Plant community response to loss of large herbivores differs between North American and South African savanna grasslands.

Herbivory and fire shape plant community structure in grass-dominated ecosystems, but these disturbance regimes are being altered around the world. To assess the consequences of such alterations, we excluded large herbivores for seven years from mesic savanna grasslands sites burned at different frequencies in North America (Konza Prairie Biological Station, Kansas, USA) and South Africa (Kruger National Park). We hypothesized that the removal of a single grass-feeding herbivore from Konza would decrease plant community richness and shift community composition due to increased dominance by grasses. Similarly, we expected grass dominance to increase at Kruger when removing large herbivores, but because large herbivores are more diverse, targeting both grasses and forbs, at this study site, the changes due to herbivore removal would be muted. After seven years of large-herbivore exclusion, richness strongly decreased and community composition changed at Konza, whereas little change was evident at Kruger. We found that this divergence in response was largely due to differences in the traits and numbers of dominant grasses between the study sites rather than the predicted differences in herbivore assemblages. Thus, the diversity of large herbivores lost may be less important in determining plant community dynamics than the functional traits of the grasses that dominate mesic, disturbance-maintained savanna grasslands.

Loss of a large grazer impacts savanna grassland plant communities similarly in North America and South Africa.

Large herbivore grazing is a widespread disturbance in mesic savanna grasslands which increases herbaceous plant community richness and diversity. However, humans are modifying the impacts of grazing on these ecosystems by removing grazers. A more general understanding of how grazer loss will impact these ecosystems is hampered by differences in the diversity of large herbivore assemblages among savanna grasslands, which can affect the way that grazing influences plant communities. To avoid this we used two unique enclosures each containing a single, functionally similar large herbivore species. Specifically, we studied a bison (Bos bison) enclosure at Konza Prairie Biological Station, USA and an African buffalo (Syncerus caffer) enclosure in Kruger National Park, South Africa. Within these enclosures we erected exclosures in annually burned and unburned sites to determine how grazer loss would impact herbaceous plant communities, while controlling for potential fire-grazing interactions. At both sites, removal of the only grazer decreased grass and forb richness, evenness and diversity, over time. However, in Kruger these changes only occurred with burning. At both sites, changes in plant communities were driven by increased dominance with herbivore exclusion. At Konza, this was caused by increased abundance of one grass species, Andropogon gerardii, while at Kruger, three grasses, Themeda triandra, Panicum coloratum, and Digitaria eriantha increased in abundance.

Convergence and contingency in production-precipitation relationships in North American and South African C4 grasslands.

Mesic grasslands in North America and South Africa share many structural attributes, but less is known of their functional similarities. We assessed the control of a key ecosystem process, aboveground net primary production (ANPP), by interannual variation in precipitation amount and pattern via analysis of data sets (15- and 24-year periods) from long-term research programs on each continent. Both sites were dominated by C(4) grasses and had similar growing season climates; thus, we expected convergence in precipitation-ANPP relationships. Lack of convergence, however, would support an alternative hypothesis-that differences in evolutionary history and purportedly greater climatic variability in South Africa fundamentally alter the functioning of southern versus northern hemisphere grasslands. Neither mean annual precipitation nor mean ANPP differed between the South African and North American sites (838 vs. 857 mm/year, 423.5 vs. 461.4 g/m(2) respectively) and growing season precipitation-ANPP relationships were similar. Despite overall convergence, there were differences between sites in how the seasonal timing of precipitation affected ANPP. In particular, interannual variability in precipitation that fell during the first half of the growing season strongly affected annual ANPP in South Africa (P < 0.01), but was not related to ANPP in North America (P = 0.098). Both sites were affected similarly by late season precipitation. Divergence in the seasonal course of available soil moisture (chronically low in the winter and early spring in the South African site vs. high in the North American site) is proposed as a key contingent factor explaining differential sensitivity in ANPP to early season precipitation in these two grasslands. These long-term data sets provided no support for greater rainfall, temperature or ANPP variability in the South African versus the North American site. However, greater sensitivity of ANPP to early season precipitation in the South African grassland suggests that future patterns of productivity may be more responsive to seasonal changes in climate compared with the North American site.