Temporal variability of partially-contaminated sediments in a strongly regulated reservoir of the upper Rhine River

¹ Laboratoire National d'Hydraulique et Environnement, EDF R&D, 6 Quai Watier 78400 Chatou France
² Saint-Venant Laboratory for Hydraulics, 6 Quai Watier 78400 Chatou France
³ Centre d’Ingénierie Hydraulique, EDF, Savoie Technolac 73373 Le Bourget du Lac France

^* e-mail: germain.antoine@edf.fr ^** e-mail: thomas.pretet@edf.fr ^*** e-mail: matthieu.secher@edf.fr ^**** e-mail: anne.clutier@edf.fr

Abstract

The upper Rhine River is a highly harnessed and regulated river. EDF (a French electricity company) is in charge of eight dams on the upper Rhine River for producing hydro-electricity. In order to increase the safety and the competitiveness of the installations, but also to reduce their environmental impact, the sediment dynamics in these reservoirs has become a key factor to control and predict. In this study, we focused on the Marckolsheim reservoir, which is located 50 kilometers upstream the city of Strasbourg. Since its construction in 1961, this reservoir has been filled continuously with cohesive sediments, partially contaminated. Two field campaigns were performed in 2015 and 2016 under two different discharge conditions, with the objectives of quantifying the complex velocity fields on this site. The numerical codes TELEMAC-2D and SISYPHE were used to simulate in 2D the hydrodynamic and the suspended sediment transport of the reservoir. A ten kilometers long model was built and calibrated with the measured data of the 2015 and 2016 field campaigns, but also with measurements of sediment parameters that have been done separately. The originality of this model consists in an explicit 3D representation of the dam gates. An algorithm was implemented in TELEMAC in order to adapt the gates position at each time step, in conformity with the real regulation rules followed by the dam operator. By using upstream measured data of discharge and suspended sediment concentration, a four months period was simulated. The comparison of the simulated results with bathymetric surveys shows good agreements if specific properties of sediments related to settling processes are taken into account. Finally, the dynamics of the contaminated sediments was simulated. A 3D spatial distribution of the contaminated sediments in the reservoir was defined at the initial state by using in situ measurements. The fully coupled hydraulic-sediment-pollutant simulation performed over a single flood event gives first interesting highlights on the resuspension conditions of the contaminated sediments.