基于元胞自动机的分布式洪水预报研究

Study of Distributed Flood Forecasting Based on Cellular Automata

洪水预报是防洪减灾的重要技术手段,鉴于我国北方半干旱半湿润地区的特殊性,以栅格DEM为计算单元,将每个栅格作为一个子流域,对基于超渗产流机制的大伙房模型进行栅格化,构建栅格型大伙房水库流域产流模型。采用元胞自动机模型实现汇流模拟,把栅格产流计算结果作为每个元胞的产流输入,将二维圣维南方程作为转换规则,充分考虑地表径流、壤中流、地下径流三种水源之间的蓄渗关系,实现元胞之间的水量交换,最终构建基于元胞自动机和大伙房模型的分布式洪水预报模型。采用30场历史典型洪水对模型进行参数率定及精度验证,利用空间信息和粒子群算法进行参数优选。研究表明:新建模型与大伙房模型具有相同的预报趋势,而且新建模型各项指标的拟合程度均比大伙房模型有所提高,说明利用新建模型进行洪水预报是可行的。

Flood forecasting is an important technical means of flood control and disaster reduction.However,due to the particularities of drought and semidrought in North China,current distributed flood model performance is poor in these areas.In this paper,the digital elevation model was used as the computing unit,with each grid cell treated as a subbasin and the Dahuofang model was built based on the mechanism of superinfiltration flow.The cellular automaton model was used to achieve concentration simulation by using the flow calculation result for runoff in each grid cell as input and the twodimensional Saint Venant equations as transformation rules.Surface runoff,interflow,and underground runoff for three kinds of water infiltration,plus water exchange between cells were fully considered and finally a distributed flood forecasting model based on the cellular automaton model and the Dahuofang model was built.The Dahuofang Reservoir was used as a case study and 30 cases of typical floods were used to determine the parameters and accuracy of the models and spatial information and particle swarm optimization were used to optimize the parameters.The results show that the new model and the Dahuofang model can produce the same trend forecast,but the new model is performed better,which indicates that use of this new model for flood forecasting is feasible.