Conclusion 



iceberg sighting ratios for 1 983 

 and 1984, the solved-for values 

 are 1 02 (W- 483 at the 95% 

 confidence level) and 430 (+/- 

 41 7 at the 95% confidence level) , 

 respectively. The data clearly 

 shows the existence of a SLAR 

 effect for both years, since the 

 number of icelsergs estimated 

 south of 48°N each year far 

 exceeds the values predicted by 

 the iceberg sighting ratio, even at 

 the upper limits of the 95% 

 confidemce interval (Figure C-2). 

 Given the small number of 

 observations (n=1 3) used in 

 performing this analysis and the 

 large confidence intervals (due in 

 part to the scatter of the data) , this 

 estimate only denranstrates that 

 the effect of SLAR on iceberg 

 reconnaisance exists and should 

 not be used to compare the 

 SLAR and pre-SLAR data 

 numerically. 



In addifion to improved 

 iceberg detection with SLAR, the 

 increase may also be partly 

 caused by misidentification of 

 non-iceberg targets as icebergs, 

 due both to unfamiliarity with the 

 new technology and the inherent 

 ambiguities of SLAR imagery. 



It is important to note that data 

 presented throughout this 

 appendix was collected and 

 analyzed overthe years by a 

 constantly changing International 

 Ice Patrol staff, using techniques 

 that undoubtedly varied 

 somewhat with the tumover in 

 personnel. Therefore, this data 

 should be examined with that 

 caution in mind. 



74 



The effect of introducing the 

 iceberg drift prediction model 

 (IBERG) to International Ice Patrol 

 operations on 1 979 was analyzed 

 by examining the relationship of 

 icebergs estimated south of 48°N 

 to iceberg sighting reports south 

 of 480N for 1960 through 1984. 

 These values had a close linear 

 relationship for the years 1965 

 through 1978 and their ratio 

 showed a marked increase with 

 the introduction of the model. 

 But since the relationship of the 

 number of icebergs estimated 

 south of 48°N due to the iceberg 

 sighting ratio appeared to be 

 unchanged after 1979, we 

 assume that the effects of the 

 side-looking airtwrne radar 

 (SLAR) on the number of 

 icebergs estimated south of 48°N 

 are much greater than those of 

 the IBERG nrwdel. 



The introduction of SLAR to 

 IIP in 1983 significantly changed 

 the nature of IIP iceberg 

 reconnaissance. 

 Reconnaissance no longer 

 depended on visibility, complete 

 coverage of search areas was 

 possible under conditions of 

 intermittent visibility, and longer 

 flights made it possible to search 

 the outlying areas of the IIP 

 region. 



These improvements, 

 together with possible 

 misidentification of non-iceberg 

 SLAR targets due to 

 inexperience and the limitations 

 of SLAR imagery resulted in 

 elevated estimates of icebergs 

 south of 48°N in 1 983 and 1 984. 

 This conclusion is supported by 

 relating pre-SLAR numbers of 

 icebergs south of 48° N with the 



■iJr U. S. GOVERNMENT PRINTING OFFICE: 1986--600-8'46 --20 , 197 



corresponding values for 1 983 

 and 1984. These values were 

 compared using the ratio of the 

 ship iceberg reports and the ship 

 SST reports, an iceberg density 

 function called the iceberg 

 sighting ratio. The apparent 

 increase in iceberg detection due 

 to SLAR is by an order of 

 magnitude in 1 983 and by half 

 that in 1984, with large 

 confidence intervals at the 95% 

 confidence level. 



The analysis of the effect of 

 SLAR reconnaissance presented 

 here provides only a preliminary 

 investigation of the issue. In no 

 way can an evaluation of two 

 years of SLAR data identify the 

 long-term effect of SLAR on IIP 

 operations. Neithercanit 

 adequately place the 1983 and 

 1 984 icet)erg seasons in the 

 context of the previous years of 

 HP history. It is clear that SLAR 

 has increased the number of 

 icebergs detected by IIP and that 

 with several more years of SLAR 

 data, and with the research 

 currently being conducted on 

 iceberg detection and 

 identification by SLAR, we should 

 be able to address this issue with 

 more confidence. 



