CFSR数据集在辉发河流域水文模拟中的应用

(1.河南师范大学, 河南 新乡 453007; 2.中国科学院 东北地理与农业生态研究所, 长春 130102)

CFSR; 径流模拟; 蒸散发模拟; SWAT; 适宜性评价

Application of CFSR Dataset to Hydrological Simulation of Huifa River Basin
HUANG Yanwei1, LI Ying2, ZHU Honglei1, PENG Xingyue 1, WANG Yudie1

(1.Henan Normal University, Xinxiang, Henan 453002, China; 2.Northeast Institute of Geography and Agroecology, Chinese Academy of Sciences, Changchun 130102, China)

CFSR; runoff simulation; evapotranspiration simulation; SWAT; suitability evaluation

备注

Climate Forecast System Reanalysis(CFSR)是美国国家环境预报中心利用全球预报系统反演的全球再分析数据产品,包含1979—2014年的气温、降水、风速、太阳辐射和相对湿度资料。以往的研究较少考虑CFSR数据集除降水外的其他气象因素对流域水循环模拟精度的影响。以辉发河流域为研究区,在比较CFSR数据集和实测气象数据精度的基础上,采用SWAT模型,从径流模拟和蒸散发两个角度评价了CFSR数据集在该流域的适宜性。结果 表明:(1)在径流模拟方面,CFSR数据集日尺度的R2和NS值范围为0.57~0.71,月尺度为0.72~0.82,均表现较好。CFSR数据集对春季径流存在一定程度的高估,相对偏差为32.61%;(2)采用CFSR数据集作为模型输入,其模拟的PET和ET均高于实测气象数据的模拟结果,即使将CFSR中的降水替换为实测降水数据,ET仍存在一定程度的高估。综上所述,在将CFSR数据集应用于流域水循环模拟时,应全面评价CFSR中各气象因子的精度及其对水循环各组分模拟结果的影响,对数据集中各要素进行全面校正可能会得到更为准确的模拟结果。

Climate forecast system reanalysis(CFSR)is a global reanalysis data product retrieved by the U. S. National Environmental Prediction Center using the global prediction system, which includes the temperature, precipitation, wind speed, solar radiation and relative humidity data from 1979 to 2014. In previous studies, the influence of meteorological factors other than precipitation in CFSR dataset on the accuracy of basin water cycle simulation was seldom considered. On the basis of comparing the accuracy of CFSR dataset with the measured meteorological data, we used SWAT model to evaluate the suitability of CFSR dataset in Huifa River Basin from the perspectives of runoff and evapotranspiration simulation. The results showed that:(1)in terms of runoff simulation, the ranges of the daily and monthly scale R2 and NS produced by the CFSR dataset were 0.57~0.71 and 0.72~0.82, respectively, both of which performed well; CFSR dataset overestimated spring runoff to a certain extent, with a relative error of 32.61%;(2)using CFSR dataset as model input, the simulated PET and ET were higher than the results of the measured meteorological data; even if the precipitation data in CFSR were replaced by the measured precipitation data, ET was still overestimated to a certain extent. In conclusion, when applying CFSR dataset to watershed water cycle simulation, the accuracy of each meteorological factor in CFSR and its impact on the simulation results of each component of water cycle should be comprehensively evaluated. Comprehensive correction of each element in the dataset may lead to more accurate simulation results.