The Analysis Of Evolution For PM2.5 In Chongqing Based On Time Serials

Smog has become a daily air pollution problem. PM2.5 is the main component of smog, and it is also an important indicator for measuring air quality, whose variation of concentration directly reflects the variable of air quality. Studying the formation and composition of PM2.5 is extremely important. However, up to now, there is not a consensus yet on it. Therefore, we should try our best from different angles to make further research on the occurrence and evolution of PM2.5.From the view of time series, the paper devotes this article to analyze the variation of PM2.5 concentrations in Chongqing with the date changing. According to some literatures, we can firstly find out that PM2.5 may be related to some known variation, such as: temperature, CO, PM10, NO2 and SO2 etc., and then we collect related data across the network and integrate the data by some preprocessing. Secondly, we study each variable sequence itself respectively to identify their individual adaptation model. Furthermore, we output PM2.5 and input the sequences which are maximum temperature, minimum temperature, CO, PM10, NO2 and SO2 to construct single input variable transfer function model and multiple input variables transfer function model to analyze their quantitative relations. Finally, we constructed an ordinary multiple linear regression model and cointegration model between PM2.5 and temperature, CO, PM10, NO2 and SO2, and further clarify their relationship by comparing the three models.Each sequence itself is not stable, and they all have their own variation. Through the establishment of transfer function model of a single input variable of PM2.5 with maximum temperature, minimum temperature, CO, PM10, NO2 and SO2, it shows that PM2.5 has a significant correlation with each input variable. From the superior fitting effect between PM2.5 and each factor, PM2.5 is related to the combination of factors. In addition, through the comparison of multiple linear regression model and cointegration model analysis, this combined effect is not a simple linear relationship, which shows that PM2.5 concentrations are not only under the influence of related factors values in the current period as well as the interference of random error, but also significantly affected by their own values and various factors early.