Apart from global economic collapse, little can stabilize climbing carbon dioxide emissions, says a new paper from atmospheric scientist Timothy Garrett.

Garrett applied basic physical rules to civilization as a whole and found that to stop the ongoing rise in CO2 emissions, we’d have to create one nuclear power plant’s worth of new green energy output every day, in addition to meeting our current demands.

In analyzing worldwide economic and energy data ranging from 1970 to 2005, Garrett saw a fixed relationship between the energy consumed by society and the production value of the global economy. This relationship worked out to be 9.7 milliwatts per inflation-adjusted U.S. dollar, meaning that for every dollar’s worth of economic output, society uses 9.7 milliwatts of power. “This is really the core result from which everything else follows,” says Garrett, who is an associate professor at the University of Utah.

The relationship he found held up through the surges, bubbles, and recessions seen in the economy over the past three-and-a-half decades. It also held over earlier data points, which were sparse and went as far back as 2,000 years.

Garrett’s result points to a feedback loop in which energy consumption leads to economic growth, which in turn leads to increased energy consumption and thus more growth.

He calculated a 2.1 percent rise in the global economy’s rate of return for 2005, corresponding to a 300 gigawatt rise in the world’s power needs for that year. These additional gigawatts have to come from somewhere, but meeting the demand with non-carbon emitting energy sources may be next to impossible. “That works out to bringing online about one nuclear power plant per day,” says Garrett, “just to stabilize emissions.”

Dan Nocera, a chemist at MIT, has come to a similar conclusion and warns of the inabilities of current non-carbon technologies to meet global energy demands. He predicts a global demand for eighteen additional terawatts of energy by the time 2050 rolls around.

If we cover every continent with wind turbines, says Nocera, we would only be able to generate two terawatts of electricity. If we grew only bio-fuel crops on all land, everywhere, we’d get another seven, and we’d be very short on food. Damming all of the planet’s rivers would only give another terawatt. For the remaining eight terawatts, we would have to build a new nuclear power plant every 1.6 days until 2050.

None of this would cut into our current carbon consumption, and of course, these options aren’t at all feasible.

Garrett’s building of a new nuclear plant per day, for example, would suck up almost the entire 3 trillion dollar annual budget of the United States. That doesn’t take into account other major expenses, like facility operations or hazardous waste disposal, nor does it consider the drastic increases in materials and specialized labor costs that would accompany such a colossal undertaking. In short, it’s not about to happen.

“Of course the other option,” says Garrett, “is that you could collapse the economy entirely.” Since energy consumption, and thus CO2 output, is linked to economic production, a complete collapse would certainly bring down emissions. “But that would not be popular,” he adds.

Garrett does not offer many solutions to the dilemma his findings raise; instead he hopes his work will lead to a greater understanding of the interactions between society and the environment.

His results also offer a simplified approach to making predictions about future emissions trends. Current projections, including those published in the 2007 report of the United Nations Intergovernmental Panel on Climate Change, include a wide range of uncertainty as they rely on diverse factors that are hard to predict, things like population and quality of life. By tacking emissions to economic worth, Garrett’s research could remove much of this uncertainty.

The inspiration for his approach came from thermodynamics. In a system like a car’s engine, the work of moving the car forward comes as a result of energy transformations in burning gasoline. Society is driven by similar energy transformations—in our car engines, in our power plants, and even within our own bodies. “So it was a rather simple concept,” says Garrett, “to think that perhaps the economic value of civilization was just a representation of energy transformations.”

But unlike a car’s engine, which drives the car forward, society uses its energy to add to itself. This is where the feedback loop arises; energy is used to feed new mouths and erect new homes, and so it creates a demand for more energy. In this way, Garrett says, society is a bit more like a growing child getting bigger from the food he eats, or a cluster of microbes multiplying in a petri dish. “Insofar as the thermodynamics is concerned,” he writes, “the difference between the child and society is really only a matter of complexity and scale.”

Due to this feedback loop, Garrett’s study implies that increases in energy efficiency might actually serve to increase overall energy consumption, as they help society to consume and grow more efficiently. This concept is known as the Jevons Paradox, and it has been the source of much debate since it was introduced in 1866 by the political economist William Stanley Jevons in reference to the British coal supply.

Still, Garrett does not see his work used as an excuse to justify indifference when it comes to CO2 emissions. “If society invests sufficient resources into alternative and new, non-carbon energy supplies, then perhaps it can continue growing without increasing global warming,” he says. Continued research may eventually lead to cheaper methods of harvesting green energy or to the discovery of new energy sources altogether.

Garrett’s paper, which is to be published in the journal Climatic Change, comes on the eve of international climate talks in Copenhagen.