This correlation is not straight forward because AOD depends on both the aerosol amount and the ability of that aerosol to scatter and absorb radiation. So you could have two samples of aerosol that had different composition and therefore different PM2.5, PM10 and TSP but the substances in the aerosol would combine to provide a very similar AOD. IT is sometimes possible to correlate these for individual locations or regions, but not on a global scale. In the sense that AOD is a measure of the total aerosol in a column, TSP is the closest match . However, depending on how the AOD is measured, it may be sensitive to all sizes (probably MODIS), or only the submicron component (instruments such as nephelometers), in which case PM2.5 might be closer. Best indicator is to look for studies in your region of interest, that span several years of data. Good luck!

Which wavelength did you use to derive MODIS AOD?

Dear All,

please find a new Simplified High Resolution (500m) MODIS Aerosol Retrieval Algorithm (SARA):

http://www.sciencedirect.com/science/article/pii/S0034425713001417

Thank you!

The correlation between aerosols at the surface and the total column AOD is highly variable over time and space. It is relatively good for some areas and extremely weak for others. The main thing is to consider is the fraction of the total AOD that is located in the mixing layer, i.e. close to the ground. For background arid conditions and low AOD, the correlation is very weak, so you should not use it.