The effect of simulated drought, elevated rainfall and increased temperature on a mesic grassland in South Africa

Climate change models predict warming, more variable rainfall and more frequent extreme events across southern Africa. Eastern South Africa, home to much of the country’s grassland biome, is expected to experience higher and more extreme rainfall and so reduce the extent of the grassland biome by up to 55%, threatening biodiversity and the agricultural economies it supports. Understanding their combined, rather than independent, effects is therefore critical for anticipating ecosystem-level change. In winter 2019, a 6-year factorial rainfall manipulation and warming trial was established in a mesic grassland in KwaZulu-Natal, South Africa. Rainout shelters reduced incoming precipitation by 55%, diverting the water to above-average rainfall plots, while open-topped chambers increased daytime temperature by ~2°C. Biomass tracked early-season rainfall more closely than total rainfall, declined under above-average rainfall, and was largely unaffected by consistent drought. Reduced rainfall did not significantly change species-level composition, but shifted dominance from tussock grasses and erect forbs towards prostrate forbs. Warming had a much larger effect, reducing total and forb species richness and most diversity indices, decreasing cover of dominant grass species, and causing a decline in rare species across multiple life forms. Warming also negated the diversity gains that altered rainfall produced in its absence, showing the two stressors interact rather than act independently. These results suggest that warming poses a greater threat to grassland diversity in mesic grassland in KwaZulu-Natal, South Africa than altered rainfall, particularly for forbs, and that the resulting loss of dominant grass cover could reduce fuel loads, altering the fire regime that normally excludes woody plants from these grasslands.