maining 18 flights neither increased nor de- 

 creased the area enclosed by LAKI. This does 

 not innply that the limits were accurate on 

 those dates because several of those patrols 

 were special purpose flights to deploy drifting 

 buoys or to verify iceberg density at interior 

 locations. A few patrols were incomplete, due 

 to sensor or aircraft problems, and had no 

 inpact on the limits. On two occasions IIP re- 

 connaissance detected ice outside LAKI. The 

 most dynamic regions, and the area most 

 important to shipping, were the southern lim- 

 its of all known ice. Typically, the southern 

 limits had the highest priority for each ice re- 

 connaissance detachment's patrol efforts. 



IIP strives to maintain accurate LAKI, neither 

 underestimating nor overestimating the threat 



icebergs pose to mariners. Effective recon- 

 naissance and proficient model performance 

 are expected to produce only nominal 

 changes in LAKI. The 1997 reconnaissance 

 data suggest that the drift and deterioration 

 models are conservative, since, in most cases, 

 the flights reduced the area enclosed by LAKI. 

 On the other hand, there were six occasions 

 in 1997 when ice was observed outside MP's 

 published LAKI. While this is a small error 

 (-2%) when compared to the 330 ice bulle- 

 tins produced by IIP in 1997, it is significant 

 because mariners were unaware of these 

 dangerous icebergs. As a result, we do not 

 wish to make our models less conservative. 

 The data emphasize the importance of regu- 

 lar, effective reconnaissance in maintaining 

 accurate LAKI. 



Table 1 



1995 - 1997 Reconnaissance Patrol Data 



47 



