城市暴雨径流模拟的参数不确定性研究  

Parameters uncertainty analysis of urban rainfall-runoff simulation

作　　者：赵冬泉[1,2] 王浩正[2] 陈吉宁[1] 王浩昌[1] 

机构地区：[1]清华大学环境科学与工程系,北京100084 [2]北京清华城市规划设计研究院环境与市政所,北京100084

出　　处：《水科学进展》2009年第1期45-51,共7页Advances in Water Science

摘　　要：基于暴雨管理模型(SWMM),构建了澳门半岛雅廉访实验小区降雨径流模拟系统,采用generalized likelihood uncertainty estimation(GLUE)方法对复杂降雨径流模型参数的不确定性进行了分析,采用均匀分布为参数的先验分布(prior distribution),选取拉丁超立方法作为采样方法,Nash-Sutcliffe效率系数作为似然评判准则。结果表明该方法可以有效的分析降雨径流模型参数的不确定性,对降雨径流模型参数识别提供了深入分析与理解模型系统的有效途径。Urban rainfall-runoff model plays an important role in water resource management. Based on the storm water management model （SWMM）, a drainage simulation system is established in Yalianfang catchment of Macao. The generalized likelihood uncertainty estimation （GLUE） approach is selected to analyze the parameters uncertainty of the complex rainfallrunoff model. Latin hypercube method is used to obtain sample sets, and Nash-Sutcliffe efficiency coefficient is used as the criteria of assessing the fitness between monitored time series and predicted one. The uniform pattern is determined as prior distribution of parameters. The results show that GLUE method is an efficient method to analyze parameters uncertainty of rainfall-runoff model, and the analysis results can offer information for further understanding to the modeling system.

关 键 词：暴雨径流模拟 不确定性 普适似然度方法 暴雨管理模型 

分 类 号：TV122[水利工程—水文学及水资源]