Harvard University researchers have resolved a conflict in estimates of how much the Earth will warm in response to a doubling of carbon dioxide in the atmosphere.
That conflict — between temperature ranges based on global climate models and paleoclimate records and ranges generated from historical observations — prevented the United Nations’ Intergovernmental Panel on Climate Change (IPCC) from providing a best estimate in its most recent report for how much doubled CO2 emissions will warm the Earth.
“Historical observations give us a lot of insight … but there is no perfect analogue for the changes that are coming,” said Professor Peter Huybers. Eliza Grinnell/SEAS
The researchers found that the low range of temperature increase — between 1 and 3 degrees Celsius — offered by historical observations did not take into account long-term warming patterns. When these patterns are introduced, the researchers found that not only do temperatures fall within the canonical range of 1.5 to 4.5 degrees Celsius but that even higher ranges, perhaps up to 6 degrees, may also be possible.

Abstract
The latest Intergovernmental Panel on Climate Change Assessment Report widened the equilibrium climate sensitivity (ECS) range from 2° to 4.5°C to an updated range of 1.5° to 4.5°C in order to account for the lack of consensus between estimates based on models and historical observations. The historical ECS estimates range from 1.5° to 3°C and are derived assuming a linear radiative response to warming. A Bayesian methodology applied to 24 models, however, documents curvature in the radiative response to warming from an evolving contribution of interannual to centennial modes of radiative response. Centennial modes display stronger amplifying feedbacks and ultimately contribute 28 to 68% (90% credible interval) of equilibrium warming, yet they comprise only 1 to 7% of current warming. Accounting for these unresolved centennial contributions brings historical records into agreement with model-derived ECS estimates.
The research is published in Science Advances.