|dc.description.abstract||Interactions between livestock, wildlife and fire elicit a strong debate in management circles and are of substantial interest to ecologists. While these interactions are quite common in most African savannas, their implications for biodiversity are poorly understood, yet most management decisions in savanna ecosystems revolve around herbivory and fire prescriptions. This dissertation presents results from three separate experiments examining interactions between livestock, wildlife and fire.
The first experiment examined the impact of cattle on wild mesoherbivores (mean adult body mass range 16 – 1000 kg) and how this may be influenced by megaherbivores (elephants and giraffe) and rainfall. Dung density data collected over a period of eight years were used as a proxy for relative use of 200 x 200m plots excluding different combinations of herbivores. Cattle had a range of effects mostly negative on common mesoherbivore species, including both grazers (zebra, oryx, and hartebeest) and mixed feeders (eland, Grant’s gazelle, and steinbuck). These effects were contingent on the presence of megaherbivores and the amount of rainfall. The second experiment examined the influence of different herbivore guilds on fuel loads, fire temperatures and post-fire survival of symbiotic acacia ant species. Small (30m x 30m) subplots were burned within each of six replicated herbivore treatment plots excluding different combinations herbivores. Fire temperatures were higher in plots excluding cattle and in plots excluding megaherbivores than plots in which they were each allowed. The impact of cattle on fire temperature were more associated with reduction in herbaceous fuel loads while the impact of megaherbivores were more related to reduction in tree density and the concomitant reduction in fine woody debris accumulating underneath trees. Survival of ants occupying shorter trees
(less than 4 m tall) was negatively correlated with mean herbaceous fuel load and mean minimum fire temperature across herbivore treatments. The third experiment examined temporal dynamics in herbivores visitation of burned areas implemented at different spatial scales. Dung survey data collected during two distinct time periods (1-2 and 6-7 years after burning) were compared. Most species displayed less discrimination of burned and unburned areas during the second sampling period, with less preference of burned areas (impala, Grant’s gazelle, and hare), or less avoidance of burned areas (elephant, eland, and warthog). Persistence of the initial impacts of burns on herbivores depended on the spatial scale of the burns. For most burns, grass had regrown to the same height as the unburned areas by the second sampling period. However, the density of taller trees (>0.8 m) was significantly lower in all the burned areas than unburned areas during the second sampling period. By highlighting important differences in response of different mesoherbivores to cattle, which may be contingent upon rainfall and presence of megaherbivores, this study presents as significant departure from the longstanding assertion that cattle simply compete with wildlife. The study also presents the first experimental evidence that different herbivore guilds indirectly influence fire temperature and severity differently. While most prior studies suggest that the impact of burning are short-lived (lasting up to two years), this study presents contrary evidence that the impacts of fire may last at least 6 years depending on the spatial scale at which fire is prescribed. Hence this study has important theoretical and conservation implications.||en_US