Correlation Analysis Of Aerosol Optical Depth And Air Pollutants In Northeast China

With the accelerated development of Chinese overall economy,urbanization and the gradual expansion of the size of the national economy,there are more and more pollutants in the atmosphere,resulting in serious deterioration of air quality and increasingly severe air pollution.In recent years,the annual average of PM2.5 in the urban air of Dalian has been greater than 50 ?g/m3,far exceeding the recommended safe value of 10 ?g/m3 by the world health organization,which will have a certain impact on human health.At present,there are10 monitoring stations in Dalian,and the coverage of atmospheric monitoring base station is very limited.By studying and making use of the correlation between satellite MODIS AOD and atmospheric PM2.5,timely predicting the overall distribution of atmospheric PM2.5 in Dalian,and reducing the management cost of the environmental protection bureau of Dalian on the atmospheric PM2.5 pollution in the whole region can effectively solve the practical problem of the lack of construction and distribution of base stations for regional atmospheric monitoring.The selection and matching of AOD and PM2.5 mass concentration data of Dalian in the summer and autumn from 2015 to 2018 were carried out,and the change characteristics of AOD data were briefly analyzed to establish a unary regression fitting model.The correlation between AOD and PM2.5 mass concentration data has been studied,providing a scientific and effective means for the distribution and diffusion of PM2.5 mass concentration and the management of atmospheric environment quality in Dalian.Based on the ENVI software platform,MODIS satellite remote sensing data is used to retrieve AOD data in Dalian area.PM2.5 monitoring network is used to obtain PM2.5 hourly mass concentration data,and effective data can be obtained after sorting,screening and matching the data.Six common functions,namely linear function,quadratic function with one variable,the unitary cubic function,logarithmic function,power function and exponential function,were used to establish a univariate regression fitting model,and a univariate regression fitting model of AOD-PM2.5 in Dalian was established and optimized through correlation evaluation parameters.The results show that,on the premise that P is less than 0.05,the R2 value of the exponent function is the largest among the six commonly used functions,which are 0.063,0.059,0.067 and 0.064 respectively.The exponent model is determined to be the optimalfitting model for 2015,2016,2017 and 2018,and the estimated standard errors of the model are 3.601,2.365,2.731 and 1.721 respectively.The multiple regression fitting model of PM2.5 mass concentration data and other atmospheric pollutant mass concentration data was established to analyze the influence of atmospheric pollutants such as PM10,NO₂,SO₂ and O3 on PM2.5 mass concentration data,exploring the main influencing factors of PM2.5 sources,and predicting the distribution of PM2.5.The correlation analysis between PM2.5 and other air pollutants was carried out on a seasonal scale from June 2015 to May 2019,and the multiple regression fitting model of PM2.5 was established.The results showed that,there was a significant linear trend between PM2.5 and PM10,and the correlation coefficient was 0.80.There was a linear trend between PM2.5 and NO₂ and SO₂,and the correlation coefficients are 0.62 and 0.52 respectively.The scatter diagram between PM2.5 and O3 is relatively dispersed,and the correlation coefficient is-0.18.O3 is negatively correlated with PM2.5.The goodness of fit test,equation significance test and parameter significance test were used to test and evaluate the model.The constructed multiple linear regression model could better predict PM2.5 to a certain extent.The 332 prediction points from June 2018 to May 2019 were used for verification,with a relative error of 6.96%.The multiple regression fitting model of PM2.5 and other air pollutants provides a basis for the effective control of PM2.5.