基于MODIS卫星火点的浙江省林火季节变化及驱动因子

Seasonal variation and driving factors of forest fire in Zhejiang Province, China, based on MODIS satellite hot spots

研究林火变化趋势和驱动因子,可为林火预防和管理提供科学依据。本研究基于MODIS卫星火点数据,结合气象(日平均风速、日平均温度、日相对湿度、气温日较差、日累计降水)、人为(到公路距离、到铁路距离、到居民点距离、人口密度、人均GDP)、地形和植被因素(高程、坡度、植被覆盖度),运用趋势分析法、Logistic回归模型,对浙江省2001-2016年林火变化趋势和驱动因子进行研究。结果表明:浙江省春、夏季林火呈显著上升趋势,秋、冬季林火呈先上升后下降趋势,秋季下降趋势显著。浙江省各季节林火预测模型拟合度均较高,模型预测准确率分别为75.8%(春季)、79.1%(夏季)、74.7%(秋季)和79.6%(冬季)。浙江省春、夏季林火发生与变化受气象、人为、地形和植被因素的显著影响;秋、冬季林火发生与变化主要受气象因素影响。在影响因素复杂、高火险区域分散的春、夏季,林火管理应重点加强人为活动管理和防火宣传教育;在秋、冬季,可通过在高火险区集中分布的西南地区增设瞭望塔和监控设备进行监测和管理。

Understanding the changes and driving factors of forest fire can provide scientific basis for prevention and management of forest fire. In this study, we analyzed the changes and driving factors of forest fire in Zhejiang Province during 2001-2016 based on trend analysis and Logistic regression model with the MODIS satellite fire point data combined with meteorological(daily ave-rage wind speed, daily average temperature, daily relative humidity, daily temperature difference, daily cumulative precipitation), human activities(distance from road, distance from railway, distance from resident, population, per capita GDP), topographic and vegetation factors(elevation, slope, vegetation coverage). The results showed that the number of forest fires in spring and summer had significantly increased, while the forest fires in the autumn and winter increased first and then decreased. Forest fire in autumn significantly declined. The four seasons’ fire occurrence prediction models had good prediction accuracy, reaching 75.8%(spring), 79.1%(summer), 74.7%(autumn) and 79.6%(winter). The meteorological, human activity, topographic and vegetation factors significantly affected fire occurrence in spring and summer, while meteorological factors were the main fire drivers in autumn and winter in Zhejiang. The focus of forest fire management should be on human activities. Fire prevention campaign should be done in spring and summer when high-risk forest fires were scattered in the study area. In autumn and winter, observatory and monitoring equipment could be built to facilitate fire management and detect in the area of high fire risk that was concentrated in the southwest region.