Can we quantify the impact of the modeler on the model?

Hydrological and hydraulic models are historically different disciplines and work on different scales.   The recent increase in computational resources allows for the two models to be combined into one model having one holistic approach. This removes the bottleneck of the data linkage between the two disciplines.

In this study, we apply the two-dimensional module of the open-source software openTELEMAC-MASCARET with the included SCS-CN method on an ungauged catchment in central Germany with an area of around 58 km². The catchment is part of the Main River tributary. We describe the excessive data preprocessing of the building and land use data, and the topography to sufficiently represent the small-scale stream geometry. This preprocessing is subjective in selecting different thresholds, such as the degree of mesh refinement in the streams and the foreland, or a minimum size for buildings to be represented in the model. Additionally, the SCS-CN method is highly sensitive to the model results, as small changes in the CN-values already significantly alter the total volume of water in the model. We collect the different sources of subjectivity and uncertainty and rank them based on the impact on the model results. The results will lead to a better view of the potential of combined hydrological-hydraulic models.