水文模型参数敏感性分析方法评述  

Review of methods of parameter sensitivity analysis in hydrologic modeling

作　　者：宋晓猛[1,2,3] 张建云[1,2] 占车生[4] 王小军[1,2] 刘翠善[1,2] 

机构地区：[1]南京水利科学研究院水文水资源与水利工程科学国家重点实验室,江苏南京210029 [2]水利部应对气候变化研究中心,江苏南京210029 [3]中国矿业大学资源与地球科学学院,江苏徐州221116 [4]中国科学院地理科学与资源研究所陆地水循环及地表过程重点实验室,北京100101

出　　处：《水利水电科技进展》2015年第6期105-112,共8页Advances in Science and Technology of Water Resources

基　　金：国家重点基础研究发展计划(973计划)(2010CB951103);国家科技重大专项(2014ZX07204-006);国家自然科学基金(51309155)

摘　　要：针对水文模型敏感性分析中存在的诸多问题,分析水文模型参数敏感性分析在模型构建及应用过程中的主要作用及其与不确定性分析和参数优化之间的联系,总结敏感性分析方法的3种分类,并探讨水文模型中常用的筛选法、回归分析法、基于方差的分析方法及基于代理模型技术的分析方法等4种关键技术方法,剖析水文模型参数敏感性分析方法的适用条件及优缺点,回顾各种方法在水文模型中的研究现状,提出水文模型参数敏感性分析的研究框架与步骤,指出水文模型参数敏感性分析的计算效率、可靠性和参数的相关性是未来的主要研究方向。With a focus on many problems that exist in parameter sensitivity analysis, this paper first analyzes the model construction of hydrological model parameter sensitivity analysis and the main functions in the process of its application. This paper also focuses on the relationship between uncertainty analysis and parameter optimization. Three methods of classification of sensitivity analysis are summarized. Four key technological methods of analysis commonly used in hydrological models are discussed. The four key technological methods include the Morris screening method, regression-based analysis, the variance-based method, and surrogate modeling approaches. At the same time, we discuss the applicable conditions for different approaches in hydrological models, summarize the advantages and disadvantages of various methods, and review the research status of various methods used in the hydrological model. In addition, a synthesized evaluation framework for sensitivity analysis for hydrological modeling is introduced. Several future research issues are also proposed.

关 键 词：水文模型 敏感性分析 参数识别 不确定性分析 参数优化 

分 类 号：P333[天文地球—水文科学]