Algorithm For Snow Monitoring Using Remote Sensing Data

Abstract

Snow cover is an important part of the Earth`s climate system so its continuous monitoring is necessary to map snow cover in high resolution. Satellite remote sensing is a science that successfully can monitor land cover and land cover changes. Although indexes such as Normalized Difference Snow Index (NDSI) has quite good accuracy, sometimes topography shadow, water bodies and clouds can be easily misplaced as snow. Using Landsat TM, Landsat +ETM and Landsat TIRS/OLI satellite images, the NDSI was modified for more accurate snow mapping. In this paper, elimination of the misplaced water bodies was made using the high reflectance of the snow in the 0.45 - 0.52 µm wavelength. Afterward, the algorithm was used for estimating snow cover through the years on the one of the highest mountains in Republic of Macedonia. The results from this study shows that the algorithms provide good accuracy than the NDSI alone, totally eliminating the misplaced water bodies, and partly the one caused from topography and clouds. Also, it was noticed that the snow cover in the study area has not been changed drastically through the years. For future studies, the algorithm should be validated on different study areas with different characteristics.

Snow cover is an important part of the Earth`s climate system so its continuous monitoring is necessary to map snow cover in high resolution. Satellite remote sensing is a science that successfully can monitor land cover and land cover changes. Although indexes such as Normalized Difference Snow Index (NDSI) has quite good accuracy, sometimes topography shadow, water bodies and clouds can be easily misplaced as snow. Using Landsat TM, Landsat +ETM and Landsat TIRS/OLI satellite images, the NDSI was modified for more accurate snow mapping. In this paper, elimination of the misplaced water bodies was made using the high reflectance of the snow in the 0.45 - 0.52 µm wavelength. Afterward, the algorithm was used for estimating snow cover through the years on the one of the highest mountains in Republic of Macedonia. The results from this study shows that the algorithms provide good accuracy than the NDSI alone, totally eliminating the misplaced water bodies, and partly the one caused from topography and clouds. Also, it was noticed that the snow cover in the study area has not been changed drastically through the years. For future studies, the algorithm should be validated on different study areas with different characteristics.