Appendix C 



A Qualitative Assessment of the Iceberg Detection and Identification Capabilities 



of RADARSAT 



LT James E. Andrews 

 U.S. Coast Guard International Ice Patrol 



ABSTRACT 



A test was conducted as part of the RADARSAT Application, Development and 

 Research Opportunity (ADRO) program to determine the ability of RADARSAT, in Wide 

 2 beam mode (W-2), to detect icebergs. Testing the ability of RADARSAT to detect 

 icebergs, determining the finest expectable resolution for point targets, and assessing 

 our ability to differentiate between icebergs and vessels are necessary prerequisites for 

 IIP in eventually employing RADARSAT operationally for remotely sensed iceberg 

 reconnaissance. 



On 28 July 1997, a visual reconnaissance by Ice Patrol personnel on board a 

 Coast Guard Hercules C-130 was conducted over the footprint area of a previously 

 arranged RADARSAT overpass. Photographs were taken by the aircrew of all icebergs 

 sighted within the footprint area, as well as of several boats observed during the flight. 

 The flight commenced at 10:00 UTC, and the last target within the footprint was detected 

 at 13:23 UTC. IIP received a RADARSAT scene from 09:43:40 UTC, 28 July 1997, 

 centered on 50°49'N, 52°10'W with a 150 km swath width and 12.5 m pixel spacing. 

 This scene contained 1 1 certain targets and exhibited a nearly equal mix of observed 

 icebergs and vessels from the underflight. We then conducted qualitative comparisons 

 between the photographed icebergs and the RADARSAT detected targets in 

 approximately similar locations. Iceberg size, image target size, "brightness" values and 

 shapes were compared between observed objects and detected targets. Of the eleven 

 targets, six were associated with icebergs and five were associated with vessels. In 

 addition, a visually observed growler (waterline length between 7 and 15 m) was 

 undetected by RADARSAT. 



We compared the cost of replacing HP's aerial reconnaissance methods with 

 coverage by RADARSAT W-2 imagery. We determined that the cost of an ice season- 

 length program of repeated scene coverage of the Grand Banks area would cost 

 approximately $1.4 million, compared to aircraft costs alone of approximately $1.7 

 million. Despite the cost savings, IIP must await further development in detecting 

 icebergs, especially growlers, in image delivery and in cost savings. It appears that an 

 ice season length program would be very beneficial in augmenting aerial reconnaissance 

 and in providing flexibility and repeat coverage; however, funding sources and a more in- 

 depth cost-benefit analysis would need to be performed. 



55 



