Human alterations of historic disturbance regimes are as pervasive as climate change for impacting ecological patterns and processes. This aspect of global change is particularly important for disturbance-dependent ecosystems, such savanna grasslands, in which fire and grazing by large herbivores are key determinants of plant community structure and ecosystem function. In most savanna grasslands today, either all or a significant fraction of the native large herbivores have been lost, often in a size-dependent manner (i.e., larger species lost first). Savanna grasslands also have experienced significant alterations in their historic fire regimes, ranging from active fire suppression that almost completely eliminates fire to annual burning. Given the importance of both fire and grazing in these systems, it is surprising how little we know about the effects of the combined alteration of these disturbances, particularly for African savanna grasslands that harbor some of the last remaining diverse, large herbivore assemblages in the world. Such knowledge is critical for predicting the impacts of modified disturbance regimes on savanna grasslands globally and for the conservation of these ecologically and economically important ecosystems.

For the last nine years, we have compared responses of mesic (subhumid) savanna grasslands (>500 mm MAP in the tropics and >600 mm MAP outside the tropics) in North America and South Africa to alterations in both fire and grazing regimes. The long-term, comparative experiment that forms the centerpiece of this cross-continental research program is located in tallgrass prairie at the Konza Prairie Biological Station (Kansas, USA) and in knob-thorn marula savanna at the Kruger National Park (Limpopo and Mpumalanga provinces, South Africa). We refer to this study as the Konza-Kruger (K-K) Experiment. At both sites, we have been manipulating grazing by removing all large herbivores (>5 kg) from research plots with permanent exclosures (each with a paired plot that grazers can freely access). These exclosures were established in replicated fire frequency experiments ongoing at each site (treatments range from >25-50 yrs of annual burning, burning every 3-4 yrs, or complete fire exclusion). In addition, at Kruger only, we have manipulated grazing in a second experiment by removing large herbivores in a size-dependent manner; that is, only the largest (> 0.8 m tall) herbivores are excluded from plots. The experiments at Kruger are particularly novel because, to the best of our knowledge, they represent the only replicated research plots in African savanna grasslands in which long-term (>50 yr) fire frequency manipulations have been combined with sized-based herbivore exclusion. These combined manipulations are highly relevant for testing fundamental ecological theory, as well as for management and conservation goals as they mimic contemporary patterns of human-caused land-use change in many African savanna grasslands.

Since the inception of the K-K Experiment, we have quantified using identical methodology herbaceous plant community and ecosystem responses to herbivore exclusion and fire frequency manipulations at both Konza and Kruger. As a consequence, we now have a uniquely comparative perspective on how alterations in these key drivers impact these mesic savanna grasslands. Our experiments have shown dramatic and expected changes in herbaceous plant communities following herbivore removal at Konza, but surprisingly very little change at Kruger. In addition, aboveground net primary productivity (ANPP) – a key ecosystem function and service of savanna grasslands – has not changed with grazing exclusion at Kruger, consistent with the lack of plant community response. In contrast, increases in ANPP at Konza have occurred as the plant community has responded to the exclusion of grazing.