基于SODA方法的HyMOD模型不确定性分析

为了改进水文建模过程中的不确定性处理,采用一种融合全局优化和数据同化(Simultaneous Optimization and Data Assimilation,SODA)的混合框架,对Hy MOD模型进行了不确定性分析,并与经典SCEM-UA方法进行了比较。SODA方法具有如下特点:(1)具备较高的参数搜索效率和寻优能力;(2)明确考虑包括输入、输出、参数以及模型结构在内的重要不确定性来源。SODA方法在渭河流域的实例应用结果表明:与SCEM-UA方法相比,SODA方法不仅显著提高了预报精度,而且推求出了性质更为优良的预报区间。SODA方法的成功应用,有助于模型概念的改进及对水文系统功能的理解。

Uncertainty Assessment of HyMOD Model Using the Method ofSimultaneous Optimization and Data Assimilation

To improve the treatment of uncertainty in hydrological modeling, a hybrid framework of simultaneous optimization and data assimilation (SODA) was adopt to assess the uncertainty of HyMOD model in this paper, and then was compared with the classical method of the Shuffled Complex Evolution Metropolis-UA (SCEM-UA).The strengths of the SODA can be described as follows: (1) high parameter search efficiency and explorative capabilities;(2) explicit treatment of the various important sources of uncertainty (i.e., input, output, parameter and model structure uncertainties) associated with the application of hydrological models.The results of the SODA applied in the Weihe River Basin demonstrate that in comparison to the performances of SCEM-UA, the SODA could notably improve the streamflow prediction efficiency, and also could derive more accurate prediction interval.The successful application of the SODA is helpful to improving model concepts and understanding of the functioning of hydrological systems.