The purpose of this study was to quantify the impact of using ancillary data from Numerical Weather Prediction (NWP) models in the derivation of cloud parameters from satellite data in the Climate Monitoring Satellite Application Facility (CM-SAF) project. In particular, the sensitivity to the NWP-analysed surface temperature parameter was studied. A one-year dataset of satellite images over the Scandinavian region from the Advanced Very High Resolution Radiometer (AVHRR) on the polar orbiting NOAA satellites was studied. Cloud products were generated by use of the Polar Platform System (PPS) cloud software and the sensitivity to perturbations of the NWP-analysed surface temperature was investigated. In addition, a study on the importance of the chosen NWP model was also included. Results based on three different NWP models (ECMWF, HIRLAM and GME) were analysed. It was concluded that the NWP model influence on the results appears to be small. An interchange of NWP model analysis input data to the PPS cloud processing method did only lead to marginal changes of the resulting CM-SAF cloud products. Thus, the current CM-SAF cloud algorithms produce robust results that are not heavily dependent on NWP model background information. Nevertheless, the study demonstrated a natural high sensitivity to the NWP-analysed surface skin temperature. This parameter is crucial for the a priori determination of the thresholds used for the infrared cloud tests of the PPS method. It was shown that a perturbation of the surface skin temperature of one K generally resulted in a change of cloud cover of about 0.5-1 % in absolute cloud amount units. However, if perturbations were in the range 5-10 K the change in cloud cover increased to values between 1 to 2 % per degree, especially for positive perturbations. Important here is that a positive surface temperature perturbation always leads to an increase in the resulting cloud amounts and vice versa.