Improving simulated soil moisture states through AMSR-E data assimilation

Model estimated soil moisture fields are often biased due to uncertainties in input parameters and deficiencies in model physics. These biases have a significant impact on the estimation of other land surface processes such as evapo-transpiration (ET) and runoff as they are calculated based on the absolute value of soil moisture in the land surface. AMSR-E satellite derived soil moisture retrievals, by design, represent the spatially averaged value of surface soil moisture in a footprint area and therefore provide observed mean soil moisture states that can be used to reduce model bias. To achieve this goal, an ensemble Kalman filter with a mass conservation constraint was developed to assimilate the actual value of AMSR-E soil moisture retrievals without any scaling or pre-processing. The mass conservation constraint not only alleviates the unbiased requirements by an ensemble Kalman filter but also provides a reliable way to relay observation information from the surface to the deeper soil profile where no measurements are available. The Little Washita watershed of Oklahoma, with an area of 611 square kilometers, was chosen as the experiment site for its abundance of in situ soil moisture measurements. Assimilation results using the Noah land surface model in the watershed showed that the mass conservation scheme reduced the model bias in the entire soil profile, while a conventional Kalman filter without any mass conservation only reduced the bias in the shallow root zone. Assimilated soil moisture fields and their impacts on ET, stream flow and soil moisture anomalies are presented.