Quantifying land-based sources of plastic pollution in South Africa

It is widely claimed that at least 80% of the plastic litter entering marine environments comes from land-based sources, yet there is little empirical evidence to support this. Most studies to date predict the flux of litter from land to sea using global models based on a handful of field studies conducted mostly in developed, Northern Hemisphere countries; others use proxies such as per capita waste generation and proportion of mismanaged waste to predict litter loads entering the sea. It thus remains unclear how accurate these predictions are, particularly for African countries where few field studies have been conducted. Due to its rapidly growing human population, Africa is expected to become a much more significant source of litter into the sea. It is therefore important to identify major land-based sources of litter from the continent so as to implement effective mitigation strategies aimed at stopping this leakage of waste plastics into the marine environment. In this thesis I investigate two possible land-based sources of plastic pollution in South Africa. In Chapter 2 I show that stormwater run-off from Cape Town, a large coastal city, is a significant land-based source of litter into Table Bay. By placing nets over three stormwater outlets, each draining a different land-use type in the city, I estimate that some 70-630 tonnes of plastic litter are released as part of urban runoff from Cape Town each year which is a similar order of magnitude as estimates based on run-off litter collected in the 1990s and stranded beach litter. Overall, 40- 78% of litter items by count and 52-64% by mass was plastic, of which most was single-use packaging. Compared to a similar survey of the same three stormwater outlets conducted in 1996, litter densities by count decreased by 50% in two of the three catchments but increased threefold in the commercial/residential area. It is clear that urban run-off is a significant source of litter into Table Bay, but the total amount is considerably less than model predictions that identified South Africa as the 11th worst source of marine plastics from land-based sources globally. While several of South Africa's large cities are on the coast, the largest industrial centre and most densely populated part of the country is inland in the Gauteng Province. Much of this province lies within the catchment of the Orange-Vaal River system, which flows west across the country to eventually drain into the Atlantic Ocean. In Chapter 3 I present the results of two extensive field surveys to assess the amounts of plastic debris transported by the Orange-Vaal River system into the sea. By sampling for micro- and macro-plastics at 33 bridges spanning the lengths of both the Orange and Vaal rivers I show that densities of larger plastic items were highest at sites in the densely populated and highly urbanised upper reaches of the Vaal River, while microfibres were particularly abundant in the arid, sparsely populated lower reaches of the Orange River at the end of the wet season. It therefore appears that the Orange-Vaal River system may be a source of microfibres to the Atlantic Ocean, but most larger plastic items are retained near point sources in the upper reaches of the Vaal River. The Orange-Vaal River system thus does not appear to be a major source of plastics into the Atlantic Ocean and there is considerably less than the 0.095 tonnes·yr-1 predicted to be washed out of the Orange River by global models. I also investigated whether dams retain microplastics and microfibres within their reservoirs by collecting bulk water and neuston net samples from above and below the walls of the five major dams in the Orange-Vaal River system. I present these results in Chapter 4 and show that microplastic and microfibre concentrations were highest in dams on the densely populated Vaal River during dry conditions, whereas the opposite pattern occurred at dams farther downstream or on the sparsely populated Orange River during wet conditions. Overall, microplastic and microfibre densities were similar at sites collected above versus below dam walls and there was no significant correlation between microplastic and microfibre densities at a site and the distance from the site to the dam wall. Dams therefore do not appear to trap floating microplastics and microfibres, although the data were noisy (average CV = 184%) and so provide only a rough estimate of differences in plastic densities among sites. Lastly, in Chapter 5 I summarise the main results from the previous chapters and present recommendations for future research. Combining the findings from Chapters 2 and 3 I make a first, very crude estimate of the amount of plastic entering the sea from land-based sources in South Africa and compare this to the 0.09-0.25 million tonnes predicted by Jambeck et al. (2015) based on national estimates of waste production and management. I conclude that actual plastic emissions are some 1-2 orders of magnitude less than the global model estimate, which concurs with estimates based on beach litter data.