| dc.description |
Satellite remote sensing is an effective tool for mapping snow covered area. However, complex terrain and heterogeneous land cover, due to vegetation and patchy snow cover, present challenges to snow cover mapping. This research compares two techniques for mapping snow covered area: binary and enhanced fractional snow cover mapping techniques. Both are implemented using data from the Moderate Resolution Imaging Spectroradiometer (MODIS). Two study regions with the differing amounts of vegetation are used in this investigation: Everest National Park, and Annapurna Conservation Area, both located in Nepal. The MODIS binary product maps a pixel as snow if roughly 50% or more is snow. It also incorporates a vegetation correction and cloud mask. The MODIS Snow Covered Area and Grain Size/Albedo (MODSCAG) product gives the fraction of snow cover in each MODIS pixel, but does not incorporate a vegetation correction or cloud mask. Landsat ETM+ images are used to assess the accuracy of each product. Results show that MODSCAG provides a more accurate mapping of snow cover in a heterogeneous landscape, where snow cover is patchy. However, the MODIS binary product is more accurate for mapping snow cover in heavily forested regions. Accuracy of the two techniques varied with elevation, MODSCAG being more accurate as elevation increased. The results presented in this paper will help improve understanding of snow cover mapping techniques and aid future planning of water resources in climatologically sensitive regions. |
|