Snow-Cloud Distinction in MODIS Images, Using Snow Map Algorithm and Cloud Mask Algorithm

The Coverage of snow is about so million Km. of Northern hemisphere and therefore, because of its high albedo and low heat conductivity, it has an important effect on balancing the whole energies on the Earth. Moreover, the most part of water reserves used by human are obtained from snow coverage, especially in mountainous regions of the world. Among the limitations of snow maps that use visible bands and thermal infra-red bands is clouds darkness’s. Therefore, identification of clouds would be necessary for two reasons: 1. Exhibiting the regions which are covered by clouds and 2. Reducing the errors produced as results of incorrect identifications. In this research, images obtained from MODIS sensor have been used because of their acceptable spatial resolution, rapid temporal revision and abundance of the spectral bands. The studied area is located on Central Alborz mountainous region of Northern Iran. The area is periodically covered by snow and encompasses several snow-gauge stations. Considering different algorithms related to identification of snow surface, finally the Snow Map algorithm has been used for the images. Liberal Cloud Mask algorithm was also used to reduce limitations of the Snow Map algorithm in identification of some clouds. Results have been compared to the images obtained from the Snow Map algorithms application, without Liberal Cloud Mask based on snow-gauge station data and height parameters. Results have showed that the Snow Map algorithm cannot distinguish high-level clouds, and clouds containing ice particles, and so it classifies clouds as snow. Investigations showed an increasing about 10 percent in precision of the Snow Map algorithm, when Liberal Cloud Mask was used.