Appendix E 



Operational Forecasting Concerns 

 Regarding Iceberg Deterioration 



LCDR Walter E. Hanson Jr., USCG 



INTRODUCTION 



Since 1971, the International Ice 

 Patrol (HP) has used computer- 

 based drift prediction models to 

 help evaluate the extent of the 

 iceberg danger to North Atlantic 

 shipping in the vicinity of the 

 Grand Banks of Nev\rfoundland. A 

 dynamic model began operational 

 use in 1979 (Mountain, 1980). 

 This model, along with a paramet- 

 ric iceberg deterioration model 

 which began operational use in 

 1983 (Anderson, 1983), has grown 

 in importance as iceberg recon- 

 naissance has gone to an every 

 other weel< schedule. During the 

 peak of the iceberg season, April 

 through June, the iceberg danger 

 covers a large area, requiring 

 reconnaissance missions to 

 concentrate on patrolling the 

 limits. Often icebergs go several 

 weeks before being resighted. 

 Resighting icebergs depends 

 heavily on these models effec- 

 tively predicting drift and deteriora- 

 tion rates. These predictions are 

 also routinely used to set the limit 

 of all known ice, as reported in the 

 IIP bulletins. 



As evidenced by the many years 

 of safe passage by trans-Atlantic 

 shipping, the IIP seems to have 

 some skill in determining the 

 extent of the ice danger. (To 

 assess the model's predictions 

 there is a need for accurate data 

 to represent the initial iceberg, 

 interim iceberg, and environmental 

 conditions.) Iceberg drift predic- 

 tion is highly dependent upon 

 iceberg mass (size and shape). 



Consequently, the ability to accu- 

 rately predict changes in iceberg 

 size for the majority of icebergs, 

 which are infrequently resighted, 

 becomes very important. 



Between 1983 and 1985, the IIP 

 studied the drift and deterioration 

 of four icebergs. Although no firm 

 conclusions could be drawn from 

 such a small data set, which rep- 

 resented an average drift of 4.5 

 days, the prediction models did 

 fairly well hindcasting the drift and 

 deterioration when observations 

 were used as inputs (Anderson, 

 1985). A similar study was per- 

 formed, using U. S. Coast Guard 

 iceberg data, for the Atmospheric 

 Environment Service of Canada 

 (El-Tahanetal, 1987). The 

 results were mixed. Thus in June 

 1987, the IIP conducted another 

 cruise to collect similar data for a 

 cluster of icebergs. 



The objectives of this study were 

 to compare iceberg deterioration 

 predictions derived from environ- 

 mental data collected in situ to 

 inputs available from operational 

 data centers. These latter inputs 

 were divided between global and 

 regional scale products. 



BACKGROUND 



The iceberg deterioration model 

 used by the IIP provides its watch 

 officers with a daily estimate of the 

 "melt" status of each iceberg 

 entered in the drift model. The 

 computer-based application 



(Anderson, 1983), which com- 

 putes the melt rate, is derived from 

 White, et al 1980. The model 

 sums the effects of the following 

 processes which are depicted in 

 order of importance in Figure E-1 : 



• solar radiation; 



• buoyant heat convection; 



• heat convection caused by 

 iceberg nrwvement relative to 

 the water mass (forced heat 

 convection); and 



• waterline wave erosion, followed 

 by calving of the resultant ice 

 overhang. 



Based on the 1980 report, warm 

 air heat convection is considered 

 insignificant and not calculated. 

 The report also identified other 

 iceberg deterioration processes; 

 however, they were only partially 

 addressed and difficult to quantify. 

 Consequently those processes are 

 not rTX)delled. 



The model calculates melt in 

 terms of length instead of mass. 

 This measure of melt accommo- 

 dates HP's operational proce- 

 dures, in which nearly all iceberg 

 dimensions are reported by size 

 categories. These categories are 

 based primarily on the maximum 

 observed length of the iceberg. 



109 



