Journal Archive

Platinum Metals Rev., 2010, 54, (4), 205
doi: 10.1595/147106710X520222

The Platinum Group Element Deposits of the Bushveld Complex in South Africa

    • By R. Grant Cawthorn
    • School of Geosciences, University of the Witwatersrand,
    • Private Bag 3, Johannesburg 2050, South Africa


Article Synopsis

There are enough platinum group element deposits in the Bushveld Complex in South Africa to supply world demands for many decades or even a century using current mining techniques. Demonstrated reserves and resources published by mining companies make detailed calculations up to a maximum of about twenty years ahead, but there is abundant and adequate geological evidence that these deposits continue far beyond where mining companies have proven according to rigorous international reporting codes. For each 1 km of depth into the Earth in the Bushveld Complex there is in the order of 350 million oz of platinum. For comparison, annual production of platinum from the Bushveld Complex currently is only around 5 million oz. The distinction between ‘reserves’, ‘resources’ and ‘deposits’ is also explained in this article.


In the minerals sector of world economics, the Bushveld Complex in South Africa (Figure 1) is renowned for its overwhelming deposits of platinum group elements (PGE) and chromium (over 80% of the world's deposits of each according to Crowson (1)). Inevitably, from time to time, the question is raised as to how reliable those estimates are. The occurrences of all other mineral deposits are scattered around the world in an erratic way. Each deposit type has well-understood geological processes that operated to form them, but those processes have operated usually all over the world and in many cases throughout long periods of Earth's 4.6 billion-year history, so that most commodities are mined in many different countries. Therefore, the PGEs, and especially platinum itself, are unusual in that they are largely concentrated in a single location.

Fig. 1.

Map of the Bushveld Complex in South Africa, showing the eastern, western and northern limbs. Major towns and cities are marked in red, operating platinum mines and projects currently underway are shown in green


Mining companies may only publish ‘reserves’ and ‘resources’ of platinum. However, as discussed below (see box) (2, 3), this figure represents only what has been rigorously quantified in the short- to medium-term mining plans of these companies, and excludes all the geologically known extensions of these deposits for which an in-depth (and expensive) evaluation is not justified. Current legally enforced definitions operative in several countries, including the USA, Canada, Australia, the UK and South Africa, define ‘reserves’ and ‘resources’ quite precisely, and exclude what geologists know to be identified deposits. In the case of those deposits in the Bushveld Complex there is simply no need to rigorously quantify them at this point in time, because there are adequate proven reserves already identified.

The South African Mineral Codes (SAMCODEs)

Rigorous definitions based on the South African Mineral Resource Committee (SAMREC) and South African Mineral Asset Valuation (SAMVAL) codes (2, 3) can be found at, but a very simplified summary follows.

A mineral ‘reserve’ is an ore body for which adequate information exists to permit confident extraction. Briefly, it requires that all aspects including adequately spaced drilling, assaying, mineralogical and metallurgical studies, mine planning, beneficiation, environmental, social and legislative issues, and financial viability have been addressed. Mining companies would typically plan their exploration and evaluation strategies such that they had a minimum of ten years of ore in this category. (Two sub-categories exist: ‘probable’ and ‘proven’ reserves.)

A mineral ‘resource’ is an ore body for which there are reasonable and realistic prospects for eventual extraction. Addressing of all the issues listed under ‘reserve’ would have been initiated, and all such results would be positive. Mining companies might aim to have a further ten years of ore in this category. (Three sub-categories exist: ‘measured’, ‘indicated’ and ‘inferred’, as a function of increasing risk.)

No mining company is likely to incur major expense in exploration beyond the combined time period of their reserves and resources of about twenty years, although completely new targets may be of interest and land tenure and acquisition around existing operations will be carefully monitored.

The ore deposits of the Bushveld Complex are so large that they far exceed the time periods mentioned above. Hence, to avoid any confusion (or over-optimistic extrapolations), the term ‘deposit’ is used in this report to indicate the continuing geological strata that contain ore mined in various areas. Reasonable geological information is available in these areas but no significant economic evaluation has been attempted. Such strata are known to continue along surface between mines, and have been identified by geophysical techniques to greater depth than current mining.

A number of estimates of the potential deposits of the PGE in the Bushveld Complex have been published in scientific journals over the last thirty years, and the quoted numbers have not changed significantly. A summary of these compilations is presented in Table I (1, 4–7). These reports have come mainly from South African-based geologists, and their figures have been reproduced or validated by independent organisations such as the United States Geological Survey (USGS), that have staff qualified to make critical assessments of these estimates.

Table I

PGE Deposits in the Bushveld Complex According to Different Authorsa

Author(s) Year Reference Reserves, million oz Resources, million oz Deposits, million oz
Pt Pd Pt Pd Pt Pd
Merensky Reef
Von Gruenewaldt 1977 (4) 318 136
Vermaak 1995 (5) 345 180
Cawthorn 1999 (6) 77 35 400 221
Vermaak and van der Merwe 2000 (7) 55 33 524 302
UG2 Chromitite
Von Gruenewaldt 1977 (4) 398 330
Vermaak 1995 (5) 379 237
Cawthorn 1999 (6) 116 69 403 354
Vermaak and van der Merwe 2000 (7) 99 65 431 281
Von Gruenewaldt 1977 (4) 123 135
Vermaak 1995 (5) 59 66
Cawthorn 1999 (6) 10 11 136 136
Vermaak and van der Merwe 2000 (7) 11 11 168 171
All of the Bushveld Complex
Crowson 2001 (1) 820 720

[i] aThe depths to which these deposits have been calculated vary. Von Gruenewaldt used 1.2 km. Vermaak's estimates vary between 1 to 2 km. Cawthorn used a depth of 2 km. Crowson gave no information

The Geology of the Bushveld Complex

Before discussing how these estimates are calculated, it is worth considering the origins and geology of the Bushveld Complex. The economic potential of many ore deposits is difficult to evaluate, but this is not so in the case of the Bushveld ores. The Bushveld Complex (see Figure 2) is an enormous irruption of magma (molten rock) sourced deep within the Earth. The extent of the magma flow was at least 300 km in diameter. In the order of 1 million km3 of magma was emplaced in a (geologically) very short period of time. As this enormous volume of Bushveld magma slowly cooled, different minerals began to solidify and accumulated in thin, parallel layers at the bottom of this huge magma ocean. The maximum thickness ultimately was about 8 km.

Fig. 2.

Block diagram of an oblique view from the southeast of the Bushveld Complex, showing the continuation of the platinum-bearing layers (Merensky Reef and UG2 Chromitite Reef) to depth. Outcrop of the Bushveld Complex on the surface is shown in dark green; its occurrence at depth in the cut-away vertical sides is shown in pale green. The layers probably continue under the entire area shown, but near the middle are deeply buried (greater than 5 km depth) by younger granitic and sedimentary rocks (called the Karoo Supergroup)


Most minerals in the Bushveld Complex have no economic importance, but two types are important here: chromite and the sulfide group of minerals. Both of these mineral types concentrate the PGE. The first economically important discovery of platinum in the Bushveld Complex was found in a single layer, associated with the sulfides, that we now call the Merensky Reef after its discoverer, Dr Hans Merensky (8). Figure 3 shows a specimen of a Merensky Reef section. The distribution of the PGE vertically through the Merensky Reef layer is somewhat variable, but mining companies would aim to extract a layer in the order of 1 m in thickness that contains the majority of the total PGE. Lower-grade ore below and above this zone has to be left behind as it is not economical to process. For that reason it is usually excluded from any resource calculations.

Fig. 3.

Photograph of a hand specimen (20 cm in height) of the base of the Merensky Reef from Rustenburg Platinum Mines. In this section the different rock types composing the reef can be seen. At the base is a white anorthosite with pale brown grains of pyroxene. Above it is a layer of coarse grains of pyroxene (brown) and plagioclase (white), which is called pegmatitic pyroxenite. It becomes finer grained upward. A thin layer of chromitite (a few mm thick) is usually present at the top of the anorthosite. The bright yellow minerals are sulfide grains. The PGE are highly concentrated with the sulfide and chromite grains. The best mineralised interval would be from about 20 cm below the base of the specimen in this photograph (in anorthosite) to about 20 cm above the top of the photograph, consisting of more fine-grained pyroxenite. This is an unusually narrow section of the Merensky Reef


Even before the Merensky Reef was discovered, the presence of platinum in chromite-rich layers in the Bushveld Complex was quite well known, but never found to be economic. There are a number of these chromitite layers in the Bushveld Complex, some reaching up to 1 m in thickness. They contain up to 3 grams per tonne (g t−1) of PGE, often with quite high proportions of rhodium and lesser amounts of iridium and ruthenium. One layer, called the Upper Group 2 (UG2) chromitite layer, is a possible hybrid of the sulfide- and chromite-hosted reefs. The distribution of the PGE associated with the chromitite layers is sharply controlled by the chromite. There is essentially no PGE above or below the chromitite reef, and so the mining operations and resource calculations are much easier to define.

Assaying of a borehole intersection of the Merensky and/or UG2 Reefs usually involves taking several consecutive lengths, each about 20 cm long, which are individually analysed, so that the best mineralised, 1 m-thick interval, especially for the Merensky Reef, can be identified. Often two or three deflections are drilled for each borehole. This involves putting a wedge into the hole some 20–30 m above a reef intersection, and redrilling. This produces another section of reef within a metre of the first (mother) hole, providing more analytical and statistical data without the cost of drilling long, deep holes. Since many boreholes are drilled to intersect the reef at up to 2 km depth such deflections provide more information extremely cheaply. Over the many years of mining, confidence in the statistical variability of reef intersections permits reliable information on predicted grade. Until ten years ago most such exploration was based around the major existing mines in the western Bushveld. Since then, because of the rise in the platinum price (9), much exploration has been undertaken on areas in the eastern Bushveld, at great depths in established mining areas, and also where there are geological complications, for both the Merensky and UG2 Reefs. Two such examples of the latter would be around the Pilanesberg intrusion in the west, and near the Steelpoort fault in the east.

Estimating Reserves and Resources

Paradoxically, although there are now more data available than previously, the ability to publish estimates of the platinum content of the Bushveld Complex is much more restricted. When the first estimates were calculated (Table I) they included predictions, projections and interpretations of geological continuity which are geologically plausible. Now, based on statutory resource codes, it is necessary to report all information only in terms of ‘reserves’ and ‘resources’, and whether proven or inferred. All the early calculations would have contained ore that would fall far outside what is now considered ‘inferred’. No exploration or mining company is going to expend unnecessary effort on proving deposits that might become mineable only a long way into the future. As a result, it appears as if the currently quoted platinum deposits of the Bushveld Complex are less than they actually are (compare Table II with Table I), purely because only ‘reserves’ and ‘resources’ may be reported and geologically known deposits are formally excluded by statutory codes. Even with this limitation, the ‘reserves’ and ‘resources’ reported by the four major mining companies in South Africa amount to at least 1200 million oz of PGE of which more than 50% is platinum (Table II).

Table II

PGE Reserves, Resources and Annual Refined Production Reported by the Major Platinum Mining Companies in South Africa for 2009a

Company Reserves, million oz Resourcesb, million oz Annual refined productionc, million oz
3PGEd + Au 3PGE + Au 5PGEe Pt
Anglo Platinum 171 632f 4.8 2.5
Implats 26g 133g 2.6 1.2
Lonmin 45 178 1.2 0.7
Northam Platinum 8 129 0.4 0.2
Totals 250 1072 9.0 4.6

[i] aThese values are for the Bushveld Complex (i.e. they exclude Zimbabwe in all cases and toll refining in the case of Anglo Platinum and Implats). All data are taken from public sources published by the respective mining companies up to the end of calendar year 2009

bUnless otherwise stated, resources are inclusive of reserves

cProduction data compiled by Alison Cowley, Principal Market Analyst, Johnson Matthey, August 2010

d3PGE = Pt, Pd and Rh

e5PGE = Pt, Pd, Rh, Ir and Ru

fResources quoted by Anglo Platinum are exclusive of reserves

gReserves and resources quoted by Implats are for platinum only

From a purely geological perspective, we can make the following calculation from the following simple assumptions. We can assume the thickness of mineable Merensky and UG2 Reefs to be 0.8 m each. Geological maps show the two reefs to occur at outcrops with a strike length of about 100 km in both eastern and western limbs, giving a measure of the horizontal extent of the reef. We know that the typical dips of the reefs are less than 15°, so the reef can be mined for 4 km down the dip before a vertical depth of 1 km is reached. The densities of Merensky and UG2 Reefs are 3.2 t m−3 and 4.0 t m−3 respectively (7). We can take a conservative grade of 2 g t−1 (0.06 oz t−1) of extractable platinum. Simple multiplication of these figures (strike length along outcrop × distance into the Earth × thickness × density × grade, using appropriate units) yields 350 million oz of platinum per km depth (see box below). I use this method of presenting the information (in oz per km depth) because it is not known what mining depths may be ultimately viable. This simple calculation ignores what are called geological losses (assorted features such as faults and potholes, as described by Vermaak and van der Merwe (7)). Taking a mining depth to 2 km, and adding what might be present in the Platreef which occurs in the northern limb of the Bushveld Complex (see Figure 2), gives the rounded figure of 800 million oz of platinum, similar to that reported by Crowson (1) in Table I. For comparison, the annual refined production of the major platinum producers in South Africa is shown in Table II, and amounts to only around 9 million oz of PGE, of which less than 5 million oz are platinum.

Example of the Calculation Principle for Estimate

Note 1: This calculation is for platinum only, and also ignores the geological losses inherent in mining and process losses during refining (7).

Note 2: This calculation is purely illustrative of the basic simplicity of the concept behind such estimations of ore. All values have been rounded off for simplicity of multiplication and should not be taken quantitatively.

Combined strike length of outcrop in eastern and western Bushveld = 230 km = 230,000 m

Distance down the dip to 1 km vertical depth at an angle of 13° = 4.4 km = 4400 m

Mined thickness of Merensky Reef = 0.8 m, and of UG2 Reef = 0.8 m, combined thickness = 1.6 m

Density of Merensky Reef = 3.2 t m−3, density of UG2 Reef = 4.0 t m−3. Average of both reefs combined = 3.6 t m−3

Grade of platinum only = 2 g t−1 = 0.06 oz t−1

Weight of platinum only to 1 km depth = 230,000 × 4400 × 1.6 × 3.6 × 0.06 = 350 million oz

Probably nearly half that figure of 350 million oz has already been mined from the Bushveld Complex and most of it from depths of less than 1 km. So we can assume that there are around 200 million oz remaining in the upper one km, and 350 million ounces in the second km. Since mining in some cases is already at more than 2 km vertical depth (as at the Northam Platinum Mine), assuming material to that depth can all be mined is technically straightforward. Mining to greater depth will encounter high temperatures and serious rock stresses. However, mining of the Witwatersrand gold reefs has progressed to a depth of 4 km. Hence, mining of Bushveld reefs to comparable depths should not be considered implausible during this century.

The Platreef

In calculating a figure of 350 million oz platinum per km depth I have excluded the Platreef. Evaluating its total PGE content is more difficult. Evaluating the mineable PGE is even harder. The PGE mineralisation is distributed over vertical intervals that can reach 100 m in some borehole intersections, but more usually is intermittent over 50 m. Tracing the best mineralised horizons from one borehole to another is difficult – grade varies, thickness varies, and their location is not at a constant geological elevation. Best mining methods then also become an issue. Currently, all mining on the Platreef is by open pit methods, but it is limited to probably 500–800 m depth. Drilling has shown that the Platreef continues to at least 2 km depth. Methods of underground mining of this wide ore body are still being developed. As a result the estimates in Table I for the Platreef show considerable variation.

Other Chromitite Layers

Every chromite-rich layer of the Bushveld Complex contains some PGE. Very little systematic work has been done to evaluate these layers since the grade is in the range 1–3 g t−1 of PGE (10). The chromite ore itself is also currently sub-economic for chromium. The layer thickness and chromium-to-iron ratio are the two most important parameters in evaluating the economic potential of chromite deposits, and the chromitite layers of the Bushveld Complex generally decrease in this ratio upward, and are in the range 2.0 to 1.2 (10). The layers with the highest values are too thin to mine. The thickest layers, which are mined, have ratios close to 1.6. Layers with a ratio of less than 1.4 are not economic at present. There is an inverse correlation between the PGE grade and the chromium-to-iron ratio in these layers. As with the PGE, there are extremely large proven reserves of chromium in the Bushveld Complex (1), even down to vertical depths of less than several hundred metres. Thus, if the chromium and PGE reefs that are currently being mined were ever exhausted one could mine the lower-grade chromitite layers for both chromium and PGE. Conservatively, we could suggest that that these chromitite layers host nearly as much PGE and chromium as is presently considered potentially economic.

The tailings dumps from these chromite mines can be reprocessed to provide additional PGE. However they contain very minor PGE and, while some recovery is taking place, it will never represent a major additional supply.

Other Deposits of the Platinum Group Elements

The PGE contained in other deposits and exploration areas around the world tend to be much harder to define. They do not occur in such well-constrained layers of uniform thickness and great lateral extent. The mineralised layer in the Great Dyke of Zimbabwe is the most similar to the Merensky Reef, but the area of the intrusion is very much smaller. The Stillwater Complex in North America is also similar to the Bushveld Complex, but the ore zone is patchily concentrated along and vertically within a thicker layer than the Merensky Reef. The length of that entire intrusion is about 40 km, which is similar to that of one single mining operation in the Rustenburg area (Anglo Platinum, Impala Platinum or Lonmin). The dip of the rocks in Stillwater is much greater, again reducing the ore tonnage to maximum mining depth.

The total potential of all the other PGE mining areas is much harder to quantify because they are not layered. Extensive exploration during the last ten years has failed to produce any new major targets or mines. Also, all occurrences in the Bushveld Complex and Great Dyke have high platinum-to-palladium proportions (platinum greater than palladium), whereas all other occurrences in the world are dominated by palladium.

Other Issues

Maintaining and increasing future production rates represent significant challenges. Resolving social, political and environmental issues, together with ensuring water and electrical supply capacities, needs ongoing monitoring and careful planning (11). These challenges are the unknowns and unpredictables in the future of platinum mining in South Africa, not the availability of the ores.


The estimates of the PGE presented here are not intended to be rigorous or quantitative. They are designed to show that the broad estimation of the PGE in the Bushveld Complex is extremely easy to make and to understand, and that the remarkably disproportionate concentration of the PGE in one geographic location, South Africa, is genuine. Even with current mining methods, technology and prices, there are many decades to a century of extractable PGE ore already known in the Bushveld Complex. With around 350 million oz of platinum per vertical km depth, the enormous deposits of PGE in the Bushveld Complex can be confidently relied upon to provide a major proportion of the demand needs for a long time into the future.

Term Definition
3PGE Platinum, palladium and rhodium
5PGE Platinum, palladium, rhodium, iridium and ruthenium
anorthosite A type of igneous rock largely composed of plagioclase feldspar, formed from intrusions of magma within the Earth's crust
beneficiation The processing of ore to produce minerals, also the further processing of minerals to metals and then to value-added products
Bushveld Complex A large, layered, saucer-shaped geological formation found in the Bushveld region of the north of South Africa; it contains deposits rich in PGE
chromite An iron chromium oxide (FeCr2O4) mineral with traces of magnesium and aluminium
chromitite A rock type containing a high concentration of chromite
deflection A secondary borehole drilled at an angle to the vertical
deposit The total quantity of an ore body contained within a geological formation
dip The angle of inclination of a reef from the horizontal
fault A discontinuity in a layered feature resulting from rock fracture and movement, with one section being displaced relative to another
feldspar An aluminium silicate mineral, containing potassium, sodium, calcium or barium
footwall The layer of rock beneath a vein or expanse of ore
grade The specific quantity of an element of interest contained within a unit mass of an ore body; for the PGEs this is most often given in grams per metric tonne
Great Dyke A linear, layered geological feature running approximately north–south in the centre of Zimbabwe; it contains deposits rich in PGE
igneous Rocks formed from the solidification of either magma in the Earth's crust or of lava on the surface
Merensky Reef A layer of the Bushveld Complex largely composed of pyroxenite that is rich in sulfide minerals; to date it has supplied most of the world's platinum group metals, and also yields significant quantities of copper, nickel, cobalt and gold as byproducts. It is mined on both the eastern and western limbs of the Bushveld Complex
mineralised horizon A layer or stratum in which minerals of interest are preferentially concentrated; this could be distinct and continuous as a reef, or more dispersed and intermittent
outcrop A section of the reef which intersects the surface of the Earth and may have been subject to weathering
pegmatite A type of igneous rock characterised by a very coarse grain structure, with crystals several centimetres across usually composed of granite (quartz, feldspar and mica)
PGE Platinum group elements (platinum, palladium, rhodium, iridium, osmium and ruthenium). This term is used in geology as the elements generally occur in mineral, rather than metallic, form within an ore
plagioclase An aluminium silicate mineral of the feldspar family, with varying relative proportions of sodium and calcium
Platreef An ore body in the northern limb of the Bushveld Complex, it is the third largest PGE deposit in the world, after the Merensky and UG2 Reefs. It consists of three broadly mineralised horizons rather than a distinct reef
pothole Circular or elliptical sections where the reef has funnelled into the footwall, leading to discontinuity and altered mineralogy
pyroxene Silicate minerals containing calcium, magnesium and iron
pyroxenite A rock type containing a high concentration of pyroxenes
reef A distinct and continuous layer or stratum in which minerals of interest are preferentially concentrated
reserves Ore bodies which have been quantified to a high degree of confidence and which can be extracted using existing methods
resources Ore bodies which are known to exist and which can be quantified to some degree of confidence. These can reasonably be expected to be extracted in the future
sedimentary A type of rock formed from solidified deposits of eroded rock material, which have usually accumulated in bodies of water
strike The line of intersection of an inclined plane with the horizontal, such as when a reef outcrops on the surface of the Earth
sulfide Minerals formed from compounds of sulfur, these are a major source of metals such as copper, nickel and lead
tailings The waste material from ore processing, usually a slurry of finely ground rock in water, from which most of the valuable minerals have been removed
UG2 Reef Upper Group 2; a layer of the Bushveld Complex rich in chromite but lacking sulfide minerals. It possibly has a larger resource of platinum group elements than the Merensky Reef. It lies below the Merensky Reef and is mined on both the eastern and western limbs of the Bushveld Complex


  1.  P. Crowson, “Minerals Handbook 2000–01: Statistics and Analyses of the World's Minerals Industry”, Mining Journal Books Ltd, Edenbridge, Kent, UK, 2001
  2.  The South African Code for the Reporting of Exploration Results, Mineral Resources and Mineral Reserves (The SAMREC Code), 2007 Edition as amended July 2009, Prepared by the South African Mineral Resource Committee (SAMREC) Working Group: on 27th July 2010)
  3.  The South African Code for the Reporting of Mineral Asset Valuation (The SAMVAL Code), 2008 Edition as amended July 2009, Prepared by the South African Mineral Asset Valuation (SAMVAL) Working Group: (Accessed on 27th July 2010)
  4.  G. von Gruenewaldt, Miner. Sci. Eng., 1977, 9, (2), 83
  5.  C. F. Vermaak (Mintek, Randburg, South Africa), ‘The Platinum Group Metals: A Global Perspective’, Internal report (unnumbered), 1995, ISBN 0-86999-926-5
  6.  R. G. Cawthorn, S. Afr. J. Sci., 1999, 95, (11/12), 481 LINK
  7.  C. F. Vermaak and M. J. van der Merwe (Mintek, Randburg, South Africa), ‘The Platinum Mines and Deposits of the Bushveld Complex, South Africa’, Internal report (unnumbered), 1999, ISBN 0-86999-944-3
  8.  R. G. Cawthorn, Platinum Metals Rev., 2006, 50, (3), 130 LINK
  9.  Platinum Today, Current and Historical Prices: on 27th July 2010)
  10.  R. N. Scoon and B. Teigler, Econ. Geol., 1994, 89, (5), 1094 LINK
  11.  B. J. Glaister and G. M. Mudd, Miner. Eng., 2010, 23, (5), 438 LINK

The Author

R. Grant Cawthorn is the Platinum Industry's Professor of Igneous Petrology at the University of the Witwatersrand, South Africa. His main research interests are the genesis of the Bushveld Complex and its chromite, platinum and vanadiferous magnetite deposits, and the Insizwa intrusion and its copper and nickel deposits. His main fields of specialisation are the origin of mafic igneous intrusive rocks and their mineral deposits. He holds a BSc from the University of Durham, UK, a PhD from the University of Edinburgh, UK, and a DSc from the University of the Witwatersrand, South Africa.

Find an article