Joint assimilation of piezometric heads and groundwater temperatures for improved modeling of river-aquifer interactions

Abstract

The ensemble Kalman filter (EnKF) is increasingly used to improve the real-time prediction of groundwater states and the estimation of uncertain hydraulic subsurface parameters through assimilation of measurement data like groundwater levels and concentration data. At the interface between surface water and groundwater, measured groundwater temperature data can provide an additional source of information for subsurface characterizations with EnKF. Additionally, an improved prediction of the temperature field itself is often desirable for groundwater management. In this work, we investigate the worth of a joint assimilation of hydraulic and thermal observation data on the state and parameter estimation with EnKF for two different model setups: (i) a simple synthetic model of a river-aquifer system where the parameters and simulation conditions were perfectly known and (ii) a model of the Limmat aquifer in Zurich (Switzerland) where an exhaustive set of real-world observations of groundwater levels (87) and temperatures (22) was available for assimilation (year 2007) and verification (year 2011). Results for the synthetic case suggest that a joint assimilation of piezometric heads and groundwater temperatures together with updating of uncertain hydraulic parameters gives the best estimation of states and hydraulic properties of the model. For the real-world case, the prediction of groundwater temperatures could also be improved through data assimilation with EnKF. For the validation period, it was found that parameter fields updated with piezometric heads reduced RMSE's of states significantly (heads -49%, temperature -15%), but an additional conditioning of parameters on groundwater temperatures only influenced the characterization of the temperature field. Key Points Assimilation of hydraulic and thermal data into a managed river-aquifer system Temperature predictions could be improved with EnKF Temperature data gave information on spatial structure of leakage parameters © 2014. American Geophysical Union. All Rights Reserved.