• 关键词：空气质量指数  空间回归模型  矩阵指数空间设定  衰减效应

• 基金项目：国家自然科学基金项目（No.71742001，41761021）；教育部人文社会科学研究青年基金项目（No.17YJC790061）；中国博士后科学基金面上资助项目（No.2017M621740）

• 姜磊
• 浙江财经大学经济学院, 杭州 310018

• 周海峰
• 浙江财经大学经济学院, 杭州 310018

• 柏玲
• 南昌大学经济管理学院, 南昌 330031

• 陈忠升
• 西华师范大学国土资源学院, 南充 637002

• 摘要：中国面临着严重的环境恶化问题，其中空气污染问题尤为突出.基于中国2014年全国城市空气质量数据，利用自然正交函数分析城市空气质量指数的时空演变特征，并采用基于衰减效应的矩阵指数空间设定模型探讨了空气污染的影响因素.研究结果表明：① AQI以京津冀为高值中心向周边地区呈衰减变化，污染核心区由京津冀逐渐向豫北地区和鲁西北地区扩散，年内AQI在1-9月呈现下降趋势，然后逐渐上升；② 150个城市的空气质量指数呈现出明显的空间集聚；③矩阵指数空间设定模型优于空间滞后模型，并且城市间空气污染呈现出显著的空间衰减效应；④人均地区生产总值的提高、SO₂排放量和PM_2.5浓度增加是导致空气污染加剧的重要原因，而外商直接投资和环保意识的提高有助于改善中国的空气质量.

• Abstract：China is suffering from the severe environmental degradation, notably air pollution. This paper uses a cross-sectional dataset of China's air quality index at the city level, applies empirical orthogonal function method to analyze spatio-temporal changes of air quality index, and then employs matrix exponential spatial specification model based on decay effects to investigate the influencing factors of air quality index. The findings are the following. ① High air quality index values are centered on the Jing-Jin-Ji region, presenting decay effects. The core area of air pollution moves from the Jing-Jin-Ji region to Northern Henan and Northwestern Shandong. Moreover, air quality index value decreases from January to September and then rises till December. ② Air quality index of China's 150 cities presents significant spatial clusters. ③ Matrix exponential spatial specification model is better fitted than spatial lag model. Moreover, air pollution presents significant decay effects across cities. ④ The rise of per capita GDP, the increase of SO₂ emission and PM_2.5 concentration are attributed to worsening air pollution while foreign direct investment and environmental awareness contribute to improving air quality in China.

• Key words：air quality index  spatial regression model  matrix exponential spatial specification  decay effects

• 
  摘要点击次数： 4300 全文下载次数： 3610