Soden, B. J., and R. Fu, 1995: A satellite analysis of deep convection, upper-tropospheric humidity, and the greenhouse effect. Journal of Climate, 8(10), 2333-2351.

Abstract: This paper combines satellite measurements of the upwelling 6.7-µm radiance from TOVS with cloud-property information from ISCCP and outgoing longwave radiative fluxes from ERBE to analyze the climatological interactions between deep convection, upper-tropospheric humidity, and atmospheric greenhouse trapping. The satellite instruments provide unmatched spatial and temporal coverage, enabling detailed examination of regional, seasonal, and interannual variations between these quantities. The present analysis demonstrates that enhanced tropical convection is associated with increased upper-tropospheric relative humidity. The positive relationship between deep convection and upper-tropospheric humidity is observed for both regional and temporal variations, and is also demonstrated to occur over a wide range of space and time scales. Analysis of ERBE outgoing longwave radiation measurements indicates that regions or periods of increased upper-tropospheric moisture are strongly correlated with an enhanced greenhouse trapping, although the effects of lower-tropospheric moisture and temperature lapse rate are also observed to be important. The combined results for the Tropics provide a picture consistent with a positive interrelationship between deep convection, upper-tropospheric humidity, and the greenhouse effect. In extratropical regions, temporal variations in upper-tropospheric humidity exhibit little relationship to variations in deep convection, suggesting the importance of other dynamical processes in determining changes in upper-tropospheric moisture for this region. Comparison of the observed relationships between convection, upper-tropospheric moisture,and greenhouse trapping with climate model simulations indicates that the Geophysical Fluid Dynamics Laboratory (GFDL) GCM is qualitatively successful in capturing the observed relationship between these quantities. This evidence supports the ability of the GFDL GCM to predict upper-tropospheric water vapor feedback, despite the model's relatively simplified treatment of moist convective processes.