2种基于历史丰歉气象影响指数的茎瘤芥产量动态 预报 方法比较  

Comparison of two methods for yield prediction of Brassica juncea var.tumida Tsen&Lee based on meteorological influence index of historical yield

作　　者：武强 韩旭 唐余学 徐倩倩[4] 阳园燕 WU Qiang;HAN Xu;TANG Yu-xue;XU Qian-qian;YANG Yuan-yan(Chongqing Institute of Meteorological Sciences,Chongqing 401147,China;Chongqing Engineering Research Center of Agrometeorology and Satellite Remote Sensing,Chongqing 401147,China;Jiangjin Modern Agrometeorology Experimental Station,Chongqing 402260,China;Hefei Meteorological Bureau,Hefei 230041,China)

机构地区：[1]重庆市气象科学研究所,重庆401147 [2]重庆市农业气象与卫星遥感工程技术研究中心,重庆401147 [3]重庆市江津现代农业气象试验站,重庆402260 [4]合肥市气象局,安徽合肥230041

出　　处：《江苏农业学报》2021年第6期1443-1450,共8页Jiangsu Journal of Agricultural Sciences

基　　金：国家自然科学基金面上项目(42175193);重庆市技术创新与应用发展专项项目(cstc2020jscx-msxmX0111);中国气象局创新发展专项项目(CXFZ2021J068、CXFZ2021J073);重庆市气象部门智慧气象技术创新团队项目(ZHCXTD-202016);重庆市气象部门业务技术攻关项目(YWJSGG-201905);重庆市气象局科技计划项目(QNJJ-201703)。

摘　　要：以涪陵茎瘤芥为例,基于历史丰歉气象影响指数,综合诊断筛选气象要素相似年型,采用大概率法与加权平均分析法建立茎瘤芥产量预报模型,并验证不同方法的茎瘤芥产量动态预报准确性。结果表明:同一年份不同起报时间的气象条件相似年型差异主要发生在热量条件的相似年,其次是光照条件相似年,而水分条件相似年在同一年份不同起报时间无变化。但同一起报时间不同年份的气象条件相似年,无明显的重复性,即年际间的气象条件具有差异性,能够综合反映气象条件对茎瘤芥产量丰歉的决定性。比较大概率法与加权平均分析法的茎瘤芥产量预报结果,在产量丰歉趋势预报方面,大概率法在较早的起报时间易出现较大的偏差,但是随着起报时间推后,产量丰歉趋势趋于正确;加权平均分析法在大多数年份的产量丰歉趋势预报准确性较高,但是个别年份会出现不同起报时间的持续性预报偏差。在单产预报准确率方面,加权平均分析法预报结果明显优于大概率法,但2种方法均在1月1日起报时间表现出最高的预报准确率,大概率法1月1日起报的单产预报准确率平均值为89.5%,加权平均分析法1月1日为起报时间单产预报准确率平均值为98.9%。在进入2月份也即茎瘤芥瘤茎膨大后期,预报准确率降低。Based on the meteorological influence index of historical yield,the similarity year of meteorological conditions was selected by integrated diagnosis.The large probability method and weighted average method were used to establish the yield prediction model of Brassica juncea var.tumida Tsen&Lee,and the accuracy of different methods was verified.The results showed that the difference of similarity year of meteorological conditions with different starting times in the same year mainly occurred in the heat condition.Secondly,the similarity year of light conditions was different in some years,while the water conditions had no change in different starting times.However,there was no obvious repeatability in the similarity years of meteorological conditions with the same starting time,that was,the meteorological conditions were different among years,which could comprehensively reflect the decisive effect of meteorological conditions on the yield of Brassica juncea var.tumida Tsen&Lee.Comparing with the results of the large probability method and the weighted average method,the large probability method was prone to large deviation in the early starting time,the yield trend tended to be correct with the delay of reporting time.The accuracy of the weighted average method was high,but the persistent forecasting deviation with different starting time could occur in some years.In terms of the prediction accuracy of unit yield,the prediction result of weighted average method was obviously better than that of large probability method,both methods showed the highest forecast accuracy on the starting time of January 1,the highest forecast accuracy of large probability method and weighted average method were 89.5%and 98.9%.In February,the late period of stem enlargement of Brassica juncea var.tumida Tsen&Lee,the accuracy of prediction decreased.

关 键 词：产量预报 大概率法 加权平均分析法 气象条件相似年 茎瘤芥 

分 类 号：S165.27[农业科学—农业气象学]