A comparison of atmospheric aerosol absorption properties from the MERRA-2 reanalysis with AERONET

Abstract

Our focus is the comparison of atmospheric aerosol absorption properties between the MERRA-2 reanalysis and the inversion product of the surface station network AERONET. We use the available Aerosol Absorption Optical Depth (AAOD) and Single Scattering Albedo (SSA) data from both sources during the period 1993–2022, for the globe. The comparison is performed for the whole dataset, but also repeated for different classifications, namely by aerosol types, Köppen climate, and continent. Moreover, we investigate the relationships between these classifications. For these medium and high optical depth cases, the MERRA-2 relative bias for AAOD and SSA is globally −3.7% and −1.4%, but there is a wide differentiation in performance for the different types, climates, and continents. The MERRA-2 SSA frequency distribution is too narrow compared to the AERONET. MERRA-2 strongly overestimates the absorption over N. America and less so over S. America and Europe, while it underestimates it greatly over Oceania. AAOD is underestimated over biomass-burning regions and overestimated over most areas rich in dust. There is evidence that the lack of brown carbon in the aerosol type classification scheme in MERRA-2 hampers its performance. The comparison over Asia dry areas (classified as Köppen climate type B) shows a large underestimation in AAOD, non-existent in other dry (B) climate types. For the relatively few observations over polar (Köppen type E) climates, MERRA-2 gives too absorptive aerosols. With respect to absorption properties, it appears that MERRA-2 would improve its agreement with AERONET by addressing some erroneous aerosol classifications, employing the brown carbon type, and introducing more detail in dry region aerosol sources.