The GlobAlbedo project aims at generating multi-sensor multi-annual global land surface albedo products. MERIS and SPOT-VGT level-1 data serves as input data to a processing chain. The function of the processing chain can be summarised as:
 processes every single input to some level in several processing steps.
 accumulates at least one year of data for each 10º x 10º tile and then calculates products every 8 days as well as every month, as well as seasonal and annual products.
 finally retrieves BRDF and albedo for each composite reporting time-step (every 8 days and every month, as well as seasonal and annual products).
Firstly, all inputs are systematically processed from L1 to surface directional reflectances
(SDR). The processing steps to retrieve SDRs are:
 pixel classification for cloud, water, etc.. detection
 retrieval of aerosol optical thickness
 atmospheric correction using retrieved aerosol optical thickness
The results are SDR values in the same granularity as the inputs. They are not written for distinct products, but just kept in memory and used as input to the next part of the processing chain. The processing steps are
 broad-band integration of these SDRs
 re-projection of these broad-band SDRs onto the MODIS Sinusoidal grid
The results of this step are re-projected broadband directional reflectances (BBDR) still in the granularity of the input products. During implementation, it has been decided to store this intermediate product of BBDR tiles on disk, but to remove them subsequently after the BRDF computation (accumulation/inversion). The steps are:
 computation of BRDFs every 8-days and every month of the broad-band BBDRs from all sensors by accumulation method, using an additional land cover mask information from Idepix
 computation of the albedo from these BRDFs
The final results are 8-day, monthly, seasonal, and annual composites of broadband
albedo.
For more information on the processing please see the product user guide and the ATBD in the docs tab.