山东省Ångström-Prescott日照辐射模型的参数优化  

The Parametric Optimization of the Ångström-Prescott Solar Radiation Model in Shandong Province

作　　者：田晓旭 田辉 张鹏[2] 刘福胜[2] 张倩 TIAN Xiao-xu;TIAN Hui;ZHANG Peng;LIU Fu-sheng;ZHANG Qian(Institute of Farmland Irrigation Research,Chinese Academy of Agricultural Sciences/Key Laboratory for Water-saving Irrigation Project under Ministry of Agriculture,Xinxiang 453000,China;College of Water Conservancy and Civil Engineering/Shandong Ageicultural University,Tai’an 271018,China)

机构地区：[1]中国农业科学院农田灌溉研究所/农业部节水灌溉工程重点实验室,河南新乡453002 [2]山东农业大学水利土木工程学院,山东泰安271018

出　　处：《山东农业大学学报（自然科学版）》2021年第6期942-948,共7页Journal of Shandong Agricultural University：Natural Science Edition

基　　金：中国农业科学院基本科研业务费项目(FIRI2016-25-01);国家重点研发计划课题(2018YFC0407703);山东省重大科技创新工程项目(2019JZZY010727);山东省重点研发计划重大科技创新工程项目(2019JZZY010710);山东省2017年度农业重大应用技术创新项目。

摘　　要：Ångström-Prescott模型是基于日照的最佳太阳辐射(Rs)模型,但由于缺乏模型参数,其应用往往受到限制。为了获得山东省准确的模型参数,本文采用FAO推荐值和根据地理位置、多年平均温度等数据拟合的四种参数化方程,在山东省有实测太阳辐射数据的3个站点(福山、济南和莒县)应用Ångström-Prescott公式计算太阳辐射值,并与实测值和校正参数预测值进行比较。结果表明:(1)3个站点中,福山参数模型估算效果最好,莒县次之,济南最差。福山、莒县和济南的决定系数R^(2)分别在在0.9559~0.9661、0.9358~0.9397及0.9083~0.9115之间,但决定系数R^(2)都在0.9以上,不同参数化方法预测的Rs与实测值相关性较高;(2)公式1参数模型估算太阳辐射值效果最好。该公式的平均偏差(MBE)、均方根误差(RMSE)及相对均方根误差(PRMSE)分别在-4.54×10^(-5)~1.23×10^(-3) MJ/(m^(2)/d)、0.021~0.078 MJ/(m^(2)/d)及15.23%~55.40%之间,和校正值比较接近,FAO推荐值估算精度较低,其估算精度远低于公式1。因此,在山东省可以用公式1(a=5.43742×10^(-4) H+0.149 b=0.00964×Փ-0.62367)即用海拔和经度较为精确的预测太阳辐射。TheÅngström-Prescott model is an optimal solar radiation(R_(s))model based on sunlight,but its application is often limited due to the lack of model parameters.In order to obtain the accurate model parameters of Shandong Province,the angstron-Prescott formula was used to calculate the solar radiation values at three stations(Fushan,Jinan and Juxian)in Shandong Province,using the recommended FAO values and four parameterized equations fitted according to the data of geographical location and annual average temperature.The results were compared with the measured values and the predicted values of the correction parameters.The results show that:(1)Among the three stations,the estimation effect of fushan parameter model is the best,Ju County is the second,and Ji'nan is the worst.The coefficient of determination(R^(2))of Fushan,Juxian and Jinan are in the range of 0.9559~0.9661,0.9358~0.9397 and 0.9083~0.9115,respectively.However,the coefficient of determination(R^(2))are all above 0.9.The correlation between Rs predicted by different parameterization methods and measured values is high.(2)Formula 1 has the best effect in estimating the solar radiation value.The mean deviation(MBE),the root mean square error(RMSE)and the relative root mean square error(PRMSE)of the formula are between-4.54×10^(-5)~1.23×10^(-3) MJ/(m^(2)·d),0.021~0.078 MJ/(m^(2)·d)and 15.23%~55.40%,respectively.And the correction value are close,and the estimation accuracy of FAO recommended value is low,which is far lower than formula 1.Therefore,formula 1(a=5.43742×10^(-4) H+0.149 b=0.00964×Փ-0.62367)by altitude and longitude can be used to accurately predict the solar radiation in Shandong Province.

关 键 词：太阳辐射 模型 参数优化 

分 类 号：S161.1[农业科学—农业气象学]