This study is an attempt to estimate the effect of a geographical distribution of clouds on climate. A method of determination of a three-dimensional cloud distribution is proposed. It is based on the solution of the inverse problem for the radiative transfer equation. By using climate data on total cloud amount, temperature, mixing ratio of water vapor, and satellite data on outgoing longwave radiation, the global distributions of high, middle, and low clouds were computed for July. The derived vertical cloud extension is in fair agreement with available data on the frequency distribution of stratiform and cumulus clouds. Two numerical experiments are carried out with an atmospheric general circulation model in which zonal and geographical cloud distributions are prescribed. The integrations are performed for 60 days with a GFDL model, and the last 30 days are analyzed. The geographical cloud distribution causes the increase of surface temperature over the continents by 2¿--4¿ and leads to a decrease of surface pressure there and an increase over the oceans. The largest changes in the surface pressure, up to ¿12 mbar, occur in the middle latitudes of both hemispheres. The largest differences in precipitation are observed in the tropics and over some coastal regions of North and South America. Arid areas in the subtropical belt become more pronounced in case of the geographical distribution of clouds. Estimates of the level of significance for precipitation and surface pressure changes reveal that they are statistically significant in some areas of the globe.