Quantifying structural uncertainty due to discretisation resolution and dimensionality in a hydrodynamic polder model C. Sehnert, S. Huang and K.-E. Lindenschmidt ABSTRACT C. Sehnert S. Huang K.-E. Lindenschmidt (corresponding author) GFZ GeoForschungsZentrum Potsdam, Section 5.4-Engineering Hydrology, Telegrafenberg, D-14473 Potsdam, Germany E-mail: kel@gfz-potsdam.de In flood modelling, the structure of conceptual models may have a large influence on the simulation results. Hence, the focus of this paper is on the structural uncertainty in hydrodynamic flood modelling systems. Three different conceptual models with an increasing order of complexity of the spatial discretisation of the flow through a polder system were compared in order to investigate the effect of spatial resolution and dimensionality on flood modelling. The hydrodynamic 1D model DYNHYD was used as a basis for the simulations. The model was extended to incorporate a quasi-2D approach and a Monte Carlo analysis was used to show the effect of structural uncertainty on the resulting flow characteristics of the diverted flood waters. Two flood events of the River Elbe were used to calibrate and test the model. The results of the velocity fields indicate that the simplest 1D model revealed more predictive uncertainty than the other two more complex models. The differences in model structure does not cause large differences in the capping of the peak discharges, but may substantially influence the results of subsequent modelling of sediment and contaminant transport. Key words | hydrodynamic modelling, model structure, polder system, structural uncertainty INTRODUCTION Polders are efficient measures for reducing flood risk and capping the flood peak discharge. After the flood event along the River Elbe in 2002, which caused huge damage in hazardous areas of the Elbe catchment, it was recognised that flood and flood risk management schemes for the Elbe basin and other German river basins needed to be updated (Petrow et al. 2006). For this reason an impetus for large research activities in Germany was given. RIMAX, founded by the Germany Ministry of Education and Research, is a consortium of projects to develop methodologies for risk management of extreme flood events (http://www. rimax-hochwasser.de). Three of the about 40 projects incorporate polders as a measure for flood management strategies. Polder control, too, is an interest in the EU-financed project FLOODsite (http://www.floodsite. net) on integrated flood risk analysis and management methodologies. A research team of the Center for Disaster Management and Risk Reduction Technology (CEDIM), founded by the University of Karlsruhe (TH) and the Geoforschungszentrum Potsdam (GFZ), is developing a modelling system to quantify the risk of extreme flooding in large river basins (http://www.cedim.de/902.php). The project also considers polder control of peak discharge capping as a possible means of flood mitigation. Computer models are important tools to test the design and efficacy of an existing or planned polder system for flood discharge capping and to predict the best control strategy for the flood hydrograph during operational flood management (Fo ¨ rster et al. 2005). Predicting the amount of solutes retained in such flood water retention systems is gaining in importance (Baborowski et al. 1999; Engelhardt et al. 1999). Quantifying the impact on the environment is doi: 10.2166/hydro.2009.038 19 Q IWA Publishing 2009 Journal of Hydroinformatics | 11.1 | 2009 Downloaded from http://iwaponline.com/jh/article-pdf/11/1/19/386321/19.pdf by guest on 01 May 2021