Experimental insight for flood flow repartition in urban areas

Variance based sensitivity analysis of 1D and 2D hydraulic models: An experimental urban flood case

“…For the 2D model, parameters ranges are chosen corresponding to uncertainty ranges of our experimental dataset (cf. [14]) (section 3.2).…”

“…This study is based on the urban flood experimental rig designed and built at the ICube laboratory in Strasbourg (France) [3,14]. It is supposed to be representative for typical urban geometries at the horizontal scale 1 /200 with street widths of 12.5cm and 5cm as well as various crossroad angles.…”

“…The streets network is composed of seven streets in the north-south direction crossed by seven streets in the west-east direction (see figure 1 -c). The inflow discharge of each street of the north and west faces is supplied by a volumetric pump, a rectangular weir controls water levels at the outlet of each street and outlet discharges are measured using calibrated weirs relationships [14]. The studied configurations correspond to steady state experiments and the corresponding boundary conditions are presented in Appendix (table 2).…”

“…Given the importance of uncertainty estimation in flood modelling and forecasting, this contribution presents a thorough sensitivity analysis (SA) of 1D Saint Venant and 2D SW models to M A N U S C R I P T A C C E P T E D ACCEPTED MANUSCRIPT their (uncertain) parameters and boundary conditions. The study focuses on the case of urban flood flows in a complex streets network, based on a fine and unprecedented laboratory scale experiment while in situ flood data are still rare ( [14]). Remark this streets network is composed of several 4 branch crossroads creating complex patterns of free surface flows confluences and diffluences.…”

“…A thorough quantitative sensitivity analysis of the widely used 1D and 2D Saint Venant SW models with a variance based method is proposed ( [35,38]). A unique experimental dataset at the horizontal scale 1 /200 is used, with urban flood flows corresponding to frequent to rare return periods ( [14]). Therefore the originality of this paper is to explore the sensitivities of 1D and 2D SW models of branched urban flood flows, involving many subcritical confluences and diffluences but also few supercritical flow zones based on a detailed experimental data-set.…”