基于局部和全局方法的SWMM敏感参数识别

为了识别SWMM（storm water management model）模型的敏感参数，从而实现参数的高效率定，本文以深圳河流域为研究对象，构建SWMM模型，分别采用修正的Morris筛选法和互信息法，从局部和全局的角度定量分析重现期1年、10年和50年设计暴雨情景下排放口洪峰流量和流域平均径流系数对各参数的敏感性。结果表明，在不同的设计暴雨情景下，两种方法的结果均显示排放口洪峰流量对透水区曼宁系数和最小下渗速率最敏感，流域平均径流系数对下渗相关参数最敏感；然而，随着暴雨强度增大，两种方法计算得到的径流系数对最大下渗速率和最小下渗速率的敏感性变化趋势不同，Morris的结果显示递减，M-I的结果显示递增。

Identification of sensitive parameters of SWMM based on local and global methods

To identify the sensitive parameters of a storm water management model (SWMM) and thus achieve their efficient calibration, this study develops a SWMM for the Shenzhen River basin. Both the modified Morris screening method and mutual information (M-I) method are separately used to quantitatively analyze the sensitivity of the peak flow at the drainage outlet and the average runoff coefficient of the basin to the model parameters under design rainstorms of return periods of 1 year, 10 years, and 50 years. The results of both methods reveal that for different design rainstorms, the peak flow is most sensitive to the Manning roughness coefficient and minimum infiltration rate of the permeable area, and the runoff coefficient is most sensitive to the parameters related to infiltration. However, with rainstorm intensity increasing, the sensitivity of runoff coefficients to the maximum or minimum infiltration rates manifests different trends calculated by different methods. While Morris method gives a decreasing trend, the M-I method gives the opposite.