中国空气质量指数AQI的空间异质性研究

Research on the Spatial Heterogeneity of AQI in China

面对经济社会高质量发展、碳达峰、碳中和的多目标需求,PM2.5引发的雾霾天气,不仅仅是环境污染问题,更是与自然、经济和社会复合生态系统密不可分的系统问题,近年来已成为人类社会面临的严峻挑战。为深入探究在近似的气象条件下,空气质量的时空异质性特征及其波动周期,研究基于2014~2019年中国335个样本城市的空气质量监测数据,利用基于大数据的函数型数据分析方法对AQI的时间与空间部分进行分离,在此基础上通过信号分解方法分析空气质量指数(AQI)的波动周期;对于空间部分,通过全局空间自相关、局部空间自相关,分析AQI的空间分异特征,检验其局部区域内的集聚和分散效应,揭示各城市及其邻近城市的空气质量之间的空间自相关关系。结果表明,空气质量指数AQI存在波动周期,具有显著的先下降后上升的年度趋势。一年中,AQI有19个月的主周期和9个月的第二主周期;考虑空间特征,空气质量指数AQI存在空间分异特征,具有显著的全局空间正相关效应,即AQI指数越高(低)的地区越容易发生聚集现象;从局部空间特征来看,AOI的空间分布变化存在差异,城市及其邻近地区的AQI多表现为同质化聚集特征,且同质化聚集型城市占多数,证明了相邻区域空气质量存在交互作用。该研究创新性地使用大数据,长周期、全地域地系统化研究空气质量指数,为治理城市空气质量问题提供参考。

Faced with the demand for high-quality economic and social development, carbon peaking and carbon neutrality, the hazy weather caused by PM2.5 is not only an environmental pollution problem but also a systemic problem inseparable from the natural, economic, and social composite ecosystem, which has become a severe challenge for human society in recent years. To deeply in-vestigate the spatial and temporal heterogeneity characteristics of air quality and its fluctuation cycle under approximate meteorological conditions, the study is based on the air quality monitoring data of 335 sample cities in China from 2014~2019, and the temporal and spatial parts of AQI are separated using a functional data analysis method based on big data, based on which the air quality index (AQI) is analyzed by the signal decomposition method. For the spatial part, the global spatial autocorrelation and local spatial autocorrelation are used to analyze the spatial dispersion characteristics of AQI, examine its clustering and dispersion effects within the local area, and reveal the spatial autocorrelation between the air quality of each city and its neighbor cities. The results show a fluctuation cycle of AQI with a significant annual trend of decreasing and then increasing. In a year, AQI has the main cycle of 19 months and a second main cycle of 9 months;considering spatial characteristics, AQI has spatially divergent characteristics with significant global spatial positive correlation effects, i.e., the higher (lower) AQI index is the more likely to have aggregation phenomenon;in terms of local spatial characteristics, there are differences in the spatial distribution changes of AOI, and AQI of cities and their neighboring areas mostly The cities and their neighboring areas show homogeneous clustering characteristics, and homogeneous clustering cities are in the majority, which proves the interaction of air quality in neighboring regions. This research innovatively uses big data to systematically study air quality indices over a long period of time and across a wide geographic area to provide a reference for managing urban air quality problems.