基于多元回归法的松墨天牛成虫发生期预测

摘要

[Objective]The aim was to forecast occurring periods of Monochamus alternatus. [Method]Based on data of trap Monochamus alternatus at 3 trapping sites in Xianju County of Zhejiang Province from 2006 to 2011, combining local meteorological data, the multiple linear regression model was established by the methods of correlation coefficients and step regression. [Result] Using this model to predict emergence period of 2012 which were not involved in the model construction, the result indicated that the model has relatively high accuracy. [Conclusion] The prediction success rate was 80%at 15 trap datas.%[目的]预测松墨天牛成虫发生期。[方法]根据浙江省仙居县2006-2011年3个试验点林间松墨天牛诱捕记录和当地的气象资料，采用相关系数法和逐步回归法建立松墨天牛发生期预测预报的多元回归模型。[结果]用未参与建模的2012年诱捕数据进行发生期预测效果检验，结果表明有较高的预测精度。[结论]15组数据的预测成功率为80%。