基于时间序列模型残差的中国东部地区空气质量指数（AQI）空间自相关特征分析

空间自相关分析可以揭示变量的空间聚集性质.基于中国东部城市群的日空气质量指数(AQI)数据,研究了AQI的空间自相关特征.同时,考虑到日AQI的空间非平稳性,分析分成两步进行.首先,对每个城市的日AQI序列建立时间序列模型,进而获得其标准残差序列;然后再基于残差项进行空间自相关分析,计算其全局Moran's I指数和局域Moran's I指数.结果表明,全局Moran's I指数体现出明显的季节变化特征,呈现出冬季高而夏季低的以年为周期的循环变化.通过分析局域Moran's I指数,发现中国东部城市群存在2个值得重视的高空间自相关区域:京津冀地区和长三角地区.由于残差项体现的是气象条件的影响,因此,大面积的高自相关区体现了邻近城市群空气质量对气象条件变化的同步响应特征.

Spatial autocorrelation analysis of air quality index (AQI) in eastern China based on residuals of time series models

Spatial autocorrelation analysis can be used to identify characteristics of spatial clustering of variables. In this study, spatial autocorrelation analysis is applied on daily air quality index(AQI) data of cities in eastern China. Considering the spatial non-stationarity in daily AQI data, the analysis is based on two steps. At first, time series models are fitted based on daily AQI series of each city and the corresponding standardized residuals series are achieved. Then, these standardized residuals are used to calculate the global Moran''s I index and local Moran''s I index. The results indicate an annual cycle of the global Moran''s I index of higher in winter and lower in summer. The local Moran''s I index shows higher autocorrelation in two regions: Beijing-Tianjin-Hebei and the Yangtze River Delta. Because the residuals of time series models reflect the impacts of meteorological conditions, higher spatial autocorrelation indicates synchronous responses of air quality in adjacent cities to meteorological factors.