基于Extend FAST方法的新安江模型参数全局敏感性分析  

Global Sensitivity Analysis of Xinanjiang Model Parameters Based on Extend FAST Method

作　　者：任启伟[1] 陈洋波[1] 舒晓娟[1,2] 

机构地区：[1]中山大学地理科学与规划学院,广东广州510275 [2]广东水利电力职业技术学院,广东广州510635

出　　处：《中山大学学报（自然科学版）》2010年第3期127-134,共8页Acta Scientiarum Naturalium Universitatis Sunyatseni

基　　金：国家自然科学基金资助项目(50479033);广东省科技计划基金资助项目(2009A030302003)

摘　　要：模型参数敏感性分析可以诊断模型结构、识别模型关键参数,是模型建立和应用的关键步骤。该文应用高效、稳健的Extend FAST方法,以流溪河水库入库洪水模拟为例分析新安江模型参数的全局敏感性。结果表明,对Nash-Sutcliffe效率系数、总水量平衡误差系数、低水流量误差系数、高水流量误差系数四个目标函数,模型参数敏感性表现不同。Extend FAST方法的部分结果与GLUE方法的结果进行对比,两者对敏感性评价的结果基本一致,Extend FAST给出的主要敏感度和参数组合作用的敏感度也可以作为模型不确定性分析的依据。Sensitivity analysis is an efficient approach to diagnose model structure and identify the dominant parameters controlling model behavior.It is also a key step in modeling practice.The global sensitivity of parameters of the Xinanjing model is analyzed using Extend FAST method in the Liuxihe reservoir catchment.Results show that the sensitivity of four object functions perform is different.The sensitivity can be influenced by many factors,such as the initial condition and flood type.The results are compared with the scatter map generated by the GLUE uncertainty method.Both methods can get comparable sensitivity analysis results.The main effect sensitivity and interaction effect sensitivity calculated by Extend FAST also supply quantitative indices to help understand the uncertainty in hydrological model.

关 键 词：EXTEND FAST 新安江模型 全局敏感性 GLUE 不确定性 

分 类 号：P349[天文地球—水文科学]