以鄱阳湖流域为研究区,构建其分布式水文模型WATLAC。以鄱阳湖流域6个水文站点2000~2005年河道径流量及地下水基流指数为多目标函数来优化耦合模型,采用PEST进行WATLAC模型参数自动率定。结果表明:6个水文站点拟合的纳希效率系数变化范围为0.71~0.84,确定性系数介于0.70~0.88,取得较好的模拟效果。基于遥感反演的实际蒸散发结果进一步验证模型,模拟值和遥感反演值在时间和空间上的变化趋势呈现较好一致性。基于地下水位及流场特征模拟效果的定性评价,表明地下水位模拟空间趋势性较好,水位与地面高程的空间变化一致,地下水主要以下游平原区和河道为主要排泄区,符合大尺度流域地下水流运动基本规律和评估结果。所提出的多目标率定方法及率定技术应用于大尺度鄱阳湖流域,也为其它水文模型在相似区域的连续模拟及参数率定方面提供借鉴。
This paper used the Poyang Lake watershed as the study area, and a distributed hydrological model was set up for this area. The observed river discharges at six main hydrological stations from 2000 to 2005 and the base flow index of groundwater were used to establish the objective functions for coupled model optimization. PEST was used as an auto-calibration tool embedded in WATLAC to improve the efficiency of model calibration. Results showed that the Nash-sutcliffe efficiency coefficients of different hydrological stations were varied from 0.71 to 0.84, and coefficients of determination were varied from 0.70 to 0.88, a better simulation effect was gained. The actual evapotranspiration (ETa) based on remote sensing was further used to validate the model. Results indicated that model simulated and remotely sensed ETa match well in terms of the spatial variation pattern and the temporal change trend. Furthermore, groundwater flow model was estimated qualitatively, and it was found that the spatial variation of water table was in consistent with the land surface elevation. The general flow direction was from the mountainous areas to the plain area, which was judged as reasonable. The model calibration strategy and technique proposed in this paper can provide high reference for similar studies in a large scale watershed simulation and parameter calibration.