基于RSMSobol方法的HIMS模型参数敏感性分析  

Parameter Sensitivity Analysis for HIMS Model Based on RSMSobol Method

作　　者：党素珍[1,2] 楚楚[3] 蒋晓辉[1,2] 李艳霞 董国涛[1,2] 

机构地区：[1]黄河水利科学研究院,河南郑州450003 [2]水利部黄土高原水土流失过程与控制重点实验室,河南郑州450003 [3]黄河水利委员会上游水文水资源局,甘肃兰州730000 [4]中国水土保持杂志社,河南郑州450003

出　　处：《人民黄河》2016年第7期11-13,共3页Yellow River

基　　金：国家自然科学基金资助项目(91325201;41301030;41301496)

摘　　要：以黑河莺落峡以上流域为例,基于PSUADE平台,采用参数敏感性快速定量评估的RSMSobol方法分析了HIMS模型参数的敏感性。以体积误差Ve、Nash-Suttcliffe效率系数、标准误差σ、相对误差Re和相关系数r等作为目标函数综合评价了HIMS模型的模拟效果。结果表明:产流参数R和汇流参数C_1是重要参数,对于不同的目标函数,产流参数R都是最敏感的;模型参数敏感性对不同目标函数存在一定差异,产流参数R和r对体积误差和Nash-Sutcliffe效率系数表现得较为重要,而汇流参数C_1、C_2对相关系数、效率系数和标准误差的影响较为明显。Sensitivity analysis of hydrological models is helpful to understand the characteristics and improve the stability of model structure.In this study,the quantitative assessment of parameter sensitivity for HIMS model was achieved by using the RSMSobol method,based onPSUADE platform,through application to the upper reaches of the Heihe River basin. It selected five objective functions（i.e. volumetricerrorVe,Nash-Sutcliffe efficiency coefficientNSE,standard errorσ,relative errorReand correlation coefficient r）to assess the model asthe output responses for sensitivity analysis. The results show that the runoff yield parameterRand the flow concentration parameterC1 areimportant model parameters,andRis the most sensitive parameter for different objective functions;there are some differences of theparameter sensitivity for different objective functions,runoff yield parametersRandrare more important for the volumetric error andNash-Sutcliffe efficiency coefficient,and the flow concentration parametersC1andC2 have obvious effect on correlation coefficient,Nash-Sutcliffe efficiency coefficient and standard error,which provides a solid foundation for the calibration and parameter optimization ofHIMS model.

关 键 词：敏感性分析 参数 响应曲面 HIMS模型 黑河上游 

分 类 号：P333[天文地球—水文科学]