The temporal windows we used to generate the products are:
-monthly composite
-10 days composite
-daily composite

Compared to the previous version V01, the new version includes daily composite at 0.5° resolution without the previous + or - 5 days running window average. These new products are the daily composite from Ascending & Descending passes for the considered day.
The Level 3 SSS algorithm is now starting using as input the L1B ESA products (while we used L1A products for the V01 version).

In addition, each product now contain also:
-ECMWF SST (GHRSST product)
-ECMWF wind speed module
These additional information in the products were first co-located with SMOS swath data and then re-gridded as the level 3 SSS space-time resolution.

-An estimate for the strong Radio Frequency Intereference (RFI) contamination probability is also included at the SSS retrieved pixel and for the level 3 SSS space-time window.

-Residual temporal drifts (at seasonal time scales) and  land contamination signals are still found in the up-to-date SMOS brightness temperature level 1 products. These are mostly due to yet unsolved calibration (thermal) & image reconstruction issues (solar effects in L-band imaging, antena patterns uncertainties). These impact

translate into
, the L3 SSS data for ascending and descending passes were bias-adjusted using a large-scale correction based on a 5°x5° spatial filter derived from the World Ocean Atlas 2001 monthly SSS climatology, temporally interpolated to the SMOS aquisition date. Not as well that only one fixed Ocean Target Transformation (used to correct for spatial image reconstruction biases) was used to generate all the products.

-a new empirial correction for the galactic glint contamination has been implemented from a Geometric Optic model which shall improve the quality of SSS retrievals in Sep-oct months for descending passes and in March-April months for Ascending passes.

-a filter at the Brightness temperature Level has been applied to remove Tb data reconstructed in the proximity of the aliased sun-disk, in order to best mitigate solar contamination effects.

-Radio Frequency Intereferences are still a big issue for SMOS SSS retrieval quality. In order to tentatively best minimize the impact of RFI in our level 3 products, a weighting average of the swath retrieved SSS based on the RFI probability at the considered pixels & time window has been applied. This probability being now incorporated in the products, the user can now chose by him or herself only those pixels with 0 probability of (detected) RFI. Pixels with RFI probability higher than 80% over the considered space-time window have been removed.