We employ a multitude of global data sets to extend recent analyses of atmospheric greenhouse effect and its dependence on surface temperature (TS) and vertical water vapor distribution. The new data encompasses a global domain including both the continents and the oceans as well as both the ascending and descending branches of the Walker and Hadley cells and the extratropical storm track regions. We adopt the radiometric definition of the atmospheric greenhouse effect, Ga, which is the difference between the surface longwave emission and the outgoing longwave radiation. We derive the global average greenhouse effect over both oceans and land areas. The east to west variations of the normalized atmospheric greenhouse effect (ga) and precipitable water (w) are just as strong as the north to south variations, thus illustrating the strong role of the dynamics in w and ga. Between 60 ¿N and 60 ¿S the lowest values of ga (0.11--0.15) are found over the Saharan and other deserts; while the largest values (0.35--0.40) are found over the warm oceans with a deep convective atmosphere. The coupling between Ga, and the vertical distribution of atmospheric water vapor and temperature, is examined from monthly mean annual cycle. When averaged from the southern to the northern latitudes, these quantities exhibit a statistically significant annual cycle. The annual cycle of TS, about 1 K for the tropics (30 ¿N to 30 ¿S) and about 4 K for the globe, is large enough to obtain a statistically significant estimate for the sensitivity parameter dGa/dTS. It is as large as 5.5--6 W m-2 K-1 for tropical mean conditions (30 ¿N to 30 ¿S) and reduces to a global mean value of 3.5 W m-2 K-1 (with a 2&sgr; range of 2.9--4.1 W m-2 K-1). Consistent with earlier studies, the tropics exhibit a strong positive coupling between TS, Ga, and water vapor distribution with large increases in the midtroposphere humidity. However, poleward of 30 ¿N, water vapor increases are about half as much as that in the tropics, and the sensitivity parameter dGa/dTS decreases. This is because poleward of 30¿, the annual cycle is dominated by land surface temperature changes which are not so effective as sea surface temperature changes in enhancing the water vapor greenhouse effect. Irrespective of the region (ocean or land) or the latitude domain (tropics or extratropics), the data presented here do not offer any support for the suggestion that increases in tropical or global mean surface temperature would lead to a decrease in water vapor greenhouse effect by drying the middle to upper troposphere. If any, the global scale sensitivity derived from the annual cycle is consistent with the magnitude of the positive feedback obtained by general circulation models. ¿ 1998 American Geophysical Union