The main objective of this study was to find the situations when the present implementation of the Cloud Mask and Cloud Type models are not able to correctly classify the cloud scenes or the surface features and to describe under which circumstances this occurs.

The method chosen was to campare the mode! output, i.e. the cloud type classification with the subjective, human interpretation of the satellite images resulting in cloud type. Until now, such a method has not been used in the validation of the CT and CMa products.

More than 600 pairs of Cloud Type Classification images and AVHRR 5 bands images and RGB combinations of them were analyzed in order to get information on the behavior of the Cloud Type Mode! in summer and winter conditions. In about 145 cases, the human interpreted cloud type was found to show significant differences requiring a more thorough analysis. It has been found that 8 classes of clouds in summer conditions (August 2000) and 11 classes in winter conditions (February-March 2001) were classified as other cloud types when compared to the outcome of the subjective analysis.