Exploring complex relationships across trophic levels in savannas: Project TREE takes root

By Dr Melissa Schmitt¹ and Dr Dave Thompson²

African savannas are highly diverse in flora and fauna, with elements of each interacting within, and across, trophic levels.

For example, the distribution and abundance of large mammalian herbivores – the browsers and grazers in the system, are influenced by lower-tier nutrition (vegetation availability and quality) and the risk of being eaten by upper-tier carnivores.

In turn, the distribution and densities of mammalian predators are influenced by lower-tier food availability and hunting success, which is impacted by manoeuvrability and visibility in the bottom-tier vegetation. Ultimately, the combination of food availability (vegetation or prey species), fear (eat or be eaten) and foraging/hunting success, which determine the spatial arrangement of species in the landscape, are all influenced by savanna habitat structure, specifically the tree:grass ratio.

Figure 1. Shifting tree:grass ratios result in diverse habitat structures that lie on a continuum from open (A) to closed-canopy (D) savanna. Each provides different opportunities for mammal species to either find, or to become, food.

Current climate change scenarios push the tree:grass ratio in favour of the trees, with bush encroachment and woody thickening being observed and further predicted for many parts of southern Africa. As climate change progresses, African savannas will likely see an increase in the amount of woody vegetation and a corresponding reduction in grasses, resulting in more closed-canopy systems (Figure 1D). Climate change and the altered savanna states it induces may therefore fundamentally influence the direct and indirect interactions between predators, their prey, and their environments (Figure 2).

In an effort to maintain the tree:grass ratio, and so maintain savanna community dynamics across trophic levels, many land managers in South Africa have implemented a number of tree suppression practices, most commonly the prescribed use of fire. But these management practices extend to the active removal of adult trees, and/or the annual mowing of the grass matrix to suppress woody plant establishment, sometimes at large scales (>100 ha).

These management regimes not only alter the vegetation composition and structure – essentially creating artificial grasslands when taken to the extreme (Figure 1A), but will also influence the behaviour, distribution and population sizes of herbivore species that track the availability of food. Ironically, conversion to closed-canopy savannas at the other end of the continuum will have different, equally complex effects elsewhere in the trophic pyramid.

Figure 2. Conceptual framework showing trophic interactions in South African savannas. These interactions are positive, negative or unknown and typically operate in both directions. For example, increasing the availability of grass positively impacts grazers, whilst increasing grazer density will negatively affect the grass. The relative strength of these interactions varies in time and may include complex feedbacks (not shown). For example, tall grass grazers facilitate feeding by short grass grazers within the grazer guild, but only at certain times of the year. Management actions which alter the vegetation (or climate change effects in the absence of intervention) – here being the ratio between trees and grass (highlighted), have direct and indirect impacts which cascade across levels up and down the trophic pyramid.

Figure 3. University of California project collaborator Dr Keenan Stears resets a motion-sensor camera trap overlooking a ‘giving up density’ (or GUD) feeding station. GUD assays measure the trade-off between food and fear experienced by herbivores. In open habitats where trees are suppressed, the fear of predation often overshadows the urge to feed.

Collaborative study

Despite dire consequences of habitat change on species densities, distributions and interactions – particularly at higher trophic levels, our understanding of the complexities of vegetation-prey-predator dynamics remains poor.

To address this, a collaboration between SAEON’s Ndlovu Node and the University of California Santa Barbara, initiated Project TREE (Tri-trophic Relationships in Engineered Environments), a research effort under the direction of postdoctoral fellow Dr Melissa Schmitt. Project TREE seeks to understand how anthropogenically induced change in savanna structure influences the relationships among plant, herbivore and predator trophic levels. Understanding the consequences of continuing woody thickening, as well as the outcomes of mitigation practices being implemented, is of critical importance to savanna conservation and the maintenance of ecosystem functioning.

The study is based on first assessing how large-scale and long-term tree removal and mowing influence the availability, quality and functional traits of the grass layer. From this, the researchers will determine: 1) how changes to the vegetation structure and composition influence the trade-off between perceived predation risk and food availability for herbivores; 2) distribution and habitat use by predators; and 3) how herbivore distribution driven by the food-fear trade-off influences the distribution of nutrients through dung deposition across the landscape.

To address the above aims, the research team are employing a diverse set of approaches, including long-term herbivore and predator sightings data, long-term kill location data, ‘giving up density’ feeding assays, empirical data on herbivore habitat associations, and annual quality and composition sampling for the plant community and the underlying soils.

1. Postdoctoral researcher affiliated with SAEON and the University of California Santa Barbara

2. Biodiversity Scientist, SAEON Ndlovu Node

Return to Index