清江流域SWAT模型多出口率定与单出口率定的比较

针对传统SWAT模型参数率定方法是以流域总出口为控制点的单出口率定,忽略了各子流域防洪要求以及不同子流域水文特征空间变异性的问题,利用清江流域26个雨量观测站及5个流量观测站1993~1995年的日资料数据,建立了基于SWAT模型的清江流域分布式水文模型,以各流量站点为子流域出口将整个流域划分为5个子流域,选取不同数目的控制点(2~5个)构成不同的多出口率定情景。与单出口率定结果对比发现,各多出口率定情景下流域总出口处的模拟精度与单出口率定结果基本相似,但在4个控制点的率定情景下,子流域出口处的模拟精度在率定期、验证期分别提高了11%~20%、2%~17%。表明多出口率定能在不降低流域总出口模拟精度的前提下提高各子流域的模拟精度;且对清江流域来说,选取4个控制点进行多出口率定时,模拟效果最好。

Comparison of Multi-gage and Single-gage Calibration of SWAT Model for Runoff Simulation in Qingjiang River Basin

Traditional way of calibrating the SWAT model was performed based on the observed flow data at a single gage which is the outlet of the basin. However, this may result in the issue that the requirements of flood control and the spatial variations of hydrological characteristics at different sites of the basin were ignored. Therefore, the Qingjiang River SWAT model was established by depending on the 26 precipitation stations and 5 flow gages with 3-year (1993-1995) daily observed records. The whole basin was divided into five sub-basins by using the 5 flow gages as the outlets of each sub-basins. Different numbers (from 2 to up to 5) of flow gages were used to perform different multi-gage calibration scenarios. Comparison of the simulation accuracies with single-gage calibration, the results show that the flow simulation accuracies at the outlet of the whole basin are similar between all multi-gage (from 2 to 5 gages) calibration scenarios and single-gage calibration; but in the 4 gages calibration scenario, the flow simulation accuracies at the sub-basin outlets are improved by a factor of 11%-20% for calibration and 2%-17% for validation. It demonstrates that multi-gage calibration can improve the flow simulation accuracies at the sub-basin outlets without reducing the simulation accuracy at the outlet of the whole basin. And the 4 gages calibration scenario is the best scenario for Qingjiang river basin.