摘要
以大汶河流域为例,借助SWAT模型,采用SUFI2方法进行参数率定,以Nash-Sutcliffe为目标函数,分别对2008~2009年大汶河流域8组不同分辨率组合的数据(DEM数据和土地利用数据)进行了16次迭代,每次迭代300次模拟,并通过添加两组分辨率不匹配的组合模拟,探究了分辨率匹配与否对模拟径流结果的影响。结果表明,各分辨率组合模拟的Nash-Sutcliffe效率系数(NSE系数)和决定系数(R2)均超过了0.85,偏差百分比(P_(BIAS))绝对值均小于10%,说明不同分辨率组合下模型对大汶河流域校准模拟结果均较好;分辨率不匹配的模拟结果比分辨率匹配的模拟结果稍差;30×30(DEM和土地利用分辨率均为30m)组合为大汶河流域模拟的最佳分辨率组合。
In this paper,the Dawenhe Basin is taken as an example.With the SWAT model,the SUFI2 method is used to determine the parameters.Nash-Sutcliffe is taken as the objective function.And the data of the eight different resolution combinations of Dawenhe Basin in 2008-2009(DEM Data and land use data)are carried out for 16 iterations calculation,and each iterating is 300 simulations.By adding two sets of resolution mismatched combinations,it explores the impact of resolution matching on the results of simulated runoff.The results show that the calibrated model simulations with different resolution are satisfactory.The NSEcoefficient and R2 are all over 0.85,PBIASabsolute percentage are all less than 10%.The simulation results of mismatched resolution are slightly worse than the matched resolution simulation results.The 30×30(DEM and the land use resolution are both 30m)is the best resolution combination for the simulation of Dawenhe Basin.
出处
《水电能源科学》
北大核心
2017年第8期27-30,44,共5页
Water Resources and Power
基金
山东省省级水利科研与技术推广项目(SDSLKY201503)
