Flood modelling in urban areas : A comparative study of MIKE 21 and SCALGO Live

Abstract: Pluvial flooding originating from extreme rainfall is problematic and an increasing issue in Sweden. Higher requirements on adapting cities in urban areas to these challenges have been placed on both municipalities and the county administration. Thus, an increased need for water modelling, both in existing and planned areas have emerged. The Danish Hydrological Institute (DHI) has developed several models and tools to simulate floods and heavy rains, of which MIKE 21 is one. MIKE 21 is a dynamic model consisting of hydrological calculations in each cell, requiring modelling skills and long simulation time, but is proven to be a good and credible model. SCALGO Live, on the other hand, is a static tool simulated by raster-based algorithms and capable of giving fast results directly on the platform. However, compared to MIKE 21, SCALGO Live is not as evaluated nor is its use as widespread for simulating floods and heavy rain events. This study aims to investigate how inundation in twenty urban areas caused by cloudbursts is simulated in both programs to examine how well the result coming from SCALGO Live, is equivalent to the result from the MIKE 21-model. The comparison is made in both depth and spread using three comparative indexes, two statistical equations and one map, showing the extension of the inundation in both models. To make the models comparable, the model in MIKE 21 is made as equivalent as possible with SCALGO Live before simulation and the purpose is to investigate whether there is any type of area where the two different models are equivalent. The result shows that the flooded areas from SCALGO Live are in good agreement in most areas with the MIKE 21-model, but that the depth in the depression zones is somewhat overestimated, compared with the highest value in MIKE 21. The MIKE 21-model has a greater spread upstream, showing flowpaths if compared directly with the flooded areas from SCALGO Live, but if activating the flow accumulation tool in SCALGO Live, the differences are reduced but cannot be quantified in this study. The differences between the models increase with a higher resolution, longer flowpaths and a larger catchment, at least for the confined catchments. To conclude, SCALGO Live is best suited for smaller confined catchments where there are no long or complex flowpaths. SCALGO Live also works well at an early stage in the planning process and as a platform for combining detailed data and results. However, for the more complex areas, MIKE 21 is better suited, since various parameters can be considered.