A wide variety of sensors are installed in satellites, each corresponding to different electromagnetic frequency bands and observational objectives. In observing sea ice conditions to provide ships with navigational support, today the most promising frequency band seems to be in microwave range. This is because microwave radiation can display conditions on the earth's surface regardless of whether skies are clear or cloudy, and can thus be used in all weather and throughout the Arctic night. Microwave sensors can be broadly divided into two groups: passive sensors and active sensors. Passive sensors pick up microwave radiations from the earth's surface and form an image; the principles by which they work and their methods of operation are rather simple. Active microwave sensors radiate microwaves from a satellite at a particular region of the earth's surface; an image is formed by the satellite as it picks up the scattered reflections. A prime example of an active microwave sensor is the Special Sensor Microwave Imager (SSM/I) used in the DMSP program in the United States. Active-type examples include the microwave scanning radiometer (MSR) installed in many satellites, such as MOS, NIMBUS and NOAA; the Scanning Multichannel Microwave Radiometer (SMMR); and the side-looking radar on Russia's OKEAN. Another type of active microwave sensor is the Synthetic Aperture Radar (SAR). Boasting an extremely high resolution of 100m, this sensor is expected to prove highly valuable in monitoring of ice conditions.
Microwave remote-sensing systems such as SAR are valuable tools for monitoring ice conditions, and indeed are used extensively by Russia to assess ice conditions in the NSR as discussed above. After the NSR was opened, several test voyages were conducted, making extensive use of satellite data to determine their actual routes. The first non-Russian vessel to traverse the NSR completely after the route was opened was L'Astrolabe, which used sea ice maps produced from SAR and SSM/I images obtained from ERS-1 and faxed directly to the ship (Johanssen, 1992). This approach was also followed in a test voyage by the Kandalaksh, organized by the SOF. In the case of the Kandalaksh, however, the SAR data was sent not by facsimile but as image files, transmitted to the ship's computer via modem (Yamaguchi, 1995). In a recent test voyage by a tanker through the Kara Sea in winter, image data from ERS-2, RADARSAT, OKEAN and METEOR were all used to decide the ship's route (Pettersson, 1999, Smirnov, 1999). These projects demonstrated the utility of microwave satellite images as a tool in assessing ice conditions, but also pointed out some of the issues that will have to be addressed in regular NSR operations in future.
* Swath width and resolution
SAR images have an excellent spatial resolution of 100m, providing brilliantly detailed information on ice conditions. The tradeoff, however, is that the swath width is confined to a mere 100km, since the SAR image resolution is dependent on swath beam, incident angle and other factors such as range and azimuth directions. If the vessel is obliged to take a course crossing at right angle away from the axis of the image at a speed of 5 knots, the ship will leave the swath width where SAR images can be received in only half a day. For fixed points, the frequency of receiving SAR images is too low. As the earth rotates, the position of the accessible region changes periodically. Under the normal operation of the ERS satellite, for example, this cycle is three days. Because the swath theoretically covers 300km at 70°, the satellite can only capture the same area, one third of the NSR, once every three days. The availability of the SAR images has a spatial limitation. In contrast, the swath width of SSM/I of passive type is wide, enabling each sea region of the NSR to be captured in a single image. The difference in swath width between these two systems can be seen in the images of ice conditions used by the Kandalaksha (Figures 5.8 and 5.9). The disadvantage of SSM/I, of course, is that its spatial resolution is as low as 20km, revealing only basic surface features such as the edges of ice and ice concentration but no indication of finer details such as presence of ice leads or variations in type of ice.
