本文分析了SWAT模型和PSO算法的原理,将PSO算法引入SWAT模型中,构建了新的SWAT模型参数自动率定模块,通过在天津蓟运河流域实例研究,发现该方法率定精度较高,收敛速度快,运行结果稳定,整体率定效果优于模型自带的参数率定模块。如果用改进后的模块在Linux平台开展自动率定,可以使模型自动率定效率提高到当前水平的7倍,适用于大型流域或长时间系列模拟。而PSO算法作为一种通用的优化算法,可广泛用于各种水文模型的参数率定。
The SWAT model is a complex distributed hydrological model and the parameter calibration is crucial for applying the model effectively. Although the latest version of SWAT model SWAT2005 has an auto-calibration module, but its computational efficiency is low which makes the auto-calibration of the model for large watershed difficult to be accomplished. In this paper, a new auto-calibration module applying the PSO(Particle Swarm Optimization) method was developed and was applied to the Jiyunhe watershed located in Tianjin. Compared with the results conducted by the original SCE auto-calibration module in the SWAT, the new auto-calibration method can calibrate the parameters faster than the SCE method while retain the accuracy and stable results, which means that the new auto-calibration module's performance is better than the SCE module. Meanwhile, if the new auto-calibration module runs in the Linux OS, the total calibration efficiency will he improved by 6 times of present auto-calibration approach, which makes the calibration of SWAT model for large area and long time series more convenient and promising. As a sophisticated optimization method, the PSO method can be used to calibrate various hydrological models.