C21B-0735
SNOW COVER MAPPING IN THE NORTHERN AREA OF PAKISTAN AND JAMMU KASHMIR (HINDU KUSH HIMALAYAS) USING NDSI, UNMIXING METHOD AND SRTM DEM DATA

Tuesday, 15 December 2015
Poster Hall (Moscone South)
Amad Ud Din1, Hyungjun Kim1, Kazuo Oki1, Wataru Takeuchi2 and Taikan Oki1, (1)University of Tokyo, Bunkyo-ku, Japan, (2)University of Tokyo, Institute of Industrial Science, Bunkyo-ku, Japan
Abstract:
Snow area measurement is very important for hydrologists, glaciologists and for climate change researchers. Field measurement is very difficult as in case of a steep and in a complex terrain such as Himalayas, therefore we rely on remote sensing (both active and passive) data. Usually snow area is calculated from reflectance data using different snow index e.g. Normalize difference snow index (NDSI) and then it is translated into snow area. However, in most cases we are actually calculating the planimetric area or grid area of every pixel. The actual snow is along the surface of the terrain and proper estimation can only be done if actual surface area is calculated along the slope within each pixel. In the past, some researchers have introduced methodologies and optimized old mechanisms. However, the orographical impact in calculating snow area (fraction), especially in steep mountainous regions, still has many problems, and many times these problems are usually ignored which leads to under estimation of total snow amount. In this study we calculated the actual surface area from SRTM version 4.1 90m (at equator) processed DEM data provided by CGIAR-CSI. MODIS Reflectance (MOD09A1 L3 Product) composite data of 500m resolution for 2010 and 2011 in the northern areas of Pakistan, Jammu & Kashmir region where great Himalayas are stretched was used to calculate snow cover using NDSI index. Threshold of NDSI>0.4 was set to classify snow or no snow for the clear pixels and for further classification, unmixing method (subjective pixel method only) was used to calculate snow fraction within each pixel. Results shows that in a complex terrain such as Himalayas, ratio of surface to planimetric snow area is more than 50%. This means that it should be taken into consideration for more realistic snow amount estimation. Seasonal snow fraction histogram from unmixing method indicates that NDSI measures snow cover area by 1.86 times more in cold season (maximum snow area) and 1.30 times more in summer season (minimum snow cover) than the snow area calculated from snow fraction using unmixing method.