Courtesy of RealClimate, here's your climatology "oh fuck" moment for the day. Particularly the last paragraph, discussing figure-B.
(...)
cheffer et al illustrate the potential feedback for the coming century in a figure which looks something like Figure A:
Temperature depends on CO2 concentration via radiative equilibrium in the blue curves, and CO2 concentration in the air is affected by temperature according to the red lines. A rise in CO2 concentration from an external source changes the equilibrium CO2 as a function of T relation toward higher CO2, to the right, labeled "forcing". The stable final equilibrium is where the two relations cross, with further CO2 degassing from the land or the ocean, so that more CO2 ends up in the atmosphere than would have if there were no feedback (a vertical red line). A climate sensitivity calculated from the coupled system is higher than one that ignores any carbon cycle feedbacks.
The situation today is complicated somewhat by a carbon spike transient. Atmospheric CO2 is rising so quickly that the terrestrial biosphere and the ocean carbon reservoirs find themselves far out of equilibrium. In attempting to keep up, the other reservoirs are taking up massive amounts of CO2. If emissions were to stop today, it would take a few centuries for the atmosphere to equilibrate, and it would contain something like 25% of our emitted CO2.
I would draw our current situation as in Figure B, with CO2 concentration wildly higher than the equilibrium red line, poised to relax toward lower concentrations if emissions stopped. The effect of the carbon cycle feedback is to change the equilibrium atmospheric CO2 that we are relaxing toward. It seems to me that the most important part of the equation for our immediate future is the decay rate of that carbon spike, rather than the equilibrium value that CO2 will relax to in hundreds of years.
http://www.realclimate.org/index.php/archives/2006/05/positive-feedbacks-from-the-carbon-cycle/