Water vapor feedback in a small ensemble of GCMs: Two approaches

Abstract

The clear-sky longwave component of the climate sensitivity parameter λCSLW is estimated for 5 diverse state-of-the-art general circulation models (GCMs). A common radiation code is used to calculate 2 alternative breakdowns. The conventional breakdown, into "surface temperature feedback", "lapse rate feedback" and "water vapor feedback" shows only the well-known result of large but mostly compensating variations in the latter two components, due to variations in the distribution of warming. In one GCM, λCSLW is rather stronger than the rest, tending to reduce the sensitivity of climate. The "partly Simpsonian" breakdown shows that this is because of a combination of having less water vapor to start with and less increase in water vapor around the tropopause on warming. Two GCMs show λCSLW rather weaker than the rest, because of weaker amplification of the warming aloft, in one case combined with stronger-than-average relative humidity increases. The net effect of greater lapse rate changes (proportionally more warming aloft) can be seen to tend consistently to reduce climate sensitivity slightly. Also, though the impact of relative humidity changes is dominated by reductions through most of the depth of the troposphere, its variation across GCMs is dominated by increases around the tropopause. While the results from so small a sample will not be quantitatively general, they suggest that applying the new breakdown to a much wider range of GCMs would give useful quantitative and physical information on the variation of λCSLW across current GCMs.