Evaluation Of Soil Moisture Data Assimilation To Improve Hydrologic Partitioning Over Agricultural Areas
Abolafia-Rosenzweig, Ronnie 1 ; Livneh, Ben 2 ; Small, Eric 3 ; Badger, Andrew 4
1 Â鶹Ãâ·Ñ°æÏÂÔØ
2 Â鶹Ãâ·Ñ°æÏÂÔØ
3 Â鶹Ãâ·Ñ°æÏÂÔØ
4 Â鶹Ãâ·Ñ°æÏÂÔØ
As data assimilation is increasingly used to merge Land Surface Model (LSM) estimates of state variables (e.g. soil moisture) with remotely sensed retrievals, it is practical to enhance the strengths of both LSMs and remotely sensed products in assimilation systems. Anthropogenic alterations to the land-surface, primarily through irrigation is frequently neglected from state-of-the-art LSM physics; whereas contemporary remotely sensed products, such as the Soil Moisture Active Passive (SMAP) satellite detects these changes. Importantly, this difference leads to diverging climatologies between observational and model time series and biases the partitioning between evaporation and transpiration. This study explores the benefits of assimilating SMAP soil moisture retrievals with the VIC model via implementing the ensemble Kalman smoother (EnKS) and particle batch smoother (PBS) over one irrigated and one non-irrigated drainage basin covering a 2-year record. The results from PBS and EnKS assimilation runs will be compared with each other using a model run with no assimilation and the resulting streamflow simulations will be validated with gauge data