摘要
根据江西省泰和县1991～2007年褐飞虱发生程度的资料及1990～2007年ENSO事件的特征值资料,采用相关分析和逐步回归分析方法,建立了泰和县褐飞虱发生程度(Y)的长期预测模型:Y′=-0.20790X1+0.22877X2+4.05618,其中X1为上一年厄尔尼诺的强度,X2为当年3月南方涛动指数(Southern Oscillation Index,SOI)。该模型1991～2007年的历史拟合符合率为82.35%,2008～2009年的预报准确率为100%,并预测2010年泰和县褐飞虱发生程度为中等(3.11级)。
According to the occurrence degree data of Nilaparvata lugens in Taihe county of Jiangxi province from 1991 to 2007,as well as the data of the characteristic values of ENSO events during 1990～2007,by using the methods of correlation analysis and stepwise regressive analysis,a model for the long-term prediction of the occurrence degree of Nilaparvata lugens in Taihe county was established as follows: Y′=-0.20790X1+0.22877X2+4.05618,thereinto X1 was the intensity of El Nio event last year,X2 was the Southern Oscillation Index(SOI) in March of the same year.The verification results indicated that the historical simulation coincident rate of this model from 1991 to 2007 was 82.35%,the forecasting accuracy from 2008 to 2009 was 100%.The constructed model predicts that Nilaparvata lugens in Taihe county will moderately occur(grade 3.11) in 2010,which awaits to be verified henceforth.
出处
《江西农业学报》
CAS
2010年第8期85-87,90,共4页
Acta Agriculturae Jiangxi
