<
doi:  10.12013/qxyjzyj2019-033
江西省油菜产量集成预测模型方法研究

Study on integrated forecasting model of rape yield in Jiangxi province
摘要点击 248  全文点击 80  投稿时间:2019-06-27  修订日期:2019-08-29
查看HTML全文  查看全文  查看/发表评论  下载PDF阅读器
基金:  国内外2019年作物产量预报专项“基于作物模型的双季稻气象灾害影响评估研究”; 江西省气象局2018年青年人才计划项目“基于构建含糖量气象模型的南丰蜜桔采摘期预测研究”.
中文关键词:  油菜产量,集成模型,关键气象因子,气候适宜度,辐热积
英文关键词:  rape yield  the integrated model  key meteorological factors  climate suitability  product of thermal effectiveness and PAR
           
作者中文名作者英文名单位
余焰文Yu Yanwen抚州市气象局
蔡 哲Cai Zhe江西省农业气象中心
姚俊萌Yao Junmeng江西省农业气象中心
蔡小琴Cai Xiaoqin抚州市气象局
引用:余焰文,蔡 哲,姚俊萌,蔡小琴.2019,江西省油菜产量集成预测模型方法研究[J].气象与减灾研究,42(3):206-211
中文摘要:
      基于气象要素建立关键气象因子模型、气候适宜度模型、辐热积模型,并根据模型预测准确率确定权重系数构建江西省油菜产量集成预测模型,对集成模型的预测效果和基础模型进行对比分析。结果表明:1) 关键气象因子模型4月中旬预测准确率低于其他模型,三种基础模型的权重系数依次为0.3、0.35、0.35,5月中旬预测准确率基本一致,权重系数依次为0.34、0.33、0.33。2) 对于集成模型趋势一致率,1991—2010年回代检验结果5月中旬最高,4月中旬仅次于气候适宜度模型;2011—2015年两个时次预测趋势一致率均为100%,达最高。3) 2011—2015年集成模型预测检验RMSE最低,预测准确率均在97%以上。油菜集成模型预测的准确性和稳定性总体优于基础模型,可应用于江西省气象业务服务。
Abstract:
      Based on meteorological factors, the key meteorological factors model (MFM), climate suitability model (CSM) and PAR model (TEPM) were established, and the integrated model (TIM) of rape yield in Jiangxi was constructed by the weight coefficient according to the prediction accuracy of the three models. Comparative analysis of those models shows that: 1) The prediction accuracy of MFM in mid April was lower than that of others, and the weight coefficients were 0.3, 0.35 and 0.35 respectively. The three basic models prediction accuracy was basically similar in mid May, and the weight coefficients were 0.34, 0.33 and 0.33 respectively. 2) The trend consistency rate of TIM was the highest in mid May from 1991 to 2010, and only second to CSM in mid April. Moreover, the trend consistency rate predicted by TIM during two periods in 2011-2015 reached 100%. 3) The RMSE of TIM from 2011 to 2015 was the lowest, and all the annual prediction accuracy was above 97%. Results show that the accuracy and stability of TIM was better than those of the basic models and could be applied to meteorological services in Jiangxi.
主办单位:江西省气象学会 单位地址:南昌市高新开发区艾溪湖二路323号
联系电话: (0791)82713175传真:0791-82713175 邮编:330096 Email:qxyjzyj@163.com
本系统由北京勤云科技发展有限公司设计