The Intergovernmental Panel on Climate Change in its first report in 1990 predicted that temperatures would warm by 0.5 degree Fahrenheit (0.3 degree Celsius) per decade if no efforts were made to restrain greenhouse gas emissions. But the panel of scientists and other experts was wrong: By 2001, the group estimated that average temperatures would increase by 2.7 to 8.1 degrees F (1.5 to 4.5 degrees C) in the 21st century, and they raised the lower end to 3.6 degrees F (2 degrees C) this year in their most recent report. In essence, neither this international team of experts nor any other can say with any certainty just how bad global warming may get.

There is a simple explanation for this, says atmospheric physicist Gerard Roe of the University of Washington (U.W.) in Seattle: Earth's climate is extremely sensitive. In other words, small changes in various physical processes that control climate lead to big results. "If nothing else changed by [warming], a doubling of carbon dioxide would ultimately lead to a temperature change of about 1.2 [degrees] C," [(2.1 degrees F)] Roe says. "In fact, because of internal processes within the climate system, such as changing snow cover, clouds and water vapor in the atmosphere, our best estimate is that the actual warming would be two to four times larger than that."

Some of these feedback processes are poorly understood—like how climate change affects clouds—and many are difficult to model, therefore the climate's propensity to amplify any small change makes predicting how much and how fast the climate will change inherently difficult. "Uncertainty and sensitivity are inextricably linked," Roe says. "Some warming is a virtual certainty, but the amount of that warming is much less certain."

Roe and his U.W. co-author, atmospheric physicist Marcia Baker, argue in Science that, because of this inherent climate effect, certainty is a near impossibility, no matter what kind of improvements are made in understanding physical processes or the timescale of observations.

"Once the world has warmed 4 degrees C [(7.2 degrees F)] conditions will be so different from anything we can observe today (and still more different from the last ice age) that it is inherently hard to say when the warming will stop," physicists Myles Allen and David Frame of the University of Oxford wrote in an editorial accompanying the article. "If the true climate sensitivity really is as high as 5 degrees C [(9 degrees F)], the only way our descendants will find that out is if they stubbornly hold greenhouse gas concentrations constant for centuries at our target stabilization level."

Therefore, waiting for more scientific certainty before acting is a mistake, Roe says. "People are comfortable with the idea that stock markets, housing prices and the weather are uncertain, and they are used to making decisions on that basis," he notes.

But this also means that targets such as stabilizing atmospheric concentrations of CO2 at 450 parts per million (nearly double preindustrial levels) to avoid more than a 3.6 degree F (2 degree C) temperature rise are nearly impossible as well. There is no guarantee that such a target would achieve its stated goal. "Policymakers are always going to be faced with uncertainty and so the only sensible way forward to minimize risk is to adopt an adaptive policy," argues climatologist Gavin Schmidt of the NASA Goddard Institute of Space Studies, "which adjusts emissions targets and incentives based on how well, or badly, things are going."

It also means that scientists and other experts are going to have to monitor measures other than just atmospheric concentration of greenhouse gases to catch catastrophic climate change developing. "It is essential that we designate the harbingers of abrupt and significant changes or, perhaps more importantly, the triggers and thresholds that could commit the planet to these changes well before their tell-tale signs appear," says economist and IPCC author Gary Yohe of Wesleyan University in Middletown, Conn. "We cannot accept the adaptive design completely without having confidence in our abilities to determine exactly what to monitor."

The IPCC has taken a crack at that, identifying 26 "key vulnerabilities" in its most recent assessment, ranging from declines in agricultural productivity to the melting of ice sheets and polar ice cover as well as determining how to judge if they are spiraling out of control. Disappearing Arctic ice is already helping to amplify global warming beyond what the IPCC had predicted in the past. "We already know about as much as we are going to about climate system's response to greenhouse gases," Roe says. "We already have the basis for making the decisions we need to make."