Correlation Analysis Between PM2.5 And Meteorological Conditions And Forecasting Of PM2.5

In recent years,air pollution is getting worse and haze occurs frequently in Jinan.PM2.5 particles can more easily penetrate into lungs and increase respiratory and mutagenic diseases.The research on the relationship between PM2.5 and meteorological conditions has a realistic and positive significance for the assessment and protection of the environment.This study mainly analyzes the correlation between PM2.5 and meteorological condition as well as how to forecast PM2.5.We determine that the daily average and daily maximum of the PM2.5 concentration have a great variation among different seasons,the concentration is highest in winter,and lowest in summer by analyzing the annual distribution characteristics of the PM2.5concentration in Jinan.The maximum of daily average and daily maximum of the PM2.5concentration appeares in January whereas the minimum appeared in August.The frequency of the concentration getting greater than 75μg/m~3 was the largest in winter.The frequency of high concentration in winter is much higher than that in other seasons,which indicates that the frequency of winter pollution events occurred frequently.Besides,the occurrence of severe PM2.5 pollution situation has decreased year by year.In order to further study the relationship between PM2.5 concentration and meteorological conditions,wind rose diagrams of the daily average and daily maximum PM2.5 concentration are drawn to study the relationship between the PM2.5concentration and wind direction.For the first time,it is found that the relationship between PM2.5 concentration and wind direction is very close to the special topography of Jinan.The southeast and northwest wind cannot promote the spread of pollutants,resulting in serious pollution,but northeast wind or southwest wind is more conducive to reduce the concentration of PM2.5.Meanwhile,correlation analysis between PM2.5 and meteorological conditions indicates that PM2.5 concentration is significantly related to wind speed(average and maximum),temperature,pressure,relative humidity,precipitation,sunshine duration,etc.Higher wind speed is beneficial to the spread of PM2.5.Higher pressure and relative humidity result in higher PM2.5 concentration.Longer sunshine duration,higher temperature and higher precipitation all lead to lower PM2.5.We choose the average and maximum of PM2.5 concentration of the 3 years as explained variable,statistical prediction model between PM2.5 and meteorological factors is established by multiple linear regression,stepwise regression,principal component analysis and partial least squares step by step and the statistical tests is done.Since PM2.5 concentration and meteorological elements varies in different seasons,the spring,summer,autumn and winter data are selected as the research object respectively,statistical forecast model of the average PM2.5 concentration and the maximum PM2.5concentration of four seasons is established through the above four methods.Finally,on the basic of these studies,a comprehensive PM2.5 forecast platform with user interaction interface is built by using the Gui interface of Matlab(2015b).After testing,the forecasting platform is running well.