Kokhanovsky, AA, Breon, FM, Cacciari, A, Carboni, E, Diner, D, Di Nicolantonio, W, Grainger, RG, Grey, WMF, Holler, R, Lee, KH, Li, Z, North, PRJ, Sayer, AM, Thomas, GE, von Hoyningen-Huene, W (2007). Aerosol remote sensing over land: A comparison of satellite retrievals using different algorithms and instruments. ATMOSPHERIC RESEARCH, 85(4-Mar), 372-394.
An inter-comparison study of the aerosol optical thickness (AOT) at 0.55 mu m retrieved using different satellite instruments and algorithms based on the analysis of backscattered solar light is presented for a single scene over central Europe on October 13th, 2005. For the first time comparisons have been performed for as many as six instruments on multiple satellite platforms. Ten different algorithms are briefly discussed and inter-compared. It was found that on the scale of a single pixel there can be large differences in AOT retrieved over land using different retrieval techniques and instruments. However, these differences are not as pronounced for the average AOT over land. For instance, the average AOT at 0.55 mu m for the area 7-12E, 49-53N was equal to 0.14 for MISR, NASA MODIS and POLDER algorithms. It is smaller by 0.01 for the ESA MERIS aerosol product and larger by 0.04 for the MERIS BAER algorithm. AOT as derived using AATSR gives on average larger values as compared to all other instruments, while SCIAMACHY retrievals underestimate the aerosol loading. These discrepancies are explained by uncertainties in a priori assumptions used in the different algorithms and differences in the sensor characteristics. Validation against AERONET shows that MERIS provides the most accurate AOT retrievals for this scene. (C) 2007 Elsevier B.V. All rights reserved.