Carbon cycle: Sink in the African jungle
http://www.nature.com/nature/journal/v457/n7232/pdf/457969a.pdf
Helene C. Muller-Landau1
Top of pageAbstractApparently pristine African tropical forests are increasing in tree biomass, making them net absorbers of carbon dioxide. Is this a sign of atmospheric change, or of recovery from past trauma?
The lush vegetation of tropical forests is a large and globally significant store of carbon1. Because tropical forests contain more carbon per unit area than any alternative land cover, cutting them down releases carbon into the atmosphere. For the same reason, growing forests take up carbon from the atmosphere. Of course, trees cannot grow for ever, and neither can forests: in the absence of disturbances that kill trees en masse — such as fires, hurricanes or logging — every forest will eventually reach a point at which tree growth and death are in equilibrium, and at which the average change in tree carbon stocks is zero.
It is thus surprising that undisturbed tropical forests currently do not seem to be at equilibrium. If you measure the size of trees in a given area, calculate their carbon stocks, and then repeat the process some years later, you will on average find that the forest holds more carbon than it did before. This was first reported for Amazonian tropical forests2, and on page 1003 of this issue Lewis et al.3 show that African forests also have increasing stocks of tree carbon.
So how much carbon are we talking about? Using data collected in Africa between 1968 and 2007, the authors find that trees have added an average of 0.63 tonnes of carbon per hectare each year. Given that approximately half the dry matter in trees is carbon, the amount of wood added annually in each hectare of African forest is equivalent in mass to a small car. For comparison, the average rate of carbon accumulation in tropical forests around the globe was 0.49 tonnes of carbon per hectare per year2, 3, 4. Extrapolating from their data3 by assuming parallel changes in the carbon pools of roots and dead trees, Lewis et al. estimate that 'old-growth' tropical forests are taking up 1.3109 tonnes of carbon per year worldwide.
There are two possible explanations for this finding. One is that the tropical forests that we think of as intact actually suffered major disturbances in the not-too-distant past, and are still in the process of growing back5. This recovery process is known as succession, and takes hundreds — or even thousands — of years. Succession involves not only initial growth to full canopy height, but also subsequent gradual shifts in species composition. The past disturbances could have been natural or anthropogenic; possible explanations include droughts and fires related to huge El Niño events, and changes in land use that allowed previously cleared land to revert to forest5.
In fact, palaeoecological and archaeological evidence increasingly documents the long disturbance histories of today's 'undisturbed' tropical forests6. There have been many large fires in Amazonian forests over the past few millennia, the timings of which are related to both climate and the size of human populations7. Far from being pristine wildernesses little influenced by their human inhabitants, many areas were cleared or otherwise intensively used in centuries past8. Given the timescales of tropical-forest succession, these disturbances are almost certainly contributing to carbon accumulation in many tropical forests today.
The second explanation for Lewis and colleagues' findings3 is that tropical forests have been knocked from their previous equilibrium by global climate and/or atmospheric change9, so that they are currently in transition to a higher carbon state. Perhaps, for example, the increase in atmospheric carbon dioxide is effectively fertilizing tropical tree growth. Under these circumstances, if tree mortality doesn't keep pace with increases in growth, then trees will on average grow larger before they die (Fig. 1), and tree carbon stocks will increase10. Carbon stocks in mature tropical forests vary enormously depending on climate, soil type and topography; temporal changes in climate and resource availability would therefore be expected to have parallel influences in the long run.
The two mechanisms that might account for Lewis and colleagues' observations3 would be expected to produce different spatial and temporal patterns of carbon uptake by trees, but our current knowledge does not allow us to predict what these patterns are, or to say which mechanism is operating in Africa. Over the course of succession, tree carbon stocks increase at an ever-slower rate as stands age. Thus, we expect tree carbon stocks and their rate of change, and stand age, to be closely related within any given forest type. In temperate and boreal forests, where stand age is generally well known, carbon stocks and fluxes do indeed show a strong relationship with stand age, even at ages many consider to be old-growth11. Examination of these relationships in tropical forests is stymied not only by lack of information about how long ago disturbances occurred, but also by limited knowledge of how growth rates and equilibrium carbon stocks are affected by rainfall, soils and other factors.
One might suppose that predictions based on the global-change hypotheses are more straightforward — after all, atmospheric CO2 concentrations are rising equally everywhere. In fact, the effects of CO2 fertilization on tree growth are expected to depend strongly on other factors that vary greatly among forests, especially the availability of soil nutrients1. And if changing climate (rather than rising atmospheric CO2) is affecting the carbon flux of tropical forests, then the outcomes will differ depending on local changes and the local baseline.
It is likely that both succession and global change have a role in explaining tropical-forest growth, with varying importance at different sites. Where recovery from disturbance drives tropical-forest change, associated changes in species composition would be expected. A study published last year4 found that tree species with slower growth rates are disproportionately increasing in biomass in nine out of ten 'undisturbed' tropical forests around the globe, as would be expected during succession. Yet Lewis et al.3 find no relationship between a species' wood density and the rate of change of its population across their African plots.
A better understanding of tropical-forest carbon dynamics is clearly needed to determine the causes of the observed increases in tropical tree carbon stocks — and, more critically, to predict the future trajectory of these stocks under global change. Furthermore, we must look not only at the trees3, but also at the soil: tropical-forest soils hold at least as much carbon as the trees. Unlike tree carbon stocks, soil carbon stocks can potentially increase indefinitely. But the prevailing prediction is that increasing temperatures will speed decomposition and reduce soil carbon stocks.
In the future, will tropical trees and soils act as carbon sinks, thereby slowing atmospheric and climate change? Or will altered climates turn them into carbon sources that accelerate further change? Standardized assessments of the main carbon pools and fluxes of tropical forests around the world — and their proposed drivers — are needed to document and understand the current trends, to inform predictive models, and ultimately to answer these pressing questions.
Top of pageReferences
Solomon, S. D. et al. (eds) Climate Change 2007: The Physical Science Basis (Cambridge Univ. Press, 2007).
Phillips, O. L. et al. Science 282, 439–442 (1998). | Article | PubMed | ISI | ChemPort |
Lewis, S. L. et al. Nature 457, 1003–1006 (2009). | Article |
Chave, J. et al. PLoS Biol. 6, e45 (2008). | Article | PubMed | ChemPort |
Wright, S. J. Trends Ecol. Evol. 20, 553–560 (2005). | Article | PubMed |
Clark, D. A. Biotropica 39, 4–19 (2007). | Article |
Bush, M. B., Silman, M. R., McMichael, C. & Saatchi, S. Phil. Trans. R. Soc. Lond. B 363, 1795–1802 (2008). | Article | ChemPort |
Mann, C. C. Science 321, 1148–1152 (2008). | Article | PubMed | ChemPort |
Lewis, S. L. Phil. Trans. R. Soc. Lond. B 361, 195–210 (2006). | Article |
Muller-Landau, H. C. et al. Ecol. Lett. 9, 589–602 (2006). | Article | PubMed |
Luyssaert, S. et al. Nature 455, 213–215 (2008). | Article | PubMed | ChemPort |
