三化螟预测模型的通径分析

利用1972-1988年三化螟发生的历史资料和气象资料,对影响因子进行相关分析,应用逐步判别分析的方法组建2代三化螟种群预测的判别模型。文中对判别模型进行通径分析的结果说明,2代三化螟种群预测系统的决定系数 R~2=0.9953,表明预报量的99.53％是由预测模型的全部7个预测因子所决定,尚有0.47％的预报量由未知因素所致,剩余通径系数 P_c=0.0684。经对历史资料的回报检验,该预测模型的判别率达到100％。经过1989和1990年实报。预报等级与实际发生量所代表的等级相符合。证实了2代三化螟种群预测模型预测因子的组成和结构符合预报的要求。

Path Analysis on Forecasting Model for Yellow Rice Borer

The previous data on the occurrence of yellow rice borer and historical records of meteorologi- cal elements over the past successive years 1972-1988 were analysed by correlation studies on influencing fac- tors.The model of forecasting population for the second generation of the pest was established on the basis of stepwise discrimination analysis.The path coefficients of the forecasting model were analysed in the paper. Path analysis indicated that in the forecasting system for the population of the second generation of yellow flee borer the determinative coefficient R~2=0.9953.It cxplained that 99.53% of prediction value were deter- mined by all seven factors in the model and only 0.047% of that were determined by unknown factors.In all unconsidered factors the path coefficient P =0.0684.The discrimination rate in the forecasting model reached 100% through examination on former data.Practical use of the model during 1989 and 1990 had ver- ified the consistency between theoretical grades for forecasting and grades representing actual population densities.Path analysis revealed the composition and structure of forecasting factors in the model for the sec- ond generation of yellow rice borer could meet requirement of prediction.