The work points out that many global models make estimates of greenhouse-gas emissions from soils based on average projected temperatures. But temperatures vary widely from those averages. That variability, along with complex biological processes, makes the issue far more complicated.

The researchers, from Oregon State University (OSU) in the US, claim that climate projections, in general, don't effectively incorporate into their calculations a major component of global warming – the enormous amounts of carbon found in decaying organic matter, which represent up to three times the amount of carbon in the Earth's live vegetation.

"We've done a pretty good job of determining how much carbon is getting absorbed by growing trees and vegetation, how much is coming in," says Mark Harmon from OSU. "However, we know much less about how carbon is released to the atmosphere through the process of decomposition, how much is going out. This is half of the equation, and there's just a huge amount we don't know about it."

It is well known that warmer temperatures will speed up the rate of decomposition of stored organic matter in soils, a process that ordinarily is slow. This faster rate of decomposition, in turn, could further increase carbon released to the atmosphere and cause even greater climate change.

"This feedback loop is one of our biggest worries with global warming, simply because the amount of carbon stored in soil is so huge," Harmon said. "Increased release of that soil carbon could offset much of what we're trying to accomplish with increased growth of live vegetation in forests. And this is a special concern in northern latitudes."

In the past, estimates of that carbon-release process were usually based on the average temperature increases that were expected. But in the real world, temperatures vary greatly, from day to night, season to season, through heatwaves and cold spells. And that variability, say the researchers, changes the biological equation considerably and can make averages misleading.

"If the response of soil respiration to temperature was a straight line, then temperature variability would not be important," says Harmon. "However, the response is curved, which means that as temperature variability increases, so does the average response. This general phenomenon is known as Jensen's inequality, but it had not previously been applied to soil respiration."

The research was not able to precisely quantify this phenomenon and more work needs to be done, say the researchers.