不同目标函数对径流模拟结果的影响分析

Impact Analysis of Different Objective Functions on Flood Forecast Results

水文模型是洪水模拟的重要工具,大部分模型包含一定数量的参数,因此,参数率定是提高水文模型径流模拟精度、保证模型结果准确性关键之一,而目标函数作为参数率定过程的寻优标准,目前缺乏详尽分类体系为目标函数的选取提供明确选择依据。为探明不同类型目标函数对水文模型参数率定结果和径流模拟精度的影响,综合考虑纳什效率系数、Kling-Gupta系数、洪量相对误差等多种不同类型目标,构建一种包含高流量目标、低流量目标、洪量目标、洪水过程目标及加权目标的目标函数分类体系,结合当前水文预报情报规范的要求,为目标函数选择提供一定标准;进一步,建立新安江水文模型,引入SCE-UA、NSGA-Ⅱ优化算法,为不同算法设置多种目标函数选择方案,以汉江流域牧马河为例进行径流模拟,分析不同目标函数方案和优化算法对径流模拟精度的影响。结果表明:采用单一目标函数,可提高该目标函数所体现水文要素的模拟精度;选取合适的目标函数是提升径流模拟精度的关键,在现有规范的要求下,选取洪水过程目标和高流量目标是保证水文模型精度的最佳选择;Kling-Gupta系数与纳什效率系数相比,其在洪量模拟精度上表现更佳,是洪水过程目标的更优选择;加权单目标与多目标方法的模拟结果在高流量部分存在差异。

Hydrological model is an important tool for flood simulation,and most of the models contain a certain number of parameters.Therefore,parameter calibration is one of the keys to improve the precision of hydrological model runoff simulation and ensure the accuracy of model results.As the optimization standard of parameter calibration process,there is currently no detailed classification system to provide a clear selection basis for the selection of objective function.In order to explore the effects of different types of objective functions on the calibration results of hydrological model parameters and the precision of runoff simulation,this paper comprehensively considered various types of objectives,such as Nash efficiency coefficient,Kling-Gupta coefficient,flood relative error,etc.,and constructed an objective function classification system that includes high flow target,low flow target,flood volume target,flood process target and weighted target.According to the requirements of the current hydrological forecast information standard,some criteria were provided for the selection of objective function.Further,the Xinanjiang model was established,and the optimization algorithms of SCE-UA and NSGA-Ⅱ were introduced.A variety of objective function selection schemes were set up for different algorithms,and the influence of different objective function schemes and optimization algorithms on the precision of runoff simulation was analyzed by taking Muma River in Hanjiang River basin as an example.The results show that using a single objective function can improve the simulation accuracy of hydrological elements embodied by the objective function;Selecting the appropriate objective function is the key to improve the precision of runoff simulation.Under the requirements of existing norms,selecting flood process target and high flow target is the best choice to ensure the accuracy of hydrological model.Compared with Nash efficiency coefficient,Kling Gupta coefficient has better performance in flood volume simulation accuracy and is a better choice for flood process target.The simulation results of weighted single target method and multi-target method are different in the high-flow part.