The majority of climate models published between the early 1970s and late 2000s were remarkably accurate, according to a systematic review published on 4 December in Geophysical Research Letters (1). The results will be included in the next climate assessment by the Intergovernmental Panel on Climate Change (IPCC) due to be released in 2021.
For years, critics have questioned the accuracy of climate models. So, to put an end to the debate, the researchers from the University of California, Berkley, the Massachusetts Institute of Technology (MIT), and the NASA Goddard Institute for Space Studies analysed 17 models used published between 1970 and 2007. Models were evaluated on how well they predicted global temperatures — and the results are striking.
The majority of predictions were “indistinguishable from what actually occurred”, said lead author and doctoral student Zeke Hausfather the Energy and Resources Group of the University of California, Berkeley. He and his colleagues found that 10 out of 17 models correctly predicted global temperatures. And an additional 4 out of 17 got the physics right but were unable to predict the future emissions.
The main challenge with climate models is, of course, is that since scientists don’t have a crystal ball, therefore, actual emissions might not exactly match the scenarios they choose to assess. For instance, some early climate models could not have anticipated the Montreal Protocol, a treaty that banned chlorofluorocarbons — potent greenhouse gases — from 1989. Furthermore, due to computational limits, there are only so many scenarios researchers can examine.
Nevertheless, “the rate of warming we are experiencing today is pretty much exactly what past climate models projected it would be,” said Hausfather.
Climate models must make two major assumptions: how much greenhouse gases are emitted into the atmosphere — mainly via the burning of fossil fuels — and how the atmosphere responds to external radiative ‘forcings’ such as carbon dioxide and other greenhouse gases. Future emissions are difficult to predict since they are driven by human behaviour, whereas most models — even crude ones — got the atmospheric physics right.
In fact, incorrect temperature predictions could be chalked down to incorrectly guessing the amount of carbon released into the atmosphere. However, when the researchers accounted for these inaccuracies in emissions estimates, they found that relationships between atmospheric CO2 and warming were still correct.
Climate deniers often spread the idea that climate models are unreliable. But climate scientists seem to have a pretty solid understanding of the Earth’s climate system: the new analysis “increases our confidence that models are accurately projecting global warming”, the authors write. Therefore, we can expect projections of future warming to be reliable. And models are only becoming more sophisticated.
The timing of the paper is ideal, as leaders of nearly 200 countries meet in Madrid, Spain for the COP25 Climate Summit. Unfortunately, if current projections are relatively accurate, the Earth is will reach 3 degrees Celcius of warming above pre-industrial temperatures by 2100, which the IPCC and others say would have catastrophic effects, increasing the likelihood of reaching several dangerous tipping points – with irreversible consequences.
(1) Hausfather, Z. et al. Evaluating the performance of past climate model projections. Geophysical Research Letters (2019). DOI: 10.1029/2019GL085378