Find in Library

Due to the advantage of high spatial coverage, using satellite-retrieved aerosol optical depth (AOD) data to estimate PM_2.5 and PM₁₀ mass concentrations is a current research priority. Statistical models are the common method of PM estimation currently, which do not require the knowledge of complex chemical and physical interactions. However, the statistical models rely on station data, which results in less accurate PM estimation concentrations in areas where station data are missing. Hence, a new hybrid model, with low dependency on on-site data, was proposed for PM_2.5 and PM₁₀ mass concentration estimation. The Gaofen-1 satellite and MODIS data were employed to estimate PM_2.5 and PM₁₀ concentrations with 100 m spatial resolution in Beijing, China. Then, the estimated PM_2.5/10 mass concentration data in 2020 were employed to conduct a spatio-temporal analysis for the investigation of the particulate matter characteristic in Beijing. The estimation result of PM_2.5 was validated by the ground stations with R² ranging from 0.91 to 0.98 and the root mean square error (RMSE) ranging from 4.51 μg/m³ to 17.04 μg/m³, and that for PM₁₀ was validated by the ground stations with R² ranging from 0.85 to 0.98 and the RMSE ranging from 6.98 µg/m³ to 29.00 µg/m³. The results showed that the hybrid model has a good performance in PM_2.5/10 estimation and can improve the coverage of the results without sacrificing the effectiveness of the model, providing more detailed spatial information for urban-scale studies.