A novel rapid flood inundation method based on GIS combined with hydrodynamics in coastal and riverine regions

Abstract Under climate change, flood-induced inundation becomes frequent in low-lying coastal and riverine regions as a result of an increasing storm surge and sea-level rise. An efficient and robust inundation assessment model is necessary for flood control, coastal management and inundation evacuation. Due to the complexity and runtime consumption of hydrodynamic models, a series of rapid flood inundation methods (RFIMs) have been developed based on high-resolution digital elevation models (DEMs) and advanced geographic information system (GIS) technology. However, a shortage of hydrodynamic mechanisms restricts the application of the RFIM. This paper proposes a novel rapid flood inundation method based on a combination of GIS and hydrodynamics for application in coastal and riverine regions. In a common-sense approach, the tidal flood flow first propagates along the river network and then diffuses from the network to the floodplain. First GIS-based algorithms are used to extract hydrodynamic model parameters from the DEM. Specifically, it includes the extraction of the flow network and the cross section bottom level on each flow segment, and the discretization of the main terrain into a triangulated irregular network (TIN) and the division of the catchments. Next, a one-dimensional hydrodynamic network model is applied to calculate flood flow propagation on the complex network. The model solves the full hydrodynamic equations to capture the dynamic process of flooding wave propagation such as wave amplification and damping. Last, a simple elevation-determined and domain-connected algorithm on the triangulated irregular network is implemented to obtain the extent of the flood inundation. The results indicate that the proposed method can estimate a reasonable flooding extent with a good degree of agreement with that by the three-dimensional hydrodynamic model. However, the runtime of the proposed method is significantly reduced by orders of magnitude at the same resolution compared to the runtime of the hydrodynamic model. Moreover, the effect of the banks, which is a difficult problem for many hydrodynamic models, can also be solved without local refinement of the grids. Therefore, our main conclusion is that the proposed RFIM is a useful tool for estimating inundation processes in coastal and riverine regions with limited input data and low computational burden.