基于卫星监测热点的林火分布及发生预报模型

Research on spatial distribution of forest fire based on satellite hotspots data and forecasting model

卫星林火监测作为森林火灾预防和扑救工作的重要方法,在我国森林防火工作中得到广泛推广和应用。基于卫星监测研究全国各省份的林火热点分布规律以及热点集中区域的气象因子与火灾频率的关系,以我国各地区2010—2015年春季森林防火期(3月1日—6月1日)的卫星监测热点数据为研究对象,分析我国各省林火的年际变化及区域分布特征;通过提取火场时空因子,结合当年气象数据,建立热点集中区域的火灾趋势回归模型。结果表明,卫星监测热点统计的森林火灾次数与当地林业局统计的火灾数据吻合度高。2010—2015年春季林火热点从整体趋势上看,森林火灾的热点数量在逐年递减,主要集中在中国的西南部区域,最多的省份为云南省,占林火热点总数的20%;其次为四川省,占林火热点总数的13%。云南省的春季林火次数多元回归方程,复相关系数R=0.838**;四川省的春季林火次数多元回归方程,复相关系数R=0.744*,回归模型都达到了显著水平。卫星监测热点用于林火监测,可以高精度统计森林火灾次数,回归模型对于林火的预测预报具有一定适应性,为森林火灾发生预测预报提供数据支持。

As an important method of forest fire forecast and fire suppression, satellite forest fire monitoring, with thecharacteristics of wide monitoring range, high frequency and short delay, has been widely introduced and applied inforest fire prevention of China. In this paper, the spatial distribution pattem of the forest fire hotspots and the relation-ship between the weather factors and forest fire frequency were studied based on the nationwide satellite monitoring.The satellite hotspots data of spring fire season（March 1st to June lst） from 2010 to 2015 were used to analyze theannual characteristics of forest fires in all the provinces. The multivariate statistical regression analysis was used to es-tablish forest fire frequency regression models in the major disaster provinces, i.e., Yunnan and Sichuan, combinedwith current weather data. The results showed that forest fires frequency provided by satellite hotspots monitoring coin-cided well with that provided by local forest administration. The number of forest fire hotspots declined year by year asa whole, and they were mainly concentrated in the southwest of China, where Yunnan and Sichuan were the provinceswith the most hotspot numbers, which accounted for 20% and 13% of the whole respectively. The multiple correlationcoefficients（R） of Yunnan and Sichuan forest fire frequency regression models were 0. 838 ＊ ＊ and 0. 744 ＊, respec-tively, both of which reached a significant level. The comparison between the model and the real values showed thatthey were much consistent. The satellite hotspots monitoring used in the forest fire monitoring could count the numbersof forest fires with high accuracy. The enhancement of fire sources management was an important and effective meas-ure to reduce the risk of forest fire. The regression model can be adapted in the forest fire prediction and provide thedata support for the decision makers in the field of forestry in China.