 

 temperatures at Upernavik, the temperature in the 1950 's is about 

 0.7°C warmer than in the 1890' s, however the details of the Upernavik 

 record are difficult to relate to the iceberg record; there are warming 

 peaks soon after the turn of the century and in the mid-1930 's 

 which appear to positively correlate about 7 years later with increasing 

 iceberg intensity on the Grand Banks. Although the Upernavik temperature 

 record does not seem to support the validity of the iceberg statistics, 

 it is of interest that if the iceberg record lags the Upernavik temper- 

 ature record by about 2 decades, they do seem to correlate inversely 

 as postulated. 



19 



