基于主成分和聚类分析的柳州市森林火灾特征研究

Characteristics of Forest Fire in Liuzhou:Based on Principal Component Analysis and Clustering Analysis

为给柳州市预防森林火灾提供参考依据,根据柳州市2010—2017年的森林火灾资料,采用常规统计方法对柳州市森林火灾的时空动态和分布规律进行研究,利用主成分分析和聚类分析对柳州市森林火灾形势进行评价。结果表明:(1)柳州市森林火灾的发生年际波动显著,2017年林火发生次数最高,其次是2011年和2014年。林火主要集中在秋末、冬季和春初;(2)从空间分布看,森林火灾主要发生在中南部的鹿寨县、柳江区、柳城县,以鹿寨县最为集中,火场总面积及受灾森林面积也最大。(3)在火灾形势方面,按照严重程度将森林火灾分为8类,其中鹿寨县的火灾形势较为严峻,级别波动明显,其次是柳城县的火灾形势呈波动恶化态势。研究结果为建立柳州市森林火险气象等级预报提供了借鉴。

The paper aims to provide references for prevention and forecast of forest fire in Liuzhou. Based on the forest fire data of Liuzhou from 2010 to 2017, we studied the spatio-temporal dynamic and distribution law of forest fire in Liuzhou by using conventional statistical methods, and evaluated the situation of forest fire by the methods of principal component analysis and cluster analysis. The results showed that:(1) the occurrence of forest fires in Liuzhou had an obvious interannual fluctuation, with the highest number of forest fire in 2017,followed by 2011 and 2014;forest fire mainly concentrated in late autumn, winter and early spring;(2) in terms of spatial distribution, forest fires mainly occurred in Luzhai, Liujiang and Liucheng in south central Liuzhou,and were the most in Luzhai, causing the largest total fire burned area and affected forest area;(3) in terms of fire situation, forest fires were divided into 8 categories according to the severity, among them, the fire situation in Luzhai was severe with obvious grade fluctuation, followed by the fire situation in Liucheng, which showed a worsening trend with fluctuation. The results provide a reference for establishing the meteorological level forecast of forest fire risk in Liuzhou.