基于物理水文模型的不确定性分析  

Uncertainty analysis of geological models based on physical hydrological model

作　　者：张平[1,2] 夏军[1,2] 邹磊[1,2] 黄金凤[1,2] 

机构地区：[1]武汉大学水资源与水电工程科学国家重点实验室,湖北武汉430072 [2]武汉大学水资源安全保障湖北省协同创新中心,湖北武汉430072

出　　处：《武汉大学学报（工学版）》2016年第4期481-486,共6页Engineering Journal of Wuhan University

基　　金：国家自然科学基金(编号:51279140);中央高校基本科研业务费专项资金(编号:2042014kf1012;2042014kf0033);国家"十二五"水专项淮河课题(编号:2014ZX07204-006)

摘　　要：基于物理基础的水文模型预测精度受不确定性影响显著,分析这些不确定性因素对模型预测精度的提高十分必要.目前对于水文模型不确定性研究比较成熟,众多专家已经提出了模型不确定分析的框架,同时设计了一些模型优化算法,但较为缺乏针对模型结构不确定性研究的系统内容.通过选择具有物理基础的水文模型——HYDRUS,分别开展模型参数和结构的不确定性研究.结果表明,详实的观测信息是模型选择不可缺少的;不同时间点的模型信息对模型的确定更为有用;在选择模型时,应充分了解真实结构信息,减少因模型结构误差所带来的模型预测不确定性.在目标函数的选择上,也应充分考虑模型结构的不确定性.要减少模型不确定性,必须加强对模型参数和结构的认识.The uncertainty of the model has a significant effect on model prediction,so to improve the model prediction accuracy analysis of these uncertainties is necessary.The uncertainty studies of the physical model is relatively mature,some experts have proposed a framework for model uncertainty analysis and design optimization of some models.Recent study of the parameters uncertainty for the model is relatively mature.While the study of model structure uncertainty is relatively rare.So this paper presents research of parameters uncertainty,model structure by selecting the hydrological model-HYDRUS model which has aphysical basis.The results show that sufficient observation information is indispensable for true model selection;the model information of the different time points is more useful.When selecting the model,we should fully understand the real structure information of model to reduce errors caused by the structure uncertainty.On the choice of the objective function we should also fully consider the uncertainty due to model structure.To reduce the model uncertainty,we must strengthen the understanding of structure and parameters of the model.

关 键 词：模型不确定性 水力参数 参数反演 

分 类 号：TV93[水利工程—水利水电工程]