浦江县褐飞虱预测预报模型选择与优化

Optimization of Forecasting Model for Brown Plant Hopper in Pujiang County

[目的]探讨更适用于现阶段的褐飞虱预测预报模型。[方法]利用浦江县2001—2016年褐飞虱的观测数据和气象数据,采用传统的逐步回归方法和BP神经网络方法,分别建立了褐飞虱发生高峰期预测预报模型。[结果]逐步回归预测模型选用的建模因子为5月上旬最高温、9月上旬湿度和6月下旬雨量,模型的预测准确率不高;BP神经网络预测模型的建模因子为始见日后40 d的平均温、最高温、最低温、雨量、湿度,模型的预测准确率达99.22%。[结论]该研究结果为今后褐飞虱预测预报模型的选择提供了参考。

[Objective]The aim was to find a more suitable forecast model for the current stage of brown plant hopper prediction. [Method]By means of the traditional stepwise regression method and BP neural network method,the peak forecasting model of brown plant hopper was established based on the observed data of brown plant hopper and meteorological data in Pujiang County from 2001 to 2016. [Result]The modeling factors of the stepwise regression model were the highest temperature in early May,humidity in early September and rainfall in late June.The prediction accuracy of the model was low. The modeling factors of BP neural network prediction model were the average temperature,the highest temperature,the lowest temperature,the rainfall and the humidity for the 40 days after the first meet. The prediction accuracy of the model was 99. 22%. [Conclusion]The results can provide a reference for the future selection of the prediction model of brown plant hopper.