Uncertainty characterization of AOD for the AATSR dual and single view retrieval algorithms

The uncertainty associated with satellite-retrieved aerosol properties is needed when these data are used to constrain chemical transport or climate models by using data assimilation. Global uncertainty as provided by comparison with independent ground-based observations is usually not adequate for that purpose. Rather the per-pixel uncertainty is needed. In this work we describe how these are determined in the AATSR dual and single view aerosol retrieval algorithms (ADV and ASV) which are used to retrieve aerosol optical properties from reflectance measured at the top of the atmosphere. AATSR is the Aerosol Along-Track Scanning Radiometer which flies on the European Space Agency Environmental Satellite ENVISAT. In addition, issues related to multi-year retrievals are described and discussed. The aerosol optical depth (AOD) retrieved for the year 2008 is validated versus ground-based AERONET sun photometer measurements with good agreement (r = 0.85). The comparison of the AOD uncertainties with those provided by AERONET shows that they behave well in a statistical sense. Other considerations regarding global multi-year aerosol retrievals are presented and discussed.