Remote Sensing Estimation Of Atmospheric Particulate Matter Concentration Based On Aerosol Optical Thickness In Liaoning Province

In recent years, with the rapid development of urbanization and industrialization,the foggy was not only restricts the economic development, resulting in financial losses, but also threat human health and life throughout the country. A large number of foreign studies have shown that fine particulate matter(PM2.5)and respirable particulate matter (PM10) could cause a lot of negative effects on health. Air pollution has become a highlight problem of our environment. It was of great significance to carry out the monitoring and prediction of PM2.5 and PM10. It provides theoretical basis and technical support for China's air pollution control and environmental management.In this paper, we selected Liaoning Province as the study area to explore the changes of PM2.5 and PM10 in urban air environment. Corresponding to the monitoring points the MODIS aerosol optical thickness (AOD) data, temperature,relative humidity,boundary layer height, zonal wind speed, meridional velocity and barometric data were extracted, and to explore the correlation between AOD, each meteorological factor and PM2.5,PM10.Then the simple linear regression model of PM2.5, PM10,simple linear regression model and stepwise regression analysis model were established step by step.To understand the spatial distribution and trend of PM2.5 and PM10 in Liaoning Province on the macro,thus could provide a basis for regional air pollution control and improvement of regional air quality.The results show that the trend concentrations of PM2.5 and PM10 were consistent.The concentrations of PM2.5 and PM10 were higher and the volatility were stronger in Spring and winter, the concentrations of PM2.5 and PM10 were low and the changes were more stable in summer and autumn.AOD,air pressure,relative humidity,boundary layer height had significant effect on PM2.5 and PM10.The regression models of PM2.5 and PM10 after correction were improved significantly compared with the direct regression models.Based on the correction of vertical and humidity coupling, the determination coefficient R2 between the predicted value and the monitored value of the PM2.5 and PM10 stepwise regression model is increased to 0.602 and 0.516.The prediction coefficient R2 between the predicted and measured values in the fall and winter models of PM2.5 and in the fall models of PM10 was higher than the R2 of the whole models,and the fitting effect of the models is good.The coarse particles in PM10 particles were insensitive to the extinction of aerosols,resulting in precision accuracy models of PM10 are lower than models of PM2.5 generally.The low inversion value areas of PM2.5 were mainly distributed in the hilly areas of Liaoning Province, and the high-value areas were mainly distributed in the central cities.The low-value areas of PM10 were mainly distributed in the eastern and southwestern cities, and the high-value areas were distributed in the central and northwest urban areas.The spatial distributions of inversion value areas of PM2.5,PM10 and AOD were consistency on the whole.The pollution degree of each city is different,and the pollution of Shenyang was the most serious, the concentration of pollutants was significantly different in at different times.The spatial distribution of inversion of PM2.5,PM10 and the distribution of monitoring were same in basically,within the acceptable range.