Two different monthly mean analyses of low, middle, and high cloud amounts for January 1977 and July 1979 are compared: 3DNEPH is a condensed version (northern hemisphere only) of the Air Force 3D-Neph analysis, which incorporates satellite data plus surface observations of clouds and auxiliary meteorological data. SFCOBS is objectively analyzed from surface observations of clouds. The SFCOBS and 3DNEPH analyses of low cloud amounts agree qualitatively in the winter extratropics. The 3DNEPH ITCZ is much more sharply defined than the SFCOBS. The sensitivity of radiative fluxes to 3DNEPH, SFCOBS, and zonal mean 3DNEPH clouds is then evaluated. The fluxes are diagnosed by a cloud-radiation model utilizing ''observed'' monthly mean temperature and water vapor fields and are verified against satellite data. The outgoing longwave radiative flux clearly verifies best for 3DNEPH clouds and worst for zonal mean 3DNEPH clouds in the tropics. It is predominantly controlled by surface temperature in the winter extratropics. Generally speaking, the shortwave fluxes do not verify as well as the longwave fluxes. Also, outside of the winter extratopics, the net radiative fluxes correlate poorly with observation. Biases in the zonal mean long and shortwave fluxes can be reduced by adjusting other cloud-related parameters Based upon the above results, it may be worthwhile to construct a monthly mean cloud climatology from a condensed version of the 3D-Neph. However, alternative strategies should be explored, such as the development of cloud analysis schemes that constrain the model-diagnosed net radiative flux to be consistent with observation.