There is a global obsession with deforestation, and not without reason, given the lessons of recent human activity in tropical systems. With this in mind, a
recent paper in Nature by Higgins and Scheiter1 poses a challenging question for African ecologists and environmentalists: do we, in the
subcontinent, face not a contraction, but a vast and inevitable expansion of subtropical tree cover, driven by levels of CO2 that have not been
seen in the past several million years? Higgins and Scheiter’s paper warrants our attention because it projects, for the first time, the continent-wide
implications of the decade-old hypothesis, originally formulated by South African ecologists, of CO2-driven woody expansion in fire-prone savannas.
2 The paper supports concerns that the expansion of woodlands and forest may be an imminent threat to ecosystem structure, function and biodiversity
across extensive landscapes in the sub-continent.3 If its projections are correct, then we stand on the brink of massive ecosystem change in the
‘savanna–complex’ vegetation (i.e. tropical grasslands, savanna and forests) of Africa. But how credible are these projections?
Unlike more intensively researched temperate ecosystems, the vegetation structure and land cover of huge tracts of sub-Saharan Africa may be highly sensitive
to increasing levels of atmospheric CO2. Vast areas of the subcontinent are currently dominated by C4 grasses – a photosynthetic mode
that owes much of its competitive advantage to the low CO2 levels of pre-industrial and, even more so, glacial times.4 Grasses do not
require the large amounts of carbon that woody plants do to support their photosynthetic tissue. This low carbon demand for growth allows grasses to outcompete
woody plants under low CO2 conditions by building up a flammable layer of grass fuel – the savanna fire trap – that immolates slower
growing woody plants and maintains the system in its grassy state. Under high levels of CO2, trees are thought to regain the advantage, escaping the
fire trap and converting the system into forest.2 This mutable balance of trees versus grasses, mediated by atmospheric CO2 levels and fire,
results in ‘bi-stable’ systems5 in which either grasses or trees could dominate. These bi-stable systems are highly prone to the infamous
‘tipping point’, an abrupt switch to an alternate stable state from which the original ecosystem does not easily recover.
C4 grassland and savanna ecosystems spread globally only in the last 8 million years6 – a spread that was primed by an extended
period of low atmospheric CO2 levels. Today, C4 grassland and savanna ecosystems contain some of Africa’s most iconic biodiversity
and support a large fraction of Africa’s human population. Together with subtropical savannas in other parts of the world, these ecosystems are vast enough
to have a major impact on the earth system. However, fossil fuel emissions have now driven atmospheric CO2 levels higher than those experienced by
plants for at least the last 800 000 years,7 and possibly several million years. If emissions continue on a ‘business-as-usual’ path,
by mid-century, CO2 levels will exceed those last seen more than 25 million years ago – far predating the rise of grasslands and savannas. Could
these elevated CO2 levels result in a widespread and abrupt shift from grassy to woody vegetation across Africa?
Higgins and Scheiter1 attempted to answer this question using an adaptive dynamic global vegetation model (aDGVM).8 Mechanistic models of
this kind use established plant physiological mechanisms and spatially explicit climate data to simulate potential vegetation through modelling the growth of
individual plants of a variety of ‘functional types’ (e.g. grasses, shrubs and trees), the outcome of competitive interactions between individuals,
and processes such as disturbance by fire. Ultimately, vegetation types are categorised according to the relative cover and/or biomass of plant functional types.
What makes their ‘adaptive’ approach unique, compared with other DGVMs, is that they model the ability of plants to adapt their phenology and growth
so as to change their allocation of carbon to different plant compartments such as roots, stems and leaves. This model is a key advance that represents, more
credibly than earlier DGVMs, the mechanisms behind the escape from the savanna fire trap. As a consequence, by turning fire on and off in their simulations, they
are able to specifically identify ‘bi-stable’ areas which could support either trees or grasses.
Using the A1B emissions scenario from the Intergovernmental Panel on Climate Change Special Report Emissions Scenarios, Higgins and Scheiter projected
large shifts in ‘savanna–complex’ vegetation types from 1850 to 2100. Key projections were:
1. A general shift to a more woody state across the continent as a result of reductions in C4 grassy ecosystems, and the spread of forests,
as a direct result of increases in atmospheric CO2.
2. A rapidly accelerating shift to forest over this century. Between 1850 and 2000, the model projects only an incipient change. However, from 2000 to 2100,
a 10-fold increase in woodiness is projected across the continent.
3. Bi-stable areas capable of sustaining both grassy and forest vegetation moving to largely non-overlapping areas between 1850 and 2100. Thus, while these
bi-stable states continue to persist, they do so in different geographical locations in the future – a major challenge for conservation and adaptation
4. Extremely abrupt changes from grassy to woody vegetation at the local scale (read ‘tipping point’), but not synchronously across the continent.
It is from this projection that the paper takes its rather ambiguous title – ‘Atmospheric CO2 forces abrupt vegetation shifts locally, but
not globally’ – meaning that the projected ‘global’ transition to woodiness across the continent will not occur as abruptly as it will at
a given locality.
Are these projections credible? While the Higgins and Scheiter paper is a significant advance in dynamically simulating future African vegetation cover, there
remain important uncertainties. For example, Higgins and Scheiter assumed constant rainfall over the 1850–2100 period because: ‘The high uncertainty
in precipitation change over Africa led us to assume that rainfall remained at ambient levels’1. Considering the fundamental effect rainfall has
on net primary productivity9 and tree mortality,10,11 and the fact that rainfall frequency and intensity is likely to change in an evolving
climate,12 their assumption seriously compromises any assessment of the sensitivity of their projections to possible rainfall change – a
remarkable omission. Another significant limitation of the model is that the aDVGM does not simulate nitrogen limitation of growth via nutrient cycling feedback
at the ecosystem level. This shortcoming, which has been resolved in other DGVMs, could exaggerate the simulation of forest expansion. Additionally, changes in
land use, including deforestation for growing crops, fire management and feedbacks between vegetation and climate, will also influence Africa’s future land
cover. The extent to which these drivers may counteract the forcing of elevated CO2 and temperature is currently unknown.
What is particularly interesting is that the dramatic CO2-driven ecosystem changes projected by Higgins and Scheiter are in stark contrast to a more
traditional, climate-centric view of the world, where African vegetation appears to be highly stable.13 Their model provides strong evidence that the
equilibrium, ‘one climate, one vegetation’ approach13 is inappropriate for Africa, where direct CO2 effects on plants appear to
be an important driver of vegetation.
Higgins and Scheiter’s study therefore prompts urgent consideration of several key questions. Are African C4 landscapes doomed under elevated
CO2? Will we see a ‘reforestation’ of subtropical Africa in the coming decades? Or are alternative drivers like changes in land use or
rainfall likely to oppose the CO2-driven trend? Currently, climate uncertainty and socio-economic uncertainty combine to create a murky view of the
future. Only through an interdisciplinary approach can these knowledge gaps be spanned.
Recently, several funding initiatives in South Africa have emerged that promise to advance our understanding of African land-cover change. In 2010, strategic
funding from the University of Cape Town initiated the formation of the Land Cover Change Consortium,14 a group of interdisciplinary scientists from
across South Africa that examines land-cover change from an experimental, observational and modelling perspective. Research initiatives such as these are now
benefitting from funding allocated through the Department of Science and Technology’s Global Change Grand Challenge, the National Research
Foundation’s ACCESS programme and funding from multinational partners (e.g. Southern African Science Service Centre for Climate Change and Adaptive Land
Management, www.sasscal.org) that will support research (and students!) in this area until at least 2020. Additionally,
there is an upswing in experimental facilities and field sites targeting global change research nationally. For example, Rhodes University has just committed
to co-funding a National CO2 Research Facility for plant sciences. With the clear imperatives for science to address societal challenges, and the
support of government and funding agencies for these initiatives, it is an exciting time to be involved in CO2 and land-cover change research in South
1. Higgins SI, Scheiter S. Atmospheric CO2 forces abrupt vegetation shifts locally, but not globally. Nature. 2012;488(7410):209–212.
2. Bond WJ, Midgley GF. A proposed CO2-controlled mechanism of woody plant invasion in grasslands and savannas. Glob Change Biol. 2000;6(8):865–869.
3. Bond WJ, Midgley GF. Carbon dioxide and the uneasy interactions of trees and savannah grasses. Philos Trans R Soc Lond B Biol Sci. 2012;367(1588):601–612.
4. Ehleringer JR, Cerling TE, Helliker BR. C4 photosynthesis, atmospheric CO2, and climate. Oecologia. 1997;112(3):285–299.
5. Staver AC, Archibald S, Levin SA. The global extent and determinants of savanna and forest as alternative biome states. Science. 2011;334(6053):230–232.
6. Beerling DJ, Osborne CP. The origin of the savanna biome. Glob Change Biol. 2006;12(11):2023–2031.
7. Luthi D, Le Floch M, Bereiter B, et al. High-resolution carbon dioxide concentration record 650,000-800,000 years before present. Nature. 2008;453(7193):379–382.
8. Scheiter S, Higgins SI. Impacts of climate change on the vegetation of Africa: An adaptive dynamic vegetation modelling approach. Glob Change Biol. 2009;15(9):2224–2246.
9. Zhao MS, Running SW. Drought-induced reduction in global terrestrial net primary production from 2000 through 2009. Science. 2010;329(5994):940–943.
10. Allen CD, Macalady AK, Chenchouni H, et al. A global overview of drought and heat-induced tree mortality reveals emerging climate change risks for forests. For Ecol Manage. 2010;259(4):660–684.
11. Fensham RJ, Fairfax RJ, Ward DP. Drought-induced tree death in savanna. Glob Change Biol. 2009;15(2):380–387.
12. Pachauri RK, Reisinger A, editors. Climate change 2007: Synthesis report. Geneva: Intergovernmental Panel on Climate Change, 2007; p. 104.
13. Bergengren JC, Waliser DE, Yung YL. Ecological sensitivity: A biospheric view of climate change. Clim Change. 2011;107(3–4):433–457.
14. Gillson L, Midgley GF, Wakeling JL. Exploring the significance of land-cover change in South Africa. S Afr J Sci. 2012;108(5/6), Art. #1247, 3 pages.