The Intergovernmental Panel on Climate Change is back in the news once again, and John Robson of the Climate Discussion Nexus has a good post which looks back at whether the work they've produced over the years has withstood the test of time. The post -- entitled Can we finally toss the hot models? -- discusses a new paper by the Italian physicist Nicola Scaffeta which details which of the IPCC's climate models have most accurately predicted real world climatological activity over the past 40 years. First, on the models themselves:
[Scaffeta] begins by noting that the IPCC currently uses 41 climate models with the rise in absolute temperature due to doubling atmospheric CO2 (known in the trade as “Equilibrium Climate Sensitivity” or ECS) ranging from 1.8 to 5.7 °C. So let’s just pause right there and reflect that in the “settled science” of climate change the models currently being used by the IPCC differ in their projected warming response to CO2 emissions by more than a factor of three. Which seems like a lot.
At the high end, if atmospheric CO2 doubles it could cause almost 6°C of warming, which would be a big problem for the world. At the low end it would cause 1.8°C, which would barely even be noticeable. So the IPCC models say CO2 emissions pose either no problem or a big problem or somewhere in between. If that’s settled science, what would the unsettled kind look like?
But which of those 41 models has been the most accurate?
The answer is, the ones with the lowest ECS values. And if we say the projections for the future should come from the models that have done the best over the past, the global warming picture suddenly looks a lot less heated.
Unsurprisingly, the most accurate models -- those which have predicted the least warming over the past four decades -- also predict the least warming going forward. Robson's conclusion: "Chill out, we will adapt and cope with the small warming predicted by the best models. The real threat isn’t climate change, it’s climate policy."