烟草青枯病早期预警模型的研究

打开文本图片集

中图分类号：S435 文献标识码：A文章编号：2097-1354（2025）03-0043-09

开放科学（资源服务）标识码（OSID）：

Study on Early Warning Model of Tobacco Bacterial Wilt

GENG Li1'²， WANG Hongtao²， HAN Yujing²， WANG Lili³，ZONG Hao³， LIU Chunju4， ZHANG Chao³， TAN Xiaolei³，

REN Guangwei²， WANG Xinwei²， CHEN Haitao'

1.Chongqing Branch of China National Tobacco Corporation，Chongqing 40oo23，China；

2.Instituteof Tobacco Research of CAAS，Key Laboratory of Tobacco Pest Monitoring Controlling

andIntegratedManagement，QingdaoShandong266101，China；

3.ShandongLinyi Tobacco Co.Ltd，LinyiShandong276o03，China；

4.ShandongWeifangTobaccoCo.Ltd，WeifangShandong261205，China

Abstract： In order to improve the level of monitoring and early warning for tobacco bacterial wilt，the effects of soil temperature，soil humidity，soil pH and number of pathogens on the disease epidemic were studied through orthogonal design. The study was conducted with a precisely controllable pot experiment in a controlled-climate room. The early warning model of tobacco bacterial wilt was established and verified by the field data. The results showed that： ① By binary logistic regression analysis，a prediction model for the disease occurrence probability was establishedTeregressionqtion ， and the overall correct percentage was 89.2% ② By stepwise regression analysis，the prediction model for disease index， disease occurrence degree， disease occurrence degree after 3O days of inoculation and the days of low-risk epidemic period were established，with the validation coeficient （R2） of o. （20 640 9， 0.715 3， 0.894 4 and O.818 4，respectively. This study laid a theoretical foundation for the realization of early warning and control of tobacco bacterial wilt.

Key words： tobacco bacterial wilt; binary logistic regression analysis;； stepwise regression analysis;warning model

烟草是我国重要的经济作物[1]，受烟草青枯病危害，每年烟叶产量损失达 8000t 以上。（剩余8582字）