摘要
根据江西井冈山森林火灾资料,利用一元回归和逐步回归分析方法,对井冈山1991—2000年森林火灾发生期间的连旱日数、降水量、相对湿度、气温、风速5个主要气象指标进行分析,将各指标分别与火灾发生次数、森林受害面积进行回归计算,建立了利用连旱日数和降水量预测森林火灾受灾面积及发生次数的数学模型。模式运算结果表明,当火灾发生前15 d连旱日数增加时,火灾次数增加;当火灾发生前15 d连旱日数增加、降水量减少时,火灾面积增加。
Based on the forest fire data in Jinggang Mountain, the univariate regression and stepwise regression analysis methods are used to analyze the five major meteorological elements, namely successive drought days, precipitation, relative humidity, air temperature and wind velocity, from 1991 to 2000 in Jinggang Mountain. Combined with the area and frequency of forest fire, each index is used in regressive calculation, and a mathematical model is established to forecast the area and frequency of forest fire by using successive drought days and precipitation. The result indicates that the frequency of forest fire increases when the successive drought days increase fifteen days before the fire, and the area of forest fire increases when the successive drought days increase and precipitation decrease fifteen days before the fire.
出处
《气象与减灾研究》
2007年第4期65-68,共4页
Meteorology and Disaster Reduction Research
基金
江西省防灾减灾研究基金项目(编号:GJZ200607)
关键词
森林火灾
气象要素
回归分析
数学模型
Forest fire, Meteorological elements, Regression analysis, Mathematical model
