贵州烟区烟蚜发生规律及其预测模型的初步研究

Preliminary studies on occurrence and prediction model of Myzus persicae in Guizhou

通过2002—2008年对贵州代表性烟区烟蚜的系统调查，初步了解了有翅蚜迁飞规律、烟田蚜虫种群消长动态及其历史演变情况，分析了烟蚜发生期内气温、雨日雨量、相对湿度和日照时数对烟蚜发生量的影响效应。在此基础上，结合以往的调查数据，应用DPS统计软件进行逐步回归分析，初步建立了烟蚜发生程度短期预测模型y=24．87＋1．27X1＋0．13X2＋1．54X4＋2．70X6＋4．06X11—3．85X13（其中，X1=移栽前期10d内日诱蚜量；X2一上年12月日气温≥10℃的积温；X4—2月均温；X6—3月均温；X11—4月均温，X13=4月雨日数），经显著性检验和预报验证，回归方程达极显著水平，回测准确率平均达93．45％，实报验证误差率仅为1．80％～5．11％，说明入选预报因子比较切合贵州实际。预测模型具有较大的实用价值。

A systematic investigation of representative Myzus persicae was carried out in Guizhou during 2002- 2008. The migratory behavior of flying aphids, population dynamics and historical evolution of M. persicae, and the effects of temperature, rainfall, relative humidity and sunshine hours on the quantity of M. persicae were pre- liminarily analyzed. The data were analyzed with stepwise regression analyses by DPS statistics software, and the short-term prediction model of occurrence conditions was Y=24.87-k 1.27X1 ＋ 0. 13X2＋1.54X4 ＋2.70X6 ＋ 4.06X11＋3.85X13 （X1=number of trapped aphids from the 1st day to the 10th day after transplanting, X2 = accumulated temperature≥10 ℃ air temperature of last December, X4 = average temperature of February, X6 = average temperature of March, X11 = average temperature of April, X13 = number of rainy day in April）. The significance test and the forecast confirmation showed that the regression equation reached the extremely remarkable level, and the rate of accuracy of reverse running reached 93.45% and the error rate of true reports was only 1.80% -5.11%. It was illustrated that the selected predictor relatively suited the Guizhou reality and the forecast model had great practical value.