基于SUFI-2算法的SWAT模型在修水流域径流模拟中的应用

为在江西修水流域高沙水文站以上汇水区域构建SWAT分布式水文模型，采用SUFI-2算法进行模型参数敏感性分析、率定、验证及不确定性分析，对修水流域1984~2003年月径流量过程进行模拟。结果表明，敏感度最强的前三个参数依次为土壤饱和导水率、土壤蒸发补偿系数和径流曲线数；率定期与验证期的纳什系数ENS值和相关系数R2值均高于090，月径流量模拟值与实测流量过程线拟合程度较好；95PPU置信区间的模拟值和实测值较为一致；不确定性区间较小，验证期小于率定期；实证了采用SUFI-2算法的SWAT模型在修水流域月径流模拟中具有一定程度的适宜性。

Application of SWAT Model Based SUFI-2 Algorithm to Runoff Simulation in Xiushui Basin

In order to simulate monthly runoff in the upstream catchment of Gaosha hydrological station of Xiushui River Basin from 1984 to 2003, a distributed hydrological model was developed with SWAT, calibrated and validated with SUFI-2 algorithm. Sensitivity and uncertainty analysis were performed to improve model performance. The results show that the first to third sensitive parameters were as follows: the soil saturated hydraulic conductivity, soil evaporation compensation coefficient and flow curve number; efficiency coefficient values ENS and the correlation of determination coefficient R2 values were higher than the value of 0.90 during the calibration and validation period; the monthly runoff simulation value with the actual flow process line fitting degree was good; in 95PPU confidence interval, the simulation value was close to the actual value; the uncertainty interval was smaller, and the validation period of uncertainty was less than the rate of calibration; the SWAT model based SUFI-2 algorithm was suitable for monthly runoff of Xiushui River Basin in a certain degree.