Flood forecasting scheme of Nanshui reservoir based on Liuxihe model

Abstract

China experiences one of the most frequent flood disasters in the world. Establishing accurate and reliable flood prediction program is the key to deal with flood disasters. Nanshui Reservoir Basin, in southern China, belongs to subtropical monsoon climate, with more rain in spring, concentrated rainstorm in summer and typhoon storm in autumn. Floods at dam site are mostly small and medium-sized floods with steep rise and slow fall as typical mountain flood. In order to explore the applicability of Liuxihe model in flood prediction of Nanshui Reservoir, this paper builds up Liuxihe model for Nanshui Reservoir based on DEM, land use and soil type data, and selects a typical flood event to optimize the parameters using particle swarm optimization (PSO) algorithm and verifies the accuracy of the model by simulating the other floods. Liuxihe model established in this paper indicates a satisfactory performance for flood prediction for Nanshui Reservoir, which can meet the accuracy requirement of flood prediction. Finally, the effects of different river grading and PSO algorithm on flood prediction are discussed. The results show that the PSO algorithm can obviously improve the accuracy of the Liuxihe model for flood forecast in Nanshui Reservoir. The simulation based on four-level channel grading has better results than that based on three-level channel, which indicates increased peak flood value, delayed peak time and closer simulation to the measured value.