It’s pretty easy to show that global temperatures are rising, or that spring is arriving earlier than it once did, but since climate has changed plenty of times in the past, that alone doesn’t prove anything. Tying climate change to human greenhouse-gas emissions – an area known as detection and attribution, or fingerprinting – is a lot harder.
That’s what makes a new study in Proceedings of the National Academy of Sciences so important. Using state-of-the-art climate models, Ben Santer of Lawrence Livermore National Laboratory and 21 colleagues have found what they call “some of the clearest evidence to date of a discernible influence on atmospheric temperature.”
Specifically, they found that while the troposphere – the lowest part of the atmosphere – has warmed over the past three decades, the stratosphere, which starts 5 to 12 above the ground, has cooled. This is exactly what you’d expect if greenhouse gases were trapping heat near the surface rather than letting it percolate upward.
“This is not a new idea,” Santer said in an interview. “We did the first fingerprinting studies of the troposphere and stratosphere back in 1996.”
The problem back then, Santer said, was that only a couple of climate models were available for studies like this. Models are crucial in this kind of research because you can’t do controlled experiments with the planet the way doctors do when they test new pharmaceuticals. With medicines, you give some patients the drug and others a placebo, or sugar pill, and see the difference in how their illnesses respond.
With the climate system, by contrast, there’s only one patient, and it’s already been dosed with extra greenhouse gases such as carbon dioxide. So scientists like Santer do simulations of how the atmosphere should look both with and without those extra gases. Unlike in 1996, Santer and his co-authors had 20 different simulations to work with for this study, all of them state-of-the-art models developed for the upcoming major report of the Intergovernmental Panel on Climate Change, due out starting in 2014.
The extra modeling power was crucial because the climate system is so complex no single model can truly capture it. They’re all approximations at best, so each has some uncertainty built in. But if you compare a number of models against each other and see the same fingerprint in most of them, you can be pretty confident that you’re seeing something real.
“After removing all global mean signals,” the authors write, referring to natural changes like volcanic eruptions and changes in the brightness of the sun, “model fingerprints remain identifiable in 70 per cent of the tests involving tropospheric temperature changes.”
In plain English, that simply means that the warming of the troposphere and cooling of the stratosphere can’t be explained in any other way than by the heat-trapping effects of human-generated greenhouse gases.
“It was surprising to me how large the signal was,” Santer said.
This is only one of the fingerprints scientists expect to see in a human-influenced climate, moreover.
“In the past we’ve looked at ocean surface temperatures changes in hurricane-forming regions, patterns in atmospheric pressure; rainfall patterns, and changes in Arctic sea ice,” Santer said.
All of these and more can be identified more easily and clearly with the new models.
“I think these simulations are like a scientific gold mine,” Santer said. “Analysts will be exploiting them for many years to come.”