Air quality evaluation on an urban scale based on MODIS satellite images

Earth-observing satellites have provided satellite image datasets for urban-scale air quality monitoring. Aerosol optical thickness (AOT) at 1 km resolution is retrieved from MODIS on Terra Satellite by using the improved algorithm. The 1 km AOT data are validated by using AOT measurements from two AERONET stations and the 10 km AOT products from MOD04_L2 in the period of October 2004. Then the validated 1 km AOT data are compared with ground-based PM2.5 mass concentration in Beijing. Four empirical models which are used to investigate the relationship between AOT data and PM2.5 mass concentration are obtained by regression analysis and their correlations are R2 = 0.818, R2 = 0.750, R2 = 0.699 and R2 = 0.629 respectively. In order to verify the models, we use these models to retrieve PM2.5 concentration from MODIS AOT on the 11th of October 2012 and then compare it with the PM2.5 concentration from the ground measurements in that day. 50%, 46.4%, 46.4% and 39.3% of the stations are within the expected errors respectively by the four models. Preliminary analysis of these four models indicates that the quadratic model has significant potential to enhance air quality monitoring on an urban scale. Although the limited daily MODIS AOT data, variability of terrain, weather conditions, and many other factors can limit the ability of predicting PM2.5 concentration, the models provide a cost-effective approach for obtaining the distribution of PM2.5 information from satellite image, which complement the defects of the limited ground-based monitoring station measurements.