Institute of Systems Engineering, Dalian University of Technology, No. 2 Linggong Road, Ganjingzi District, Dalian, 116024, China.
Institute of Systems Engineering, Dalian University of Technology, No. 2 Linggong Road, Ganjingzi District, Dalian, 116024, China.
J Environ Manage. 2020 May 15;262:110341. doi: 10.1016/j.jenvman.2020.110341. Epub 2020 Mar 3.
Serious PM2.5 air pollution has persistently plagued and endangered most urban areas in China in recent years, and targeted policies are necessary to improve urban air quality ranging from macro policy (national level) to medium policy (city level) to micro policy (site specific). However, the macro-pattern study of air pollution between Chinese cities is inadequate, and not conducive to the formulation of macro-policy. To bridge this gap, we applied a sequential pattern mining algorithm to analyze the spatial-temporal patterns of PM2.5 pollution across Chinese cities during the period 2015 to 2018. The sequential patterns were collected from three levels of granularity on geographic areas and ten temporal scenarios covering time intervals from 10 to 100 h. Many underlying associative relationships were revealed between different cities by the mined patterns. The patterns were heterogeneous and presented five characteristics (i.e., clustering, symmetry, imbalance, decay, and stability). Each of the urban areas under investigation at different granularities was analyzed to identify the occurrence of associative relationships between it and other urban areas; moreover, we determined the degree of severity of such relationships. Our research results provide solid data that can be used as a reference by the various levels of Chinese governments for decision-making; overall, they can be used to improve the design of joint policies to prevent and control PM2.5 pollution in Chinese urban areas.
近年来，严重的 PM2.5 空气污染持续困扰和威胁着中国大多数城市，需要采取有针对性的政策来改善空气质量，范围从宏观政策（国家层面）到中观政策（城市层面）再到微观政策（特定地点）。然而，中国城市之间的宏观空气污染模式研究还不够充分，不利于制定宏观政策。为了弥补这一差距，我们应用了一种序列模式挖掘算法来分析 2015 年至 2018 年期间中国城市 PM2.5 污染的时空模式。从地理区域的三个粒度级别和涵盖 10 至 100 小时时间间隔的十个时间场景中收集了序列模式。通过挖掘的模式揭示了不同城市之间许多潜在的关联关系。这些模式具有异质性，呈现出五个特征（即聚类、对称、不平衡、衰减和稳定）。对不同粒度的调查城市区域进行分析，以识别与其他城市之间关联关系的发生情况，并确定这种关系的严重程度。我们的研究结果提供了可靠的数据，可以作为中国各级政府决策的参考；总体而言，这些数据可用于改进联合政策的设计，以防止和控制中国城市的 PM2.5 污染。
