The short-term validation plan of the POLDER-2 "ERB, WV & Clouds" products is schematically presented in figure 1. It is based on three types of validation:

Checking of consistency between the different output data.
This first step is important for all the parameters, mainly for the cloud cover. For example, the different "directional" values of the cloud fraction should not present a bias as a function of the viewing angle.
The angular consistency is also essential for the cloud optical thickness. In fact, at this stage, we do not try to validate the retrieved values of cloud optical thickness but rather the cloud modeling used for the retrieval. The different "directional" and spectral values of cloud optical thickness have to be consistent between them. (Note that the cloud optical thickness is calculated at three wavelengths and for each direction).
This step is also very important for the microphysical index. This index describes the angular behavior of the measured ice cloud polarized reflectance. Thus one can’t cast doubt on it. However, it is quite important to verify that the choice of the different classes of angular behavior is judicious. An analysis of the geographic distribution of the different classes should be informative.

Comparison with groundbased measurements.
Concerning the total column water vapor content, the comparison with the radiosonde measurements is obviously necessary. Concerning the cloud parameters, the radar-lidar cloud measurements can be very informative. During the ADEOS-1 period, only the SGP (Southern Great Plains) ARM data were available; they were used for checking the cloud thermodynamic phase and the cloud pressures (Oxygen pressure and Rayleigh pressure) derived from POLDER-1 data. For POLDER 2, data from ARM/SGP and CloudNET/SIRTA sites have been used.

Comparison with products derived from other satellite radiometers.
For the validation of the cloudiness, we will privilege the method previously used for POLDER-1 by comparing our outputs to the cloud classification based on the dynamic cluster analysis method. To do that, geostationary satellite data such as METEOSAT, GOES (and MSG/SEVIRI when available) will be treated.
Concerning the total water vapor content, the DMSP-SSM/I product will be used over ocean as for POLDER-1.
As there will be some good coincidences in time and space between ADEOS-2 and EOS-Terra, the POLDER water vapor and cloud products will be confronted to Terra-MODIS products such as the cloud fraction, the water vapor content, the cloud phase, the cloud optical thickness, the cloud top pressure and the cloud droplet effective radius.
A particular point concerns the shortwave (SW) albedo. The validation of the spectral integration (i.e. the derivation of the SW albedo from the spectral albedoes derived from POLDER measurements) was not possible for POLDER-1. Indeed, it needs a comparison between POLDER data and ERB-scanner measurements. For POLDER-2, this will be possible by using Terra-CERES measurements.
 

The validation of the "ERB, WV & Clouds" products will be performed by Laboratoire d'Optique Atmosphérique (LOA), Laboratoire des Sciences du Climat et de l'Environnement (LSCE) and Laboratoire de Météorologie dynamique (LMD). It is supported by CNES (Centre National d'Etudes Spatiales), CNRS (Centre National de la Recherche Scientifique) and Région Nord-Pas de Calais.
For any question, please contact Pr. J.C. Buriez (at LOA)

---

© CNES Copyrights 2005