How do global change and functional traits influence savanna woody plant encroachment?

The degradation of tropical ecosystems by deforestation and global environmental change is a major global issue that harms biodiversity and people's livelihoods. Degradation can be defined as the erosion of ecosystem functions and services. In tropical forests, it arises largely from deforestation. However, the situation is different in tropical savannas because many of the services they provide depend on the grassy vegetation growing between scattered trees. In these ecosystems, degradation can therefore be caused by tree planting or natural tree invasion.

Increasing tree densities in savannas lead to degradation because they shade out the grasses. In the iconic savannas of Africa, satellite data shows that, all across the continent, savannas are being naturally invaded by trees, excluding grasses to make dense thorny thickets. This is a process of degradation because people need grasses to graze their animals, trees use more water than grasses and deplete drinking water supplies, and dense thickets are bad for the savanna wildlife that thrives in open grassy vegetation.

Our project will study how global environmental change is causing the invasion of trees into African savannas. It focuses on the root cause of this continental transformation of iconic ecosystems, aiming to resolve fundamental unknowns about global change drivers, especially how rising atmospheric CO2 interacts with water availability. We will carry out a large-scale field experiment, investigating how global change causes tree encroachment into savannas, and particularly evaluating two observations.

First, satellite data and field surveys have shown that tree encroachment is accelerating in wet savannas but is slow in dry savannas. We will test a possible explanation for this observation, by looking at how rising CO2 interacts with rainfall, and testing the idea that rising CO2 benefits grasses more than trees in dry conditions, causing grasses to out-compete trees.

Secondly, field surveys have shown that only particular tree species are responsible for encroachment into savannas. We will test the idea that rising CO2 enhances the growth characteristics of these trees that enable them to escape burning by fires, recover after being eaten by herbivores, and compete effectively with grasses. Conversely, we don't expect to see the same level of enhancement in trees that don't encroach into savannas.

We will run the experiment in a unique South African facility that uses 3-metre open-top chambers to enrich the atmosphere in CO2 to a level expected in the near future. Our experiment will compare savanna tree and grass growth under current and near-future CO2 concentrations, under competition when rooted in large pots filled with a savanna soil. To test our ideas about the interactions between trees, grasses, CO2 and water, we will grow the trees with and without grass competition, and with a high or low soil water availability.

The idea that rising CO2 benefits grasses more than trees in dry conditions is new and we do not yet appreciate its broad significance. Therefore, in parallel with the experiment, we will set up a field monitoring programme to look at when and where savanna grasses and trees are limited by water. We will use this information to interpret our experimental results. How widespread and frequent is water-limitation, and where / when is the interaction between water-limitation and rising CO2 most likely to become important?

Because CO2 has already risen from historical levels, any effects on grasses should also be evident in long-term savanna management experiments where the same fire regime has been applied for decades. We will therefore analyse data from these experiments to look for changes in grass production.

Overall, our work will bring an important advance in understanding of how global change causes savanna tree encroachment, and the ecological mechanisms involved.

Who might benefit from the research?

WPE potentially threatens the livelihoods of a diverse range of stakeholder groups, by threatening a range of savanna ecosystem goods and services. Services that depend on the grassy ground layer of savannas are particularly important in this context, for example grazing for domestic livestock. However, tree invasion also has the potential to influence groundwater recharge and, by excluding grass-dependent wildlife, to impact on the income from wildlife tourism. Potentially beneficial effects of tree invasion include flood risk mitigation and landscape carbon sequestration.

Our research is basic ecological science but will provide information about the likely future impacts of WPE and management strategies that may be effective. In the long-term, the findings of this work could therefore be beneficial for a diverse range of stakeholders. These include: a) commercial ranchers, subsistence herders and communities that manage common rangeland; b) wildlife managers and governmental agencies with interests in biodiversity (e.g. the South African National Biodiversity Institute, SANBI); c) local, regional and national institutions responsible for managing water supplies; and d) government agencies responsible for meeting commitments under the Paris climate change agreement to sequester carbon through land use change. The challenge of how to manage tree encroachment into rangelands impinges on SDGs 15 (Life on Land), 6 (Clean Water) and 13 (Climate Action), with a range of important implications for local economies and food production.

The work will also be beneficial for: e) modellers employed by governmental and non-governmental bodies such as NCAR in the USA and the Met Office in the UK. Drylands in general, and savannas in particular, are not currently represented well in models of terrestrial vegetation and the carbon cycle, and this poor representation has important potential implications for Earth System feedbacks to the climate system.

How might they benefit?

Four aspects of the proposed research could provide beneficial information.

i) Experimental evidence about the ways in which water availability interacts with rising CO2 could provide important data about where, when and why WPE is occurring, offering an ecological explanation for spatial variation in the rate of change.

ii) Experimental evidence about the plant traits causing some tree species encroach into savannas could provide novel insights into how encroachment works, suggesting potential strategies for managing the phenomenon and the appropriate mechanisms that need to be represented in models of savanna ecosystem structure and function.

iii) Field observations about when and where water is limiting for savanna ecosystem production is potentially beneficial for understanding the most appropriate seasons for harvesting tree biomass or grazing livestock in each region.

iv) Historical evidence of increasing grass production is potentially of interest for managing stocking densities in rangelands.