摘要
以2001—2018年的浙江省MODIS-MCD64A1森林火灾数据为研究基础,运用逻辑斯蒂和随机森林方法结合气象、植被、地形、经济、人口和基础设施等因子对浙江省的林火时空变化及影响因子进行研究。结果表明:(1)2001—2018年,浙江省共发生森林火灾1 866次,年均发生林火104次,林火时空分布不均匀,空间上集中在温州中部和东部及与丽水市交汇区域、丽水市中部和北部区域、衢州中部、台州和宁波东部区域及金华、杭州和湖州的小面积区域;(2)时间上,森林火灾的发生次数在年际间的波动较大,总体呈先增加后降低的趋势,月火点分布不均衡,林火在10、11、12月和次年的1、2、3和4月的火点占全年总火点的95.28%以上;(3)基于随机森林算法筛选出9个林火发生重要因子,按重要性由大到小排序依次为年均温度、高程、年降水量、河流密度、月植被覆盖度、坡度、前一年植被覆盖度、人均国内生产总值、公路密度;运用逻辑斯蒂模型方法计算出林火发生与日平均气压、月植被覆盖度、坡向指数、坡度、日最大地表气温、公路密度、人均国内生产总值、居民点密度、河流密度之间呈显著正相关关系,与年降水量和高程之间呈显著负相关关系。
Based on MODIS image data of forest fires during 2001-2018 in Zhejiang province,temporal and spatial variation was studied by logistic and random forest model,combined with meteorology,vegetation,terrain,economy,population and infrastructure.The results showed that there were 1866 forest fires in the province,with an average annual of 104 times.The spatial distribution of forest fires was concentrated in the southern Zhejiang,such as central and eastern parts of Wenzhou and the intersection with Lishui city and the middle of Lishui.Forest fires mainly occurred in October,November,December,January,February,March and April,accounting more than 95.28%of the total.The most important factors for forest fire occurrence by random forest were ordered by annual average temperature,elevation,annual average precipitation,river density,monthly vegetation coverage,slope,vegetation coverage of previous year,GDP and road density.There were positive relation of forest fires with daily average air pressure,monthly vegetation coverage,aspect index,slope,daily maximum surface temperature,road density,GDP,residential density,and river density,while significant negative correlation with annual precipitation and elevation.
作者
李永和
靳全锋
LI Yong-he;JIN Quan-feng(Jinyun Forestry Bureau of Zhejiang,Lishui 323000,China;Lishui Vocational and Technical College,Lishui 323000,China)
出处
《浙江林业科技》
2021年第4期62-69,共8页
Journal of Zhejiang Forestry Science and Technology
基金
浙江大花园建设项目(No.DHYB1008)
浙江省教育厅一般项目(No.Y201840513)。
关键词
浙江
林火
时空变化
影响因子
Zhejiang province
forest fire
temporal and spatial variation
influence factor
