Evaluation of remotely sensed snow cover product in Central Asia

Central Asian countries depend highly on water resources from snow and glacier melt, which has to be studied thoroughly to estimate water availability. However, the observation network in Central Asia is poor to carry out such studies in detail. Observations from space using remote sensing techniques might fill this observation gap, which needs to be validated. Therefore, this study evaluates the Moderate Resolution Imaging Spectroradiometer (MODIS) daily snow cover product in Central Asia. For the evaluation, in situ snow depth data from 30 meteorological stations and higher resolution Landsat data are used. The results show an overall snow agreement between MODIS and ground observations of 93.1 and 92.7% for MODIS Terra and MODIS Aqua snow products, respectively. The agreement between MODIS and Landsat is 91.9% when considering snow and land agreements. The snow fraction product from MODIS is also validated using Landsat data, and varying accuracies are obtained. The main disadvantage of the MODIS snow cover product are the cloud induced gaps. Therefore, cloud covered pixels are eliminated using the ModSnow algorithm. Using the in situ data, the snow agreement of cloud removed snow cover data is checked, and an accuracy of 84.4% is achieved.