This paper presents theoretical and modeling evidence suggesting that the atmospheric poleward heat transport can lead to a polar warming amplification (i) by redistributing part of the extra energy intercepted by the low-latitude atmosphere to high latitudes, and (ii) by strengthening the water vapor feedback in high latitudes. For an anthropogenic radiative forcing of 4 Wm/2, we illustrate that the dynamical amplifier contributes to about 1/4 (1/10) of the total high-latitude (global) surface warming in winter in a simple coupled atmosphere-surface moist radiativetransportive climate model. Budget analysis of the radiation fluxes at the top of the atmosphere derived from IPCC AR4 CGCM climate simulations seems to support the dynamical amplifier theory for the larger warming in high latitudes.
The most interesting part of this is part (ii) - the strengthening of the WV feedback in high latitudes. Its interesting because although many people think its why polar amp occurs, this is the first paper I've seen to assert it. The least interesting is part (i), because thats been done already (and better, from a brief scan) by Alexeev in Cli Dyn (see my prev post). The problem is that Cai is using a 4-box coupled atmosphere-surface moist radiative-transportive climate model to illustrate the dynamical amplification of the high-latitude surface warming due to an anthropogenic forcing. And the WV feedback isn't done properly, The water vapor feedback has been crudely parameterized using an ad-hoc formula Equation 5 to mimic the strong dependency of the atmospheric absorption of longwave radiation to the amount of water vapor. Specifically, the total atmospheric effective emissivity in the model is made of two parts: a constant part e0 representing the absorption due to other gases, such as CO2, and a part that varies as a function of the total amount of water vapor in an atmosphere column. And worse than that, I can't see where they seperate out effects (i) and (ii), so at the end of it all I can't see where they say how important the WV feedback is (if we trusted their eq 5).
[Updated to fix funny char (ctrl-A?) in the Wm/2 that was doing naughty things to the feed, apparently]