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

---

Vincent Wolfs, Patrick Willems

Wednesday 1 july 2015

9:45 - 10:00h at South America (level 0)

Themes: (T) Flood risk management and adaptation, (ST) Flood risk assessment

Parallel session: 8J. Floodrisk - Assessment

---

The Dender catchment in Belgium is regularly hit by floods. It is likely that the situation will even worsen due to further urbanization and climate change. Water managers face the task of developing a long-term strategy to reduce the flood damage. To evaluate and iteratively optimize such strategy, multiple long term simulations are required with river models. Such simulations allow for the quantification of the decrease in flood level for given return periods and at multiple locations. Detailed full hydrodynamic models are computationally too expensive for performing such simulations. Therefore, simplified conceptual models are proposed, which emulate the behavior of the detailed models, but only require a fraction of the calculation time. This study evaluates such an approach for the Marke River in Belgium by quantifying the effect of expanding existing and installing additional retention basins along the river. First, planned structural adjustments to the Marke River and retention basins that will be built in the near future were implemented in a full hydrodynamic InfoWorks RS model. Next, equivalent grey box conceptual models were identified and calibrated based on a set of short-term simulations with the detailed model. The conceptual models take all dike levels explicitly into account and contain all hydraulic structures. These models were used to simulate a 36-year rainfall series for the current and future state of the river and statistical analyses were performed on the simulation results. The results show that the retention basins can reduce the flood frequency and magnitudes along the river significantly. The impact on further downstream reaches is, however, rather limited. The proposed approach based on conceptual models that complement full hydrodynamic models can also be employed for other applications that require a large number of simulations, such as other types of optimization applications, uncertainty analysis, real-time optimization and control.