A Method for Atmospheric Correction Based on the Meris Spectral Variability

Atmospheric correction is necessary to estimate the surface reflectance required within biophysical algorithms used to estimate canopy characteristics. Aerosol characterization is obviously one of the main problem in atmospheric correction because aerosol may vary rapidly with time and space. The objective of this study is to develop an autonomous aerosol correction method exploiting the information content in MERIS images. The spectral variation of the radiance signal, when enough sampled by the sensor, allows decoupling aerosol effects from that of the surface because of the very different spectral features exhibited. We thus propose to use 13 over the 15 MERIS bands and the geometry of the scene to estimate the aerosol optical thickness (AOT) assuming only continental aerosol type in this prototype algorithm. For this purpose, several dedicated neural networks were trained to retrieve aerosol AOT from the top of atmosphere signal contained in MERIS level 1B products. The training database was generated with radiative transfer model simulations, SMAC coupled to SAIL and PROSPECT. Performances evaluated over a significant number of AERONET stations shows a RMSE of 0.0958.