目标函数不确定性对汉江安康以上流域水文模拟的影响

Influence of Uncertainty of Objective Function on Hydrological Simulation Process in Upper Ankang Basin of Hanjiang Basin

不同优化算法和目标函数组合将导致不同的模拟结果,进而影响流域水文预报精度。通过设定不同目标函数组合方案,对比汉江上游安康以上流域在多目标优化算法(NSGA-Ⅱ和MOPSO)下对HyMod模型性能评估的影响,并识别季尺度目标函数统计规律与模拟精度的本质联系。结果表明,NSGA-Ⅱ算法更适合于汉江流域(年尺度、季尺度模拟精度更优);模拟值与实测值的拟合误差控制函数和高流量误差控制的函数(F2、F3)为多目标模型参数校准的首选目标函数,其重要性大于水量模拟的误差控制函数和枯水流量的误差控制函数(F1、F4)。

Different optimization algorithms and objective functions will lead to different simulation results,which could affect the accuracy of hydrological forecasts in the basin.By setting different objective function combination schemes,the impact of multi-objective optimization algorithms(NSGA-Ⅱand MOPSO)on the performance evaluation of HyMod model was compared.And then the essential relationship between seasonal scale objective function statistics and simulation accuracy was identified.The results show that the NSGA-Ⅱalgorithm is more suitable for the Hanjiang Basin,and the simulation accuracy of the annual and seasonal scale was better;The fitting error control function of the simulated and the measured values and the high-flow error control function(F2,F3)were the preferred objective functions for the calibration of the multi-objective model parameters,which are more important than the error control function of water volume simulation and the low-flow error control function(F1,F4).