Why Temperatures Could Eventually Go Higher Than We Think

by Michael D. Lemonick

I went to a talk last summer by a geologist who’s skeptical about climate change. “Unlike climate scientists,” he started out, “we geologists look at what happens over hundreds of millions of years,” arguing that any trends in global temperatures over the past hundred years, or the next hundred, are pretty much background noise.

Setting aside for the moment the fact that this background noise could wreak terrible havoc on ecosystems, economies and people (the very existence of people itself, of course, being “background noise” against the 4½-billion-year history of Earth), his statement about climate scientists is just wrong.

That was driven home to me once again with the publication yesterday of a paper in Nature Geoscience that looked 3 million years into the past to project temperatures hundreds of years, at the very least, into the future. Daniel Lunt, of the University of Bristol, England and the British Antarctic Survey, along with several colleagues, were looking at something called the “Earth System Sensitivity”—that is, the long-term, global temperature response to increases in carbon dioxide. Double the amount of CO2, and how much warmer does it get?

If the Earth were a featureless ball, with no water or ice or mountains or valleys or living organisms, this calculation would be easy. Since it’s not, climate modelers have to factor in all sorts of feedback mechanisms. Rising temperatures melt sea ice, for example, exposing darker, more heat-absorbent ocean, which leads to even more temperature rise. Warmer temperatures lead to more evaporation, adding water vapor (itself a greenhouse gas) to the atmosphere. But that also creates more clouds, which might reflect more sunlight, tamping down on warming.

Put all of these so-called fast feedbacks together and you get the IPCC’s latest projection of a 4C (7F) temperature rise by century’s end if we keep using fossil fuels intensively.

But there are slower feedbacks as well, including the melting of ice sheets in Greenland and Antarctica, and changes in vegetation such as forests migrating toward the poles and replacing tundra. They tend to be left out of the models because adding them takes a lot of computer time, and also because their effects take longer to kick in.

A couple of years ago, James Hansen tried to model these slower feedbacks, and concluded that adding slow feedbacks pretty much doubled the temperature increase you’d expect from fast feedbacks. That’s why you hear the number 350 being bandied about so much; it’s Hansen’s upper limit, in parts per million of CO2, that we can afford for any length of time to avoid disaster (for the record, we’re currently at about 390).

Now, using data from the mid-Pliocene warm period, about 3 million years ago, Lunt’s team has come up with a refined figure: given a doubling of CO2, temperatures should go up between 30% and 50% more than current climate models project. It’s not as extreme as Hansen’s projection—and it isn’t by any means definitive. As with a lot of the climate science that’s now considered very solid, this will have to go through a lot of refinement, with other teams weighing in, before anyone finally agrees on a number.

The other uncertainty has to do with how long it takes the climate to reach equilibrium. Levels of CO2 are currently rising, but if we cut off all emissions today, it would take a while for all the fast feedbacks to play out—and even longer for the slow feedbacks. “We’re quite vague about the timescale,” Lunt told me in a phone conversation yesterday, “because we really don’t know. The million-dollar question at the moment,” he says, “is how fast the ice sheets move. Current models say that Greenland would take 3-5000 years to melt. It may be faster than that. But clearly, in just 100 years, the ice sheets won’t be gone.”