Satellite microwave observation to monitor and predict ecohydrological conditions: a land data assimilation approach

Satellite microwave observations such as AMSR2, SMOS, and SMAP can monitor surface soil moisture and vegetation water content in all weather conditions. The integration of these observation data and a land surface model (LSM) has high potential to accurately monitor the unobservable land conditions (e.g., root-zone soil moisture) and predict their future. Here we present the development and application of our land data assimilation system, called the Coupled Land and Vegetation Data Assimilation System (CLVDAS) to monitor and predict water and vegetation dynamics. CLVDAS can optimize the LSM's unknown parameters (e.g., soil hydraulic conductivity) and sequentially adjust state variables (i.e., soil moisture and vegetation water content) by assimilating passive microwave brightness temperature. CLVDAS has been used for drought quantification at the global scale and for regional drought monitoring and prediction in the Horn of Africa, North Africa, and Northeastern Brazil. Our findings from those applications are: (1) assimilating satellite microwave observation significantly improves the skill of the LSM to simulate both water and vegetation dynamics; (2) simultaneous monitoring and prediction of soil moisture and vegetation dynamics provide the holistic view of drought propagation; (3) the general circulation model-based seasonal prediction is useful to accurately predict vegetation conditions in 2 to 3 months lead time; (4) initial conditions of land water and vegetation have an important role in predicting agricultural drought, which reveals that it is important for drought prediction to accurately monitor land surface conditions by satellites. Finally, we would like to emphasize the importance of land data assimilation as a platform to integrate a large volume of data from past and future remote sensing.