利用主成分回归方法预估连云港地区水稻气象产量

Estimation of the meteorological yield of rice in Lianyungang area using principal component regression method

根据连云港地区1974-2016年5个站点的气象资料和水稻产量资料,利用主成分回归方法研究该地区水稻不同生长期内,气象要素与水稻气象产量的关系,构建预估模型并对其进行检验。结果表明,抽穗开花期和灌浆成熟期的气温、积温等热力因子对连云港地区水稻气象产量影响最大;主成分分析的前3个分量为“抽穗开花期因子”、“灌浆成熟期因子”和“移栽返青期因子”,可以解释连云港地区85%左右的水稻相对气象产量变化;回归模型组合出3个新的因子,分别对应抽穗开花期的热力条件、灌浆成熟期的热力条件和移栽返青期的温差,该模型可以较好地预估连云港地区水稻的气象产量。

The relationship between meteorological elements and the meteorological yield of rice during different growing periods of rice in Lianyungang region was studied based on the meteorological data and rice yield data of five stations from 1974 to 2016,using principal component regression method.A prediction model was constructed and tested.The results showed that,thermal factors such as air temperature and accumulated temperature at heading-flowering stage and filling-maturity stage had the greatest impact on the meteorological yield of rice in Lianyungang region.The first three components of principal component analysis were“heading-flowering period factor”,“filling-maturity period factor”and“transplanting-returning green period factor”,which could explain about 85%of the relative meteorological yield changes of rice in Lianyungang area.Three new factors were combined by the regression model,which corresponded to the thermal conditions of the:“heading-flowering period”the thermal conditions of the“filling-maturity period”and the temperature difference in the“transplanting-returning green period”,respectively.The model can well predict the meteorological yield of rice in Lianyungang area.