河北沧州夏玉米产量动态预报的多方法对比研究  

Comparative Study on Different Methods for Dynamically Forecasting Summer Maize Yield in Cangzhou,Hebei

作　　者：秦俊灵 王培娟[2] 王琼 魏秀梅 任妙春 刘炳杰 李琛 QIN Jun-ling;WANG Pei-juan;WANG Qiong;WEI Xiu-mei;REN Miao-chun;LIU Bing-jie;LI Chen(Hebei Cangzhou Meteorological Bureau,Cangzhou 061000;Chinese Academy of Meteorological Sciences,Beijing 100081;Hebei Xiongan New Area Meteorological Service,Xiongan 062551,China)

机构地区：[1]河北省沧州市气象局,河北沧州061000 [2]中国气象科学研究院,北京100081 [3]河北雄安新区气象局,河北雄安062551

出　　处：《玉米科学》2025年第3期58-67,共10页Journal of Maize Sciences

基　　金：国家自然科学基金项目(32171916);河北省科研项目(21ky23)。

摘　　要：利用1987-2021年河北沧州气象资料和夏玉米产量数据,采用丰歉指数法、关键气象因子法、气候适宜度法和权重线性集成法分别构建夏玉米产量动态预报模型,利用预报准确率、正确率和标准差评估模型的预报效果。结果表明,集成模型的预报效果和稳定性明显优于单模型,且权重系数赋值时预报效果排序为相关系数>专家经验>预报准确率。单模型中关键气象因子法和气候适宜度法的预报效果和稳定性明显优于丰歉指数法。最优起报日期集成模型为9月1日,气候适宜度法为8月21日,关键气象因子法和丰歉指数法为8月1日。综合考虑,在河北沧州夏玉米产量预报中,按相关系数权重赋值建立3种单模型的集成模型预报效果最佳,且9月1日可作为最优起报日期。Based on the meteorological observations and yield of summer maize in Cangzhou City,Hebei Prov⁃ince from 1987 to 2021,the dynamic prediction models of summer maize yield were constructed using the abun⁃dance-loss index method,the key meteorological factors method,the climate suitability method,and the weighted linear ensemble method,respectively.The prediction accuracy,correct rate and standard deviation were used to evaluate the prediction effect of models.The results showed that the forecasting effect and stability of the integrated model were obviously better than that of the single model,and the order of forecasting effect is:correlation>expert experience>forecast accuracy.The forecast effect and stability of key meteorological factor method and climate suit⁃ability method in single model are obviously better than those of abundance index method.The optimal start date is September 1 for the integrated model,August 21 for the climate suitability method,August 1 for the key meteorologi⁃cal factors method and the abundance index method.Overall,in forecasting summer maize yield in Cangzhou,He⁃bei,the integrated model with weights assigned based on the correlation coefficient provided the best forecasting re⁃sults,with September 1 being the optimal start date.

关 键 词：夏玉米 产量预报 丰歉指数法 关键气象因子法 气候适宜度法 权重线性集成法 

分 类 号：S513.01[农业科学—作物学]