- That's a sensitivity, not a prediction. If we cause the atmospheric concentrations to more than double, the temperature perturbation goes up still more.
- There are other human perturbations besides CO2, some of them significant greenhouse gases, and all of them wildcards in relatively complicated climate dynamics
- It's the sensitivity of what has arbitrarily been called the "climate system" for the purposes of certain scientific reasoning processes. What we should be concerned about is the sensitivity to emissions, not the sensitivity to concentrations. The sensitivity to concentrations is pretty well constrained to be near 3 C per doubling, but we have a far less well-constrained idea of how much net emission corresponds to such a change in concentration.
We know from paleoclimate evidence that there must be multiple feedbacks in the system. Here's how the argument goes:
Solar forcing is synchronous with much larger climate changes than it can account for directly. We see massive retreats of ice sheets more or less synchronized with peak summer insolation at the southern edge of the ice sheets. That is, when the earth's orbit wobbles so that the sunshine is greatest in, say, the middle of Canada, it is hard for the ice to maintain itself from one season to another, and the ice retreats. This makes qualitative sense.
But whoops! It doesn't make quantitative sense. The thermodynamics don't work out. A little ice should melt, and the ice should retreat a little, but it abruptly almost all goes away. Something gives the system a very big kick, the warming trend runs away, and the ice vanishes relatively abruptly.
SO we notice that CO2 increases synchronously with the ice retreat. Work that in, and it all works out. The extra CO2 enhances the greenhouse effect. The enhanced greenhouse effect rapidly warms the edge of the ice sheet. The ice sheets melt. The theory works out in detail, the models replicate the paleobotanical evidence, and Bob's your uncle.
OK, so problem solved, right? Well, no, not really. Nobody knows how the extra carbon got into the atmosphere. We don't really understand the fluxes of carbon in the system on the time scales of rapid natural climate change.
What we can be pretty sure of is that there is a mechanism whereby increased temperature leads to increased carbon, and we already have a mechanism whereby increased carbon leads to increased temperature. Now fortunately there must be some limit to the process or we would have all been throughly boiled and then steamed long before we had this conversation. On the other hand, that process must exist.
So we need to worry about whether that process has any extra ammunition.
In an excellent series of articles on Grist, Joe Romm points out the serious consequences of this error. Like me, Joe suspects land surface processes are the key, though he doesn't refer his argument to paleoclimate evidence. There are reasons based in isotopic evidence that argue against this, and the majority of the paleoclimate community therefore believes in an all ocean process as the missing feedback.
For now it doesn't matter. We are giving a tippy system which we don't fully understand an extra push; an extra carbon pulse is a very likely result.
This feedback is entirely unrepresented in most climate models. Joe's latest article refers to preliminary results from the first model to model the climate system and the carbon fluxes in a unified system. That news is not good.
I think there are all kinds of scientific reasons why coupling the carbon cycle into a climate model to create what is being called an ESM (Earth System Model) is premature. Many of the cirticisms unfairly leveled at GCMs and CGCMs will be leveled quiasi-fairly at ESMs. The results of these models are not at all trustworthy and it's very hard to say if and when they will be.
Presumably the Mojojojo set, who for some godforsaken idiotic peabrained psychotic reason would prefer to have more money in a lousy world than less money in a nice one, will argue that since ESMs are unreliable, they should be ignored. The contrary is true. If we cannot well constrain the carbon cycle, we need to account for the possibility that a great deal more carbon is lying in wait for us, be it under Siberia, or Greenland, or Antarctica, or the bottom of the Atlantic, or the swamps of Amazonia. Wherever it came from in the past, there might be more of it.
The feedback from temperature to carbon is not well understood but that does not mean it does not exist.
As usual the mojojojoists start with a nugget of scientific truth (the temperature inflection seems to lead the carbon inflection in the ice cores) and promote exactly the wrong conclusion. They conclude that it proves that CO2 does nothing.
(Of course, since they are determined to spew all the CO2 in their nefarious plot to destroy the Power Puff Girls and take over Townsville and then rule the world, every single piece of evidence anywhere is presumed to prove that CO2 does nothing).
In reality this is another piece of evidence that indicates that nature is likely to happily match our contributions to our impending calamity. That "temperature is the cause of the CO2 rise" argument does not prove the falsity of "CO2 is the cause of the temperature rise". They must both have been true in the past.
If they are both true now our boat could be much leakier than the usual understanding of a 3 C sensitivity indicates.