基于状态转移网络的PM2.5时间变化特征与相似性分析

Analysis of Time Variation Characteristics and Similarity of Urban PM2.5 Based on State Transfer Network

摘要: 揭示城市雾霾演化规律对雾霾浓度预测和雾霾防治具有重要的现实意义。本文选取2019~2020年中国长三角地区四座城市上海、杭州、宁波、苏州的PM2.5日均浓度指数数据，用可视小图方法分别建立了这四座城市的PM2.5状态转移网络。通过分析这四个状态转移网络，发现全连接的基本图均为这四个转移网络中的重要节点。进一步计算了该节点的位置序列的扩散熵指数后发现该节点的出现存在长程关联。最后，通过矩阵相似性等方法对比了城市间PM2.5浓度演化的相似程度，发现宁波和苏州的PM2.5浓度演化规律相似度较高。

Abstract: It is of important practical significance to reveal the evolution law of urban haze for the prediction of haze concentration as well as the prevention and control of haze. In this article, we selected the dai-ly average concentration index data of PM2.5 in four cities in the Yangtze River Delta region of China from 2019 to 2020, including Shanghai, Hangzhou, Ningbo, and Suzhou. Then, we used the visuali-zation method to establish the PM2.5 state transition network of each city. The four state transition networks were analyzed, and it was found that the fully connected basic graph is the important mo-tif in these four transition networks. The diffusion entropy index of the position sequence of the mo-tif was further calculated, and it was found that there was a long-range correlation between the ap-pearance of the motif. Finally, the similarity degree of PM2.5 concentration evolution between cities was compared by methods such as matrix similarity, and it was found that Ningbo and Suzhou had high similarity in PM2.5 concentration evolution law.