基于逐步回归法对重庆市秀山县褐飞虱发生量的预测

A Stepwise Regression-Based Study on the Forecasting of the Occurrence of Nilaparvata lugens in Xiushan County of Chongqing

利用重庆市秀山县1990-2006年褐飞虱的田间、灯下虫情及气象资料,以田间发生高峰期的调查数量为因变量,以灯下诱集虫量和气象因子为自变量,用多元逐步回归法组建了褐飞虱在不同时期的发生量预测模型,并利用该模型对重庆市秀山县褐飞虱主害期的发生量进行预测检验.结果表明:褐飞虱田间发生高峰期的发生量不仅与气候条件有关,且与灯下虫量有关,并得出与近期灯下总虫量成正相关,而与具有繁殖能力的雌成虫呈负相关.

Brown planthopper, Nilaparvata lugens, is a major immigration pest of rice fields in causes great losses to rice production. In order to forecast the occurrence orological data and the records of the pest occurrence investigated in the yof N. lugens, China and the mete- d with light-trap in Xiush an, Chongqing from 1990 to 2006 were used to construct a forecasting model based on stepwise regression analysis, with the data of occurrence peak in field as the independent variable and the data of light-trapped and weather as the dependent variables. Then this forecasting model was used to predict the occurrence of N. lugens to verify its validity. The result showed that the occurrence peak of N. lugens in field was not only associated with the weather condition but related to the immigration quanti ted positively with the recent total amount of light trap catches and negatively w adult which has the ability of reproduction. as well. It was correla immigrated female adult which has the ability of reproduction.