The really scary thing about climate change is not simply that humans may fail to get their emissions under control. It's that at some point, the Earth could take over and start adding even more emissions on its own.

A new study underscores this risk by looking closely at Indonesia, which has a unique quality - some 70 billion of tons of carbon that have built up in peatlands over millennia. In this, Indonesia is much like the Arctic, where even larger quantities of ancient carbon are stored in permafrost, and are also vulnerable.

In each case, if that carbon gets out of the land and into the atmosphere, then global warming will get worse. But global warming could itself up the odds of such massive carbon release. That's a dangerous position to be in as the world continues to warm.

In the new study in Geophysical Research Letters, a team of researchers led by Yi Yin of the French Laboratoire des Sciences du Climat et de l'Environnement look at the potential of peat bogs in equatorial Asia - a region that includes Malaysia, Papua New Guinea and several other smaller countries but is dominated by Indonesia and some of its largest islands, Kalimantan and Sumatra - to worsen our climate problems. It's timely, considering that last year amid El Niño-induced drought conditions Indonesian blazes emitted over 1.5 billion tons of carbon dioxide equivalents into the atmosphere. That's more than the annual emissions of Japan (or, needless to say, of Indonesia's fossil fuel burning).

And the research finds that over the course of this century, that could keep happening. "The strong nonlinear relationship found between fire emissions and cumulative water deficit suggests a high future risk of peat carbon loss due to fire given that future climate projections indicate a twofold increase in the frequency of extreme El Niño," the researchers write.

The situation arises because of the unique qualities of peat: In peat bogs, wetlands accumulate large amounts of organic matter - dead plant life - over many, many years. If those bogs are then drained, and fires are allowed to burn on them and deep into them, then it is possible to light up huge stores of ancient carbon and to put it back in the atmosphere much more rapidly than the speed at which it accumulated originally.

The study finds a tight relationship between El Niño events and large peat fire emissions over the past 19 years. And it projects that under a high global emissions scenario, a warming climate will trigger more intense El Niños that, in turn, will correspond to more intense blazes (barring, that is, some major policy or political change that stops humans from starting fires to begin with). And under more moderate warming, there could still be major peat fire emissions during the century, the research finds.

"Most climate models predict a little bit stronger, not more El Niños, but more intense El Niños," says Guido van der Werf, one of the study's co-authors and a fire emissions researcher at VU University in Amsterdam.

And so much carbon could be lost in this way that it could affect the global atmosphere in what the paper calls a "significant positive feedback to global warming."

More specifically, the study forecasts that as many as 25 petagrams, or billion tons, of carbon could be released in a high warming scenario - and 13 billion tons in a more modest warming scenario, which would require significantly changing the globe's emissions trajectory and shifting it downward, while still likely missing the goal of limiting warming to just 2 degrees Celsius above pre-industrial levels.

If these numbers are converted to carbon dioxide, that would correspond to between more than 47 billion tons in the more moderate warming scenario and about 91 billion tons of carbon dioxide for the high emissions scenario. (In fires, most of the greenhouse gas emissions would be in the form of carbon dioxide, though there would also be some methane and nitrous oxide emitted.)

The carbon math is clear: The world simply can't have such an additional source of emissions. For instance, the so-called "carbon budget" for holding warming below 2 degrees Celsius above pre-industrial levels was about 1,000 billion tons of carbon dioxide, as of 2011. And that's for human burning of fossil fuels - not for releases from peat fires. If these fire emissions happen, then the already extremely narrow carbon budget becomes even narrower.

"It's 2 to 3 years of current fossil fuel emissions globally," Van der Werf says.

Granted, as always is the case in science, there are some key caveats here. We might get ourselves off of a high-emissions pathway, and thus, not have to deal with a worst-case scenario above. Furthermore, it could be the case that models are wrong about ongoing global warming intensifying the El Niño phenomenon, which is in turn related to the droughts that drive the worst Indonesian fires. If so, then these numbers might also be, again, too negative.

"This assumption, up to 25 petagrams, really hinges on how El Niño is going to change in the future," Van der Werf says. "And I don't think anybody really knows for sure what is going to happen with El Niño in a warmer world," though he notes that most models predict more intense El Niño events, when they do occur.

And finally, something could occur to cut down on the human role in the fires - draining peat bogs and using fire to clear land. Last year showed that these human land-use practices, combined with drought in an El Niño year, can cause catastrophic fires. "If somehow Indonesia is able to lower the amount of fire, the emissions will go down, of course," van der Werf says.

The research suggests that, along with worrying about all the carbon stored in Arctic permafrost, which could greatly add to emissions in this century, we definitely have to worry about the carbon in peat, too.

"There's definitely more carbon stored in boreal regions, but it's not as vulnerable as this stuff," van der Werf says. "If you get another few El Niño, you get a big pulse. Permafrost is a much more gradual process."