Ice Age Climate Test Confirms Concerns about Future Warming

How much global warming will result from the man-made accumulation of carbon dioxide in the atmosphere? A group of climate scientists from the Potsdam Institute for Climate Impact Research (PIK) in Germany has found a novel approach to answering this question: they turned to the last great Ice Age for a crucial test. The conclusion is that a doubling of carbon dioxide concentration would cause a global temperature increase of around 3 ºC - in line with previous estimates using other methods, mainly computer models.

Ice Ages are caused by changes in the Earth's orbit, but lower carbon dioxide concentration is one of several factors that make Ice Ages so cold. If the different factors leading to the big chill can be disentangled, then valuable information on how strongly carbon dioxide affects climate can be gathered. To extract this information, Thomas Schneider von Deimling, Hermann Held, Andrey Ganopolski and Stefan Rahmstorf combined climate models with a multitude of data about Ice Age climate. Their results are published this month in the journal Climate Dynamics.

In their study, the scientists accounted for model uncertainties by creating a set of 1,000 climate model versions - each one with a somewhat different behaviour of clouds, ocean currents or various other uncertain processes and feedbacks affecting climate. With each of these 1,000 models, a global warming scenario with doubled carbon dioxide concentration was computed. As expected, the amount of global warming (in case of CO ₂ doubling called the "climate sensitivity") differed substantially between different model versions - as, in fact, it does between different climate models developed by different research groups. The question is: what is the real sensitivity of the climate system to CO₂ ? Rather than waiting for the future to provide the ultimate answer, we can have a look at the climatic past.

Therefore, all the model versions were subjected to a critical test to separate the wheat from the chaff: how well can they simulate the climate of the last great Ice Age? Those models that are too sensitive to carbon dioxide would tend to simulate an Ice Age that is too cold, and vice versa. After a careful analysis of all sources of uncertainty, the researchers concluded that a climate sensitivity of less than 1.2 ºC or greater than 4.3 ºC would be inconsistent with what we know about the great Ice Age, with the most likely value being near 3 ºC. Notabene: Climate sensitivity is a simple measure of how much climate would warm in the long run if CO ₂ concentration was doubled. So this number should not be confused with the global warming at a given time (say, the year 2100), which depends on the actual time evolution of the CO ₂ concentration and on other climate factors.

"One positive result is that we can practically rule out an extremely high climate sensitivity of six, seven or even more degrees Celsius, which some colleagues have speculated about," says Thomas Schneider von Deimling, lead author of the study.

Uncertainty about the future climatic effect of our carbon dioxide emissions consists of two parts: First, to what future carbon dioxide concentration will a certain amount of emissions lead? And second, how much warming can we expect from a given carbon dioxide concentration? The new study addressed the second part and thereby complements another recent study looking at the first issue, the carbon cycle feedback, also with the help of past climate data (Scheffer, M., V. Brovkin, and P. Cox (2006), Geophys. Res. Lett., 33; see our press release of 22 May). Taken the findings of both studies together, the results suggest that current estimates of the most likely future global warming may be somewhat too low - not because the climate sensitivity has been underestimated, but because the carbon cycle could act to amplify the warming later this century.

Contact
thomas.schneider@pik-potsdam.de, phone: +49-331-288-2630
stefan.rahmstorf@pik-potsdam.de, phone: +49-331-288-2688

Original Article:
Thomas Schneider von Deimling, Hermann Held, Andrey Ganopolski and Stefan Rahmstorf: Climate sensitivity estimated from ensemble simulations of glacial climate, Climate Dynamics (2006), Vol. 27, 149-163.
Abstract available at: http://www.springerlink.com/content/a70n722515243682/