Estimation and Spatial-temporal Distribution Characteristics of PM2.5 Concentration by Remote Sensing in China in 2015

Air pollution characterized by PM2.5 pollutants poses severe challenges to the sustainable development of society and human health. Therefore, it is of great significance to clarify the spatial-temporal distribution and evolution of PM2.5 pollutants in China for regional joint prevention and control of PM2.5 pollutants. Based on the MODIS satellite aerosol products, meteorological basic data and PM2.5 pollutant monitoring site monitoring data, a geographically weighted regression model was established to simulate and estimate PM2.5 pollutant concentration in China in 2015 on the basis of aerosol and meteorological data pre-processing. In addition, the spatial distribution pattern, the seasonal evolution characteristics of PM2.5 pollutant concentration were analyzed. The results showed that: (1) the PM2.5 concentration values in China in 2015 as a whole showed obvious spatial zonal differentiation characteristics. The concentration of pollutants in the north is significantly higher than that in the south, and the areas with high PM2.5 concentrations are mainly concentrated in the Beijing-Tianjin-Hebei region, the Jianghuai plain, the Sichuan basin, and the Takaramalkan desert. The area has a wide spatial distribution and significant continuity; (2) The PM2.5 concentration in the fourth quarter showed obvious seasonal adaptive evolution characteristics. The PM2.5 concentration changed significantly in the season. PM2.5 pollution was the heaviest in the fourth quarter, followed by the first quarter of the third quarter and the lowest in the second quarter. The maximum occurred in the fourth quarter (165 μg/m³). The minimum appeared in the second quarter (4.3 μg/m³). Seasonal changes in PM2.5 concentrations were influenced by meteorological factors and human social activities; and (3) The accuracy of the inversion of PM2.5_{concentration by a multi-factorial, geographically weighted regression model was higher, with relative errors in the four quarters being 10.2%, 7.0%, 9.3%, and 8.6%, respectively.}