The Kruger National Park (KNP) in northeast South Africa harbors a unique assemblage of Pleistocene megafauna, increasingly threatened by protracted drought. Managers are concerned that climate change has destabilized this diverse food web by reorganizing competition for surface water resources and shifting ranges. Such destabilization could appear as responses to vegetation and rainfall variation. In particular, locally rare antelope populations (tsessebe Damaliscus lunatus, eland Taurotragus oryx, sable Hippotragus niger, roan Hippotragus equinus) have plummeted over the past six decades, bringing about a decline in the biodiversity that is a cornerstone of tourism and ecosystem resilience.
Management of a diverse, destabilized community necessitates an understanding of how species individually and collectively respond to their environment. Such communities susceptible to disturbance are a challenge for traditional species distribution models (SDMs), whose static data lose much of the information present in increasingly available long-term datasets. We utilize decades of census, vegetation, and climate data from the KNP in a dynamic generalized joint attribute model, GJAMTime, extending a traditional Lotka-Volterra framework to include density-independent growth and interspecific density dependence. Our assessment of how changing vegetation, surface water, and biotic interactions have influenced biodiversity in the KNP provides a framework for future adaptive management decisions.
Results/Conclusions
Outputs consistent with prior knowledge of biotic and environmental interactions confirm that our model works well in predicting species distribution and behavior. Our model estimated similar responses to environmental variables for species known to respond similarly (i.e. blue wildebeest Connochaetes taurinus and plains zebra Equus quagga). The model attributes little variation in species abundance to movement, which is consistent with almost-full fencing around the park. We found that rare antelope distributions are mostly attributable to density-independent growth. Roan and eland have negative responses to temporal anomalies in grass biomass, and roan and tsessebe respond positively to nutrient-rich clay soils. Sable, however, responded positively to vegetation anomalies and negatively to clays, which suggests an affinity for conditions not shared by the other three.
We find that climate change does affect this savanna community, but that consequences for rare antelope are not uniform, indicating a need for more targeted, adaptive management practices.