基于全局和局部方法的城市雨洪模型参数敏感性——以济南市主城区为例  

Parameter Sensitivity of Urban Storm Models Based on Global and Local Optimum Methods—A Case Study in the Main Urban Area of Jinan City,China

作　　者：唐清竹 徐宗学[1,2] 王京晶 赵刚 李鹏[1,2] TANG Qing-zhu;XU Zong-xue;WANG Jing-jing;ZHAO Gang;LI Peng(Beijing Key Laboratory of Urban Hydrological Cycle and Sponge City Technology,Beijing Normal University,Beijing 100875,China;College of Water Sciences,Beijing Normal University,Beijing 100875,China;School of Geographical Sciences,University of Bristol,Bristol BS81TH,UK)

机构地区：[1]北京师范大学城市水循环与海绵城市技术北京市重点实验室,北京100875 [2]北京师范大学水科学研究院,北京100875 [3]布里斯托大学地理科学学院,布里斯托BS81TH

出　　处：《地球科学与环境学报》2022年第4期713-722,共10页Journal of Earth Sciences and Environment

基　　金：国家重点研发计划项目(2017YFC1502701)。

摘　　要：开展模型参数敏感性分析是进行参数率定及提高模型模拟效果的重要前提。以山东省济南市主城区为研究对象,基于SWMM模型构建城市雨洪模型,从全局和局部角度分别采用GLUE方法与修正Morris筛选法,以流域出口流量数据为模型率定依据,对不同暴雨情景下SWMM模型中11个参数进行敏感性分析。结果表明:①GLUE方法筛选出河道糙率系数(RoughnessR)、衰减常数(Decay Constant)、不透水区洼蓄量(S-Imperv)为敏感参数;②采用修正Morris筛选法时,不同暴雨情景下各参数对径流总量的影响大致相同,对洪峰流量影响不尽相同,对径流总量表现出中高敏感性的参数有河道糙率系数、排干时间(Drying Time)、衰减常数、最大入渗率(MaxRate)、最小入渗率(MinRate)、透水区洼蓄量(S-Perv)、管道糙率系数(RoughnessP),对洪峰流量表现出中高敏感性的参数有河道糙率系数、衰减常数、透水区曼宁系数(N-Perv);③两种方法均可准确识别SWMM模型参数敏感性,均显示河道糙率系数与衰减常数对模拟结果的影响较为显著,具体区别在于GLUE方法可考虑参数间相互作用,判断参数是否为敏感参数,而修正Morris筛选法可对参数敏感性进行量化分级并排序。Sensitivity analysis of model parameters is an important prerequisite for urban flood modeling.An urban hydrological model named Storm Water Management Model(SWMM)was developed in the main urban area of Jinan city,China.The generalised likelihood uncertainty estimation(GLUE)method and the modified Morris method were used for sensitivity analysis respectively from both global and local perspectives,and the discharge series at the watershed outlet was used as the calibration and validation data.The sensitivity of SWMM model parameters under different storm scenarios was analyzed.The results show that①RoughnessR,Decay Constant,S-Imperv are selected as sensitivity parameters by using the GLUE method;②the influence of each parameter on the total runoff by using modified Morris method is roughly the same under different storm scenarios,but the influence on the peak flow is different;RoughnessR,Drying Time,Decay Constant,MaxRate,MinRate,S-Perv,and RoughnessP show high sensitivity to the total runoff,and RoughnessR,Decay Constant and N-Perv show high sensitivity to the peak flow;③both GLUE and modified Morris methods can accurately identify the high sensitive parameters of SWMM model;the uncertainty of RoughnessR and Decay Constant has a significant influence on the simulation results;the difference is that the GLUE method can be used to determine whether the parameters are sensitive,and the modified Morris method can be used to quantitatively rank the parameter sensitivity.

关 键 词：SWMM模型 参数 敏感性 GLUE方法 Morris筛选法 暴雨 流量 山东 

分 类 号：TV122[水利工程—水文学及水资源]