Automatic assessment of uncertainties in the case of urban flood modelling

Urban flood modelling has long been attracting attention from hydraulic modelllers. Due to the complicated floodplain topography, inaccurate boundary conditions, etc, uncertainties will arise from urban flood simulations and should be addressed properly. The present study delas with a novel contribution to the field of urban flod modellling with the aim to quantify uncertainties by examining the response of a finite volume two-dimensional hydraulic model, RUBAR 20, used for simulating urban floding in a cross road. Two methods have been launched to assess uncertainties and the confident range of the solutions; the first method is the very well known Divided Differences (DD) approximation. The second method is a recently developed technique, based on the computation of derivatives using Automatic Differentiation techniques (AD). For the second method, Tapenade software has been used to calculate derivatives of the water depth with respect to the boundary conditions and roughness coefficients. Benchmark tests were carried out against wll-known flows using a high quality and high spatial resolution data set collected from laboratory open channels. Boyh methods provided good solutions for a reasonable set of parameters. Moreover, the AD adjoint model is found to be useful in case of large number of the input parameters; / Une méthode innovante pour étudier les incertitudes dans le cas d'inondation urbaine est développée sur la base d'une modélisation 2D d'un écoulement dans un carrefour, phénomène reproduit dans un laboratoire. La méthode développée s'appuie sur la différentiation automatique par le logiciel Tapenade. Dans ce cas, le modèle adjoint fournit des informations utiles pour le cas de nombreux paramètres.