基于Sobol方法的SWMM模型参数敏感性分析

Sensitivity analysis on SWMM model parameters based on Sobol method

SWMM模型已被广泛应用于城市区域雨洪模拟及海绵城市设计,为提高参数优化效率,减少模型不确定性的影响,有必要对参数敏感性进行定量分析。以济南市黄台桥排水片区为例,选用基于方差分解的Sobol方法,定量分析不同量级洪水情景及不同预报精度评价标准下SWMM模型参数的敏感性。研究结果表明:(1)SWMM模型参数的敏感性在不同量级洪水和不同目标函数的情况下表现均不相同,模型参数具有较大的不确定性;(2)不透水区洼蓄量、最小入渗率、河道及管道糙率系数对Nash效率系数作用显著;(3)洪水量级相对较小时,SWMM模型参数间相互作用现象较为明显,但随着洪水量级的增加,参数相互作用逐渐减小;(4)Sobol方法能定量分析模型参数敏感性及参数间相互作用,适用于城市水文模型敏感性相关研究。

The SWMM model has been widely used in urban rainfall-runoff simulation and design of sponge cities. A quantitative analysis on the parameters of this model is crucial to parameter identification and uncertainty reduction. In this study, a variance-based Sobol method and different accuracy assessment standards are used to quantitatively analyze the sensitivity of the SWMM model parameters in a case study of the Huangtaiqiao catchment under floods of different levels. Results show that （1） the parameters are different in sensitivity when accuracy assessment standard or flood level is varied, indicating a great uncertainty in model parameters. （2） The Nash coefficient depends much on Dstore-Imperv, MinRate, Roughness （rivers）, and Roughness （conduits）. （3） Interaction between the parameters is frequent in small flood cases, while it diminishes with a trend toward concentration as flood level is raised. （4） The Sobol method, suitable for the sensitivity analysis of urban hydrological models, can provide the sensitivity of each parameter and the interaction between different parameters.