Extreme droughts are predicted to occur more frequently under climate change. A recent drought in South Africa – the worst on record in some regions – provided an opportunity to test conflicting hypotheses about savanna vegetation responses to such events. Trees in savannas are limited not only by rainfall, but also by grass competition, fire, and herbivory, complicating predictions of drought responses. Savanna trees could suffer greater mortality and reduced growth under drought, due to physiological stress or to impacts of herbivores diet-switching from grass to trees. Alternatively, drought could present opportunities for tree growth and recruitment, as grass competition and fire frequency and intensity are drastically reduced after dry years. Here, we measured individual- and community-level responses of trees to the 2015-2016 drought in Kruger National Park, South Africa across existing long-term fire and herbivory experiments. We re-sampled tree community structure during and after drought across various fire treatments of the Kruger Experimental Burn Plots, in two regions differing in drought severity, to examine drought and fire interactions. We also monitored dendrometers on trees inside and outside of experimental herbivore exclosures and compared with baseline data from previous years (2006-2013) to examine drought and herbivory impacts on individual growth and mortality.
Results/Conclusions
Where drought was severe, tree densities in the Experimental Burn Plots decreased in all size classes within six months of the 2016 drought, and had not recovered one year later across most plots, with no clear correlation to fire treatment. By contrast, where drought was milder, tree densities held steady and then increased over the subsequent year, with increases greatest in unburned control plots. Monitoring of individual trees in an herbivore exclosure experiment also revealed dramatic responses to severe drought. Individual tree mortality following drought was extremely high relative to previous years (20-46% vs. typical 0.4-0.8% annual mortality), and average monthly diameter growth declined overall. Mortality was highest in partial exclosures (only elephants and giraffes excluded). Moreover, while previous work here found the presence of herbivores facilitated greater tree diameter growth (perhaps by reducing tree-grass competition), tree growth post-drought was instead consistent across herbivore treatments, suggesting that changes in herbivore abundance or feeding behavior had negative consequences for trees. Contrary to expectations that drought could present a recruitment opportunity for savanna trees released from grass competition and fire, our results suggest that drought strongly limits savanna trees, both directly and via altered herbivory patterns.