Current and future vulnerability of South African ecosystems to perennial grass invasion under global change scenarios

Invasive alien grasses are a group of increasing importance which respond to global changes and alter native plant community structure and ecosystem processes. To date there have been few studies that assess the ecological drivers of emerging alien grass invasions and their impacts on arid and semi-arid ecosystems of South Africa. This thesis is focused on the early stages of an invasion process, using fountain grass, Pennisetum setaceum as a model system to understand ecological factors promoting its current spread in South Africa. This is an invasive alien perennial C4 grass from North Africa and the Middle East. The study also investigates the interaction between its invasion and ongoing climate change, and its impact on native ecosystems. P. setaceum spreads along the edges of roads on the outskirts of most towns, and is common on mine dumps in many areas throughout South Africa. Occasionally, it escapes into natural vegetation along drainage lines or after fires. This grass can be a costly problem for agriculture and biodiversity conservation as it is unpalatable and increases fire risk. In this study I assessed how P. setaceum overcomes different invasion barriers in South Africa as an emerging invader. First, I determined the current distribution and habitat preferences of this grass in arid and semi-arid parts of South Africa. I show that the grass performs better on road-river interchanges, although it is found in other parts of the landscape. I suggest that these interchanges should be targeted for management and control of this species. Second, I assessed environmental resources and habitat conditions that promote its invasive potential, and found that nutrient addition and extra soil moisture promoted performance of established grass seedlings. I suggest that management and control should focus on areas with high nutrients and extra water, as these areas facilitate growth and reproduction. Third, I then determined the influence of habitat and competition on its performance along a rainfall seasonality gradient. I demonstrated that established P. setaceum seedlings do not perform well under competition from native vegetation, an aspect which was similar across this gradient. From this finding, I suggest maintaining a high indigenous cover along road verges. Fourth, I assessed differences in reproductive output and potential along this rainfall seasonality gradient, and found no differences between populations at different localities. This suggests high local adaptation of this grass especially in areas of low resources where it persists until favourable conditions return. Predicting future distribution of this grass is important for its management. Fifth, I therefore determined its potential future distribution based on current distribution and spread dynamics using climate-matching and dynamic probabilistic spread models. This predicted most biomes to be at increasing risk of invasion by this grass, except the succulent karoo, in 2050 and 2100. I found that disturbance is a major factor promoting its invasion into semi-natural areas away from roadsides. I finally determined the response of an arid ecosystem to P. setaceum invasion and fire promotion. I show that this ecosystem function and structure will be adversely altered by P. setaceum invasion. My work has emphasised the need for early detection and rapid response of emerging invasive grass species in South Africa, and made recommendations for future research.