冬小麦产量分阶段预测模型  

Winter Wheat Yield Forecast Multi-stage Model

作　　者：何亚娟[1,2] 汪庆发[1] 裴志远[1] 王连林 马志平 潘学标[2] 

机构地区：[1]农业部规划设计研究院,北京100125 [2]中国农业大学资源与环境学院,北京100094 [3]玉田县发展计划局,唐山064100

出　　处：《农业机械学报》2012年第1期89-93,133,共6页Transactions of the Chinese Society for Agricultural Machinery

基　　金：农业部农业信息预警财政专项资助项目(2130111)

摘　　要：为了解决冬小麦估产的时效性和运行化问题,通过对河北省玉田县2007～2009年冬小麦的连续监测,在不同生育期(抽穗期、灌浆期和收获期)对其产量构成三因子(穗数、粒数和粒质量)进行实地抽样测定,并结合冬小麦各个生长发育期的生理生态特点,建立相应的分阶段单产预测模型。试验发现,单因子模型的应用,可使冬小麦估产的预报时间提前到抽穗期,其拟合精度可达到88%以上。双因子模型的应用可使预报时间提前到抽穗后期至灌浆期,模型拟合精度大于90%;结果表明,冬小麦分阶段预产模型可以作为县级区域农业遥感业务化运行系统的基础,增强农业遥感监测产量的预警能力。In order to search for more timeliness and accuracy of forecasting methods,Yutian County,Hebei Province was chosen as the pilot site.A three-year field experiments were conducted on winter wheat.During different growing periods,three yield-contribution factors were measured in situ and the analysis was done.The results revealed that single-factor model of application allowed the prediction of time ahead of winter wheat yield to the heading stage;the prediction accuracy can reach to 88%.Prediction of two-factor model can be advanced to the heading stage to the early filling stage,and the AI is more than 90%.This study provided an effective method namely the multi-stage prediction model to effectively predict crop yield.It can overcome the regular operation problems and provide important references for agriculture warning system by remoter sensing at all levels of agricultural crop production forecast.

关 键 词：冬小麦 产量构成 多阶段预测模型 

分 类 号：S165.27[农业科学—农业气象学] S512.11[农业科学—农业基础科学]