基于主成分回归模型的漯河市小麦相对气象千粒重的模拟模型

Simulation Model of Relative Meteorological 1000-Grain Weight of Wheat of Luohe Based on Principal Component Regression

根据2010-2021年漯河市3个地区的气象资料及小麦千粒重,对小麦不同灌浆阶段内的气象因子与相对气象千粒重进行相关分析,运用主成分回归模型、逐步回归模型建立包含关键气象因子的相对气象千粒重模拟模型。结果表明,灌浆快增期的平均气温和最高气温以及缓增期的最高气温是影响漯河市小麦相对气象千粒重的重要因子;主成分分析的前3个分量“快增期和缓增期最高气温因子”、“快增期平均气温因子”和“快增期光照因子”可以解释相对气象千粒重89.44%的主要变化;与逐步回归模型相比,3个地区的相对气象千粒重和千粒重在主成分回归模型下的预估值与实际值模拟效果更好,尤其是五里岗。因此,主成分回归模型对漯河市相对气象千粒重的预估更具合理性和准确性;当前培育并筛选耐后期高温品种是提高漯河市小麦千粒重的有效办法。

The correlation between meteorological factors and 1000-grain weight of wheat during different filling stages of wheat in Luohe was studied based on the meteorological data and 1000-grain weight data of three stations from 2010 to 2021,then an empirical model with key meteorological factors used to simulate relative meteorological 1000-grain weight(Yr)of wheat was established by using of the principal component regression model and stepwise regression model.The results showed that thermal factors such as average temperature and maximum temperature in rapid-increasing stage and maximum temperature in slight-increasing stage had the greatest impaction on the Yr of wheat in Luohe.The first three components of principal component analysis were“the maximum temperature factor in rapid-increasing stage and slight-increasing stage”,“the average temperature factor in rapid-increasing stage”and“the sunshine factor in rapid-increasing stage”,which could explain 89.44%of the Yr changes of wheat in Luohe.Compared with the stepwise regression model,the predicted and actual values of Yr and 1000-grain weight of wheat by using of the principal component regression model in three stations were better,especially in Wuligang.Therefore,the principal component regression model was more reasonable and accurate in estimating the Yr of Luohe.At present,it is an effective way to increase the 1000-grain weight of wheat in Luohe by breeding and screening high-temperature resistant varieties in late stage.