BAER Algorithm

Algorithm Specification

The BAER method (Bremen AErosol Retrieval) is an algorithm for remote sensing of aerosols from MERIS data over land. The algorithm has been developed to monitor the aerosol optical thickness (proportional to the aerosol total loading), over most of part of the continents. The aerosol information is used in a second step to perform atmospheric corrections, using either the SMAC processor on a pixel-by pixel basis or the UBAC processor, to derive the remotely sensed surface reflectance over the land.

The actual MERIS Level 2 product provides reflectance data with an incomplete atmospheric correction over land. The atmospheric correction is made for Rayleigh scattering only and the variable aerosol influence is not considered. Thus, an additional step of atmospheric correction for L2 data over land is required, considering the effect of the atmospheric aerosol. For the retrieval of the AOT, the BAER method is used and modified for the use with MERIS L2 data.
The original approach has been developed to retrieve AOT over land from SeaWiFS L1 data. It determines the spectral aerosol optical thickness (AOT) from nadir looking multi-wavelength radiometers. The method is based on the determination of the aerosol reflectance over 'dark surfaces', using the UV and short-wave-VIS range below the red-edge of the vegetation spectrum. This requires a proper separation of the variable surface effects, other atmospheric effects and aerosol effect.
For L2 data over land, the variability of the vegetation cover and the kind of the vegetation will be considered dynamically by means of a surface reflectance model tuned from the satellite scene self by the NDVI. The aerosol reflectance is obtained by removing the estimation of the surface effect. Look-up-tables of the relationship between AOT - aerosol reflectance and the use of constraints enable the determination of the AOT for 7 MERIS channels in a spectral range of 0.412 - 0.670 µm. AOT is extrapolated, using Angström power law with parameters estimated from the retrieved AOT. Others terms of radiative transfer (aerosol reflectance, total transmittance and hemispheric reflectance) are computed once the AOT known to correct the Top Of Aerosol reflectance from aerosol effect.

Once the aerosol optical thickness estimated, it is used as input of the atmospheric correction method (either SMAC or UBAC) to perform the aerosol correction and to provide the surface reflectance in the 13 MERIS channels. If SMAC is used, the AOT at 550 nm required in input is estimated using the angstrom law from the first MERIS channel.

BAER requires as input in addition to the measured top of aerosol reflectances in the 13 MERIS bands, the sun and view zenith angles, the relative azimuth and the atmospheric pressure.