Atmospheric science

Drought and fire change sink to source

Aircraft have captured the 'breath' of the Amazon forest — carbon emissions over the Amazon basin. The findings raise concerns about the effects of future drought and call for a reassessment of how fire is used in the region. See Letter p.76

The Amazon forest accounts for 40% of the aboveground biomass stored in the world's tropical forests1, but we do not know whether this crucial but threatened biome will be a sink or a source of atmospheric carbon in the coming decades2. Given the need to predict future climate scenarios, it is essential to refine our understanding of tropical forests' ability to sequester or release carbon3. The profiling of air columns over such forests by aircraft offers a much-needed window onto the major fluxes of tropical carbon. On page 76 of this issue, Gatti et al.4 report the first estimate of carbon fluxes from the Amazon basin obtained in this way over the course of two years. Their findings suggest that the combined effects of drought and fires can cause the Amazon forest to become a net source of atmospheric carbon.

The authors sampled air masses several kilometres above the forest canopy at four Amazon locations, creating a patchwork of atmospheric profiles of carbon dioxide and carbon monoxide that spans the entire Amazon basin. They conducted these measurements during a major drought year (2010) and a relatively wet year (2011) for the region.

The researchers found that, during the drought year, burning of vegetation associated with land use and reduced photosynthesis resulted in 0.48 ± 0.18 petagrams of carbon (Pg C; 1 Pg is 1015 grams) being lost from the Amazon forest biome (uptake by the biome was 0.03 ± 0.22 Pg C per year; fire emissions were 0.51 ± 0.12 Pg C per year). During the wetter 2011, however, the Amazon was effectively carbon neutral: biome uptake (0.25 ± 0.14 Pg C per year) very nearly cancelled the fire emissions (0.30 ± 0.10 Pg C per year). Temperatures were above average in both years, but similar, suggesting that a moisture deficit reduced photosynthesis rates in 2010, rather than the crossing of a temperature threshold.

The growth rate of atmospheric CO2 levels observed over the past five decades at Mauna Loa, Hawaii, and at the South Pole was recently shown5 to be highly sensitive to year-to-year variability in tropical temperatures, and is further moderated by moisture conditions. This finding, taken together with Gatti and colleagues' study, implies that a shift in the terrestrial carbon cycle may be occurring because of the sensitivity to drought of tropical forests globally.

The world's vegetation takes up about 2.6 ± 0.7 Pg C per year, compared with around 9 Pg C per year emitted to the atmosphere, mostly as CO2, from fossil-fuel combustion and cement production6. The Amazon forest accumulated an average of 0.4 Pg C per year in the two decades before 2005 — a range of 0.3–0.6 Pg C per year, estimated through repeated sampling of nearly 100 permanent plots across the basin7 — and so has had a substantial role in offsetting global anthropogenic emissions of greenhouse gases. Whether this annual uptake will persist and compensate for emissions related to drought and land use in the future remains uncertain.

Gatti and colleagues' approach captures bi-weekly atmosphere–biosphere gas exchange across millions of square kilometres, the first time that this has been done at such a scale and for so long. Their method surpasses the spatial and temporal restrictions of, as well as some of the assumptions associated with, other methods such as plot-level inventories or modelling based on satellite data.

Atmospheric profiling using aircraft is a crucial tool in our understanding of Amazon carbon fluxes, and has the potential — if a pan-tropical network of aircraft observations can be established — to determine how tropical forests worldwide are responding to the combined threats of increasing drought and land-use pressures. A big advantage of this method is that it integrates emissions and uptake from naturally occurring land and river processes with land-use emissions to give a regional picture of total carbon fluxes. However, understanding the drivers and mechanisms behind these fluxes is key for the future management of carbon in tropical regions.

Given the importance of fire in shifting the Amazon basin from a sink to a source of carbon, one of the next steps is to reconcile the different fire types that contribute to the authors' regional estimates of fire emissions. Gatti and co-workers' vertical profiling detected carbon monoxide, which could have been caused by fires used for deforestation (Fig. 1), land management (pasture burning and 'slash-and-burn' agriculture, for example) and escaped understory wildfires8. More than 85,000 square kilometres of otherwise intact forests burned in understory fires in the southern Amazon during the 2000s, and, in dry years, the area affected can exceed the area deforested for agriculture and pasture9. These fires kill 8–64% of mature trees across Amazon forest sites10, and burn biomass11, thereby reducing forest carbon stocks. Teasing out the different land-use drivers that contribute to overall fire emissions is essential to aid fire-prevention and fire-management strategies that could help to reduce those emissions.

Figure 1: Deforestation fire in the southeastern Amazon.


Gatti et al.4 report that a combination of severe drought and fires associated with land use can shift the Amazon region from being a sink to a source of atmospheric carbon.

Because drought frequency and intensity in the Amazon might increase in the future12, the authors' results are concerning. Furthermore, during the period of the study, deforestation rates were the lowest they had been since the records of Brazil's National Institute for Space Research began in 1988. The substantial fire emissions documented by the authors during their study therefore imply that efforts to reduce deforestation must also address the use of fire as a land-management tool. In sum, if drought and fire frequencies increase in the future, they may override the Amazon's function as a carbon sink.


  1. 1

    Baccini, A. et al. Nature Clim. Change 2, 182–185 (2012).

  2. 2

    Davidson, E. A. et al. Nature 481, 321–328 (2012).

  3. 3

    IPCC Climate Change 2013: The Physical Science Basis (Cambridge Univ. Press, 2013).

  4. 4

    Gatti, L. V. et al. Nature 506, 76–80 (2014).

  5. 5

    Wang, X. et al. Nature (2014).

  6. 6

    Le Quéré, C. et al. Nature Geosci. 2, 831–836 (2009).

  7. 7

    Phillips, O. L. et al. Science 323, 1344–1347 (2009).

  8. 8

    Balch, J. K., Nepstad, D. C., Brando, P. M. & Alencar, A. Science 330, 1627 (2010).

  9. 9

    Morton, D. C., Le Page, Y., DeFries, R., Collatz, G. J. & Hurtt, G. C. Phil. Trans. R. Soc. Lond. B 368, 20120163 (2013).

  10. 10

    Barlow, J. & Peres, C. A. in Emerging Threats to Tropical Forests (eds Laurance, W. F. & Peres, C. A.) 225–240 (Univ. Chicago Press, 2006).

  11. 11

    Balch, J. K. et al. Global Change Biol. 14, 2276–2287 (2008).

  12. 12

    Malhi, Y. et al. Science 319, 169–172 (2008).

Download references

Author information

Correspondence to Jennifer K. Balch.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Balch, J. Drought and fire change sink to source. Nature 506, 41–42 (2014).

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.