基于DCCA方法的成都市市区与周边城镇大气污染长程相关性分析

利用2008年1月～2012年12月期间成都市市区与其周边14个城镇的逐日空气污染指数（API）数据,采用去趋势互相关分析法（DCCA分析）研究区域城市间大气污染的相关性及其随时间演变规律。结果表明:成都市区与其周边城镇大气污染的空气污染存在一定程度的相关性,并且在不同的月份,相关性会随之发生变化。这种相关性具有长期持续特征,具体表现为在一定的时间尺度上成都市区与其周边城镇大气污染的相关性随时间的变化并不遵循经典的马尔可夫过程,即不随时间呈现指数快速衰减,而是以幂律形式随时间缓慢衰减。进一步,结合不同月份间区域大气平均流场的分布特征,探讨了DCCA分析结果的科学性。研究结果说明,在特定地理环境和气象条件控制下,成都市及其周边城市之间已经存在明显的相互输送和耦合作用,成都地区大气污染已由局地性污染转变为区域性大气污染。城市间污染物输送的长期相关影响机制极有可能在特定月份加重成都市区的空气污染状况,这在成都市区及其周边城镇的规划建设中必须加以特别重视

LONG RANGE CORRELATION OF AIR POLLUTION BETWEEN CHENGDU CITY AND ITS SURROUNDING TOWNS

We aims to report the cross correlation characteristics and the temporal evolution properties of air pollution between Chengdu City and 14 surrounding towns. In this study, we collected the daily air pollution index (API) data of Chengdu City and 14 surrounding towns provided by Sichuan Provincial Environmental Monitoring Center. The 5 years regular monitoring data, from Jan.1 2008 to Dec.31 2012, were analyzed by using the detrended cross correlation analysis (DCCA) method. The DCCA exponent quantitatively measures the cross correlation of two series in terms of long term cross correlation exists. The results show that the relationships of air pollution between Chengdu City and its surrounding towns display a long range cross correlation at the time scale ranging from one day to 5 years. The long range cross correlation suggests that the correlations of air pollution between Chengdu City and its surrounding towns do not obey the classical Markov type stochastic behavior (exponential decrease with time), but display more slowly decaying correlations in a power law fashion with time. For example, there is a varied tendency in air pollution of Chengdu City (its surrounding towns) to be followed by another varied tendency in air pollution of its surrounding towns (Chengdu City) at a different time in a power law fashion. It has been found that the long range cross correlation properties will follow changes in different months. In order to further clarify this result, the cross correlations properties were analyzed based on monthly mean wind field. It is found that the results of DCCA analysis and the monthly mean wind field between Chengdu City and its surrounding towns have a good synchronization. These results correspond to atmospheric circulation and regional transport hypothesis. It suggests that in specific geographical and meteorological conditions, it exists inter city transport of air pollutants between Chengdu City and its surrounding towns. The atmospheric pollution type of Chengdu Plain may have been changed from local pollution into regional air pollution. The inter city transport has been the important factor of air pollution in the Chengdu Plain, which provides conditions for the formation and aggravation of heavy air pollution in Chengdu City. Therefore, in the air pollution forecasting model of Chengdu City, the long range cross correlation properties of air pollution between Chengdu City and its surrounding towns should be considered. This paper helps to understand the nature of regional air pollution in the Chengdu Plain. The prevention and control of air pollution in the Chengdu Plain should be developed towards regional jointly prevention and control in the planning and construction of Chengdu City and its surrounding towns