The clear-sky greenhouse effect sensitivity to a sea surface temperature change

The clear-sky greenhouse effect response to a sea surface temperature (SST or Ts) change is studied using outgoing clear-sky longwave radiation measurements from the Earth Radiation Budget Experiment (ERBE). Considering geographical distributions for July 1987, the relation between the SST, the greenhouse effect G (defined as the outgoing infrared flux trapped by atmospheric gases), and the precipitable water vapor content (W), estimated by the Special Sensor Microwave Imager (SSM-1), are analyzed first. A fairly linear relation between W and the normalized greenhouse effect g, defined as G/sigmaTs4, is found. On the contrary, the SST dependence of both W and g exhibits nonlinearities with, especially, a large increase for SST above 25 degrees C. This enhanced sensitivity of g and W can be interpreted in part by a corresponding large increase of atmospheric water vapor content related to the transition from subtropical dry regions to equatorial moist regions. Using two years of data (1985 and 1986), the normalized greenhouse effect sensitivity to the sea surface temperature is computed from the interannual variation of monthly mean values. Although subject to uncertainties, results show a smooth variation over the 0 degrees - 32 degrees C temperature range. A maximal sensitivity of g (approximately 10 X 10(-3) K-1) is found for both extreme temperature ranges (0 degrees - 4 degrees and 28 degrees - 32 degrees C), while a minimal sensitivity (approximately 6 X 10(-3) K-1) is found in the 12 degrees - 16 degrees C temperature range. The enhanced greenhouse effect sensitivity in the warmest temperature intervals is tentatively explained by increased convection that injects water vapor into the middle and upper atmosphere. In the coldest temperature ranges, the atmosphere is dry and implies more nonsaturated absorption bands and, therefore, higher sensitivity to water vapor content. These values are related to a small interannual variation of outgoing longwave flux with SST (approximately 1 W m-2 K-1), while without water vapor feedback, this sensitivity would be on the order of 4 W m-2 K-1.