基于GWR模型的中国城市雾霾污染影响因素的空间异质性研究

基于全国城市的PM_2.5监测数据,识别PM_2.5的时空分布特征,并着重利用地理加权回归模型分析自然和社会经济因素对PM_2.5影响的空间异质性。结果显示：2015年全国PM_2.5的年均浓度为50.3 μg/m3,浓度变化呈现冬高夏低,春秋居中的“U型”特征;PM_2.5的空间集聚状态明显,其中京津冀城市群是全国PM_2.5的污染重心。地理加权回归结果显示：影响因素除高程外,其余指标均呈现正负两种效应,且影响程度具有显著的空间差异性特征。从回归系数的贡献均值来看,自然因素对城市PM_2.5浓度影响强度由高到低依次是高程、相对湿度、温度、降雨量、风速、植被覆盖指数;各类社会经济指标对城市PM_2.5浓度影响强度排名依次是人口密度、研发经费、建设用地比例、产业结构、外商直接投资、人均GDP。由于各指标对城市PM_2.5浓度变化的影响程度存在着空间异质性,因此在制定大气治理对策时可以考虑不同指标影响程度的空间差异,从而使得治霾对策更具针对性。

Spatial heterogeneity of driving factors of urban haze pollution in China based on GWR model

Based on the PM_2.5 monitoring data of China's cities, we identified the spatial and temporal distribution characteristics of PM_2.5 concentrations, and used the geographically weighted regression (GWR) model to analyze emphatically the spatial heterogeneity of the influence of natural factors and socio-economic factors on PM_2.5 concentrations. The results showed that: in 2015, the average annual concentrations of PM_2.5 in China was 50.3 μg/m ³, and the monthly concentration change presented a "U-shaped" pattern with a higher level in autumn and winter while a lower one in spring and summer. In addition, PM_2.5 concentrations were high in cities of eastern and northern China, but low in cities of southern and western China. Beijing-Tianjin-Hebei urban agglomeration was the center of PM_2.5 pollutions in China. The results of geographically weighted regression showed that: (1) in terms of natural factors, elevation had a negative correlation with the urban PM_2.5 concentrations, while positive and negative correlations exist for other indexes, and negative correlation effect dominated, which is conducive to reducing PM_2.5 concentrations in most cities. Thus it can be seen that the influence indexes of PM_2.5 concentrations have significant spatial difference characteristics. From the mean contribution of the regression coefficient, the ranking of the influence intensity of natural indexes on PM_2.5 concentrations were: digital elevation model, relative humidity, temperature, rainfall, wind speed, normalized difference vegetation index. (2) In terms of socio-economic factors, all the indicators showed positive and negative effects, with significant spatial heterogeneity. Among them, the build-up and GDP per capita were conducive to reducing PM_2.5 concentrations in most cities, while population density, foreign direct investment, industrial structure and research and development expenditure can aggravate the air pollution in regions. The ranking of the influence intensity of socio-economic factors on PM_2.5 concentrations were: population density, research and development expenditure, built-up, industrial structure, foreign direct investment, GDP per capita. (3) Due to the spatial heterogeneity of the influence of various factors on urban PM_2.5 concentrations, the spatial difference of the influence of various indexes can be taken into account in the formulation of atmospheric governance countermeasures. Moreover, although natural factors have a more significant influence on PM_2.5 concentrations, since it is difficult to change the natural conditions of cities artificially, specific strategies should be proposed from the perspective of social and economic factors in tackling haze.