OK, the new meme among the denialists is that the tide is with them, that their evidence is overwhelming, that we must be religious zealots not to be able to see the overwhelming evidence that, um, that "very not what the IPCC says". I think we need to talk about the balance of evidence, someday, but I'd like to address the "falsifiability question", i.e., what evidence would it take to shake me of my firm "belief" in AGW.
There is some implication that there is an "AGW theory" and that there is an argument in its support, and that said argument is a cohesive thread starting with Fourier and ending at the dreaded-extremist-boogeyman-Gore, and that failure of any chain in said argument necessarily implies "see, so no carbon policy is necessary". (I'm missing a few steps in their reasoning here, too, but that's another topic still.)
I claim there is no "AGW theory" in the sense that there is an argument that four colors suffice, or more fairly, that stars follow an evolutionary path based on their mass. AGW is not an organizing principle of climate theory at all.
Hypotheses, organizing principles, of this sort emerge from the fabric of a science as a consequence of a search for unifying principles. The organizing principles of climatology come from various threads, but I'd mention the oceanographic sysyntheses of Sverdrup and Stommel, the atmospheric syntheses of Charney and Lorenz, paleoclimatological studies from ice and mud core field work, and computational work starting with no less than Johnny von Neumann.
The expectation of AGW does not organize this work. It emerges from this work. It's not a theory, it's a consequence of the theory.
Admittedly it's a pretty important consequence, and that's why the governments of the world have tried to sort out what the science says with the IPCC and its predecessors. That tends to color which work gets done and which doesn't, and I think it should. As Andy Revkin pointed out, it may be time to move toward a service-oriented climatology, or what I have called applied climatology. The point is that this amounts to application of a theory that emerged and reached mathematical and conceptual maturity entirely independent of worry about climate change.
So attacks on climate change as if it were a "theory" make very little sense. Greenhouse gas accumulation is a fact. Radiative properties of greenhouse gases are factual. The climate is not going to stay the same. It can't stay the same. Staying the same would violate physics; specifically it would violate the law of energy conservation. Something has to change.
The simplest consequence is that the surface will warm up. That this is indeed most of what happens is validated pretty much in observations, in paleodata, in theory and in simulation. Further, all those lines of evidence converge pretty much about how much warming: about 2.5 C to 3C for each doubling of CO2. (It's logarithmic in total CO2, not in emitted CO2, guys, by the way.) There's no single line of reasoning for this. There are multiple lines of evidence.
If you want to convince me that the sensitivity is less than 2 or more than 4, you will have to provide quite a good deal of evidence, but I don;t think this is what the denialists have in mind when they ask me what would "falsify the hypothesis". In fact, though, they haven't defined their terms. If the sensititivity is less than 1, is the supposed hypothesis falsified? What if it is more than 6? If the onset time is a hundred years rather than ten?
They want to know what it would take to pry my free of my "beliefs", but they are not beliefs, they are estimates. Estimates of the sensitivity (2.8 C per doubling). Estimates of the built-in delays (about twenty years for full effect of current concentrations). Estimates of the threshold of excessively high social risk (some range here but I go with 2 C ~ 450 ppmv).
What would it take to change my opinion of the threshhold to 451 ppmv? A nice dinner at Fonda San Miguel, margaritas included, would surely do it. If that constitutes a falsification, bring it on.
Really, though, I don't understand the question. If these numbers wobble around a bit that might shift the optimum policy a bit, but we're so far from the optimum now that it's not worth putting much thought into it yet. The numbers, however, are never going away. There will always be a sensitivity, a response curve, a risk threshold. If you are asking what evidence could make me believe that there are no such numbers, I can't actually imagine it.
That's not because I have a blind attachment to some theory. It's because the numbers must exist, and we have lots of evidence as to what they might be.
What would it take to change my estimate of the numbers? That depends by how much. What would it take to convince me that the meteorology and oceanography I have learned is wrong? I don't know. What would it take to falsify any mature quantitative science?