Calculation and visualization of flood inundation based on a topographic triangle network

Abstract

Summary Simulations of flood inundation have important practical significance for flood control and disaster alleviation. Accuracy and efficiency are two major indicators in evaluating the performance of flood inundation models. Although a flood inundation model based on hydrodynamics can use detailed topographic data to accurately evaluate flood inundation processes, it is difficult to build a reasonable model for complex topography and variable flood conditions. More importantly, the simulation time for these types of models is generally too long and cannot be accomplished in real time. This paper proposes a new method, named the flood-connected domain calculation (FCDC) method, to improve the efficiency of flood inundation models in regions of complex topography. The main concept of the method is that inundation can only occur in areas where the water level is higher than the terrain elevation and connected with the river; this requires the key development of an algorithm to search for connected domains. The FCDC method uses a topographic triangle network to define the DEM source data and topological relationships are constructed to undertake the flood-connected domain search. Two cases from the Songhua River in Harbin City, China have been examined to expand the usage and effect of the method. The results show that the FCDC method can simulate source flooding such as inundation by river flooding or dike breach flooding and can characterize inundation under a range of conditions. Also, it can use a range of resolutions for topographic data and water level as model input, which in turn significantly reduces simulation time while providing reasonable simulation results. In addition, the FCDC method can be used to obtain a smooth display effect for flood inundation processes. Overall, the FCDC method is useful for estimating flood inundation and preparing for emergencies because of its time-efficient performance and low input and hardware requirements.