基于主成分分析降水量指标预测棉铃虫百株卵量模型研究

Principal Component Analysis Forecast Model of Helicoverpa armigera Based on Precipitation Index

棉铃虫是中国棉花主要害虫之一,降水量对其种群数量具重要影响。采用连续10a棉铃虫第2代和第3代的百株卵量与同期降水量指标资料,通过相关系数法筛选出10个预测第2代、第3代棉铃虫的降水量指标,采用主成分分析法建立第2代、第3代棉铃虫百株卵量的特征预测模型。检验结果表明,第2代棉铃虫的预测模型信度较好,历史符合率为70%;而第3代棉铃虫的预测模型信度较差,历史符合率为40%。由于入选因子均为前兆因子,因此,第2代棉铃虫的预测模型可用于区域棉铃虫发生程度的监测和预警研究。

Helicoverpa armigera （Hübner） is one of the main pests in cotton grown in China, meteorological factors have very important implications with its population dynamics. Using observational data of H. armigera and meteorological factors from 1961 to 1970, this paper adopted correlation coefficient method to select ten precipitation index and as forecasting index of the amount of second and third generation H. armigera per one hundred plants, and using principal component analysis method to establish predictive models. The results show the model of second generation accuracy is 70%, and the third generation is 40%. Forecast based the model has strong timeliness and favorable effect, therefore, it can be used for monitoring and early warning of occurrence degree of H. armigera.