Interpretation of satellite retrievals of PM2.5 over the southern African Interior

A case study is presented for using an unsupervised classification (Self-Organizing Map) of a global PM2.5 data product assembled from satellite retrievals (SeaWIFS, MODIS Terra and MODIS Aqua) and ground observations. The PM2.5 data products are available, on a daily basis, from August 1997 to the present, with 10 km resolutions and global coverage. In this study, a sub-set of the PM2.5 retrievals (collected over the southern African Interior) has been averaged over ten-day intervals for a period of ten years. These averaged sub-sets have been clustered using self-organizing maps to generate spatial and seasonal "PM2.5 climates" and air quality interpretations over southern Africa. Results are an indirect validation of the satellite based data product against available regional ground-based and airborne studies. The final PM2.5 aerosol climatology shows that the data product provides credible PM2.5 estimates for a region that is lacking routine aerosol monitoring data. (C) 2015 Elsevier Ltd. All rights reserved.