基于HYDRUS-1D生物滞留设施模型的参数敏感性分析及率定

Parameters Sensitivity Analysis and Calibration of HYDRUS-1D Model for Bioretention

为验证HYDRUS-1D模拟生物滞留设施水文调控效应的可靠性,通过物模试验,采用修正Morris法对模型主要参数的敏感性进行分析筛选,结合模拟值和实测值对筛选后的参数进行率定验证。结果表明:径流总量、径流峰值及其各削减率输出结果的敏感参数不尽相同,其中降雨量均为高敏感参数;利用变异系数法,判断参数的综合敏感程度,其主要敏感参数包括降雨量,蓄水层高度,种植土层和基质层的饱和含水率、饱和导水率和孔径分布参数;验证期内模型模拟和物模实测结果的均方根误差(root mean square error,RMSE)最高为0.131、相对误差(relative error,RE)均低于6.92%、纳什效率系数(Nash-Sutcliffe efficiency coefficient,NSE)和决定系数R2均为0.9以上,匹配度高,模拟结果可靠,为利用HYDRUS-1D改进生物滞留设施规划设计方法提供理论支撑。

In order to verify the reliability of hydrologic regulation effect of HYDRUS-1D simulated bioretention facilities,the sensitivity of main parameters of the model was analyzed and screened by the modified Morris method through the physical model experiment,and the model parameters were calibrated by combining simulated and measured values.The results show that the output sensitive parameters of the total runoff,peak runoff and each reduction rate are different,which rainfall(q)are all highly sensitive parameters.The variation coefficient method was used to determine the comprehensive sensitivity degree of the parameters which main sensitive parameters include rainfall(q),aquifer height(H),as well as saturated water content(θ_(s)),saturated hydraulic conductivity(K_(s))and pore connectivity parameters(n)of planting soil and filling layer.During the validation period,the maximum root mean square error(RMSE)of simulated and measured results is 0.131,the relative error(RE)is lower than 6.92%,and the Nash-Sutcliffe efficiency coefficient(NSE)and the coefficient of determination R^(2)are both above 0.9,indicating high matching degree and reliable simulation results.It providing theoretical support for improving the planning and design method of bioretention facilities by using HYDRAS-1D.