Coastal Zone Monitoring with RADARSAT-1
RADARSAT Standard mode imagery (25 m nominal resolution) of Songkhla Lake, Thailand, acquired November 15th, and 18th., 1996, clearly showed the benefit of RADARSAT's variable incidence angle, and how the correct choice of incidence angle was selected to enhance a given application. The November 15 image (Fig. 1 ) was acquired in Standard mode 2 (24° -31° incidence angle range and the November 18 (Fig. 2 image was acquired in Standard 7 (45° -49° incidence angle range). Water borne aquaculture nets (left side of image), which, due to small incidence angle, were not visible in the Standard 2 image, were clearly evident when imaged at the larger incidence angle of Standard 7. In contrast, there was evidence of water-slick discrimination in the Standard 2 image (center of image), but at the larger incidence angle of Standard 7, evidence of water-slick discrimination was not observed. Meteorological data were acquired during both acquisitions, and wind speed, which is a key factor for discrimination of water-surface features, was in the 2m/s -3m/s range.
Figure 1. Standard 2 Subscene (12km x 12km) Songkhla Lake, Thailand acquired November 15,1996. Surface slicks are visible in the central part of the image. © Candian space Agency/Agence spatiale canadienne 1996.
Figure 2. Standard 2 Subscene (12km x 12km) of Songkhla Lake, Thailand acquired November 18,1996Aquaculture nets are visible in the left-hand side of the image © Canadian space Agency/Agence spatiale canadienne 1996.
3.2 Coastline Mapping
Australia is currently developing a GIS-based maritime boundaries information system. O key interest in the system is coastline mapping, and RADARSAT imagery, both Fine beam (8 m nominal resolution) and Standard beam (25 m nominal resolution) was used to evaluate the potential of RADARSAT for this application. Imagery was acquired at the lowest astronomical tide (LAT); the LAT is used to establish territorial sea baseline.
A fundamental issue of this study was the difference in information content of the two beam modes. Resolution proved to be the biggest advantage of the fine beam compared with Standard beam. The Fine beam permitted increased accuracy of the shoreline delineation (Fig. 3). For both beams, it was possible to detect the land-water interface to about 203 pixels. The variance in land-water discrimination translates into 16-24 meter Standard mode and 6-9 meters for Fine beam. Another consideration the incidence angle. Both scenes in this study had a relatively large (>40oC) incidence angle which is optimal for land-water discrimination. However, the choice of incidence angle may depend on acquiring imagery at high or low tide, and the choice of an optimal incidence angle may not be an option.
In addition to the choice of beam mode and incidence angle, the morphology of the shoreline was also an a factor in the land-water delineation. Radar does not penetrate water, but the presence of offshore reefs can be detected and potentially cause confusion between the actual coastline and the reef. The presence of a reef can be detected if the between the actual coastline and the reef. The presence of a reef can be detected if the reef, at low tide for example, protrudes from the water, or the reef may be submerged at all tide levels, but still detected due to radar sensing the breaking waves. As the wave approach the reef and break, the turbulent water increase the surface roughness relative to the (smoother) surrounding water.
Fig. 3. Fine mode RADARSAT subscene (5 km x 9 km) acquired January 9, 1998. Delineation of the territorial sea based on image acquisition at lowest astronomical tide.
3.3 Ship Surveillance
The juxtaposition of the heavy exploitation of available fish stocks, the paucity of new sources of supply in the wild capture fisheries, and the increasing demand for seafood has placed a premium on the access to sources of fish and shellfish around the world. Numerous countries have established Exclusive Economic Zones (EEZ), quotas, limited entry, and other management systems to protect the inherent value of these living resources, but legislation and regulations cannot by themselves control access to valuable fish stocks and deter illegal fishing activity.
Effective fisheries management strategies must be supported by an meaningful surveillance program (Freeberg et al., 1995). To date, these surveillance programs have relied on conventional methods such as patrol vessels and aircraft. The cost and limited spatial coverage of conventional systems has, however, restricted their effectives as a countermeasure against poaching and non-legal fishing operations. RADARSAT, as a component of a comprehensive fisheries surveillance program, provides wide-area surveillance using ScanSAR mode, or high resolution imaging using Fine mode. RADARSAT has been proven as an effective solution for fisheries surveillance in numerous vessel detection validation studies.
RADARSAT imagery was successfully used for a ship detection validation study off the west coast of Canada using 50 m ScanSAR Narrow imagery (Staples et al., 1997). Vessels including tug and barge combinations, fishing boats, and passenger ferries were detected. The smallest vessel detected was a fishing boat with a length of 27 m, and the largest was a bulk carrier with length of 225 m. For fisheries monitoring, the size of the smallest detectable vessel is of more concern than the size of the largest detectable vessel. The focus of much of the ship detection research using RADARSAT is on the establishment of the minimum detectable ship length as a function of beam mode, incidence angle, wind speed, and other ship parameters (Vachon et al., 1997).
A dramatic example of the capability of RADARSAT to detect target (ships and oil rigs in this case) and oil slicks was evident in the Wide 1 image of the Persian Gulf acquired February 18, 1998 (Fig. 4). A second image was acquired on February 28 and used to discriminate ships and oil rigs; the assumption was made that a stationary target that had no change in spatial orientation an oil rig.
