many icebergs that model entry represented. 

 (When IIP received reports with multiple ice- 

 bergs in the same location, only one entry 

 would be made into the model and the number 

 of icebergs that entry represented would be 

 noted in the model input log.) The technician 

 would also ensure that the target number was 

 not included in last month's count. The num- 

 ber of chances for errors within the manual 

 counting scheme increased with the number 

 of icebergs entered into the model. 



Beginning in 1 982, the computer model 

 used by International Ice Patrol created a file 

 containing the sighting and final drift position 

 foreach sighting entered into the model. Only 

 sighting reports received by Ice Patrol with 

 positions within the bounds of the model (40 to 

 52 North and 39 to 57 West) were entered into 

 the model. This means all sighting reports 

 after 1 982 outside the bounds of the model are 

 NOT included in the data base. The numberof 

 reports received outside of the model area is 

 estimated to be less than 1 00 peryear with the 

 majority of the reports being north of 52 North. 

 After 1 982, sighting reports received when the 

 model was not being njn are not included in 

 the data base. 



1989-1991 



Development of a computerized tech- 

 nique to "count" the number of icebergs cross- 

 ing 48 North began in 1 989. This method was 

 used to generate the numbers published in the 

 1989 to 1991 bulletins. Development was 

 completed in 1 991 . The new technique takes 

 into account the number of icebergs each 

 entry represents and provides a break down 

 by sighted versus drifted but never sighted 

 south of 48 North. The new computerized 

 technique uses the model iceberg sighting 

 history file. This file contains the sighting date 

 and position, and final drift position for each 

 target entered into the model, including 

 resights. The counting program examines all 

 of the entries for each target number that was 

 an iceberg (not a radar or growler) to see if: (1 ) 

 an iceberg entry was sighted south of 48 North 

 or (2) if an entry was drifted south of 48 Nor+h 

 without being sighted south of 48 North. 



A "Count" Review 



A summary of the iceberg sightings 

 from the database for the period 1960 to 1981 

 is shown in Table 1 . Also included in Table 1 

 is the estimate of the icebergs to have crossed 

 south of 48 North for each year. In comparing 

 the database summary to the published esti- 

 mate of icebergs drifting south of 48 North, 

 some observations can be made. 



In 1966, Ice Patrol published an esti- 

 mate of zero icebergs drifting south of 48 

 North yet the database includes 12 iceberg 

 reports representing 1 4 icebergs sighted south 

 of 48 North. No explanation for this observa- 

 tion is offered. 



In 1969, 1979, 1980, and 1981, the 

 number of icebergs represented by the data- 

 base iceberg sighting reports sighted south of 

 48 North is less than the published estimate of 

 the number of icebergs drifting south of 48 

 North. A possible explanation for this obser- 

 vation would be the inclusion in the published 

 estimate of icebergs predicted to have drifted 

 south of 48 North by the prediction scheme 

 usedduringthoseyears.ln1960through1964, 

 the number of icebergs represented by the 

 database sighted south of 48 North is larger 

 than the published estimate of the number of 

 icebergs drifting south of 48 North. This could 

 be accounted for by the resighting of icebergs 

 already reported and regular reconnaissance 

 flights of the same area. 



The new computerized counting tech- 

 nique was applied to the model iceberg sight- 

 ing history files from 1 982 to 1 991 . A summary 

 of this application is in Table 2. A more 

 detailed break down by year is included in 

 Attachment 2. For the period 1982 to 1984, 

 the sighting source was not entered into the 

 model. For the period 1982 to 1985, the 

 number of icebergs each entry represented 

 was not entered into the model. The new 

 computerized counting technique was applied 

 consistently to all of the data. The results 

 show differences with the previous methods 

 used. The computerized technique is be- 

 lieved to be a more accurate reflection of the 

 icebergs WHICH IIP INCLUDED IN ITS 

 MODEL that were either sighted south of 48 

 North or drifted south of 48 North. 



61 



