福建林火发生的驱动因子及空间格局分析  

Driving factors and spatial distribution patteren of forest fire in Fujian Province

作　　者：苏漳文 刘爱琴[1] 郭福涛[1] 梁慧玲[2] 王文辉[1] 林芳芳[2] 

机构地区：[1]福建农林大学林学院,福建福州350002 [2]福建农林大学计算机学院,福建福州350002

出　　处：《自然灾害学报》2016年第2期110-119,共10页Journal of Natural Disasters

基　　金：国家科技支撑计划项目(2014BAD15B02);福建省教育厅资助省属高校专项(JK2014012);福建省自然科学基金项目(2015J05049)~~

摘　　要：福建省是我国的林火高发区,但目前国内关于该地区林火发生特性等方面的研究还不够深入。国内外研究表明,林火的发生受到气象、植被、地形以及社会经济等因素的综合影响。运用Arcgis10.0等相关软件对林火潜在影响因子(气象、植被、地形等)进行空间信息提取,并与火点和对照点结合通过Logistic回归分析和receiver operating characteristic(ROC)检验方法分析了福建地区林火发生的驱动因子,并获得了林火预测模型。以预测模型为基础对福建地区火险进行了区划。结果显示:"点到铁路的距离","高程","人口密度","日最高地表气温"等7个因素对当地林火发生有显著的影响。此外,模型整体拟合水平和预测精度较高,对当地的林火管理工作具有一定的参考价值。Fujian province is a high - frequency region of forest fire. However, the study related to occurring char- acteristics of the forest fire is relatively rare. Studies at home and abroad showed that, forest fire is affected compre- hensively by climate, vegetation, topography and socio - economy factors. In this study, Arcgis 10.0 and other software were used to extract the spatial information of potential influence factors such as climate, vegetation, topog- raphy and socio- economy factors. Combining with five points and contrast points, logistic regression method and receiver operating characteristic （ROC） test method were used to analyze the driving factors of forest fire in Fujian region and establish forest fire prediction. Based on the prediction model, the fire risk in Fujian regions were zoned. The results showed that, the seven driving factors, such as distance to railway, elevation, density of popu- lation, daily maximum surface temperature, etc, had significant influence on the local fire occurrence. In addition, the fitness and prediction accuracies of the model were high, which could provide some reference for the local forest fire prediction management.

关 键 词：福建 人为火 逻辑斯蒂回归 驱动因子 模型预测 

分 类 号：S762.2[农业科学—森林保护学]