Assimilating passive microwave remote sensing data into a land surface model to improve the estimation of snow depth

Accurate spatiotemporal snow data are crucial for understanding climate systems and managing water resources in cold regions. This paper describes a snow data assimilation system that employs the ensemble Kalman filter to directly assimilate passive microwave brightness temperature data into a snow process model. In the system, the Common Land Model coupled with a snow grain size growth algorithm was adopted to predict layered snow state variables. The forcing data were derived from the Japan Meteorological Administration-Global Spectral Model (JMA-GSM) operational global data assimilation system. The Microwave Emission Model of Layered Snowpacks (MEMLS) was used to convert the snow state variables to brightness temperatures. The snow data assimilation system was one-dimensionally tested at a Siberian cold region reference site of the Coordinated Enhanced Observation Project (CEOP). The validation experiment indicates that the data assimilation system can improve depth estimates during the accumulation period but not the ablation period. The assimilation method proposed herein can be easily applied to an operational weather forecasting system to improve snow depth estimations.