运用马尔科夫链模型预测张家界永定区松毛虫发生面积

Prediction of occurence area of Dendrolimus in Yongding District of Zhangjiajie using Markov chains

以湖南省张家界永定区2003—2018年马尾松毛虫、云南松毛虫发生情况的历史数据为材料,探讨了不同等级划分体系对应用马尔科夫链模型预测马尾松毛虫、云南松毛虫混合发生面积的历史符合率的影响,并对2019年张家界永定区的马尾松毛虫、云南松毛虫发生等级及发生面积进行预测,从而指导防治工作科学、及时、有效地开展。结果表明,当原始数据集中于某一级别时,用3等级分级可以增加历史符合率。本文数据分为3个等级,历史符合率为76.9%。利用概率矩阵对2019年马尾松毛虫、云南松毛虫混合发生面积进行预测,结果为1级,即发生面积为0~1 033.3 hm 2。

Based on the archive data of occurrence of Dendrolimu spunctatus and Dendrolimus houi in Yongding District, Zhangjiajie, Hunan Province from 2003 to 2018, we explored the effects of different grade hierarchies on the historical coincidence rate of occurrence area prediction using Markov chain model, and predicted the occurrence grade and mixed occurrence area in 2019, in order that the control work could be implemented scientifically, duly and efficiently.The results showed that the use of 3-grade could increase historical coincidence rate when primary data were centralized in one of the grades.Therefore, the 3-grade hierarchy was used in this paper, resulting in a historical coincidence rate of 76.9%.Using probability matrix to predict, the result showed mixed occurrence area of Dendrolimus punctatus and Dendrolimus houi in 2019 was at level 1 and the mixed occurrence area was 0~1 033.3 hm^ 2.