泰安地区金纹细蛾成虫发生期及发生量预测模型研究

为适时有效防治金纹细蛾，2008—2012连续5年，在泰安肥城苹果园使用信息素诱捕法检测了金纹细蛾成虫的发生动态。金纹细蛾一年发生5代，越冬代成虫发生量较低，果园第1、2代成虫发生量均远远低于防治指标，一般选择第3代作为防治的关键时期。据检测金纹细蛾成虫的发生量动态数据，以温度、湿度、降水、日照等气象资料作为气象因子，采用简单逐步回归建立了金纹细蛾第3代发生高峰期的预测预报模型：Y=-26．07719441＋0．03172086213X1＋0．26477681472X3；采用简单逐步回归和多因子互作回归建立了金纹细蛾世代发生量的预测预报模型：YI = 76. 7670083＋1.7896678278X1＋ 3.189238346X4 ; YII =- 518. 557711 - 3.220994915X1 ＋ 0.22643837696X1X2 ＋0.14479769093X2X4 ＋0.023406769049X3X5。经检验，简单回归模型能够较为准确预测出金纹细蛾第3代发生高峰期，多因子互作回归模型和简单回归模型均能准确预测金纹细蛾的世代发生量，其中以多因子交互回归模型预测结果更可靠。

Research on Prediction Model for Occurrence Period and Quantity of Lithocolletis ringoniella Adults in Taian Region

To effectively prevent and control the occurrence of Lithocolletis ringoniella in apple orchards of Feicheng in Taian region, the occurrence dynamics of adults were researched by pheromone trap method in the past five years from 2008 to 2012. Five generations of Lithocolletis ringoniella occurred in one year; the occurrence quantity of overwintering generation adults was lower and the occurrence quantity of the first and second generations were far less than the control standard, so the third generation was chosen as the key period of prevention and control generally. Based on the monitoring data of trapped pests collected from 2008 to 2012 combined with the weather factors including temperature, humidity, rainfall and illumination, the prediction model was established by simple regression for the occurrence peak of the third generation of LithocoUetis ringoniella, which was Y = -26. 07719441 ＋0.03172086213X1 ＋0. 26477681472X3. The prediction models for the occurrence quantity of every generation were also established by simple regression and multiple factor regression, and they were Y I =76. 7670083 ＋ 1. 7896678278X1 ＋ 3. 189238346X4 and Yn = - 518. 557711 -3. 220994815Xl ＋ 0. 2264387696XIX2 ＋ 0. 14479769093X2X4 ＋ 0. 023406769049X3X5. The results showed that all the models had better prediction effects. The prediction model of occurrence quantity established by multiple factor regression was more credible compared with that by simple regression.