A river channel terrain reconstruction method for flood simulations based on coarse DEMs

To mitigate the unavailability of accurate river channel terrain data for flood modeling, this work develops a new approach to construct channel terrain based on available global coarse digital elevation models (DEMs). Compared to other established methods, the approach constructs the river surface using the sparse cross-section data from available coarse DEMs and not measurements. For a 134 km river reach of the Yangtze River, the absolute percentage error of simulated peak discharge with the coarse DEMs is 76.77%, and the delay time of peak discharge is 13 h; however, with the reconstructed terrain data are 12.13% and 0 h. They are 9.70% and -1 h on a 382 km river reach. The simulation results show that the approach can be used to reliably predict the flood propagation process without fine terrain data. The proposed approach can be used to improve flood management in areas without fine river terrain data.