中国农业气象

• 论文 • 上一篇    下一篇

渝西地区森林虫害的气象预报模型

何永坤;阳园燕;罗孳孳;   

  1. 重庆市气候中心,重庆市气候中心,重庆市气候中心 重庆401147,重庆401147,重庆401147
  • 出版日期:2008-04-10 发布日期:2008-04-10
  • 基金资助:
    中国气象局气象新技术推广项目(CMATG2006M44)

A Meteorological Forecast Model for Forest Pests in Western Chongqing

HE Yong-kun,YANG Yuan-yan,LUO Zi-zi(Chongqing Climate Center,Chongqing 401147,China)   

  • Online:2008-04-10 Published:2008-04-10

摘要: 对永川区1950-2002年逐年森林虫害发生面积的周期分析表明,渝西地区森林虫害大发生存在5~7 a周期,80年代呈准2 a周期高位震荡;通过对森林虫害与气象要素的数理统计分析得出,春季、盛夏期间温度、湿度及初夏强降水是影响渝西地区森林虫害的主要气象要素。在相关分析的基础上,利用人工神经网络方法建立了当地森林虫害面积的气象预测模型,1985-2002年的模拟预报的平均误差为4.85%,回代检验效果较好。

关键词: 森林虫害, 气象因子, 相关分析, 预测模型, 渝西地区

Abstract: Based on the analysis the observation data of the area of the forest pest occurrences from 1950 to 2020 in the Yongchuan District of Western Chongqing by period analysis method,the results showed that there existed a 5~7 years period for forest pest occurrences in the last 53 years and a quasi-2 years period in the 1980s.Through the statistical analysis of the forest pest occurrences and the meteorological elements,the temperature,atmospheric humidity in the spring and the midsummer,and the strong precipitation in the early summer were the main meteorological elements affecting forest pest occurrences in the Western Chongqing.Based on the correlation analysis,a meteorological forecast model for the forest pest occurrences was established by using Artificial Neural Network(ANN) method.The average errors of the forecast results during 1985 and 2002 were 4.85%.The validation by the back substitution showed a preferable result.

Key words: Forest Pest, Forest Pest, Meteorological element, Correlation analysis, Forecast model, Western Chongqing