Table 1. What-lf model permutations conducted during this project. 



Results 



In general, the final positions of the What-If modeled icebergs were within 20 nautical 

 miles (NM) of the true position of the tracked Ice Island. The 20 NM threshold is interesting 

 because it represents the error circle radius presently used by Ice Patrol for an iceberg that has 

 been in the model for 3 days. From that perspective, it can be stated that the model is a 

 reasonable representation of the real ocean for the area being considered (on the Bank, away from 

 more complex regions near the tail). Operationally, this result provides IIP with good support for 

 the model error estimates currently employed in the system. 



Counter-intuitively, the modeled very large tabular iceberg did not behave most like the 

 tracked Ice Island. The very large tabular was greatly affected by cunent and to a lesser degree 

 by the wind. In the case of the very large tabular drifted with wind and currents, the modeled 

 iceberg was only within 20 NM of the actual position following 42% of the model runs. When 

 the same iceberg was drifted with winds only, it was within 20 NM after 52% of the runs. 

 Examples of the model results for the very large are presented in Figure 2. 



The growler modeled with no currents provided the most accurate representation of actual 

 Ice Island drift. Following 92% of the model runs, the wind-driven growler was within 20 NM of 

 the Ice Island's actual position. With currents and wind however, the resultant growler was only 

 42% accurate. 



Figure 2. Model results from What-lfs drifting a very large tabular iceberg. The left panel displays the very large 

 drifted with winds and currents and the right displays drift with winds only. The blue symbols represent the modeled 

 iceberg while the brown represents actual Ice Island position. Note the growth of the error circle as time in the model 

 elapses from 1 day to 3 days (5 NM, 10 NM, 20 NM). 



Figure 3 presents some examples of What-lf/growler results. While not surprising, this result is a 

 nice confirmation of the general assumption that Ice Island drift will tend to be dominated by 

 wind effects because of their relatively shallow draft. 



53 



