耦合SWAT与RIEMS模拟黑河干流山区径流

以黑河干流山区为研究区,采用1∶100 000植被类型图、1∶1 000 000土壤类型图和气象水文观测数据,耦合SWAT水文模型与RIEMS高分辨率区域气候模式,模拟1995-2010年月径流变化过程,探讨水文模型气象驱动数据的优化方法和气候水文模型耦合的区域适宜性。RIEMS气候模式输出精度较高,降水、温度、湿度、风速的相关系数均在0.80以上,均通过了0.01显著性水平检验,时空分辨率达到6 h和3 km。构建虚拟气象站点,弥补气象观测站点稀少且分布不均匀的不足,对水文模型气象驱动数据进行优化;遵循多时间尺度、多变量和多站点的原则来校准模型。结果表明,径流模拟值与观测值的过程趋势拟合程度较好,NSE均在0.60以上,PBIAS介于±20%之间,R2达到0.70以上。径流模拟在枯水期表现较好,在丰水期存在一定的误差,主要是受降水驱动数据偏高的影响,气候模式模拟能力需要提高,水文模型空间插值方法和气候水文模型耦合方案需进一步完善。总体来看,耦合SWAT模型与RIEMS模式能够较好地模拟黑河干流山区水文过程,可为流域水资源的预测和管理提供科学依据。

Coupling SWAT and RIEMS to simulate mountainous runoff in the upper reaches of the Heihe River basin

Weather stations are scarce and unevenly distributed in the inland river basins of northwest China,which limits the application of hydrological simulation.A high-resolution regional climate model can export accurate climate data to improve the weather-driven data of a hydrological model.Therefore,coupling a climate model and a hydrological model is necessary to quantify water resources in inland river basins.In this study,we selected the mountainous region in the upper reaches of the Heihe River basin (HRB) as the study area to simulate the monthly runoff process from 1995 to 2010 by coupling the Soil & Water Assessment Tool (SWAT) hydrological model and the Regional Integrated Environmental Model System (RIEMS) climate model.We studied the optimization scheme for the weather-driven data and the suitability of the coupling of the climate and hydrological models.The data exported by RIEMS exhibited high accuracy.The correlation coefficients of precipitation,temperature,humidity,and wind speed were all higher than 0.80 and achieved a significance level of 1％.Spatial and temporal resolution reached 6 h and 3 km,respectively.Virtual weather stations built to compensate for the shortage of climate stations,which were scarce and unevenly distributed in the study area,optimized the weather-driven data from SWAT.A model calibration method was adopted using multi-temporal scales,multiple variables,and multiple sites that incorporated the natural characteristics of the basin.Results showed that the simulation values were close to the observation values.The Nash Sutcliffe efficiency was higher than 0.64,the percent bias was controlled within 20％,and the determination coefficient value was higher than 0.70.Runoff simulation demonstrated good performance during dry seasons,but failed to provide accurate results during wet seasons for some of the years because of high precipitation from RIEMS.Thus,the simulation capability of RIEMS and the model coupling method still require further improvement.Overall,the coupling of SWAT and RIEMS was suitable for simulating mountainous runoff in the upper reaches of HRB.The model can accurately describe the hydrological processes of the basin.The results can provide references for forecasting and managing the water resources of the basin.