中国农业气象

• 论文 • 上一篇    下一篇

利用地面气象资料建立四川省日总辐射计算模型

蔡元刚,王明田,蔡怡亨,刘雅琳,陈东东   

  1. 1.绵阳市气象局,绵阳 621000;2. 四川省气象台,成都 610072;3.南方丘区节水农业研究四川省重点实验室,成都 610066;4.南京信息工程大学,南京 210044;5.四川省农业气象中心,成都 610072
  • 出版日期:2019-09-20 发布日期:2019-09-20
  • 作者简介:蔡元刚(1967?),高级工程师,研究方向为农业气象与气候。E-mail:mycyg@126.com
  • 基金资助:
    农业农村部西南山地农业环境重点实验室开放基金(AESMA-OPP-2019006);高原与盆地暴雨旱涝灾害四川省重点实验室科技发展基金项目(省重实验室2018-重点-05-01)

Using Surface Meteorological Data to Establish Daily Total Solar Radiation Calculation Model for Sichuan Province

Cai Yuan-gang,WANG Ming-tian,CAI Yi-heng,LIU Ya-lin,CHEN Dong-dong   

  1. 1.Mianyang Meteorological Bureau,Mianyang 621000,China;2.Sichuan Meteorological Observatory,Chengdu 610072;3.Water-Saving Agriculture in Southern Hill Area Key Laboratory of Sichuan Province, Chengdu 6l0066;4.Nanjing University of Information Science & Technology,Nanjing 210044;5.The Agrometeorological Center of Sichuan Province,Chengdu 610072
  • Online:2019-09-20 Published:2019-09-20

摘要: 利用四川省6个辐射观测站2016?2018年日总辐射和地面气象资料,应用“个案排秩”、一元线性回归和逐步回归方法,建立四川省日总辐射计算模型(模型Ⅰ),并按日照时数是否为0建立有日照总辐射计算模型(模型Ⅱ)和无日照总辐射计算模型(模型Ⅲ)。结果表明:模型Ⅰ、模型Ⅱ和模型Ⅲ均通过0.01水平显著性检验;模型Ⅰ、模型Ⅱ和模型Ⅲ回代检验的MAPE分别为12.62%、10.02%、16.34%,NRMSE分别为16.17%、12.23%、28.40%;4个典型日应用这3个模型的MAPE分别为7.59%、4.50%、36.53%,NRMSE分别为9.22%、5.93%、40.98%;对于4个典型日在日照时数为0时不用模型Ⅲ而改用模型Ⅰ、日照时数不为0时用模型Ⅱ,其MAPE为5.79%、NRMSE为7.47%,比全部资料用模型Ⅰ模拟分别提高1.80个和1.75个百分点。建立的3个日总辐射计算模型均具有应用价值;四川省日总辐射最佳模拟方法是日照时数为0时用模型Ⅰ计算,日照时数不为0时用模型Ⅱ计算;海拔、天气状况和日照长短决定四川各地日总辐射量的大小,其中海拔和天气状况是造成四川各地日总辐射差异的主要因素。

关键词: 四川, 日总辐射, 正态得分, 气象因子, 数值模拟

Abstract: Based on the daily total solar radiation and the surface meteorological data from 2016 to 2018 collected of 6 radiation observation stations in Sichuan Province, the daily total solar radiation calculation model (Model I) of Sichuan Province was established by using “case ranking”, linear regression and stepwise regression method, and another total solar radiation calculation model (Mode II) and a no-sunlight total radiation calculation model (Mode III) were established according to whether the sunshine hours are 0 or not. The results showed that the outcomes of Mode I, Mode II and Mode III achieved significance level of 0.01. The backtest’s MAPE for the model I, model II and Mode III were 12.62%, 10.02% and 16.34%, respectively, and the NRMSE were 16.17% and 12.23%, 28.40%, respectively. The MAPEs of these three models for the 4 typical days are 7.59%, 4.50%, and 36.53%, the NRMSEs were 9.22%, 5.93%, and 40.98%, respectively. As for 4 typical days, if sunshine hours are 0, it’s better to use Mode Ⅰ instead of Mode III, the model II simulation could be used when the sunshine hours are not 0, and then the MAPE was 5.79% and the NRMSE was 7.47%, which were 1.80 and 1.75 percentage points higher than the values calculated by model I simulation. All three established models had application value in calculating daily total radiation; the best simulation method for daily total radiation in Sichuan Province is using Mode I when the sunshine hours are 0, and the model II is used when the sunshine hours are not 0; Altitude, weather conditions and the length of sunshine determine the total amount of daily radiation in all parts of Sichuan. The altitude and weather conditions are the main factors caused the daily total solar radiation differences in Sichuan.

Key words: Sichuan, Daily total solar radiation, Normal score, Meteorological factors, Numerical simulation