Study Sites. We established large herbivore manipulations in long-term fire experiments at the Konza Prairie Biological Station (Konza), northeastern Kansas, USA, and the Kruger National Park (Kruger), northeastern South Africa. Both savanna grasslands are dominated by C4 grasses with a mix of herbaceous forbs and woody vegetation. Konza is a native tallgrass prairie managed for research since 1977. To address the role of native grazers and fire/grazing interactions, bison were reintroduced in 1987 to a 1000 ha fenced area that includes 10 replicate watersheds burned in the spring (mid-April) at 1-, 2-, 4- and 20-yr intervals. At Kruger, the Experimental Burn Plots (EBPs) were established in 1954 to determine the effects of fire seasonality and frequency on forage production. Similar to Konza, experimental burns occur in the spring (August) at 1-, 2-, 3-, and 6-year intervals, along with unburned controls, each applied to ~7 ha plots. Our research focuses on the Satara, Marheya, and N’wanetsi blocks of the EBPs in south-central Kruger where precipitation, soil type, and the relative abundance of herbaceous and woody plants are similar to Konza. A total of 12-14 megaherbivore species including elephant, rhinoceros, wildebeest, zebra, and impala commonly graze on the EBPs.
Core Experimental Design. To manipulate the presence of large herbivores, we established replicate, 38.5-m^2 (7m diameter circular) herbivore exclosures prior to the growing season in 2005/06 in unburned, intermediate burned (3 or 4 years), and annually burned areas at both Kruger and Konza. We established three blocks of seven exclosures with co-located paired plots open to grazing in each of the three fire treatments (n = 21 exclosures/treatment/site). Exclosures and paired plots were not located beneath trees or in dense shrub patches as we are focusing on the dynamics of the herbaceous plant community.
Differential Effects of Herbivore Size Guilds Experimental Design. Given the wide size range of herbivores in Kruger from the 3000kg elephant to the 11kg steenbok, we have established a second experiment to examine the effect of different size guild of herbivores on plant communities. Across the different fire regimes, we established full herbivore exclosures that exclude all herbivores and partial exclosures that exclude only herbivores larger than an impala (85cm shoulder height). This series of exclosures separates the effects of the common medium-sized grazers such as zebra and wildebeest from the effects of the most abundant herbivores, the mixed-feeding impala.
Vegetation Sampling Methods. Each year, we survey herbaceous plant community structure at the beginning and end of the growing season (Konza: June and August; Kruger: January and March) to capture peak abundance of early and late-season species, respectively. We sample vegetation in a permanent 2 x 2 m plot located within each of the full exclosures, partial exclosures, and paired plots. The 4-m2 plot is divided into four 1-m2 subplots, and within each subplot, we estimate percent cover (to the nearest 1%) for each species rooted inside. We also sample herbaceous plant biomass in two ways. First, we measure ANPP each year at the end of the growing season inside each full exclosure by clipping all aboveground biomass in three 0.1 m2 quadrats. Second, we measure aboveground biomass at the end of each growing season in full exclosures, partial exclosures, and open paired plots using a disc pasture meter. Canopy density is estimated in full exclosures, partial exclosures, and open paired plots by measuring PAR at the ground surface and above the canopy in four locations per species composition plot.
Herbivore Abundance Sampling Methods. We sample the herbivore communities across each of the fire regimes in Konza and Kruger using two complementary methods. First, we use visual surveys of herbivore distribution from a vehicle. We drive the perimeter of each of the different burn plots during the day approximately twice weekly during the growing season and identify and count the different herbivores utilizing the different burn regimes. In Kruger, we have also incorporated nocturnal surveys because many of the herbivores often feed at night and because predators such as lions are most active at night potentially resulting in much different diurnal vs. nocturnal animal distribution. Second, we use surveys of herbivore dung on the different burn regimes to get a time integrated metric of habitat use as dung surveys incorporate both diurnal and nocturnal animal distribution.