Evaluation of the IMS snow map in southern China based on high spatial resolution imagery

The Interactive Multisensor Snow and Ice Mapping System (IMS) combines multiple data sources to map global snow cover. IMS can discriminate snow cover beneath clouds using time sequential imagery from geostationary satellites and passive microwave observations. During the snow disaster of 2008 in southern China, the IMS snow cover data identified more snow than those retrieved from passive microwave remote sensing data. In this study, the IMS snow cover mapping accuracy was assessed against the Landsat Enhanced Thematic Mapper Plus (ETM+) imagery. Land cover and terrain effects on the snow cover accuracy were considered. Good agreements (> 90%) was observed over flat cropland surfaces. For forested, mountainous areas, a pronounced disagreement was observed between the two datasets. The accuracy of the IMS in these regions was less than 50%. Due to the sparse distributed snow cover in these regions, the effects of snow cover fragmentation degrees were studied.