Short- and long-term prediction models of rubber tree powdery mildew disease index based on meteorological variables and climate system indices

文献类型： 外文期刊

第一作者： Bai, Rui

作者： Bai, Rui;Wang, Jing;Chen, Renwei;Bai, Rui;Li, Ning;Wang, Jing;Li, Ning;Li, Ning

作者机构：

关键词： Rubber tree powdery mildew; Disease index; Climatic variable; Machine learning; Prediction model

期刊名称：AGRICULTURAL AND FOREST METEOROLOGY （ 影响因子：5.6； 五年影响因子：6.3 ）

ISSN： 0168-1923

年卷期： 2024 年 354 卷

页码：

收录情况： SCI

摘要： Rubber tree powdery mildew is the most serious disease of rubber trees in China. Climate is the most important factor influencing the occurrence and development of rubber tree powdery mildew. However, few studies combined the short- and long-term prediction models for rubber tree powdery mildew disease index to prevent and control rubber tree powdery mildew in advance. The study collected daily meteorological data, monthly climate system indices, and rubber tree powdery mildew disease index from 1951 to 2021 in China's rubber planting area. The meteorological variables and climate system indices affecting short-term (1-14 days prediction in advance) and long-term (1-5 months prediction in advance) rubber tree powdery mildew disease index were screened with factor expansion method. Random forest (RF) and artificial neural network (ANN) were used to construct the short- and long-term prediction models of rubber tree powdery mildew disease index. The results showed that a hybrid approach using RF and ANN could effectively predict rubber tree powdery mildew disease index. The maximum temperature had the highest importance in the short-term prediction model of rubber tree powdery mildew disease index. In the long-term prediction model of rubber tree powdery mildew disease index, trade wind-related climatic variables had the highest importance in Hainan rubber planting area, while subtropical high-, polar vortex- and sea surface temperature-related climatic variables had the highest importance in Yunnan rubber planting area. The short-term prediction of the rubber tree powdery mildew disease index was recommended to predict the 6th day in advance, while the long-term prediction was recommended in November of previous year. The study fills in the knowledge gaps in developing short- and long-term prediction models for rubber tree powdery mildew disease index and provides a method for early warning and prediction of rubber tree powdery mildew.

分类号：

  • 相关文献
作者其他论文 更多>>

微信小程序