摘要
为了探究我国火灾空间聚集特征与影响因素的空间异质性,采用全局莫兰指数、局部莫兰指数、逐步回归模型、地理加权回归模型、地理探测器方法对我国地级市单元进行研究。研究结果表明:我国火灾发生率具有显著的聚集性;我国火灾发生率较低的"冷点"区域有1个,火灾发生率较高的"热点"区域有4个;人均GDP、城镇居民人均可支配收入、人口密度、年平均气温4个因素的影响效应具有空间异质性。人均GDP与火灾发生率为正相关,另外3影响因素对火灾发生率的作用表现出正负2种相关关系;2因素交互作用要比单因素作用于火灾发生率时影响力更显著,各影响因素的交互作用类型有非线性增强型和双因子增强型2种。
In order to explore the spatial accumulation characteristics and the spatial heterogeneity of influencing factors of fire in China,the units of prefecture-level city in China were studied by using the methods of global Moran’s I,local Moran’s I,stepwise regression model,geographically weighted regression(GWR)model and GeoDetector.The results showed that the fire incidence in China presented the significant aggregation.There existed one“cold spot”area with low fire incidence and four“hot spot”areas with high fire incidence in China.The influence effect of four factors including GDP per capita,urban per capita disposable income,population density and annual mean temperature had the spatial heterogeneity.There was a positive correlation between GDP per capita and fire incidence,while the action of other three influencing factors on the fire incidence presented both the positive and negative correlation.The interaction between two factors had more significant influence on the fire incidence than single factor,and the interaction types between the factors included the nonlinear enhancement and the bi-factor enhancement.
作者
杨尧
李功权
YANG Yao;LI Gongquan(School of Geosciences,Yangtze University,Wuhan Hubei 430000,China)
出处
《中国安全生产科学技术》
CAS
CSCD
北大核心
2018年第9期158-163,共6页
Journal of Safety Science and Technology
关键词
火灾发生率
空间异质性
莫兰指数
地理探测器
地理加权回归
fire incidence
spatial heterogeneity
Moran’s I
GeoDetector
geographically weighted regression(GWR)