Use of remote sensing and GIS in a risk assessment of gold and uranium mine residue deposits and identification of vulnerable land use

• Main Author: Sutton, Malcolm William
• Format: Others
• Language: en
• Published: 2013
• Subjects: Gold mines and mining - South Africa - Witwatersrand.; Uranium mines and mining - South Africa - Witwatersrand.; Acid mine drainage.; Mineral industries - Environmental aspects - South Africa.; Pollution - South Africa - Witwatersrand.
• Online Access: http://hdl.handle.net/10539/12692

A research report submitted to the Faculty of Science, University of the Witwatersrand, Johannesburg, in partial fulfilment of the requirements for the degree of Master of Science in Environmental Science Johannesburg, November 2012. === Acid rock drainage (ARD) and dust are potential consequences of gold and uranium mine residue deposits (MRDs) on the Witwatersrand basin. Urbanisation has taken place around mines and, with the curtailing of mining activities and clearing of land previously covered by MRDs, there is pressure to use this land for residential, industrial and agricultural purposes. However, mining companies historically were not required to provide pollution control measures and there is evidence for contamination of land and water. Thus, there is a need to prioritise contamination sources for mitigation and to understand the extent of contamination and potential risks associated with different categories of land-use on mining land. The aim of my study was to conduct a first-order risk assessment to aid in identifying vulnerable land use in the vicinity of gold and uranium mining, and prioritising MRDs, including footprints, for mitigation. To achieve this I constructed a Geographical Information System (GIS) using publicly available spatial data, and then tested the usefulness of historical aerial photographs and remote sensing imagery for mapping MRDs and impacts of MRD origin under Highveld conditions (i.e. a seasonal climate with summer rainfall and annual evapotranspiration of >2.5 times mean annual precipitation). The Ekurhuleni Metropolitan Municipality (EMM; 1923 km2) is an area of extensive historical mining with major urbanisation, while retaining areas for agricultural land use; thus it was selected as a representative study site. I used a numerical rating scheme, which combined a number of parameters in two separate stages to calculate a risk index. The first stage involved the classification of hazards associated with MRDs while the second involved an assessment of land use vulnerability based on exposure pathways and proximity. Historical aerial photographs (1938, 1964 and 2003) and the Chamber of Mines (CoM) Dump Indexes were used to identify and classify MRDs in terms of basic geotechnical properties, current status and historical failure. Multi-spectral data, acquired over two years (2002 and 2003) in two seasons (spring and summer) by the TERRA satellite’s Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) sensor, were used to compile thematic images, indicating potential contamination of surrounding land. It was intended that a zone of influence could be distinguished for each MRD enabling me to rate the hazard severity. The thematic images I selected included primary minerals (pyrophyllite and chlorite), secondary minerals (copiapite and jarosite), an indicator of uranium-bearing ore (referred to as mincrust) and the normalised difference vegetation index (NDVI). These minerals were chosen as potential indicators of different transport routes of contaminants and I tested their associations with different features and land use. I also tested for seasonal differences in the detection of these minerals, and used NDVI to examine the masking effect of active vegetation. I found GIS to be well suited for combining the various forms of spatial data and providing information about MRDs, aqueous pathways, proximity to vulnerable land uses and impacted areas. However, I found that the potential severity of the hazards posed by each MRD, as indicated by a zone of influence, could not be determined from aerial photographs and ASTER alone. I therefore utilised the findings expressed in the literature survey to assign ratings for the different classes of MRDs. The vulnerability assessment was also supplemented by literature review to rate land uses based on human exposure pathways. I determined that MRDs (including footprints) cover 4.1% of EMM, with slimes dams, totalling 3.5%, occupying the majority of this area. I found that 64% of slimes dams had failed prior to 2003 and I plotted a further 0.6% of EMM covered by visible mine residue spillage. Fifty three percent of MRDs were situated within 100 m of drainage lines or old wetlands, while 52% of these (i.e. 27% of the total) had been constructed in the watercourse. I also found that 15% were constructed on dolomites. Informal settlements were located on or bordering 6% of MRDs, with 41% of MRDs within 1 000 m. Eighty eight percent of MRDs were found within 1 000 m of formal residential areas, 71% within 500 m, and formal settlements were located on or bordering 5% of MRDs. Twenty three percent of MRDs were located within 500 m of agricultural land, while 35% were within 1 000 m; and industrial land use was on 9% of MRDs (footprints), with 40% of MRDs being within 500 m of industrial areas and 61% within 1 000 m I found that chlorite did not provide a ‘signature’ of gold and uranium mine residue, whereas the other four minerals did. I also found that, of the two seasons examined (spring and summer), the best time to take an ASTER image to detect mineral signatures of gold and uranium mine contamination is after a few dry days following the first spring rains. For this reason, I used the ASTER taken in late October (spring) 2003 to examine associations with pathways and land use. I found more pyrophyllite and copiapite on industrial and business land use than background, which I suggest is associated with the settling of windborne dust on large and flat roofs; although, in the case of copiapite this could be related to the oxidation of settled wind blown pyrite material. I found jarosite to be a reliable indicator of mine residue, which, together with mincrust, helped me identify contamination in former agricultural holdings, which are now a township. Although, chemically undefined, mincrust was a useful indicator of contamination, as I found it to be reliably detected on MRDs (including footprints), mine residue spillage, wetlands and other contaminated sites, and absent from known uncontaminated sites. Furthermore, it was not necessarily masked by active vegetation, whereas copiapite, jarosite and pyrophyllite were. Mincrust was also detected on irrigated agricultural land with an odds ratio of between 10 to 36 times greater than for rain-fed. Consequently, the most likely pathway for mincrust is the aqueous. The mincrust signature, together with historical aerial photographs, also assisted me to identify historical mining along Black Reef outcrops, through detection in a wetland upstream of known mining activities. The culmination of my study was a risk class and index for MRDs from which ‘risk maps’ were produced. These maps provide a guide to the level of risk posed by each MRD to the surrounding land use. Of the total 287 MRDs (including footprints) identified in the EMM, 50% were classified lower-risk; 40% medium-risk; 10% higher-risk and 0% as much higher risk. The lower-risk MRDs were predominantly rock dumps, whereas the higher-risk MRDs were slimes dams. The findings from my study will contribute to meaningful recommendations for future land use and enable mining companies, landowners, developers and government to allocate their resources judiciously (i.e. appropriate to the level of risk). The results of this study have been published as: Sutton, M.W., Weiersbye, I.M., Galpin, J.S and Heller, D., 2006. A GIS-based history of gold mine residue deposits and risk assessment of post-mining land uses on the Witwatersrand Basin, South Africa. In: A. B. Fourie and M. Tibbett (eds.), Mine Closure 2006: Proceedings of the 1st International Seminar on Mine Closure, Perth, ISBN: 0-9756756-6-4, pp. 667–678 (Appendix I). Sutton, M.W. and Weiersbye, I.M., 2007. South African legislation pertinent to gold mine closure and residual risk. In: A.B. Fourie, M. Tibbett and J. Wiertz (eds.), Mine Closure 2007: Proceedings of the 2nd International Seminar on Mine Closure, Santiago, ISBN: 978-0-9804185-0-7, pp. 89–102 (Appendix II). Sutton, M.W. and Weiersbye, I.M., 2008. Land use after mine closure – Risk assessment of gold and uranium mine residue deposits on the eastern Witwatersrand, South Africa. In: A.B. Fourie, M. Tibbett, I.M. Weiersbye and P.J. Dye (eds.), Mine Closure 2008: Proceedings of the 3rd International Seminar on Mine Closure, Johannesburg, ISBN: 978-0-9804185-6-9, pp. 363–374 (Appendix III).