Appendix F 



IIP Iceberg Resights 

 1992 and 1993 Seasons 



Bruce E. Viekman 



International Ice Patrol receives ice- 

 berg data from several sources. Reports from 

 HP's own reconnaissance, from Canadian Ice 

 Patrol and Department of Fisheries and Oceans 

 flights (conducted by Atlantic Airways), and 

 from ships are combined to produce iceberg 

 warnings to mariners. An important aspect of 

 MP's use of ice reports is a process called a 

 'resight' which is where new reports are corre- 

 lated with icebergs already 'on plot'. The intent 

 of this process is to avoid interpreting every 

 reported iceberg as an iceberg that has never 

 been seen before. If the reported iceberg is 

 determined to have been seen before, the 

 forecast position of the iceberg is updated 

 using the new data. Otherwise, the iceberg is 

 entered as a new iceberg. 



IIP began use of the iceberg Data fvlan- 

 agement and Prediction System (DMPS) at 

 the start of the 1993 season. This system 

 affords llPthe ability to resight icebergsgraphi- 

 cally. The DMPS displays the icebergs on plot 

 along with their reported size, error circle, and 

 drift track. The system displays new sightings 

 simultaneously with the icebergs already 'on 

 plot.' The IIP watch officer then decides to 

 either correlate the most recent iceberg 

 sightings with those on plot based on size 

 comparisons, position differences, and drift or 

 add a new iceberg. The system used before 

 the 1992 season required manual plotting of 

 old icebergs to correlate them with the latest 

 sightings. Effective use of multiple iceberg 

 reports was therefore difficult, particularly in 

 areas of heavy iceberg concentration. Using 

 the new system, operators should be able to 

 resight icebergs much more effectively. 



The procedures used to resight ice- 

 bergs have several effects. First, the process 

 influences the total number of icebergs re- 

 ported by IIP for the year. If resights are 

 conducted more frequently, the number of 

 new icebergs added to the model will de- 

 crease, thereby lowering the totals. More 

 frequent resights should also increase the 

 quality of IIP drift predictions of icebergs near 

 the limits. Murphy and Anderson (1 985) found 

 that drift prediction errors increase with time in 

 a roughly linear manner. Therefore, decreased 

 time since sighting should result in decreased 

 predicted position error. The experiments 

 comparing MP's model with observed drift were 

 conducted by tracking a known iceberg. Op- 

 erationally, the correlations are done without 

 such positive identification. 



To evaluate the changes in the resight 

 process caused by the use of the DMPS, an 

 investigation of resights was conducted using 

 records from the 1 992 and 1 993 ice seasons. 

 The iceberg sighting data base was used to 

 compare the changes in resighting frequency 

 between the 1993 ice season and the 1992 

 season. The geographicdistribution of resights 

 was also computed. 



The 1993 ice season was much more 

 severe than 1992. IIP uses the number of 

 icebergs which drift south of 48N as an indica- 

 tion of season severity. In 1993, the third- 

 heaviest season on record, IIP estimated that 

 1 753 icebergs crossed this latitude, while 876 

 did so in 1992. The 10-year average (1983- 

 1992) of icebergs crossing 48N is 927. 



87 



