Characterizing Aerosol Optical Properties Using the Aerosol Robotic Network (AERONET) Version 3 Aerosol Optical Depth and Inversion Products

The Aerosol Robotic Network (AERONET) is a surface-based network of Sun/sky radiometers measuring aerosol optical depth (AOD) or aerosol loading above a specific location (Holben et al. 1998). The network has expanded to over 500 sites worldwide over the past two decades and it provides fundamental remotely sensed measurements of AOD and retrievals of aerosol characteristics such as the size distribution, complex index of refraction and single scattering albedo. AERONET data are used for satellite retrieval evaluation (e.g., MODIS, VIIRS, MISR, OMI), atmospheric correction (e.g., MODIS, LANDSAT), data synergism (e.g., MPLNET, SPARTANS), aerosol forecast model and reanalysis verification (GOCART, NAAPS, ICAP, MERRA-2), numerical weather prediction model analysis (e.g., NCEP, ECMWF, UKMET, GOES-5), and field campaign support (e.g., DISCOVER-AQ , KORUS-AQ, ORACLES). These data are publically available as Level 1.0 (unscreened), Level 1.5 (automatically cloud screened and quality controlled), and Level 2.0 (quality assured) data sets. The release of the Version 3 AOD product employs modified cloud screening methodology that preserves fine mode aerosol (e.g., smoke, haze) and more efficiently removes optically thin cirrus clouds from contaminating the aerosol measurement. Furthermore, the AOD product, for the first time, utilizes a fully automated data quality control algorithm to provide an objective near-real time data product and historical database.

The Version 3 AERONET aerosol retrieval code inverts AOD and almucantar sky radiances using a full vector radiative transfer called Successive ORDers of scattering (SORD; Korkin et al., 2017). The full vector code allows for potentially improving the real part of the complex index of refraction and the sphericity parameter and computing the radiation field in the UV (e.g., 380nm) and degree of linear depolarization. Rather than single scattering, the Version 3 inversion now utilizes multiple scattering assumptions which improve the sky residual fit for high aerosol loading conditions and therefore product accuracy. The latest version of the AERONET inversions includes the effective lidar ratio and depolarization ratio as additional products. Estimated uncertainties are provided for each retrieved product based on the random error and biases due to the uncertainty in AOD (±0.01), absolute sky radiance calibration from a NIST-traceable integrating sphere (±5%), retrieved MODIS surface reflectance for non-snow and snow covered surfaces (±5% absolute), and instrumental pointing errors (±0.1°). The Version 3 inversion products generally follow the same data quality assurance criteria as explained by Holben et al. 2006. The entire AERONET inversion database was processed using the NASA High End Computing resources at NASA Ames Research center and NASA Goddard Space Flight Center.

The aerosol characteristics derived from the Version 3 AOD and inversion products are compared to the previous Version 2 database at selected sites representing dust, biomass burning smoke, urban pollution/haze, and maritime aerosols. The comparison shows that the Version 3 data for the tropical sites of Nauru and Singapore tend to increase the number of available measurements overall with the Angstrom exponent (440-870nm) either remaining the same or increasing due to the removal of optically thin cirrus clouds when compared to Version 2. Further, Angstrom Exponent (440-870nm) especially at very low aerosol optical depth has improved spectral dependence due to temperature characterization applied to spectral channels greater than and equal to 440nm as shown by the Lulin mountain Taiwan climatology. The Palangkaraya site in central Kalimantan, Indonesia, shows very high aerosol loading during the ENSO episode during 2015. These Palangkaraya data demonstrate that AERONET Sun/sky radiometers measure very high AOD (up to ~7.38) while operational satellite aerosol retrieval algorithms do not observe this very high aerosol loading due to algorithm constraints. AERONET Version 3 almucantar inversions are shown to be comparable to the Version 2 almucantar inversions except for high AOD where the use of multiple scattering improves the retrieval results. Version 3 inversions provide uncertainties for all spectral channels and the calculated single scattering albedo uncertainty at 440nm (~0.03) is similar to the estimated uncertainty reported by Dubovik et al. 2002.

Dubovik, O., B.N.Holben, T.F.Eck, A.Smirnov, Y.J.Kaufman, M.D.King, D.Tanre, and I.Slutsker (2002): Variability of absorption and optical properties of key aerosol types observed in worldwide locations, J.Atm.Sci., 59, 590-608.

Korkin S., A. Lyapustin, A. Sinyuk, B. Holben, A. Kokhanovsky (2017), Vector radiative transfer code SORD: Performance analysis and quick start guide, J. QSRT, 200, pp. 295-310, https://doi.org/10.1016/j.jqsrt.2017.04.035.

Holben, B. N., T. F. Eck, I. Slutsker, A. Smirnov, A Sinyuk, J. Schafer, D. Giles, O. Dubovik, 2006: Aeronet’s Version 2.0 quality assurance criteria, Proc. SPIE 6408, Remote Sensing of the Atmosphere and Clouds, 64080Q, doi:10.1117/12.706524

Holben B.N., T.F. Eck, I. Slutsker, D. Tanre, J.P. Buis, A. Setzer, E. Vermote, J.A. Reagan, Y. Kaufman, T. Nakajima, F. Lavenu, I. Jankowiak, and A. Smirnov, 1998: AERONET - A federated instrument network and data archive for aerosol characterization, Rem. Sens. Environ., 66, 1-16.