ADAPTIVE FILTERING FOR (SOIL MOISTURE) DATA ASSIMILATION

Data assimilation (DA) techniques provide the means to integrate observational information into dynamical models. The success of such methods requires accurate knowledge of model and observation uncertainties, which are seldom available. Adaptive DA techniques estimate these uncertainties as part of the system. In this study, we apply the recently developed Monte Carlo based adaptive Kalman Filter (MadKF) to assimilate SMOS brightness temperature (Tb) measurements into the Catchment Land Surface Model for soil moisture state updating. The MadKF yields robust Tb uncertainty estimates and significant skill improvements relative to model-only soil moisture estimates, when evaluated with in situ measurements.