Climate experts declared last year that they have strong evidence that human activities have warmed Earth's climate by half a degree over the past century. But computer models predict that the warming, caused by buildup of greenhouse gases in the atmosphere, should be twice that amount. Now, a group at Columbia University's Lamont-Doherty Earth Observatory reports that the shortfall may be due in part to a cooling effect in the oceans driven by a seemingly unlikely source: greenhouse warming itself.
The group reports in tomorrow's issue of Science* that the eastern tropical Pacific, where the periodic warming of El Niño is found, has actually cooled on average over the past century. That's contrary to the predictions of the global climate models used to gauge the effect of greenhouse gases. But it's consistent with a Lamont model of the tropical Pacific Ocean and the layer of atmosphere immediately above it, which predicts that the eastern part of the ocean cools as the greenhouse effect builds.
Lamont team member Richard Seager says that his group's model presents a more realistic view of the Pacific, which allows the model to respond more strongly to shifts in atmospheric circulation caused by greenhouse warming. In their model, stronger winds draw more cold water to the surface in the east, as happens during El Niño's opposite phase called La Niña. "I'm pretty sure that will delay the warming," says Seager. "There's a good chance it will ultimately reduce the warming, but we haven't shown that yet."
Not everyone is convinced, however. The doubts are strongest among those who run the rival global models. "We realize that we don't simulate [El Niño] as well as we should," says Thomas Knutson of the Geophysical Fluid Dynamics Laboratory in Princeton, New Jersey, but "I'm not real enthusiastic about using their regional type of model for examining the [global] climate-change issue." The regional model should be connected to the rest of the world, he says, and the whole atmosphere, not just the lowermost layer, should be included. Knutson argues that the ocean most likely just amplifies the response the atmosphere would have on its own, while Seager maintains that a fair test must include a more realistic ocean than Knutson's. The two groups are discussing their differences, but in the meantime, Knutson confesses, "we just disagree."