京石两地“雾霾”天气的相关性研究——基于Copula函数的方法

Study on the Correlation of"Fog and Haze"Weather in Beijing and Shijiazhuang:Based on Gopula Function

运用Copula函数研究北京和石家庄与雾霾相关的空气污染物的相关特征,通过平方欧式距离检验,认为应该利用Gumbel Copula函数对数据进行拟合,PM10、SO_(2)与NO_(2)的Kendall秩相关系数分别为0.62、0.52和0.5,上尾相关系数分别为0.60、0.69和0.58,而下尾相关系数均为0,污染物的传播表现出显著的非对称性,即某地空气污染严重时会带动另一个地区污染物的上升,但一个地区良好的空气质量却不能引起另一个城市污染物的下降。

The Copula function is used to study the correlation characteristics of air pollutants related to smog in Beijing and Shijiazhuang.Through the square European distance test,it is believed that the Gumbel Copula function should be used to fit the data.The Kendall rank correlation coefficients of PM10,SO_(2)and NO_(2)are 0.62,0.52 and 0.5 respectively,the upper tail correlation coefficients are 0.60,0.69 and 0.58 respectively,and the lower tail correlation coefficients are 0.The propagation of pollutants shows significant asymmetry,That is,when the air pollution in one place is serious,it will drive the increase of pollutants in another area,but the good air quality in one area cannot cause the decline of pollutants in another city.