【摘要】：為了建立南方雙季稻低溫災害綜合預測預警技術(shù)體系,基于南方雙季稻種植區(qū)1961 2010年708個氣象站的逐日氣象資料、水稻生育期資料和低溫災害發(fā)生的氣象行業(yè)標準,采用Fishe r判別分析法、因子膨化法、相關(guān)性分析法,利用SPSS軟件構(gòu)建早稻春季低溫災害高風險區(qū)(Ⅰ區(qū))未來10 d、晚稻寒露風高風險區(qū)(Ⅰ區(qū))、主災區(qū)(Ⅱ區(qū))未來5 d的低溫災害發(fā)生等級逐日滾動預警模型。其中,1961 2009年資料用于模型構(gòu)建和回代檢驗,2010年資料用于模型的外延預測。結(jié)果表明:早稻、晚粳稻、晚秈稻Ⅰ區(qū)平均外延預測基本一致準確率分別達到90.5%,74.2%,80.3%,晚粳稻、晚秈稻Ⅱ區(qū)平均外延預測基本一致準確率分別為89.4%和80.3%。構(gòu)建的南方雙季稻低溫災害逐日滾動預警模型的外延預測基本一致準確率多超過80%,等級預測檢驗誤差總體上在1個等級以內(nèi),模型評價效果較好。
[Abstract]:In order to establish the comprehensive prediction and early warning technology system of low temperature disaster in southern double cropping rice, based on the daily meteorological data of 708 meteorological stations in southern double cropping rice growing area in 1961 and 2010, the data of rice growth period and the meteorological industry standard of low temperature disaster, Fishe r discriminant analysis, factor expansion method and correlation analysis were used to construct the early rice spring low temperature disaster high risk area (area I) for the next 10 days. A daily rolling early warning model for the occurrence of low temperature disasters in the high risk area of late rice Cold Dew (area 鈪，

本文編號：2509740