Scientists have long known that countries in the tropics are cutting down their rain forests and digging up peatlands to make room for agriculture — and at a rapid rate. These trees and soil contain huge reservoirs of carbon, and, as they get cleared and burned, they release heat-trapping greenhouse gases into the air. But how much carbon gets kicked up, exactly, has always been a bit of a puzzle.

An illegal logger prepares to cut trees from a tropical rain forest in Riau province, Sumatra, Indonesia. (Ulet Ifansasti - Getty Images)

In its 2007 report, the Intergovernmental Panel on Climate Change estimated that tropical deforestation accounted for roughly 20 percent of the world's global-warming emissions. But there was an asterisk next to these estimates. They were based on self-reported data from countries themselves. Some nations, like Brazil, do a solid job of tracking deforestation. Other countries, particularly poorer ones in Africa, lack the resources and manpower to monitor their rain forests. They essentially have to guess, extrapolating from data that's often years out of date.

So, more recently, scientists have turned to satellite monitoring to offer a clearer picture of which forests are actually being chopped down. In a new study published in Science, a team of nine researchers — led by Nancy Harris of Winrock International — pored over satellite data and estimated that tropical deforestation accounted for about 10 percent of all greenhouse gas emissions between 2000 and 2005. That's lower than previous estimates.

More valuable still, the researchers were able to create a more precise map of where the most intensive deforestation was occurring. Roughly 55 percent came from Brazil and Indonesia. (Note that this map doesn't measure the loss of peatlands, which can account for half the greenhouse-gas emissions in Indonesia in some years):

From Harris et al., "Baseline Map of Carbon Emissions from Deforestation in Tropical Regions"

But this method still isn't yet perfect. The trouble is that even with better satellite data, there's still some uncertainty about how much deforestation is occurring in certain countries. Indeed, two earlier studies based on satellite monitoring, published in 2011 and in 2012, concluded that emissions from deforestation were twice as high as the Science study. (See this graph here for a comparison.)

In an accompanying essay in Science, Dan Zarin of the Climate and Land Use Alliance pointed out that the differences can likely be chalked up to methodology, though some uncertainty still remains: "[T]he explanation is not immediately apparent, and the magnitude of the difference remains surprising." Here's a map from the recent Science study on where it's most difficult to measure the pace of deforestation:

From Harris et al., "Baseline Map of Carbon Emissions from Deforestation in Tropical Regions"

In a phone interview, the lead author of the study, Nancy Harris of Winrock International, explained that there were still limits to current satellite analyses (for instance, her team explicitly didn't examine what happened when new, younger forests grew back in areas that had been cut down by logging). As always, further research will prove necessary.  What's useful, however, is that this study can provide a more reliable baseline against which to measure changes in deforestation.

So why does any of this matter? It's not terribly important for our understanding of global warming per se — climate scientists are able to measure the increase in carbon dioxide in the air directly, at observatories like the one in Mauna Loa, even if it's not clear whether 10 percent or 20 percent is coming from deforestation.

But it is important for climate policy. Currently, the U.N. is trying to set up various programs to reduce deforestation, under which wealthier nations would pay tropical countries such as Brazil or Indonesia or Congo to protect their forests. In theory, this could prove one of the cheaper ways for the world to reduce carbon emissions (though it won't be simple). But the program won't work as well unless scientists can create an accurate picture of where deforestation is actually occurring — and has a solid baseline to measure progress.