“Each model is slightly different in its starting conditions (the weather on the day the model is started), the fields it is forced with (for instance, different scenarios for how carbon dioxide develops this century), and most crucially in the approximations in the physics,” said Dr. Sylvia Knight of the atmospheric, oceanic and planetary physics department at Oxford. “All climate models have to make approximations, because the climate is so complicated.”

“Thanks to chaos theory, we can’t predict which versions of the model will be any good without running these simulations, and there are far too many for us to run them ourselves,” said Dr. Myles Allen of the University of Oxford. “Together, participants’ results will give us an overall picture of how much human influence has contributed to recent climate change, and of the range of possible changes in the future.”

David Stainforth, the experiment’s chief scientist, said: “While many model studies in the past have made plausible predictions of climate change, it hasn’t been possible to quantify our confidence in these predictions. We hope to be able to say, for the first time, what the climate probably will and, more importantly, probably won’t do in the future.”