Data assimilation of in-situ and satellite remote sensing data to 3D hydrodynamic lake models

Abstract

Abstract. The understanding of lakes physical dynamics is crucial to provide scientifically credible information for ecosystem management. We show how the combination of in-situ data, remote sensing observations and three-dimensional hydrodynamic numerical simulations is capable of delivering various spatio-temporal scales involved in lakes dynamics. This combination is achieved through data assimilation (DA) and uncertainty quantification. In this study, we present a flexible framework for DA into lakes three-dimensional hydrodynamic models. Using an Ensemble Kalman Filter, our approach accounts for model and observational uncertainties. We demonstrate the framework by assimilating in-situ and satellite remote sensing temperature data into a three-dimensional hydrodynamic model of Lake Geneva. Results show that DA effectively improves model performance over a broad range of spatio-temporal scales and physical processes. Overall, temperature errors have been reduced by 54 %. With a localization scheme, an ensemble size of 20 members is found to be sufficient to derive covariance matrices leading to satisfactory results. The entire framework has been developed for the constraints of operational systems and near real-time operations (e.g. integration into http://meteolakes.ch).

Multi-column modelling of lake Geneva for climate applications

“…Delft3D was run continuously since 2010 for Lake Geneva. It realistically replicated Lake Geneva’s thermal behavior and dynamics 38 . It provided lake surface temperatures and water temperature profiles every 3 hours via the Meteolakes plateform 39 .…”

“…A constant air density of 1.24 was considered in both models. The absorption coefficient was deduced from the measured Secchi disk depth 43 Finally, in both models the surface water albedo was assumed constant at = 0.09 and the thermal infrared emissivity fixed at 0.985 38 .…”

Assessment of a Multi-column Deep Lake Model for Climate Applications: The Case of Lake Geneva

“…Delft3D was run continuously since 2010 for Lake Geneva. It realistically replicated Lake Geneva's thermal behavior and dynamics 35 . It provided lake surface temperatures and water temperature profiles every 3 hours via the Meteolakes plateform 36 .…”

“…Finally, in both models the surface water albedo was assumed constant at α = 0.09 and the thermal infrared emissivity fixed at 0.985 35 . We relied on the bathymetry derived from the digital elevation model DHM25 of the Swiss Federal Office of Cartography Swisstopo 41 .…”

“…Finally, in both models the surface water albedo was assumed constant at α = 0.09 and the thermal infrared emissivity fixed at 0.985 35 .…”