A determination of Cape flats sand fynbos restoration success in the Western Cape, South Africa, based on vegetation establishment and pollinator activity

Abstract

Invasive plant species as well as altered land-use often lead to the degradation of natural areas by displacing native species or by disturbing ecological processes. The consequent restoration of these degraded areas is often associated with high costs. Pollinator visitation can ensure the long-term success of vegetation restoration by facilitating successful plant reproduction and allowing for genetic variability. A diversity of pollinators can also increase stability and colonization rates of newly established plant communities. The Fynbos vegetation of the Cape Floristic Region (CFR) in South Africa is dominated by small-leaved, ever-green shrubs. The region is characterised by a Mediterranean climate and is highly diverse with a high level of endemism. My study area, Milnerton Racecourse, is a small urban nature reserve (22ha) that consists of critically endangered, Cape Flats Sand Fynbos, a vegetation type that is dependent on fire, but has been severely fragmented in the area where Cape Town is today. Determining the synergistic effect of vegetation restoration treatments, with and without fire, and alternative methods requiring minimal labour (such as topsoil translocation), could assist to identify the most successful and cost-effective treatments to restore native plant communities. Few studies have included pollination as an ecological function to be evaluated as part of active vegetation restoration programmes, even though the long-term survival of plant species is dependent on it. To determine the most successful active restoration methods as well as to understand the effect that the restoration of degraded areas have on pollination, I had two primary aims for the study. Firstly, to determine the success and cost-effectiveness of several vegetation restoration treatment combinations; and secondly, to determine if active vegetation restoration also restores pollination as an ecosystem function. Over a period of two years, I have tested how successful and cost-effective six combinations of vegetation restoration treatments were. Treatments used were Soil-plant, Soil-mulch-sow-plant, Remove-grass-sow-plant, Burn, Burn-sow, and Burn-sow-plant. Topsoil and mulch were sourced and translocated from an intact natural site. Mechanical methods were used for grass removal. Soil-plant, Soil-mulch-sow-plant and Burn-sow-plant led to plant richness resembling near-pristine plots at 30% higher than Burn. Native shrub cover was the highest for treatments which included planting. Survival for species planted was 36 - 41% higher for Burn-sow-plant. Overall, the contribution of sowing was low, with only 9 - 16% of species sowed, being present across all treatments. Topsoil translocation from a near pristine area can be successful, if combined with planting. Other than topsoil translocation, Burn-sow-plant was the most successful treatment at a reasonable cost, despite the addition of planting, which is less frequently used and expensive, while Burn and Burn-sow were less successful, but cheaper. Remove-grass-sow-plant was the least successful and most expensive treatment. In terms of pollination, I compared the recently restored and fragmented site with two reference sites (fragmented and intact). I aimed to determine whether pollination was restored by comparing visitation frequency and species richness of pollinators as well as the seed set of seven plant species. The specialization of the tested plant species was also determined, based on the observed pollinator species. Generalist plant species all produced seed set equivalent to that produced from optimal pollination, whether they were capable of autonomous selfing or not. Specialist plants that were incapable of autonomous selfing were absent from fragmented study sites and even when introduced at these sites, pollination was low or non-existent. While species richness (45-100%) and visitation frequency of pollinators visiting the generalist species Moraea flaccida and Ornithogalum thyrsoides were both higher at the intact site, natural seed set for the species indicated optimal pollination at all sites. Ferraria crispa, a generalist plant species primarily pollinated by Dipterans, had a higher species richness and visitation frequency of pollinators at the restored site (55%) and the fragmented reference site (55%). I identified several factors in addition to fragmentation and vegetation restoration that can affect pollination in natural sites that seem to be complex and different for individual plant species. These factors included surrounding land-use, distance to the urban edge and specialization. My results show that generalist plant species have a greater potential to facilitate ecosystem recovery in terms of pollination. Providing sustained nectar sources, high floral abundance and stepping-stones for specialist pollinators, such as sunbirds in South African Fynbos, might aid their movement to restored habitat fragments. This study shows that a combined approach of vegetation restoration ecology and pollination biology, with a plant species-specific approach to accommodate pollinator species should be used when planning active restoration.