Abstract |
Land Surface Models (LSMs) simulate terrestrial water, energy, and carbon cycles and they are one of the essential components of earth system models and disaster monitoring and prediction systems. The behavior of LSMs highly depends on their unknown parameters (e.g., characteristics of soil, micro topography, and ecosystem traits) and their initial state conditions (e.g., soil moisture, temperature, and biomass). Therefore, to obtain the LSM's unknown parameters and initial conditions, many Land Data Assimilation Systems (LDASs) have been developed to assimilate in-situ and/or satellite observations into a LSM. Since passive microwave brightness temperatures observed by AMSR-E, AMSR2, SMOS, and SMAP are highly sensitive to surface soil moisture, the LDAS based on these datasets has been intensively investigated. In this talk, I will firstly show the capability and limitation of passive microwave land observation data based on our recent field verification studies. Then, I will introduce our LDAS, the Coupled Land and Vegetation Data Assimilation System (CLVDAS). The CLVDAS is the first satellite-based LDAS which can improve the skill of the LSM to simulate not only soil moisture but also vegetation growth and senescence by assimilating passive microwave brightness temperatures. Our recent verification of the CLVDAS and its application to mega-drought monitoring and prediction in Africa will be presented. Lastly, I would like to discuss the future direction of LDAS in the exascale high performance computing and 'big data' era.
|