A test of two mechanisms proposed to optimize grassland aboveground primary productivity in response to grazing

Mesic grasslands have a long evolutionary history of grazing by large herbivores and as a consequence, grassland species have numerous adaptations allowing them to respond favourably to grazing. Although empirical evidence has been equivocal, theory predicts that such adaptations combined with alterations in resources can lead to grazing-induced overcompensation in aboveground net primary production (ANPP; grazed ANPP > ungrazed ANPP) under certain conditions. We tested two specific predictions from theory. First, overcompensation is more likely to occur in annually burned grasslands because limiting nutrients that would be lost with frequent fires are recycled through grazers and stimulate ANPP. Second, overcompensation of biomass lost to grazers is more likely to occur in unburned sites where grazing has the greatest effect on increasing light availability through alterations in canopy structure. We tested these nutrient versus light-based predictions in grazed grasslands that had been annually burned or protected from fire for > 20 years. We assessed responses in ANPP to grazing by large ungulates using both permanent and moveable grazing exclosures (252 exclosures from which biomass was harvested from 3192 quadrats) in a 2-year study. Study sites were located at the Konza Prairie Biological Station (KPBS) in North America and at Kruger National Park (KNP) in South Africa. At KPBS, sites were grazed by North American bison whereas in KNP sites were grazed either by a diverse suite of herbivores (e.g. blue wildebeest, Burchell's zebra, African buffalo) or by a single large ungulate (African buffalo). We found no evidence for overcompensation in either burned or unburned sites, regardless of grazer type. Thus, there was no support for either mechanism leading to overcompensation. Instead, complete compensation of total biomass lost to grazers was the most common response characterizing grazingANPP relationships with, in some cases, undercompensation of grass ANPP being offset by increased ANPP of forbs likely due to competitive release. The capability of these very different grass-dominated systems to maintain ANPP while being grazed has important implications for energy flow, ecosystem function and the trophic dynamics of grasslands.