Author(s): Sylvain Neelz; Gareth Pender
Linked Author(s):
Keywords: Floodplain flow; Inundation; Parameterisation; Urban flooding
Abstract: This paper investigates the possibility to use the roughness parameter in 2D coarse square grid inundation models to take into account sub-grid scale effects caused by buildings in the urban area. A method is proposed where roughness is assigned on a cell-by-cell basis depending on the percentage of surface area covered by buildings in each cell. The method is very straightforward to apply in practice provided that GIS-based high-accuracy land-use maps are available and given the modern capabilities of Geographic Information Systems. The method is applied to the modelling of a hypothetical flooding event affecting a coastal floodplain. Results from Coarse Grid Models (CGM) at resolutions 10m and 50m are compared against benchmark results from a Fine Grid Model (FGM) at resolution 2m using statistics based on the timing of inundation throughout the modelled geographic domain and the best possible agreement is sought using a single calibration parameter k. The main conclusion is that an optimised value of this parameter is clearly identifiable at least for the 10m CGM. This varies only to a limited extent with the magnitude of the flood discharge, and the exact definition of flood arrival time used in the comparison of flood timing. It is also shown that the proposed method allows a significant reduction of the discrepancy between the CGM and the FGM compared to a much simpler method where roughness is allocated depending on land use type at cell centers. Results from the 50m CGM lead to similar conclusions, with optimised parameter values being found in the same range as for the 10m CGM. A useful finding is also that the choice of the parameter value k only needs to be made approximately for the above conclusions to remain valid.
Year: 2007