Not all of the carbon dioxide emitted into Earth's atmosphere remains there. So where does it go? Some is taken up by the oceans, but the destination of a considerable amount remains uncertain, and has been labelled the 'missing carbon sink'. Tree-monitoring projects in ten African countries weren't set up to tackle this problem, but because their measurements were consistent Simon Lewis and his colleagues were able to pool the data and use them to address the question of where the missing carbon might be. Their estimates, along with those made in other tropical regions, indicate that intact tropical forests remove enough CO2 from the atmosphere to account for about half of the missing sink.

Lewis, an ecologist at the University of Leeds, UK, and his team used data collected over various time frames during the past 40 years on 79 inventory plots in west, central and east Africa. Within each plot, local forest managers and scientists had tagged all of the trees that were at least 10 centimetres in diameter and had recorded those trees' diameter, location and species.

Over the course of a decade and a dozen field trips to Africa, Lewis and his co-workers hunted down records as well as the people who had surveyed the forest plots. “We had to be a bit like detectives to find some of the old plots,” he says. “The collaborations were essential because local scientists and the villagers on the ground know the forests much better than I do.”

In one case, Lewis was able to restore historical records to the locals. In Liberia, all of the forest-plot records had been destroyed during the civil war. Lewis found a Dutch scientist who had electronic copies, and passed these on to Liberia's Forestry Development Authority. The work benefited all collaborators, Lewis says: “They have all the data to answer local questions, and I can use them to answer big, overarching global-change questions.”

Ultimately, Lewis went into the field to try to find most of the 79 forest plots and to positively identify individual trees. “The raw measurements came from more than 70,000 trees spread across 10 countries,” says Lewis, who also made new measurements. “The difficulty is that it can take months of work to add just one data point by the time you find a plot, develop the collaboration, digitize the historical data, and then remeasure the trees.”

And if that wasn't enough of a challenge, Lewis and his colleagues also had to try to identify each species of tree from the thousands of possibilities in tropical Africa. Tracking species is key to interpreting the possible causes of the growth rate that the team found. Growth in forests can be spurred by various factors — for example, by past disturbances, such as when wind blows down a tract of forest, opening up the canopy and favouring sun-loving species. If this had occurred in an area, Lewis would expect to find certain species in decline and others on the increase. Instead, his data show “an across-the-board increase in carbon storage for many different species”. Lewis suspects that the rate of growth might be a response to the increased CO2 in the atmosphere (see page 1003).

The study highlights the importance of collecting and archiving standardized data, says Lewis, because you never know what future question they might answer. “No one started this looking to monitor the effects of global environmental change. But that's what these data have ended up contributing to.”