中文摘要：

利用地理加权回归模型（GwR），实证考察了经济发展、产业结构、能源消费结构、城镇化发展、建筑施工、汽车保有量、环境规制、产业集聚、技术进步、能源价格、对外开放、自然条件、生态建设等因素对PM2．5年平均浓度影响的区域空间差异。实证结果表明：所有影响因素的估计系数符号与预期一致，大部分影响因素的估计系数统计量显著，其对PM2．5的影响程度在区域之间呈现空间差异，且表现出一定的空间集聚特征；环境规制、对外开放及空气流动性等因素对PM2．5影响程度的空间差异不大，均对雾霾减少有显著的正向影响，其他因素对PM2．5的影响程度则呈现出一定的空间梯度分布。

英文摘要：

Using the geographically weighted regression （GWR） model, this paper examines various factors＇ influence on the re- gional annual average concentration of PM2.5 empirically. The empirical results show that, the estimated coefficients of all the influ- ence factors are in line with expectations,most of the influence factors are statistically significant,and influence on PM2.5 present space differences between regions and also show certain degrees of spatial agglomeration phenomenon, environmental regulation, ex- panding the opening to the outside world and natural factors that strengthen the air flow cause nearly no spatial difference to PM2.5, and they have significant positive effects on haze reduce. While the other influence factors show some spatial gradient distribution im- oact on PM2.5.