Savannas occur globally across a perplexingly broad range of environmental gradients, and tree cover within can be dynamic. Fire and ungulate herbivory by browsing and grazing animals are known to exert strong influences on key demographic transitions from juvenile to adult stage in savanna trees. Interactions between these processes, however, have largely been overlooked. For example both browsers and grazers selectively forage in burned areas due to increased palatability of re-sprouting vegetation, but how herbivore community composition affects re-sprouting juvenile trees size following fire is unknown. We investigated post-fire size responses of a widespread monodominant East African tree Acacia drepanolobium. We hypothesized that the post-fire re-sprout size of trees top-killed by fire would depend on 1) the community composition of the ungulate herbivores and 2) the presence and identity of obligate acacia-ant mutualists known to confer varying levels of defense against browsers. Using the Kenya Long Term Exclosure Experiment (KLEE), we compared post-fire tree responses among six herbivore treatments that included factorial combinations of cattle, meso-wildlife species (large ungulates 15–1000 kg), and mega-herbivores (elephants and giraffes). We assessed height (an indicator of a tree’s likelihood of surviving future fires) and stem length (a proxy for biomass, irrespective of height).
Results/Conclusions
Trees top-killed by fire were subsequently browsed by cattle, meso-wildlife, and mega-herbivores, all of which reduced re-sprout height. Cattle had a small negative effect on re-sprout height, but only in the absence of wildlife. Meso-wildlife significantly reduced re-sprout heights, and mega-herbivores further reduced heights. Height was also related to the presence and identity of four species of tree obligate ant mutualists. Tree heights were significantly lower where ants were absent and for trees inhabited by ant species predetermined to have relatively low effectiveness in deterring herbivory. Total stem length (a proxy for biomass) was reduced by both meso-wildlife and mega-herbivores. Despite their shorter post-fire height, trees exposed to mega-herbivores had the greatest stem lengths (i.e., more short-statured horizontal branch tissue), suggesting that trees exposed to elephant browsing prior to fire had larger pools of starch reserves (i.e. re-sprouting ability) available per unit above ground biomass. Our results suggest that wild ungulates reduce the potential for re-sprouting trees to escape from the “fire trap” created by trees growing in the flammable grass layer, but that ant mutualists may assist in this escape.