D.Phil. / Over a century gold exploration and extraction on the central Witwatersrand has left a legacy of mine residue deposits. Although there have been numerous complaints and claims of health effects associated with dust from these residue deposits, mostly these have been dismissed as mere nuisance. This study hypothesized that gold mine tailings on the Witwatersrand pose significant health risks to the nearby communities due to respirable airborne tailings material. Preliminary review of information on the tailings led to the insight that tailings source material and atmospheric properties at receptor sites are poorly characterized from the point of view of particle size-mass distributions. For years, routine monitoring of emissions from tailings storage facilities has been limited to sampling of dust fallout (settlable particles ≥ 30 μm). A suite of fifty four source samples (thirty-six slime, six newer slime and twelve sand) were collected from tailings storage facilities along the mining corridor covering deposits in the Carltonville area through to Springs. Size class characterization of source material was performed in the diameter range 0.05 μm to 900 μm using a Malvern® MS-14 Particle Analyser with 64 channels, from which the respirable (dp< 5 μm) and thoracic (dp < 10 μm) components were measured. Secondly, source materials were sieved (using a sieving cloth) in dp < 5 and < 10 μm fractions and the data derived was used in validating the size class results from the Malvern® MS-14 Particle Analyser before subjecting the samples to chemical analyses (elemental, mineralogical and radioactivity). Two years of continuous dust fall samples were collected in two different residential areas close to selected tailings storage facilities and subjected to similar size class characterization with the Malvern® MS-14 Particle Analyser. Further, continuous size-mass characterization of airborne aerosols was conducted in two different locations in the vicinity of tailings storage facilities with the Grimm® aerosol particle counter, in the diameter range 0.25 μm to 32 μm. The Grimm® aerosol particle counter was collocated with the MicroVol® PM10 filter sampler ambient monitor. The risk zones around selected tailings storage facilities were established by conducting dispersion modeling with American Meteorological Society/Environmental Protection Agency Regulatory Model (AERMOD). The spatial evolution of tailings storage facilities and built-up areas is assessed using multiple aerial photographic images, covering four periods from 1952 to 2008. Overall, detailed spatial analysis was conducted and human inhabitants are now at greater risk than before. We have quantified the number of inhabitants living within the high-risk zone surrounding selected tailings storage facilities. The combined areas occupied by tailings storage facilities have increased from ~4 km2 (1952) to 10 km2 (2002); and residential areas ~4 km2 (1952) to 27 km2 (2002) within a buffer of 2 km. Population grew between 1952 and 2002 by a factor x29 at Crown Gold Recoveries; x43 at Durban Roodepoort Deep (from 1976); and x25 at East Rand Proprietary Mines.

Msc. / This study focuses on the functional evaluation of a gold Tailings Storage Facility (TSF), the Fleurhof TSF (designated slimes dam 2L3), which is representative of gold slimes (tailings) dams in the Witwatersrand. The aim is to evaluate the rehabilitation status (in terms of species composition and landscape function) of TSF surfaces that were planted ~20 to ~40 years ago using the most common rehabilitation practice, namely pasture grassing. Pasture grassing aims to obtain rapid dust control. Mine closure was not considered at the time, and the vegetation was hence not designed with ecological processes or long-term objectives in mind. However, over time, grassing practices became routine in the TSF vegetation industry, and the contribution that they have made towards surface rehabilitation needs to be assessed against the current site closure objectives. On the selected site, evaluations of vegetation cover, species composition, and substrate physical and chemical properties were conducted in 1975 (Thatcher, 1979). The aim of the current mid-summer study in 2009 was addressed by assessing the same parameters of vegetation cover, species composition, and soil physical and chemical properties, and comparing these to the results of the previous assessment in 1975. Landscape Function Analysis (LFA – an index-based method of evaluating the retention of resources by surfaces and soil condition), was applied to determine the landscape function (stability, infiltration and nutrient cycling that are derived from the eleven soil surface assessment indicators), and compared with the landscape function of natural analogues in the same locality. Landscape organisation (also an indicator of LFA) was used to investigate the contribution of vegetation overall, and of discrete plant functional groups to landscape functionality. A total of 35 plant species were observed in the quadrats (56 species on the TSF overall), with 51% of species overall being alien. Perennial vegetation and cryptogam covers contributed the most to stabilising the TSF slopes and flat (top) surfaces. The percentage of vegetation basal cover on the TSF (30% – 70%) was lower than for analogue grassland sites on a nearby hill (80% – 100%). Lower basal vegetation cover on the TSF than on the analogue site, were associated with lower LFA indices of stability on all TSF slope aspects (P = 8.19 x 10-13 for all aspects), iii lower nutrient cycling for the northern and southern TSF slopes (south aspect P = 4.20 x 10-6; top aspect P = 0.03), and lower infiltration on the southern slope of the TSF (P = 2.68 x 10-6). However, a higher infiltration on the northern TSF aspect (P = 0.02) than on the analogue site, was attained. Perennial tussock grass patch types (which included perennial forbs) made the largest contribution to the soil surface assessment (SSA) indicators on the TSF and analogue sites, and therefore to the LFA indices that were derived from these SSA indicators. The contribution of woody patches to function was not evaluated.