Estimating the spatial distribution of PM2.5 concentration by integrating geographic data and field measurements
- Chinese Academy of Surveying and Mapping, Lianhuachi West Road 28, Haidian District, Beijing, 100830, China
Keywords: Air pollution, remote sensing, dust surface, industrial polluting source, correlation analysis, multivariate regression analysis
Abstract. Air quality directly affects the health and living of human beings, and it receives wide concern of public and attaches great important of governments at all levels. The estimation of the concentration distribution of PM2.5 and the analysis of its impacting factors is significant for understanding the spatial distribution regularity and further for decision supporting of governments. In this study, multiple sources of remote sensing and GIS data are utilized to estimate the spatial distribution of PM2.5 concentration in Shijiazhuang, China, by utilizing multivariate linear regression modelling, and integrating year average values of PM2.5 collected from local environment observing stations. Two major sources of PM2.5 are collected, including dust surfaces and industrial polluting sources. The area attribute of dust surfaces and point attribute of industrial polluting enterprises are extracted from high resolution remote sensing images and GIS data in 2013. 30m land cover products, annual average PM2.5 concentration values from the 8 environment monitoring stations, annual mean MODIS AOD data, traffic and DEM data are utilized in the study for regression modeling analysis. The multivariate regression analysis model is applied to estimate the spatial distribution of PM2.5 concentration. There is an upward trend of the spatial distribution of PM2.5 concentration gradually from west to east, of which the highest concentration appears in the municipal district and its surrounding areas. The spatial distribution pattern relatively fit the reality.