Feasibility of Improving Groundwater Modeling by Assimilating Evapotranspiration Rates

Water table dynamics are linked to the atmosphere through evapotranspiration and recharge. The assimilation of observed evapotranspiration values should thus improve groundwater model results. This paper presents a method for assimilating evapotranspiration data into a coupled unsaturated zone and groundwater model using the ensemble Kalman filter. The method is tested for a synthetically generated losing stream system in climatic conditions common in South Australia. The study focused on areas with a deep (recharge area) and a shallow (extraction area) water table and an intermediate area showing seasonal variations between these two (transition area). The data assimilation algorithm consistently improved the model states in the three areas when the ensemble spread was adequate. Improvements were also obtained in the calculation of net recharge and modeled actual evapotranspiration. The results indicate the potential to improve groundwater models using remote sensing observations of actual evapotranspiration values.