特早种茶叶开采期预报模型的建立与误差分析  

Construction and Verification of Picking Date Prediction Model for Extremely Early Budding Tea Based on Meteorological Conditions

作　　者：武强 王旭 方丽 江姣 孙恩虹 韩旭 陈思英[5] WU Qiang;WANG Xu;FANG Li;JIANG Jiao;SUN Enhong;HAN Xu;CHEN Siying(Chongqing Institute of Meteorological Sciences,Chongqing 401147,China;Banan Meteorological Bureau,Chongqing 401320,China;Changshou Meteorological Bureau,Changshou Chongqing 401220,China;Jiangjin Modern Agrometeorology Experimental Station,Jiangjin Chongqing 402260,China;Anhui Science and Technology Research and Development Center,Hefei 230088,China)

机构地区：[1]重庆市气象科学研究所,重庆401147 [2]重庆市巴南区气象局,重庆401320 [3]重庆市长寿区气象局,重庆长寿401220 [4]重庆市江津现代农业气象试验站,重庆江津402260 [5]安徽省科技研究开发中心,合肥230088

出　　处：《西南大学学报（自然科学版）》2022年第5期50-57,共8页Journal of Southwest University(Natural Science Edition)

基　　金：国家重点基础研究发展计划项目(2013CB430205);重庆市气象部门业务技术攻关项目(YWJSGG-201905,YWJSGG-201906);重庆市气象部门智慧气象技术创新团队项目(ZHCXTD-202016).

摘　　要：以特早种茶叶品种“巴渝特早”为研究对象,应用阶段积温预报法、逐步回归预报法、集成预报法分别建立开采期气象预报模型,比较不同预报方法的预报性能.结果表明:阶段积温预报法对于“巴渝特早”开采期的预测偏差较大,预测值与观测值的剩余均方差为5.225,拟合优度为0.539;逐步回归预报法相较于阶段积温预报法预测结果的准确性明显提升,预测值与观测值的剩余均方差为2.826,拟合优度为0.749;集成预报模型综合考虑了积温条件以及影响春茶开采的关键时段的关键气象因子,优化单一预报方法的不稳定性,预测值与观测值的剩余均方差为2.729,拟合优度0.765,可以应用于特早种茶叶开采期气象预报服务.Early picking is the characteristics of extremely early budding tea.The meteorological factors affecting its picking date are different from conventional varieties.Prediction of picking date for extremely early budding tea is important for economic benefits of spring tea and study of the mechanism of climate influence on tea picking date.In this study,Bayu Tezao,the extremely early budding tea was taken as the research object.The picking date prediction models of extremely early budding tea were established by using the method of stage accumulated temperature,stepwise regression of meteorological factors and ensemble forecast,respectively.The results showed that the deviation of stage accumulated temperature prediction method was within 5 days.Its residual mean square variation and determination coefficient of linear equation between the predicted value and the observed value was 5.225 and 0.539,respectively.The deviation of stepwise regression prediction method was within 3.5 days.Its residual mean square variation and determination coefficient of linear equation between the predicted value and the observed value was 2.826 and 0.749,respectively.The ensemble forecast prediction method had the best prediction accuracy,the predicted deviation within 3.4 days.Its residual mean square variation and determination coefficient of linear equation between the predicted value and the observed value was 2.729 and 0.765,respectively.The ensemble forecasting method can be used for forecasting the picking date of extremely early budding tea.

关 键 词：“巴渝特早” 开采期 预报模型 误差分析 

分 类 号：S571.1[农业科学—茶叶生产加工]