Megaherbivory, plant defenses, and landscape change in savanna ecosystems

Elephants and other large ungulates are important agents of mortality for savanna trees, and browsing maintains the tree-grass codominance typifying savannas. However, plant defenses may buffer against browsing as a driver of tree population dynamics. We have combined time series of satellite images, free-choice feeding experiments, and long-term monitoring of tree growth to show that abundances of a myrmecophyte (Acacia drepanolobium) are decoupled from variation in elephant populations over a six-year period. Other acacia that are not protected by ant symbionts decline following increases in elephant numbers through time. Our findings demonstrate that ant symbionts deter catastrophic herbivory on trees, thereby slowing landscape change following the extinction of elephants and other ungulates. Further, because adult A. drepanolobium are buffered from top-down control, tree recruitment inevitably assumes a critical role in driving population dynamics. Consequently, small mammals impose demographic filters by consuming tree seeds and seedlings, and have impacts on tree populations that rival those of their more conspicuous mammalian counterparts. Finally, we are beginning to test how the spread of an invasive ant (Pheidole megacephala) affects tree cover by disrupting the ant-acacia mutualism.

Current collaborators: Corinna Riginos, Todd Palmer, Alejandro Pietrek, Brandon Hays

Ungulate Herbivory Under Rainfall Uncertainty: the UHURU experiment

Interaction webs are pervasive features of ecological systems, and studies of these systems often attempt to reduce nature to pairs or subsets of species. However, the contextual nature of interaction webs merits closer inspection if we are to generate a mechanistic understanding of when and where the loss of species is likely to have consequences for community structure and ecosystem function in a world of increasing environmental change. Established in 2008 with colleagues Rob Pringle and Todd Palmer, the Ungulate Herbivory Under Rainfall Uncertainty (UHURU; the Kiswahili word for "freedom") experiment is unique both in size and in scope. Using a series of 36 1-ha plots situated across a pronounced rainfall gradient, UHURU differentially excludes or permits access by various combinations of wild herbivores to mimic extinction and climate change scenarios in East Africa. Through UHURU, we are trying to address a number of questions in community ecology, including functional redundancy among mammalian grazers, effects of extinction on ecosystem processes (e.g., nutrient cycling, soil-water infiltration), and the extent to which herbivory can stabilize productivity under different levels of anthropogenic disturbance.

Current collaborators: Rob Pringle, Todd Palmer, Alois Wambua, Allison Louthan, Dan Doak

Diet generalism, dispersal, and abundance-occupancy relationships

Abundance-occupancy relationships are among the most pervasive patterns in community ecology, with species tending to be locally abundant and widespread, or rare with restricted distributions. The mechanisms underlying this pattern are rooted in long-standing debates over the relative importance of niche- versus dispersal-assembly. Using a species-rich guild of small, herbivorous mammals, we are testing the hypothesis that abundant, widespread species are ecological generalists, capable of persisting on a variety of foods. In contrast, rare, restricted species may be ecological specialists, highly selective in their diets and habitats.

Current collaborators: Tyler Kartzinel, Seth Newsome, Deborah Boro, Rhiannon Jakopak, Alex Buerkle

Trophic dynamics in mammal-dominated ecosystems

By consuming herbivores, carnivores have the potential to bolster the abundance and diversity of plants via trophic cascades. Due largely to highly publicized work in North America involving wolves, elk, and trees, trophic cascades have captured the attention of ecologists and the general public alike. Although the hypothesis of trophic cascades makes good sense, we lack a broader knowledge of if and how they work in other ecosystems, and how human activities such as livestock production and predator control might affect them.

Current collaborators: Caroline Ng'weno, Adam Ford, Abdullahi Hussein Ali, Matt Kauffman