The iceberg drift model affects 

 all aspects of MP's operations. 

 It is used to set the limits when 

 no reconnaissance is being 

 conducted. The model pre- 

 dictions are also used to help 

 identify resightings of previ- 

 ously detected icebergs. With- 

 out this technique, every 

 iceberg sighting would be 

 viewed as an initial 

 detection, thus greatly inflat- 

 ing the iceberg census num- 

 bers. 



The accuracy of 

 the iceberg drift model predic- 

 tions is strongly influenced by 

 the mean currents in the IIP 

 database. Every week during 

 the iceberg season IIP uses 

 the drift of several satellite- 

 tracked buoys to make tem- 

 porary modifications to the 

 current data base in the im- 

 mediate vicinity of the buoys 

 (Summy and Anderson, 1983). 

 When the buoy leaves an area, 

 the currents revert to their 

 mean values over a period of 

 two weeks. Despite this effort 

 to use near real time current 

 data from drifting buoys, the 

 size of the IIP operations area 

 (40-52 N, 39-57 W) and the 

 length of the iceberg season 

 (5-6 months) make it impos- 

 sible to use only observed data 

 to run the model. Hence, the 

 mean currents of the IIP data 

 base have a strong impact on 

 HP's iceberg predictions. Ice 



Page 78 



Patrol uses approximately 12 

 buoys per year. 



HISTORY OF THE IIP 

 CURRENT DATA BASE 



The mean current 

 field forms the basis of Ice 

 Patrol's efforts at predicting 

 iceberg movement. The grid 

 spacing varies according to 

 location, with most of the area 

 divided into segments of 20 

 minutes of latitude by 20 min- 

 utes of longitude. The area of 

 the offshore branch of the La- 

 brador Current has a finer lon- 

 gitudinal grid spacing (20' by 

 10') in an attempt to resolve 

 this narrow, southward-flow- 

 ing current. This is also the 

 region of the densest (in space 

 and time) hydrographic 

 sampling. 



The foundation of 

 the mean current field is 

 hydrographic data collected 

 during over 1 00 surveys from 

 1934 to 1978. At least one 

 survey was conducted 

 annually during the period, 

 with the exception oftheyears 

 of World War II. From these 

 surveys the distribution of 

 mean dynamic typography 

 was computed for each of 

 four months (April - July) by 

 Soule (1964). These charts 

 were updated by Scobie 

 and Schultz (1976). Figure 



C-2 presents the mean dy- 

 namic typography for the 

 month of April. 



Until 1979, Ice Patrol 

 used four separate current 

 files, one for each of the four 

 months. However,the monthly 

 variability in the current files 

 was small and in 1979 they 

 were averaged into a single, 

 time-invariant mean current 

 field (Murray, 1979). This field 

 is based primarily on geostro- 

 phic currents calculated from 

 the gradients in the dynamic 

 topography. However, in 

 some cases, notably in the 

 core of the offshore branch of 

 the Labrador current in the 

 region south of Flemish Pass, 

 the magnitude of the current 

 was increased from the val- 

 ues calculated from 

 geostrophyto reflect some lim- 

 ited drifter data. There is no 

 documentation for these 

 changes. 



The rationale for in- 

 creasing the current magni- 

 tudes was based on the con- 

 cern that calculating current 

 speeds from the hydrographic 

 data would lead to serious un- 

 derestimates of the actual cur- 

 rent speeds. For example, 

 Soule (1964) argued that the 

 uniform grid spacing of the 

 normal (mean) charts tended 

 to smooth the gradients in 

 dynamic height, resulting in a 

 low estimate of the current 



