SWAT模型径流模拟的校正与不确定性分析

    Calibration and uncertainty analysis of runoff simulation by SWAT model

    • 摘要: 为比较SUFI-2和GLUE两种常用的SWAT模型校正与不确定性分析方法的优劣,以三峡库区大宁河流域为研究对象, 应用ArcSWAT 2012对大宁河流域进行径流模拟,并分别基于SUFI-2和GLUE方法进行校正和不确定性分析,进而从模型模拟结果的不确定性、拟合效果、运算效率和参数不确定性分析等方面对SUFI-2和GLUE进行比较。结果表明:SUFI-2和GLUE都可以较好地应用于大宁河流域SWAT模型的校正和不确定性分析,但SUFI-2方法能通过较少的模拟次数获得更小的模拟不确定性带以及更佳的拟合效果,而GLUE方法在参数不确定性分析方面的表现更佳。对于SWAT模型径流模拟,建议用SUFI-2方法进行校正和模拟结果不确定性分析,用GLUE方法进行参数不确定性分析。

       

      Abstract: SUFI-2 and GLUE are two widely used calibration and uncertainty analysis techniques for SWAT model in recent years. To compare the strengths and weaknesses of these two techniques, we choose Daning River watershed as the study area to simulate the runoff in watershed by using ArcSWAT2012 and conduct calibration and uncertainty analysis based on SUFI-2 and GLUE respectively. Furthermore, SUFI-2 and GULE were compared in terms of the uncertainty of SWAT output, the fitting effect, the computational efficiency and the parameter uncertainty. The results indicated that both SUFI-2 and GLUE can be well applied to the calibration and uncertainty analysis of SWAT model of Daning River watershed. However, the calibration and uncertainty analysis based on SUFI-2 could achieve smaller prediction uncertainty band and better fitting effect by less simulation numbers, and the parameter uncertainty analysis based on GLUE performs better than SUFI-2. For runoff simulation by SWAT model, the calibration and uncertainty analysis of simulation output is suggested to be conducted based on SUFI-2, while the parameters uncertainty analysis is suggested to be conducted based on GLUE.

       

    /

    返回文章
    返回