versão On-line ISSN 2411-9717
versão impressa ISSN 0038-223X
J. S. Afr. Inst. Min. Metall. vol.115 no.2 Johannesburg Fev. 2015
Thermophysical properties of rocks from the Bushveld Complex
School of Geosciences, University of the Witwatersrand, Johannesburg
This paper presents a compilation of physical properties of rocks from the Bushveld Complex. The database consists of more than 900 measurements each of thermal conductivity and density. The data are well distributed from localities around the Complex and most rock types are well represented. Thermal conductivity and density are shown to vary widely in the ranges 1.8-4.2 W m-1K-1and 2600-4200 kg m-3 respectively. Although only 190 heat capacity measurements are available, this parameter is quite uniform for most rock types present, 800-900 J kg-1K-1, except for chromitite, which has a lower average, 750 J kg-1 K-1. Rocks encountered in deep platinum mines are particularly well characterized and this has important implications for prediction of mine refrigeration requirements. The heat flux into a semi-infinite region with properties typical of the Bushveld Complex as a function of time is substantially lower than an equivalent model for the Witwatersrand Basin.
Keywords: thermophysical rock properties, platinum mines, mine cooling.
Geothermal research provides insights into a wide variety of geological problems, including tectonic studies (Jones, 1988, 1992, 1998), maturation of oil and natural gas (Royden et al., 1980), exploration for geothermal energy (e.g. Martinelli et al., 1995), and even climatology (e.g. Jones et al., 1999). In South Africa, the most important application is in mine engineering. South Africa leads the world in deep mining, and knowledge of virgin rock temperatures, geothermal gradients, and thermal properties of rocks is essential for planning refrigeration and ventilation requirements of deep mines (Jones, 1988; Jones and Rawlins, 2002; Rawlins et al., 2002; Jones, 2003a, 2003b).
One of the most important challenges of mining is to control the environmental conditions, particularly temperature, of deep underground workings. The primary cause of elevated temperatures is heat derived from exposed rock surfaces, whose temperature is determined by the natural increase of virgin rock temperature (VRT) with depth (Rawlins et al., 2002; Jones 2003a). Knowledge of VRT is therefore fundamental for calculating refrigeration requirements. Knowledge of rock thermal properties is also fundamental for refrigeration studies because (1) they determine the rate at which heat flows from the rock into underground excavations, and (2) they control the thermal gradient in strata overlying and underlying these excavations and hence their VRT (Jones, 2003a, 2003b).
Because of its long history of gold mining, which has progressively approached greater depths, the Witwatersrand Basin has been the subject of geothermal research for more than 75 years. Measurements of rock temperature and thermal conductivity dating back to 1938 (Weiss 1938; Bullard, 1939; Krige, 1939) provided the foundation for later studies and eventually an intensified collaborative research programme (from about 1980-2000) between the University of the Witwatersrand (Wits) and the Chamber of Mines Research Organisation (COMRO), and later, CSIR Mining Technology. The net product was a database of VRT, thermal gradient, and rock properties that probably surpasses any in the world (Jones, 1988, 2003a, 2003b).
In the latter period, platinum mining depths in the Bushveld Complex (to the north of the Witwatersrand Basin) increased significantly. Earlier measurements at two localities showed that the geothermal gradient in the Bushveld Complex is substantially higher than that in the Witwatersrand Basin (Carte and van Rooyen, 1969), and it was obvious that there was a need for a detailed geothermal investigation of the Complex. Subsequent research collaboration between Wits, COMRO, and other key stakeholders from industry ensued. New measurements of the geothermal gradient were made in many of the existing and potential platinum mining areas, and these measurements confirmed that the thermal gradients are high (generally 20-25 K km-1) (M.Q.W. Jones, unpublished data, 2014). These data will be presented for publication elsewhere.
During the past 30 years, the number of measurements of thermal conductivity, density, and heat capacity on Bushveld rocks gradually grew and the database currently rivals that from the Witwatersrand Basin in size and coverage of rock types. It is the purpose of this paper to present a summary of the Bushveld database, characterize the thermal properties of various rock units constituting the Bushveld Complex, and discuss the implications for deep mining.
The Bushveld Complex is an enormous igneous province occupying an area of approximately 65 000 km2 to the north, east, and west of Pretoria (Figure 1). The Complex hosts the world's largest reserves of platinum, chromium, and vanadium, and consequently there is a vast literature. This brief summary is largely based on reviews provided by SACS (1980) and Cawthorn et al. (2007). For ease of reference, Table I contains a glossary of major Bushveld rock types discussed below.
The Bushveld Complex consists of coarse-grained igneous rocks with a wide range of composition from ultramafic to felsic. It is temporally, and possibly genetically, related to volcanic rocks of the Rooiberg Group (Cawthorn et al., 2007) (Figure 1). However, this paper deals solely with the plutonic rocks that were intruded into the Transvaal Supergroup and older rocks 2060 Ma ago. The Complex is formally classified into three distinct units: the ultramafic to mafic Rustenburg Layered Suite and the felsic Rashoop Granophyre and Lebowa Granite Suites (Figure 2).
The average thickness of the Rustenburg Layered Suite is approximately 6 km, and it crops out in four main 'limbs'. The geology of the eastern limb (northwest to southwest of Burgersfort) and the western limb (Thabazimbi to Pretoria) is reasonably well known, whereas large parts of the southern (or Bethal) limb and northern (or Potgietersrus/Mokopane) limb are covered by younger Karoo and Waterberg strata. Different rock units can be traced for hundreds of kilometres and, although the succession is seldom complete, geophysical evidence suggests that at least the eastern and western limbs are connected at depth (Cawthorn et al., 1998).
The Rustenburg Layered Suite consists of five main 'zones', which is the most convenient classification for discussing similar rocks from similar levels (Figure 2). Although specific rock types are listed below and in Table I, it should be noted that there is often a gradational change from one rock type to another. For example, there can be an almost continuous variation from anorthosite to pyroxenite and, as will be seen, this results in a gradation of thermal properties.
The lowermost 'Marginal Zone' consists predominantly of uniform, relatively fine-grained norite with minor amounts of pyroxenite. The overlying 'Lower Zone' is ultramafic in composition and dominated by pyroxenite, but includes thick layers of harzburgite with some dunite.
From a mining point of view, the 'Critical Zone' is the most important because it is the host of chromium-rich and platiniferous ores. The lower part of the Critical Zone is essentially pyroxenite with some olivine-bearing rocks, whereas the upper section is represented by cyclic layers involving pyroxenite, norite, and anorthosite. Numerous chromitite layers may be well developed in the Critical Zone, and these are divided into the Lower Group (LG1-7), Middle Group (MG1-4), and Upper Group (UG1-3). The LG6 and MG1 layers have been extensively mined for chromium. The main sources of platinum and associated platinum group elements (PGEs) are the UG2 chromitite layer and pyroxenites within the Merensky Reef (Figure 2).
The 'Main Zone' is a thick uniform sequence of norite and gabbronorite with a few layers of anorthosite and pyroxenite. It is overlain by the 'Upper Zone' in which magnetite appears as an accessory mineral in many rocks, which are predominantly anorthosite, gabbronorite, magnetite gabbro, and olivine diorite towards the top. Magnetite may accumulate in up to 24 layers, some of which exceed 2 m in thickness. Vanadium is associated with all the magnetitite layers and can reach concentrations of up to 2%. The Main Magnetite Layer in the lower part of the Upper Zone (Figure 2) is mined for vanadium.
The last stage of Bushveld magmatism involved the emplacement of large volumes of granophyric and granitic rocks of the Rashoop Granophyre Suite and the Lebowa Granite Suite, which occupy the interior of the Complex (Figure 1). Although they are often referred to collectively as 'Bushveld granite', there are significant variations in the proportions of major minerals (quartz and feldspar) as well as the content of minor mafic phases (mainly hornblende, biotite, and pyroxene). This results in subtle variations of thermal properties.
Measurement of thermophysical properties
The methods used to measure the thermophysical properties are essentially the same as those reported in this journal by Jones 10 years ago (Jones, 2003b). The reader is referred to that paper for details, and only the most salient points will be repeated here. The measured parameters are thermal conductivity (K, units W m-1 K-1), density (p,kg m-3), and heat capacity (C, J kg-1K-1). A fourth parameter, thermal diffusivity (k,m2 s-1) is calculated from the relation K=K/Cp. All measurements reported here were made in the heat flow laboratory at Wits.
Thermal conductivity measurements were made using a divided bar apparatus that is specifically designed for precise analysis of competent rocks such as those encountered in the Bushveld Complex. This is a steady-state method in which the conductivity of a rock sample is measured relative to that of a material of known conductivity (the 'substandard'). Precisely machined rock discs (30 or 38 mm in diameter and approximately 20 mm in length) and discs of the substandard are placed in the divided bar. Heat is supplied and abstracted at either end of the divided bar using temperature-controlled water circulators. When a condition of linear heat flow along the stack of discs is established, the conductivity of the sample disc relative to that of the substandard is determined by measuring the temperature gradient across each disc and applying Fourier's Law of heat conduction, q=KdT/dx, where q is heat flux (W m-2) and dT/dx is thermal gradient (K m-1). The contact resistance between discs constituting the divided bar is minimized by applying an axial pressure of 5 MPa, and radial heat loss is minimized by carefully insulating the stack. All samples were saturated with water prior to measurement and average sample temperatures were close to 25°C in all experiments. The thermal conductivity of the substandard was calibrated against gem-quality quartz cored perpendicular to the c-crystallographic axis, for which an international standard conductivity was used. The overall error in determining conductivity, including reproducibility and calibration errors, using this method is estimated to be less than 0.1 W m-1 K-1.
Most densities were determined from accurately measured masses (using a precision balance) and volumes (using a micrometer and digital vernier) of samples prepared for thermal conductivity measurement. Measurements, first dry and then saturated with water, on 30 samples across the spectrum of rocks analysed showed no significant difference (<5 kg m-3on average); this is to be expected because the porosity of the rocks is essentially zero. Repeat measurements on the same 30 specimens also yielded differences averaging at approximately 5 kg m-3, and the maximum uncertainty in measuring density is estimated at 10 kg m-3. The densities of damaged conductivity samples and other irregular specimens were determined from their measured masses while suspended in air and water and applying Archimedes' Principle. Measurements were made on 10 control samples using both methods and the average difference was found to be less than the above overall uncertainty.
Heat capacities of samples, crushed and sieved to sand-sized grains, were determined by calorimetry using the method of mixtures. The temperature of measurement in all experiments was as close as possible to ambient temperature, but in most cases a small correction for Newtonian cooling was necessary. The uncertainty in heat capacity measurement, based on repeat measurements on 10 samples, is estimated at approximately 25 J kg-1 K-1.
During the past 30 years, routine thermal conductivity measurements have been made for both mine engineering applications and for tectonic studies of the Bushveld Complex. Initially, density measurements reported here were made primarily for mine refrigeration studies, but these were supplemented during the past year with measurements made on specimens originally prepared for conductivity analysis. The total number of measured values for each parameter exceeds 900.
The distribution of boreholes and mines from which samples were derived is indicated in Figure 1. Average thermal conductivities and densities of different rock types in different stratigraphic units of the Bushveld Complex are listed in Table II. It is clear that the database is representative of the Complex as a whole and that all important rock types constituting the Complex are well characterized. Conductivity and density variations are discussed in the next section. Additional detailed density measurements are available from boreholes and mines in the Complex (Cawthorn and Spies, 2003; Ashwal et al., 2005; Davis et al., 2007).
Heat capacity measurements were made specifically for mine refrigeration investigations. The data are summarized in Table III. Although only 190 values are available, most rock types encountered in the deep platinum mines are reasonably well represented. Fortunately, heat capacity is a relatively uniform rock parameter, and the data in Table III suffice for most purposes. Heat capacity data are discussed in the last section of this paper.
Thermal conductivity and density
Variations of thermal conductivity and density of rocks within different stratigraphic units of the Bushveld Complex are illustrated in the form of histograms and plots of conductivity versus density (Figures 3-11). Histograms for different rock types, identified from hand specimens, are all drawn to the same horizontal and vertical scales so that the results can be compared directly. Conductivity and density of rocks are essentially controlled by their constituent minerals (Table I), which have distinct thermal properties (Table IV). As noted previously, the distinction between different rock types is not always clear-cut, and there is continuous gradation from one rock type to another. This is particularly relevant in the Bushveld Complex, where rock density variations may reveal the relative proportions of the major rock-forming minerals (Cawthorn and Spies, 2003; Davis et al., 2007). It was not possible to generate conductivity-density plots with the same horizontal (density) scale, so data for the Critical Zone and Main Zone (Figure 6a) are reproduced in all such diagrams (Figures 4, 7, 9, and 11) to facilitate comparison.
Rocks sampled from the Lower Zone are all ultramafic, consisting essentially of pyroxene with at least small amounts of olivine. Because both of these minerals have a high thermal conductivity and density (Table IV), the respective values for the rocks are the highest in the Complex (Table II, Figures 3 and 4), with the exception of chromitite from the Critical Zone and magnetitite from the Upper Zone.
Critical Zone and Main Zone
Thermal characterization of the Critical Zone and Main Zone (particularly its lower part) is crucial for investigations of refrigeration requirements of platinum mines and potential deep-level chromium mines. These zones are the best studied and are discussed together here for convenience. The most important rocks vary in composition from anorthosite (almost pure plagioclase feldspar) through norite and gabbro to pyroxenite (almost pure pyroxene) (Table I). Table III, the histograms (Figure 5), and the conductivity-density plot (Figure 6) show the effect of the increasing pyroxene content. Figure 6a shows that the gradation of thermal properties is not quite complete, as reflected in the paucity of data at approximately 3000 kg m-3. Nor is the relationship between the parameters purely linear. Average conductivities for successive 50 kg m-3 density intervals suggest that a relationship is better defined by two lines, one dominated by plagioclase and the other by pyroxene (Figure 6b). This may be understood by the fact that density is controlled simply by the percentages of plagioclase and pyroxene, whereas conductivity is controlled by more or less conductive paths associated with connectivity of these minerals, as well as their relative abundance. Measuring density is quick and easy, and the correlation between conductivity and density is potentially useful for estimating conductivity where measurements of this parameter are not available.
Results for chromitite (Table II, Figure 7) are shifted substantially toward much higher density. All data are from the UG2 Reef. The data adequately characterize this important platinum-bearing horizon and serve as best estimates for Lower and Middle Group layers exploited for chromium. Few samples are known to be derived from the Merensky Reef, but this layer consists essentially of pyroxenite, and the data from the Critical and Main Zones in Table II are appropriate for mine engineering purposes.
Thermal characterization of the Upper Zone is complicated by the presence of significant amounts of magnetite in many of the rocks. Most samples are norite and/or gabbro, with fewer samples of anorthosite and pyroxenite, and all data are plotted in the same histograms, the only distinction being between rocks that are apparently depleted in magnetite (Figure 8a and Figure 8c) and those that are obviously magnetite-bearing (Figure 8b and Figure 8d). Approximately half of the results fall close to the trend defined by anorthosites-pyroxenites from the Critical and Main Zones (Figure 6), but the other half suggests a different trend (Figure 9a). The latter samples are all magnetite-bearing and are displaced to higher densities as indicated by the least-squares line fitted to successive 50 kg m-3 density intervals for this subset of the data Figure (9b). Although magnetite has a high conductivity (Figure 9b) (Table IV), its relatively low abundance in most rocks means that the connectivity between magnetite grains is incomplete and apparently not efficient enough to substantially enhance thermal conductivity. Density, on the other hand, is purely a function of the amount of magnetite present. Vanadium mining on the Main Magnetite Layer is shallow at present, but the data are potentially useful for engineering purposes if deeper levels are exploited in the future.
Bushveld granites and granophyres
Granitic rocks of the Rashoop Granophyre Suite and Lebowa Granite Suite have not been differentiated. Table II and Figures 10 and 11 indicate that these rocks are thermally quite distinct from the mafic and ultramafic rocks of the Rustenburg Layered Suite. The thermal parameters are typical of granites, except for a few magnetite-bearing specimens that have higher densities (Figure 11a). Although there is no strong relationship between conductivity and density in Figure 11a, a plot of average conductivity versus average density for different sample localities suggests a negative correlation (Figure 11b).
Heat capacity and thermal diffusivity
Heat capacity measurements were made on 190 samples and results are summarized in Table III. Most of the samples were derived from the Critical Zone and lower part of the Main Zone because data on rocks from these levels is most important for applications in mine refrigeration. Although heat capacity increases with increasing pyroxene content (Table I and Table III) this effect is small, and an overall average value for the Critical and Main Zones (850 ±50 J kg-1K-1) will suffice (in the absence of measured values) for calculating thermal diffusivity for mine refrigeration purposes. Chromitite has a lower average heat capacity (750 J kg-1K-1, Table III), which should be used for calculations involving mining of the UG2 chromitite layer. The value for Main Zone pyroxenite is recommended for similar analyses involving the Merensky Reef. The calculated parameter, thermal diffusivity, varies from 0.8 x 10-6 to 1.4 x 10-6 m2 s-1.
Estimating the heat load from the country rock in underground workings is a complicated process because it depends on several factors including mine geometry, VRT, rock properties, and boundary conditions at newly exposed rock surfaces. This is beyond the scope of this paper, but the effect of thermal properties alone reported here can be illustrated by calculating the heat flux at the surface of a semi-infinite region initially at a temperature of Ti and instantaneously exposed to a constant surface temperature of Toat time t=0:
Figure 12 shows the surface heat flux as a function of time for Ti=50°C, To=25°C, and various values for the thermal properties. The heat flux is up to 50% lower if the half-space is characterized by thermal parameters appropriate to the Bushveld Complex (blue) compared with an equivalent situation in gold mines of the Witwatersrand Basin (red). Estimation of the controlling effect of rock properties on heat flux into underground workings obviously requires specific and in-depth calculations (C.A. Rawlins, personal communication, 2014).
The extensive thermophysical rock property database from the Bushveld Complex permits reliable estimation of the average thermal conductivity, density, heat capacity, and thermal diffusivity of most rock types in various stratigraphic units constituting the Complex. The results provide important inputs for calculations of the heat load on deep platinum mines and would be potentially important in chromium and vanadium mines if such mining proceeds to deeper levels. A positive correlation between conductivity and density of rocks in the Critical Zone and Main Zone may be useful as a conductivity estimator where conductivity data are not available. The high magnetite content of many rocks in the Upper Zone results in a correlation that differs from the Critical and Main Zones and the rest of the Upper Zone; the density is substantially higher but the conductivity is largely unaffected. Illustrative calculations indicate that the generally lower thermal conductivity and thermal diffusivity of rocks in Bushveld platinum mines results in a lower heat flux into underground workings compared with gold mines in the Witwatersrand Basin.
The research leading to this publication was made possible by financial and logistical support from Bluhm Burton Engineering (Pty) Ltd, the Chamber of Mines Research Organisation (later CSIR Mining Technology), General Mining Union Corporation Ltd, the Geological Survey of South Africa (now Council for Geoscience), Gold Fields of South Africa Ltd, Johannesburg Consolidated Investment Company Ltd, and Impala Platinum Ltd. Individual geologists who assisted in field surveys and sample collection are too numerous to list, but their efforts have not been forgotten. John Sorour's assistance in measuring thermal conductivity is greatly appreciated. The author gratefully acknowledges reviews by Grant Cawthorn, Alex Rawlins, and Russel Ramsden.
Ashwal, L.D., Webb, S.J., and Knoper, M.W. 2005. Magmatic stratigraphy in the Bushveld northern lobe: continuous geophysical and mineralogical data from the 2950 m Bellevue drillcore. South African Journal of Geology, vol. 108. pp 199-132. [ Links ]
Bullard, E.C. 1939. Heat flow in South Africa. Proceedings of the Royal Society of London, Series A, vol. 173. pp. 474-502. [ Links ]
Cairncross, B. and Dixon, R. 1995. Minerals of South Africa. Geological Society of South Africa. Johannesburg. [ Links ]
Carte, A.E. and van Rooyen, A.I.M. 1969. Further measurements of heat flow in South Africa. Special Publication of the Geological Society of South Africa, vol 2. Johannesburg. pp. 445-448. [ Links ]
Cawthorn, R.G., Cooper, G.R.J., and Webb, S.J. 1998. Connectivity between the western and eastern limbs of the Bushveld Complex. South African Journal of Geology, vol. 101. pp. 291-298. [ Links ]
Cawthorn, R.G., Eales, H.V., Walraven, F., Uken, R., and Watkeys, M.K. 2007. The Bushveld Complex. The Geology of South Africa. Johnson, M.R., Anhaeusser, C.R., and Thomas, R.J. (eds). Geological Society of South Africa/Council for Geoscience, Johannesburg/Pretoria. pp. 261-281. [ Links ]
Cawthorn, R.G. and Spies, L. 2003. Plagioclase content of cyclic units in the Bushveld Complex, South Africa. Contributions to Mineralogy and Petrology, vol. 145. pp. 47-60. [ Links ]
Cermak, V. and Rybach, L. 1982. Thermal conductivity and specific heat of minerals and rocks. Landholt-Bornstein Numerical Data and Functional Relationships in Science and Technology. New Series, Group V, vol. 16. Angenheister, G. (ed.). Springer-Verlag. Berlin. pp. 305-343. [ Links ]
Clark, S.P. 1966. Thermal conductivity. Handbook of Physical Constants. Memoir 97. Clark, S.P. (ed.). Geological Society of America, New York. pp. 459-482. [ Links ]
Davis, M. D., Wilson, A. H., and van der Merwe, A. J. 2007. The use of density as a stratigraphic and correlative tool for the Bushveld Complex, South Africa. Proceedings of Exploration 07: Fifth Decennial International Conference on Mineral Exploration. Milkereit, B. (ed.). Decennial Mineral Exploration Conferences, Toronto. pp. 1193-1197. [ Links ]
Deer, W.A., Howie, R.A., and Zussman, J. 1966. An Introduction to the Rock- Forming Minerals. Longmans, London. [ Links ]
Horai, K. 1971. Thermal conductivity of rock forming minerals. Journal of Geophysical Research, vol. 76. pp. 1278-1308. [ Links ]
Jones, M.Q.W. 1988. Heat flow in the Witwatersrand Basin and environs and its significance for the South African shield geotherm and lithosphere thickness. Journal of Geophysical Research, vol. 93. pp. 3234-3260. [ Links ]
Jones, M.Q.W. 1992. Heat flow in South Africa. Handbook 14. Geological Survey of South Africa. Pretoria. [ Links ]
Jones, M.Q.W. 1998. A review of heat flow in southern Africa and the thermal structure of the lithosphere. South African Geophysical Reviews, vol. 2. pp. 115-122. [ Links ]
Jones, M.Q.W. 2003a. An update in virgin rock temperature analysis of the Witwatersrand Basin. Journal of the Mine Ventilation Society of South Africa, vol. 56. pp. 107-112. [ Links ]
Jones, M.Q.W. 2003b. Thermal properties of stratified rocks from Witwatersrand gold mining areas. Journal of the South African Institute of Mining and Metallurgy, vol. 101. pp. 173-185. [ Links ]
Jones, M.Q.W. and Rawlins, C.A. 2002. Thermal properties of backfill from a deep South African gold mine. Journal of the Mine Ventilation Society of South Africa, vol. 54. pp. 100-106. [ Links ]
Jones, M.Q.W., Tyson, P.D., and Cooper, G.R.J. 1999. Modelling climatic change in South Africa from perturbed borehole temperature profiles. Quaternary International, vol. 57/58. pp. 185-192. [ Links ]
Krige, L.J. 1939. Borehole temperature measurements in the Transvaal and Orange Free State. Proceedings of the Royal Society of London, Series A,vol. 173. pp. 450-474. [ Links ]
Martinelli, G., Dongarrà, D., Jones, M.Q.W, and Rodrigues, A. 1995. Geothermal features of Mozambique - country update. Proceedings of the World Geothermal Congress 1995, vol. 1. Barbier, E., Frye, G., Iglesias, E., and Palmason, G. (eds). International Geothermal Association. Auckland. pp. 251-273. [ Links ]
Rawlins, C.A., Phillips, H.R., and Jones, M.Q.W. 2002. The use of backfill to control heat ingress in deep level mining. Mine Ventilation. Proceeding of the North American 9th US Mine Ventilation Symposium, Kingston, Ontario, Canada. De Souza, E. (ed.). Balkema, Lisse. pp. 355-361. [ Links ]
Royden, L., Sclater, J.G., and von Herzen, R.P. 1980. Continental margin subsidence and heat flow: important parameters in formation of petroleum hydrocarbons. Bulletin of the American Association of Petroleum Geologists, vol. 64. pp. 173-187. [ Links ]
SACS (South African Committee for Stratigraphy). 1980. Stratigraphy of South Africa. Part 1. Lithostratigraphy of the Republic of South Africa, South West Africa/Namibia, and the Republics of Bophuthatswana, Transkei and Venda. Kent., L.E. (compiler). Handbook 8. Geological Survey of South Africa. Pretoria. [ Links ]
Weiss, O. 1938. Temperature measurements with an electrical resistance thermometer in a deep borehole on the East Rand. Journal of the Chemical, Metallurgical and Mining Society of South Africa, vol. 39. pp. 149-166. [ Links ]