Assimilating Passive Microwave Brightness Temperature Data into a Land Surface Model to Improve the Snow Depth Predictability

This paper introduces the application of the ensemble Kalman filter (EnKF) technique for the assimilation of passive microwave remote sensing observations into a landsurface model, to improve the snow depth (SD) predictability. A new landsurface model, currently developed at the Japan Meteorological Agency (JMA), which is based on the simple biosphere model (SiB), is used as a forward model to predict the change of the snow pack. The microwave emission model of layered snowpacks (MEMLS) is used as observation operator, to transfer the model prediction into the corresponding satellite brightness. The assimilation system was applied using data from the coordinated enhanced observation period (CEOP) Asia-Australia monsoon project (CAMP) Eastern Siberia Taiga region for the period from November 2002 to March 2003. The data sets includes JMA-GSM model output, which is used as forcing data, satellite brightness temperature observation from the advanced microwave scanning radiometer (AMSR-E) and in-situ snow depth (SD) observation and the current AMSR-E snow depth product for comparison. The assimilation results are in good agreement with the data from the snow depth observation sites in this region and improve the forecast of the land-surface model. Furthermore, comparison with the AMSR-E SD product showed, that the assimilation results are also in better agreement with the in-situ snow depth observation.