Sea Ice Mapping Algorithm for Future Operational Geostationary Satellites

Sea ice plays a major role in local and global climates. Spatial and temporal information on sea ice has various end users such as local residents of the ice covered regions to transportation companies, mariners, army, etc. Our work is on the development of an automated ice-mapping algorithm which would make maximum use of the next generation GOES-R ABI's improved observing capabilities. High spatial resolution in the thermals in addition to high temporal resolution of GOES-R will provide an enhanced sampling of the scene and hence will allow for improved image classification and ice cover identification. Two Earth orbiting environmental sensors have been used as proxy. Data collected by SEVIRI onboard the Meteosat Second Generation (MSG) satellites and MODIS on board the Aqua and Terra satellites have been used as prototypes. The Northern region of the Caspian Sea was initially selected for the algorithm development and calibration, using MSG SEVIRI data. The approach used in the algorithm development includes daily cloud-clear image compositing as well as pixel-by-pixel image classification using spectral criteria. Data in available spectral channels (reflectance and infrared brightness temperature) have been tested and used to develop the algorithm. The algorithm has been applied over Caspian Sea and demonstrates an exceptional ability to accurately classify land, water, cloud and ice scenes over the region. The ice cover maps generated from MSG SEVIRI data have been assessed using the interactive maps of snow and ice cover produced within NOAA Interactive Multi-Sensor Snow and Ice Mapping System (IMS). In addition, the algorithm has been applied over the Great Lakes region and compared against ice analysis charts from the U.S. National Ice Center.