Quantification of impact of retention basins on river floods in the Dender catchment in Belgium using computationally efficient models

Developing effective and sustainable water management plans for river basins is difficult due to the complex river dynamics of river systems and the large number of objectives, factors and policies that have to be considered. The evolution towards larger spatial scales and more integral systems makes the use of simulation models to support decisions indispensable. Applications such as the (iterative) optimization of management strategies, the assessment of the impact of numerous types of uncertainties or the statistical analysis of effects based on long-term simulations all require very fast simulation models. Conventional detailed full hydrodynamic models are computationally too expensive for such applications. To overcome this problem, the presented research uses an innovative generic surrogate modeling methodology to emulate the results of detailed models, but at a fraction of the calculation time. These surrogate models are employed to quantify the reduction of flood magnitudes and frequencies due to the installation of retention basins and application of different filling and emptying strategies for a Belgian case study. This reduction was computed by performing a statistical analysis after a 36-year rainfall series was simulated in the surrogate models. The results show that the flood countermeasures can reduce the flood depth at a critical location by 60 centimeters for larger return periods, but do not succeed in decreasing the peak flow in a downstream receiving river. The developed conceptual models can be employed for optimizing retention basin parameters and gate operations while accounting for antecedent conditions, different kinds of uncertainties and objectives at river basin scale.status: publishe