The cloud parameterization in the regional atmospheric model SN-REMO (Spectrally Nudged Regional Model) was validated using satellite data from the ISCCP (International Satellite Cloud Climatology Project). There is an overall good agreement between the cloudiness of SN-REMO and ISCCP in terms of temporal and spatial means. However, with further investigation a deficiency was localized regarding the simulation of cloud amount: Too many clouds are simulated. This overestimation occurs especially during the night. It is connected with a poor simulation of the cloud diurnal cycle. Clouds at low-level emissivity heights (1000--475 hPa) are causing this overestimation. The magnitude of the overall overestimation is also affected by the underestimation of simulated cloud amount at high-level emissivity heights (<475 hPa) and its diurnal variation. The overestimation of the simulated cloud amount is caused by subgrid-scale cloudiness. Since the simulation of subgrid-scale clouds in the regional model SN-REMO is described by a relative humidity parameterization, these deficiencies are connected with this parameterization.