1D-2D Model Validation for the Assessment of Urban Flooding: A PCSWMM 2D Application

The investigated catchment is a small structured urban basin, with many narrow streets and small lots and buildings in the city of Graz (Austria). This led several difficulties in data collection. Hence a step-wise approach for the 1D-2D model-set up was chosen, providing the possibility to obtain intermediate results for evaluation.

The first step was to set up a 1D model in PCSWMM and calibrate it with observed rainfall and sewer flow data in the investigated area. Next, node flooding was evaluated with Euler II type design storms. Based on the node flooding results, several regions within the investigation area were identified for implementing a dual drainage approach.

The dual drainage model was then implemented in the 1D model using a bidirectional flow exchange approach with two weirs for every manhole (Concha Jopia and Gomez Valentin 2010). As the dual drainage model served only to gather rough information about the flow paths along the streets, no street inlets were implemented in the 1D-1D model. The results from the 1D-1D model were then used for assigning hot-spots. Around these hot-spots a detailed 1D-2D approach was implemented. The main goal of this detailed 1D-2D model is to simulate heavy storm events and assess urban flooding as well as flood control measures.

In addition to the step-wise model approach a citizen survey for gaining knowledge of surface water levels and flow paths on the streets and inside private estates as well as information about flood caused damages in the past was conducted. For the citizen survey site-based questionnaires were created. The response rate to the survey was more than 50 %. The responses were then used to validate the simulation results with ‘soft data’.

The outcome of the calibrated 1D model was a clear boundary for a more detailed investigation in further modelling steps. The 1D model results highlighted flooded nodes within the investigation area. Flow paths were simulated within the next modelling step – the 1D-1D model. Within the dual drainage model the simulated flow paths were observed and the hot-spots for the application of PCSWMM 2D were defined.

The evaluation of the 1D-2D model result focusses on the assessment of such an approach in urban flooding analyses respectively on the evaluation of possible flood control measures applicable for suddenly occurring storm events. The simulation results indicated that using a resolution varying from 0.5 to 2.5 m is necessary otherwise the mesh would falsify the flow paths on the surface.

The results of the 1D-2D model were compared with the information gained by the citizen survey. The simulated surface water levels and the caused infrastructure and houses show good accordance with the ‘soft data’ collected from the respondents. Furthermore, urban flooding events observed documented by the project team could be covered reasonable well.

The results show that a mesh resolution of 0.5 to 2.5 m likely leads to stable results for the investigated catchment. The utilisation of such mesh resolutions led to long simulation times. That’s why the previous choice of hot-spots turned out to be important, resulting in “manageable” simulation times.