Fire and grazing play key roles in determining the distribution, structure and function of savanna grassland ecosystems worldwide. Since these ecosystems cover ~40% of the terrestrial land surface and provide many important ecosystem services it is important to understand the effects of fire and grazing on these systems. It has been argued that in South Africa (SA) with its longer evolutionary history with fire and grazing, generally lower soil fertility, and a greater diversity of plants and larger herbivores that fire and grazing will influence savanna grassland structure and function in fundamentally different ways compared to North America (NA). This study tests that hypothesis by comparing sites in Kruger National Park (KNP), SA, a 2 million ha protected area containing 14+ species of large mammalian herbivores, and in the Konza Prairie Biological Station (KPBS), NA, a 3,487 ha research station grazed by bison. In 2006, fenced exclosures were established in plots that were unburned, burned annually or burned every 3-4 years at KNP and KPBS, to exclude all large herbivores. Additionally, exclosures were established at KNP within an enclosure which contains only a single grazer, water buffalo. Species composition was sampled annually in all exclosures and corresponding paired grazed plots.
Results/Conclusions
In grazed sites, plant species richness and diversity were lower in unburned sites in NA, but higher in SA. Additionally, community composition differed among the fire regimes in SA due to turnover of the dominant grass species, but there was no significant difference in species composition attributable to fire regime in NA. Furthermore, while initial differences in richness and diversity across the different fire regimes were maintained in SA, these differences disappeared in NA. Meanwhile in ungrazed sites richness and diversity declined rapidly in NA across all fire regimes, whereas richness declined only in the absence of burning in SA. Additionally, herbivore exclusion causes shifts in species composition in NA, but not in SA. However, some of these observed differences between the NA and SA sites appear to be due to the higher herbivore diversity in SA, as more similar trends in the response of composition, richness, and diversity to fire and grazing were observed when comparing the NA sites to the SA site with only a single grazer. This suggests that fire and grazing can impact savanna grasslands in different ways in SA and NA, and thus care should be taken when generalizing results across savanna grassland ecosystems.