摘要
SWAT模型是基于水文过程的、具有很强物理机制的流域分布式水文模型.SWAT2003以SWAT2000为基础,增加了灵敏度分析、自动校准和不确定分析3个全新模块.该文主要介绍SWAT2003模型灵敏度分析模块的原理,并在罗玉沟流域应用与验证.通过灵敏度分析,辨析出影响该流域产流产沙模拟结果精度的主要参数因子:径流曲线数(CN2)和土壤可利用水量(SOL-AWC)对罗玉沟流域模拟产流产沙值的影响最为显著,且对产沙值的影响大于对径流值的影响;而基流α系数(ALPHA-BF)、土壤蒸发补偿系数(ESCO)、坡度因子(SLOPE)和坡长因子(SLSUBBSN)对模拟产沙值的影响也十分显著,但对径流值的影响并不明显.依据上述结论调整参数值,采用连续15年的实测月流量数据对模型进行了校准和验证,结果表明,模型对月流量模拟的相对误差在模型校准期和验证期均小于15%,相关系数R2和Nash-Suttcliffe系数Ens均高于0.8,SWAT2003模型对罗玉沟流域产流的模拟结果良好.因此,灵敏度分析模块的应用极大地提高了模型在黄土丘陵沟壑区的可用性,具有一定的推广意义.
The soil and water assessment tool(SWAT) is a physically-based continuous event hydrologic model.Besides all the features included in SWAT2000,the latest SWAT2003 version also includes sensitivity analysis,automatic model calibration and uncertainty analysis.The sensitivity analysis of the SWAT model was expatiated and applied to simulate the rainfall and runoff relationship in Luoyugou Watershed in the Loess Plateau.Based on the sensitivity analysis, the main parameters affecting the simulated values in this watershed were determined.The results showed that the CN2 and(SOL_-AWC) parameters had the most significant influences on the simulated values of runoff and sediment yield in Luoyugou Watershed.Besides,the effect on sediment yield was higher than that of runoff.In addition,other parameters,such as(ALPHA_-BF,) ESCO,SLOPE and SLSUBBSN,had striking influences on the simulated sediment yield either,but little on the simulated runoff.Adjusting the parameters based on the analyzed results,satisfied simulation results were obtained with both of Nash-suttcliffes coefficient(E_(ns)) and correlation coefficients R^2 exceeding 0.8,and the relative error was under 15% over 15 years continuous simulations on monthly discharges.Therefore,the application of sensitivity analysis module of SWAT model in typical watershed in the Loess Plateau has improved the accuracy of simulation dramatically.
出处
《北京林业大学学报》
CAS
CSCD
北大核心
2007年第S2期238-242,共5页
Journal of Beijing Forestry University
基金
国家重点基础研究发展规划资助项目(2002CB111502)