Snow-Covered Area Determination Based on Satellite-Derived Probabilistic Snow Cover Maps

Abstract

Snow-covered area (SCA) is an important component in hydrological cycle, and its importance increases as the snowmelt runoff percentage in the annual runoff volume increases. Satellite-based remote sensing can help monitor SCA. However, in operational simulations or in forecasts, either the satellite-based snow cover map may not be readily available or is affected by cloud blockage. In this study, a statistical methodology is proposed to estimate the SCA in both cases. The methodology utilizes Interactive Multi Sensor Snow and Ice Mapping System (IMS) snow cover maps and is developed over Turkey. Based on the long-term datasets of the IMS snow product, probability of snow (PS) for each IMS pixel is calculated. Probabilities that yielded minimum errors in SCA detection are used in SCA estimation. SCA map for 1 March 2013 is obtained using the PS values and is compared with the actual IMS snow cover maps. Out of 219 ground stations, 210 (95.89 %) indicated same land cover type (snow/no snow) between PS and IMS-based snow cover maps. Only nine stations (4.11 %) did not match with the actual IMS snow cover map. Among these nine stations, five (2.28 %) indicated underestimation and the remaining four (1.83 %) showed overestimation. High agreement (95.89 %) among the land cover types between two snow cover maps indicates the usability of proposed methodology in snow-covered area forecasting. © 2016, Saudi Society for Geosciences.