昆虫种群动态时空回归预测方法及应用研究  

Spatio temporal explicit interactive model for forecasting the pine caterpillar population dynamics

作　　者：周国法[1] 徐汝梅[1] 李天生[2] 

机构地区：[1]北京师范大学生命科学院,北京100875 [2]中国林业科学院森林保护研究所,北京100091

出　　处：《生态学报》1999年第4期464-467,共4页Acta Ecologica Sinica

摘　　要：根据昆虫种群内个体空间相互作用的特点，提出分析昆虫种群时空相关的三维相关图方法及种群动态预测的时空混合回归模型。对马尾松毛虫幼虫密度分布的预测采用了时空自回归加时空趋势面的形式，预测的马尾松毛虫幼虫的平均密度和实测的平均密度在时空变化的趋势上是完全一致的，而且预测结果给出了马尾松毛虫幼虫密度分布的图象形式，给防治决策提供了方便。Based on the spatial temporal characteristics of interaction of pine caterpillar larvae population,a three dimensional correlogram and spatially mixture regression model was proposed.The correlogram includes both spatial and temporal correlation analysis,the forecast model includes three components,temporal and spatial auto regression,and spatio temporal trend surface.As for the forecasting of pine caterpillar density distribution,both temporal correlation and the intensity of spatial interaction were determined by correlogram.The forecasted larvae density of pine caterpillar has the similar trend,both temporally and spatially,as the sampled density in the field.The resultant distribution of larvae density can provide an easy using image for IPM.

关 键 词：时空相关 时空混合回归 马尾松毛虫 预测 

分 类 号：S763.421[农业科学—森林保护学] Q968.1[农业科学—林学]