Despite a steady decline in carbon emissions since the 1990s in the continent as a whole, South East Asia continues to see high emissions due to deforestation. But emissions due to land-use change are still hard to estimate accurately because of poor land-use definitions in the scientific literature, the team believes.

"A lot of the disagreement in our estimates simply comes from definitions," said Benjamin Poulter of Montana State University, US.

Some 10% of total carbon emissions are thought to be due to changes in land cover or land use. The Intergovernmental Panel on Climate Change (IPCC) reports that the figure has been in decline over the past two decades – partly because of checks on deforestation in the Amazon, and partly because emissions from fossil fuels have risen. Still, there is great uncertainty as to the magnitudes of change and to where they can be attributed, with Asia in particular being an unknown quantity.

Poulter and colleagues – who are based at other institutions in the US, the UK, Australia, Japan, France and Germany – addressed this uncertainty by drawing together several independent data sources to estimate emissions from land-use change. The sources include the UN Food and Agriculture Organization, dynamic global vegetation models and remote sensor-based assessments. "A challenge in such a comparison is to document in detail how [land use] definitions vary between data sources, and what sorts of biases in models may drive uncertainties," said Poulter.

The researchers split the data sources into three categories according to their type, and documented the differences in how the land-use terms were defined. They then summarized the emissions estimates according to each Asian sub-region and decade.

The estimates showed that up to 40% of global carbon emissions since the 1980s have come from land-use change in Asia, with up to 25% in Southeast Asia alone – mostly because of deforestation, leading to carbon release from peatlands. Most estimates showed these emissions to have declined in the 1990s.

The fact that not all estimates agreed largely came down to definitions – the varying meaning of "forest", for instance – and differing inclusions of carbon fluxes, such as wood harvests, degradation and regrowth. "We knew these differences existed a priori, [but we] wanted to highlight [them] and to recommend what can be done in the future," said Poulter.

One of those recommendations is to think harder about definitions. "The [land use] scientific community must continue to pay close attention to how land use is defined, and what losses and gains are considered in each estimate," Poulter said.

The researchers are now investigating how atmospheric greenhouse-gas observations can help reduce the uncertainty of emissions arising from land-use change using carbon cycle models.

The study is published in Environmental Research Letters (ERL).

Related links

Related stories