Since at least the inception of modern records of atmospheric carbon dioxide levels in the 1950s, there has been a small global excess (about 2%) in the amount of CO2 taken up by land plants for photosynthesis, compared with the amount emitted as a result of the decomposition of organic material. This land carbon sink has absorbed around 25% of all fossil-fuel emissions since 1960 (ref. 1)1, offsetting some global warming. Tropical forests have been a major component of the land carbon sink, and the largest intact tropical forest is in Amazonia. Writing in Nature, Gatti et al.2 report extensive direct sampling of the atmosphere over this region. Their data reveal that western Amazonia is still a relatively weak carbon sink, but suggest that deforestation and warming over eastern Amazonia have degraded — or even reversed — regional uptake of carbon by the forest.
The sink arises from a combination of the increased vegetation growth that occurs in response to rising levels of CO2 and other nutrients, changes in land management and ecosystem responses to climate change3. Tropical forests are Earth’s most productive ecosystems, but they are not recovering from past human disturbance as their counterparts at mid-latitudes have done, nor are they benefiting from the markedly longer growing seasons associated with climate change, as are boreal and arctic ecosystems.
Carbon has accumulated in the biomass of Amazonian forests for decades4, but studies in the past few years suggest that Amazonian carbon sinks are threatened by deforestation and forest degradation5, progressive drying of the climate, and fires6. However, it is difficult to make direct measurements that reflect the local carbon balance of many Amazonian ecosystems, because access to those regions is limited. And it is hard to extrapolate available local data7,8 across the whole region, because Amazonian ecosystems vary enormously.
Satellite measurements of atmospheric levels of CO2 and carbon monoxide (a tracer of combustion), and of solar-induced fluorescence from vegetation (a proxy for photosynthesis), show that the year-to-year carbon balance in Amazonia is quite sensitive to drought and fire9. However, persistent cloud cover in this region complicates the acquisition of such measurements, and the collection of these data began only about a decade ago. Direct measurements of the atmosphere could constrain estimates of regional carbon balance, but are sparse for Amazonia.
Enter Gatti and colleagues, who have directly measured the atmosphere in four regions across Amazonia for nine years (2010–18; Fig. 1). The authors used aircraft to collect air samples from near the surface up to an altitude of 4.5 kilometres at each region, then analysed the samples to produce a vertical profile of the concentrations of a suite of gases, including CO2 and CO. The authors produced 590 vertical profiles during their study, with a typical sampling period of about twice a month at each region.
Gatti et al. also used data from several sites on remote islands and coastal headlands around the South Atlantic Ocean to establish background concentrations of gases. This allowed them to compute spatial gradients of CO2 and CO concentrations between the background sites and each of the profiled regions in Amazonia. The authors analysed their data by season and by year, and determined how the spatial patterns of CO2 and CO concentrations varied according to the region over which sampled air had passed before collection. Finally, they used the seasonal CO2 concentration gradients to estimate regional carbon flux associated with forest growth and decay, and estimated carbon emissions produced by fires from the CO gradients.
Northwestern Amazonia is almost always very wet, and shows little seasonal variation in growth and decay. Gatti and colleagues’ atmospheric profiling indicates that this region was close to carbon balance during the period of the study — about as much carbon was taken up by plants for growth as was emitted from decay processes.
However, the moisture and fertility of Amazonian forests changes substantially farther south and east. Dry seasons (periods with rainfall of less than 100 millimetres per month) get progressively longer, eventually lasting for 5 months or more as the forest grades into savannah10. Gatti and colleagues find that the drier forests in the northeastern and southeastern regions studied were close to carbon balance during the wet season, but that carbon release from decomposition and fire tended to exceed carbon uptake by photosynthesis during the dry season. The observed regional and seasonal patterns of carbon uptake in the northwest transitioning to carbon release in the drier east were consistent with the year-to-year variability of the data — which revealed that greater carbon releases, associated with decomposition and fire, occurred during hotter and drier years.
Gatti and co-workers show that the transition of eastern Amazon forests from carbon sink to carbon source during the dry season is associated with strong regional warming trends. Eastern Amazon sites have warmed by as much as about 0.6 °C per decade during the dry season over the past 40 years. This is more than three times the rate of global warming and about the same rate as for the Arctic. Wet-season and western Amazonian forests have warmed, too, but at a much slower rate. Warming rates in the dry season for eastern Amazonia might have been amplified by deforestation and forest degradation. Gatti et al. conclude that increases in fires, and in physiological stress, mortality and decomposition of trees in this area, are associated with increasing carbon loss from regional ecosystems.
The authors have documented the accelerating transition of forests from carbon sinks to sources using direct measurements of large-scale gradients of atmospheric gas concentrations. The overall pattern of deforestation, warmer and drier dry seasons, drought stress, fire and carbon release in eastern Amazonia seriously threatens the Amazon carbon sink. Indeed, the results cast doubt on the ability of tropical forests to sequester large amounts of fossil-fuel-derived CO2 in the future.
For decades, ecologists have been surprised that the fraction of fossil-fuel emissions absorbed by land ecosystems has remained fairly constant11, even though these emissions have increased. Forests at high latitudes have continued to accumulate carbon because their growing seasons have lengthened as a result of climate change. Mid-latitude forests have done so because they have been recovering from past clearance, and because they have benefited from the increased availability of nutrients (produced as a result of human activities, or mobilized in soils by climate warming).
By contrast, increased carbon sequestration by tropical forests must be driven largely by an increase in photosynthesis associated with rising CO2 levels — but regional atmospheric profiling12 suggests that this carbon sink is threatened by forest degradation and warming. Another complication is that fossil-fuel emissions must be quickly reduced to meet international climate targets, but it is not clear how the CO2-driven carbon sinks of tropical forests will respond to a rapidly warming world in which CO2 levels are no longer rising13. The future of carbon accumulation in tropical forests has therefore long been uncertain. Gatti and colleagues’ atmospheric profiles show that the uncertain future is happening now.
Nature 595, 354-355 (2021)
Friedlingstein, P. et al. Earth Syst. Sci. Data 12, 3269–3340 (2020).
Gatti, L. V. et al. Nature 595, 388–393 (2021).
Fung, I. Annu. Rev. Earth Planet. Sci. 48, 1–20 (2020).
Pan, Y. et al. Science 333, 988–993 (2011).
Song, X.-P. et al. Nature 560, 639–643 (2018).
Aragão, L. E. O. C. et al. Nature Commun. 9, 536 (2018).
Hayek, M. N. et al. Biogeosciences 15, 4833–4848 (2018).
Malhi, Y. et al. Biol. Conserv. 253, 108889 (2021).
Liu, J. et al. Science 358, eaam5690 (2017).
Marengo, J. A., Liebmann, B.,Kousky, V. E., Filizola, N. P. & Wainer, I. C. J. Clim. 14, 833–852 (2001).
Ballantyne, A. P., Alden, C. B., Miller, J. B., Tans, P. P. & White, J. W. C. Nature 488, 70–72 (2012).
Hubau, W. et al. Nature 579, 80–87 (2020).
Kaushik, A. et al. Eos 101, https://doi.org/10.1029/2020EO140276 (2020).
The author declares no competing interests.