Performance of MODIS aerosol products at various timescales and in different pollution conditions over eastern Asia

High-quality aerosol optical depth (AOD) data derived from MODIS is widely used in studying spatiotemporal trends of fine particulate matter (PM_2.5) concentrations in eastern Asia. However, the differences of MODIS-AOD (3/10 km DT; 10 km DB) under four pollution situations (No-Po; Sl-Po; Mo-Po; Se-Po) are rarely considered. In this study, the MODIS-AOD and AOD-Difference spatial distributions from 2008 to 2017 are analyzed through annual/seasonal mean AOD maps generated at 0.1°×0.1° resolution. The MODIS-AOD performances are validated using AERONET AOD data for various pollution situations and aerosol types. Annual validations indicate that the 10-km DB algorithm provides the best performance, followed by 3-km DT and 10 km DT. The DB algorithm can provide spatially continuous AOD data for all seasons, whereas the DT algorithm often fails to yield valid data during winter. The validations under different pollution conditions illustrate that severe pollution significantly affects the validity of data obtained by the DB algorithm. However, the accuracy of DT-derived AOD data is robust against interference. Under the same pollution conditions, the correlation coefficient of the DB algorithm is smaller than that of the DT algorithm. The quantity of valid data in the DB product is higher than those in DT products for all pollution conditions, especially under Se-Po.