Seminaire Du Gap Biophotonics "xbaer: a Generic Algorithm for the Retrieval of Aerosol Properties from Satel-lite Observations"

Because of a long-term synergy with other observations from ENVISAT, there is advan-tage to derive aerosol optical thickness (AOT) using the MEdium Resolution Imaging Spectrometer (MERIS) instrument. Although limited in spectral coverage, it is possible to derive accurate AOT from MERIS. Here, we introduce a new AOT retrieval algorithm for MERIS over land surfaces, henceforth called eXtensible Bremen AErosol Retrieval (XBAER). XBAER is similar to the dark-target retrieval algorithm used for Moderate Imaging Resultion Spectroradiometer (MODIS), in that it uses a lookup table (LUT) approach, and that minimum differences between observed and LUT spectral reflectance represents the aerosol solution. However, instead of global parameterization of surface spectral reflectance, XBAER uses a set of spectral libraries to prescribe surface properties. Thus, XBAER retrieves AOT over both dark surfaces (vegetation) and bright surface (desert, semiarid, and urban areas). Cloud mask is an essential step for XBAER. However, establishing an accurate and adequate cloud mask is a challenging task for passive aerosol remote sensing from space base instrumentation. This is especially true for those instruments without infrared channels. Aerosol retrievals require a particular cloud masking approach since a mask based on too conservative thresholds and assumptions will screen out strong aerosol episodes and a less conservative approach could allow cloud contamination significantly affecting the subsequently retrieved aerosol optical properties (e.g. AOT). In this context, a cloud mask algorithm utilizing MERIS instrument characteristics has been developed for the MERIS operational aerosol retrieval algorithm XBAER. The XBAER Cloud Mask (XBAERCM) algorithm is a threshold-based method that does not require any auxiliary data. MERIS Top Of Atmosphere (TOA) reflectances have been evaluated not only taking into account the brightness of the scenes, but also of its texturing/variability and of the cloud altitude.