中文摘要：

面对严重的雾霾天气以及雾霾污染边界不断扩张的严峻挑战,加快创新大气污染联防联控体系以形成跨区域协同治污合力势在必行.本文基于京津冀、长三角、珠三角、成渝、长中游等五大地区96个城市2015年的空气质量指数（AQI）以及PM2.5、PM10、SO2、CO、NO2、O3等6种分项污染物的逐日数据,从时间序列数据“预测能力”的视角,在向量自回归模型框架下识别雾霾污染的城市间动态交互影响效应,运用社会网络分析方法刻画雾霾污染空间关联的网络结构特征.在此基础上,运用二次指派程序从分项污染物视角考察雾霾污染空间关联的关键诱因,并利用双变量Moran指数揭示雾霾污染与其影响因素之间的空间相关性.研究发现,城市雾霾污染之间存在普遍的动态关联关系且呈现出联系紧密、稳定性强、带有明显特征的多线程复杂网络结构形态.不论在地区内部还是在全部样本城市当中,均不存在孤立的城市节点,这意味着面对雾霾污染的空间关联网络,任何一个城市都不能独善其身,均受到来自地区内部和地区以外其他城市以及它们构成的空间关联网络的影响.在六种分项污染物中,PM2.5的空间关联是导致雾霾污染空间关联的最主要诱因.城市雾霾污染与其影响因素尤其是城市人口密度、投资强度、工业污染排放之间存在显著的空间相关性.基于上述结论,中国应当加快构建以防控PM2.5为重点的跨区域雾霾污染协同治理机制,并将其融入城市群发展战略以及区域发展战略之中,最终实现包含雾霾污染协同治理在内的全方位的区域协同发展.

英文摘要：

Facing severe challenge of regional haze pollution and its boundary expansion,it is imperative to form interregional pollution governance force through the innovation of joint prevention and control system.Based on 96 cities＇ daily data of air quality index （AQI） and six sub-pollutants such as PM2.5,PM10,SO2,CO,NO2,O3 of Beijing-Tianjin-Hehei region,the Yangtze River Delta,Pearl River Delta,Chengdu-Chongqing region and the middle reaches of Yangtze River,the article identified dynamic interaction effects among urban pollution with the VAR framework from the perspective of forecast capability of time series and revealed the network structure characteristic of spatial correlation of haze pollution by the method of Social Network Analysis （SNA）.Meanwhile,the article used quadratic assignment program to reveal the key incentives of spatial correlation of haze pollution within sub-pollutants,and exploited bivariate Moran index to describe spatial correlation between haze pollution and its influential factors.The main results were as follows：AQI and six sub-pollutants among cities all had multithreading and dynamic correlation which presentsed complex network characteristic with close linkage and strong stability.Meanwhile the spatial correlation network of urban haze pollution had a significant small-world characteristic.There is no isolated urban node both within the region and in the entire sample cities,faced with haze pollution spatial correlation network,which means that no city could be immune to haze pollution,and would be all influenced by other cities that are built the spatial correlation network from inside and outside the region.In the six sub-pollutants the spatial correlation of PM2.5 was the main cause of urban spatial correlation of haze pollution.Urban haze pollution and its influence factors,especially the urban population density,investment intensity,industrial pollution emissions,had a significant spatial correlation.Based on the above conclusion,China should accelerate to build a cross