Variation of Uncertainty of Drainage Density in Flood Hazard Mapping Assessment with Coupled 1D-2D Hydrodynamics Model

Coupled 1D-2D hydrodynamic models are widely utilized in flood hazard mapping. Researchers have explored several uncertainties in flood hazard mapping, but have not addressed the uncertainty of drainage density. Drainage density is equal to total length of the drainage divided by the catchment area. The model sets denser the tributary drainages for higher drainage density values. This study uses a designed case and a real case, Yanshuixi Drainage in Tainan, Taiwan, to assess the uncertainty of drainage density in flood hazard mapping. Analytical results indicate that under the same return period rainfall, reduction in tributary drainages in a model (indicating a lower drainage density) results in an underestimate of the flooded area in tributary drainages. This underestimate causes higher peak discharges and total volume of discharges in the drainages, leading to flooding in certain downstream reaches, thereby overestimating the flooded area. The uncertainty of drainage density decreases with increased rainfall. We suggest that modeling flood hazard mapping with low return period rainfalls requires tributary drainages. For extreme rainfall events, a lower drainage density could be selected, but the drainage density of local key areas should be raised.