分布式水文模型的参数率定及敏感性分析探讨

参数率定与敏感性分析是分布式水文模型应用和发展中的难点问题,论文对当前典型的、应用比较成功的全局最优化参数率定和敏感性分析方法进行归纳和分析,包括:遗传算法(Genetic Algorithm)、SCE-UA算法(Shuffled Complex Evolution)、贝叶斯方法(Bayesian Method)、RSA方法(Regionalized Sensitivity Analysis)、GLUE方法(Generalized Likelihood Uncertainty Estimation)等等。并对计算机自动优化方法和人工参数调试方法的利弊进行讨论,展望了分布式水文模型的参数率定与敏感性分析的发展方向。

Comments on Sensitivity Analysis, Calibration of Distributed Hydrological Model

The sensitivity analysis and calibration is a key and difficult issue to the application and development of the distributed hydrological model.In this paper,several typical and effective global optimization calibration and sensitivity analysis methods are inducted,which include the flowing methods:Genetic Algorithm(GA),Shuffled Complex Evolution(SCE),Bayesian Method (BM), Regionalized Sensitivity Analysis(RSA),Generalized Likelihood Uncertainty Estimation(GLUE) and so on. The advantages and disadvantages between the computer automatic calibration and the artificial calibration are discussed, and then we prospect the future development of the sensitivity analysis and calibration of the distributed hydrological model.