A multi-isotope (S-Sr-Nd) investigation of the Flatreef, Northern Limb, 

Bushveld Complex: 

Petrogenetic implications and comparison with the Merensky Reef 

 

 

 

 

Jarlen Jocelyn Keet 

 

 

 

 

 

 

 

 

Thesis 

Submitted in fulfilment of the requirements in respect of the Doctor of Philosophy 

degree in the Department of Geology, Faculty of Natural and Agricultural 

Sciences, University of the Free State. 

 

 

Bloemfontein, 2022



 
I 

 

 

 

 

 

DECLARATION 

I, Jarlen Jocelyn Keet, declare that this thesis (inclusive of published articles) that I 

herewith submit for the Doctor of Philosophy degree at the University of the Free State, 

is my independent work, and that I have not previously submitted it for a qualification at 

another institution of higher education. 

 

 

 

 

Jarlen J. Keet 

28/11/22 

 

 

 

 

 

 

 

 

 

 

  



 
II 

 

ABSTRACT 

 

Historically, research on the Bushveld Complex (BC) in the northern limb mainly focused on the 

near surface Platreef. The Platreef is well-known for being complicated and erratic due to 

significant degrees of country rock assimilation and contamination along strike. A correlation of 

the disrupted magmatic stratigraphy of the Platreef to the Upper Critical Zone of the western and 

eastern limbs has thus proven difficult. The discovery of the Flatreef, the down-dip, sub-horizontal 

extension of the Platreef, on the farms Turfspruit and Macalacaskop, opened new avenues for 

enquiry. The proportion of assimilated country rock within the Flatreef is significantly less than in 

the near surface Platreef, with a stratigraphy that is less disrupted and affected by footwall 

interaction than the Platreef, such that Bastard Reef and Merensky Reef correlates may be 

identified in the Flatreef stratigraphy.  

In this study Sr, Nd and S isotopic compositions across the stratigraphic units of the Flatreef as 

intersected by drill hole UMT-393 on the farm Macalacaskop, and to a lesser extent, drill hole 

UMT-276 on the farm Turfspruit is reported. The initial 87Sr/86Sr ratio (Sri) (also indicated as 

87Sr/86Sri) results show a significant shift from about 0.706 to 0.707 in the immediate footwall, to 

values > 0.709 near the top of the Main Reef. This isotopic shift matches the isotopic shift 

described for the Merensky Reef of the eastern and western BC. The mineralized units of the 

Flatreef, namely the Main Reef and Upper Reef, have 𝛿34S values ranging between -0.2 to 1.5‰ 

(with the exception of three outliers), and 0.52% and 11.2‰ in UMT-393, respectively. The S-

isotope data for UMT-276 are lighter with 𝛿34S values ranging between -0.96 and 2.24‰ for the 

Main Reef and 3.19‰ was recorded for an Upper Reef sample.  These near-mantle 𝛿34S values 

fall within the range of about 0 to 3‰ reported for the Merensky Reef in the eastern and western 

BC. The initial epsilon Nd (εNd
i) values for the Flatreef were found to range between –5.2 to -7.6, 

strongly overlapping with εNd
i values of the Upper Critical Zone and the lower reaches of the Main 

Zone from the eastern (with εNd
i values ranging between -4.8 and -8.5 for the Upper Critical Zone) 

and western limbs (with εNd
i values ranging between -6.3 and -7.6 for the Upper Critical Zone, and 

-6.3 and -7.4 for the Main Zone) of the Bushveld Complex. 

It is evident from our findings that the Flatreef was less affected by interactions with footwall rocks 

compared to the Platreef, with isotopic signatures overlapping significantly with those reported for 

the upper reaches of the Upper Critical Zone in the remainder of the BC. Our findings provide 

further support for the contention that the Flatreef is the stratigraphic equivalent of the Upper 

Critical Zone inclusive of the Merensky Reef and the Bastard Reef in the remainder of the BC. 



 
III 

 

The similar isotopic composition of the hanging wall and footwall units of the Flatreef suggests 

that these units may have been continuous and were separated by subsequent magma pulses of 

the remainder of the Flatreef. Due to significant overlap between the εNd
i values of the Flatreef 

and local potential contaminants occurring at the base of the Northern Limb, it is proposed that 

the Sr-Nd isotopic composition of the magmas that gave rise to the Flatreef are most likely 

attributable to the interaction of mantle-derived magma with upper and lower crustal rocks of the 

Kaapvaal Craton within a sub-Bushveld staging chamber, with possible syn- to post-emplacement 

modification thereof as a result of interaction with locally available dolomitic footwall rocks.  The 

reef horizons of the Flatreef may have been emplaced as sills fed from the proposed staging 

chamber, which would explain the similarities in lithology and geochemistry of the hanging wall 

and footwall units as mentioned above. This would imply that the Main Zone could not have been 

the source of PGEs within the Flatreef, and by extension, the Merensky Reef.  

  



 
IV 

 

 

 

 

DEDICATION 

 

To my nieces, nephew and cousins. I may be the first in the family, but I am not the last. We are 

not defined by where we come from… there is nothing you cannot do. 

  



 
V 

 

ACKNOWLEDGEMENTS 

 

I would like to thank Prof Frederick Roelofse, not only for his committed supervision of this project, 

but also for his patience, all his support, suggestions and encouragement during the course of 

this study; it is highly appreciated. I would also like to thank Prof Christoph Gauert who helped 

conceive this project, thank you for your input and support from afar. This project would not have 

evolved to what it is today without the guidance of my supervisors. 

This project was funded in part by the Inkaba /Iphakade funding, the nGAP grant (UID 119698) 

and NRF grants to my supervisors. 

Ivanplats Ltd. is also thanked for access to the core yard for logging and sampling, especially to 

Dr Danie Grobler and the geologists who assisted during my visit. Dr. Grobler’s expertise on the 

northern limb was of great assistance in this project. 

Ms. Henriette Ueckermann is thanked for assisting with Laser Ablation Multi-Collector Inductively 

Coupled Plasma Mass Spectrometry (LA-ICP-MS) housed at the University of Johannesburg. Dr. 

Linda Iaccheri at the Wits Isotope Geoscience Laboratory is thanked for the preparation of 

samples for Nd-isotopic analysis. Mr. Mike Butler at iThemba Laboratories is thanked for S isotope 

analyses. Mr Marlin Patchappa from the EarthLab at the University of Witwatersrand is thanked 

for the geochemical analyses. Mr. Deon van Niekerk is thanked for assisting with the Jeol JXA 

8230 Superprobe that is housed at Rhodes University. This instrument is sponsored by NRF/NEP 

grant 40113 (UID 74464). 

My sincere thanks also go to the Geology Department’s laboratory staff members at the University 

of the Free State. Mr. Jonas Choane and Mr. Pelele Lehloenya, thank you for preparing the fusion 

discs and pressed pellets; Mr. Daniel Radikgomo, thank you for preparing the thin sections and 

epoxies used in this study and for teaching me how to prepare my own; Mrs. Megan Welman-

Purchase, thank you for calibrating the XRF and SEM instruments, and rerunning fusion discs 

countless times. Special thanks also to Mrs. Rina Immelmanꝉ, Mrs. Charlene Van Der Vyver and 

Mr. Andries Felix for their administrative support throughout this project. I would also like to thank 

my colleagues at the department for their continuous support. 

Lady-Bridget One (née Akuffo) and the Akuffo family, thank you so much for opening your home 

to me whenever I was in Johannesburg, whether it was for analyses, or visa application, 

everything you did for me is highly appreciated. Several other people have contributed to my 

academic journey and growth over the past years. It is difficult to acknowledge each and every 



 
VI 

 

contribution; but to everyone who played a role in this project as well as my personal development, 

I am so grateful, thank you.  

I would also like to thank my parents, my parents-in-law and my church family for all their support 

and encouragement throughout my academic journey.  

To my husband, Danrich Keet, words are not enough to express my gratitude to you. You have 

supported me through every season, going above and beyond to ensure that I finish strong. Thank 

you, for everything. I would not have made it this far without you. 

Most importantly, I would like to thank my Saviour, Jesus Christ, without whom none of this would 

have been possible.  

  



 
VII 

 

TABLE OF CONTENTS 
 

DECLARATION .......................................................................................................................... I 

ABSTRACT ................................................................................................................................ II 

DEDICATION ............................................................................................................................ IV 

ACKNOWLEDGEMENTS .......................................................................................................... V 

LIST OF FIGURES ................................................................................................................... XI 

LIST OF TABLES ................................................................................................................... XVII 

LIST OF APPENDICES ........................................................................................................ XVIII 

CHAPTER 1 : INTRODUCTION ................................................................................................. 1 

1.1 Study context .................................................................................................................... 1 

1.2 Research aims and objectives .......................................................................................... 3 

1.3 Thesis Layout and Author Contributions ........................................................................... 4 

1.4 Methodology ..................................................................................................................... 6 

CHAPTER 2 : GEOLOGICAL BACKGROUND .......................................................................... 7 

2.1 The Bushveld Complex ..................................................................................................... 7 

2.1.1 The Rustenburg Layered Suite ...................................................................................... 9 

2.1.1.1 Marginal Zone (MZN) ......................................................................................................... 11 

2.1.1.2 Lower Zone (LZ) .................................................................................................................. 12 

2.1.1.3 Critical Zone (CZ) ................................................................................................................ 12 

2.1.1.4 Main Zone (MZ) ................................................................................................................... 13 

2.1.1.5 Upper Zone (UZ) ................................................................................................................. 13 

2.1.2 Mineralization in the Rustenburg Layered Suite of the western and eastern limbs ........14 

2.1.2.1 The Merensky Reef ............................................................................................................ 15 

2.1.2.2 The UG-2 Chromitite .......................................................................................................... 16 

2.2 The Northern Limb of the RLS .........................................................................................17 

2.2.1 The Platreef ..................................................................................................................20 

2.2.2 The Flatreef ..................................................................................................................22 

CHAPTER 3 : STRATIGRAPHY OF THE FLATREEF ..............................................................24 

3.1. Drill Core UMT-393 .........................................................................................................24 

3.2. Drill Core UMT-276 .........................................................................................................31 

CHAPTER 4 : ARTICLE 1 Strontium isotope variations in the Flatreef on Macalacaskop, 

northern limb, Bushveld Complex: Implications for the source of platinum-group elements in the 

Merensky Reef ..........................................................................................................................33 

4.1 Abstract ...........................................................................................................................33 



 
VIII 

 

4.2 Introduction ......................................................................................................................33 

4.3 Regional geology .............................................................................................................35 

4.3.1 Stratigraphy of the northern limb .......................................................................................... 36 

4.3.2 The Platreef and Flatreef at Turfspruit and Macalacaskop .............................................. 36 

4.4 Sample locations and descriptions...................................................................................38 

4.5 Analytical Methods ..........................................................................................................43 

4.6 Results ............................................................................................................................44 

4.6.1 Sr isotope composition of plagioclase ................................................................................. 44 

4.6.2 An content of plagioclase ...................................................................................................... 45 

4.7 Discussion .......................................................................................................................49 

4.7.1 Correlation of the Flatreef with the UCZ – MZ transition interval in the remainder of 

the Bushveld Complex ..................................................................................................................... 49 

4.7.2 Implications for the source of PGE mineralization in the Merensky Reef ...................... 51 

4.8 Conclusions .....................................................................................................................52 

4.9 Acknowledgements .........................................................................................................52 

4.10 References ....................................................................................................................52 

CHAPTER 5 : ARTICLE 2 A comparative study of the sulfur isotope variations within the 

Flatreef and Merensky Reef of the Bushveld Complex, South Africa .........................................59 

5.1 Abstract ...........................................................................................................................59 

5.2 Introduction ......................................................................................................................60 

5.3 Regional Geology ............................................................................................................63 

5.3.1 Stratigraphy of the Northern Limb ........................................................................................ 64 

5.3.2 The Flatreef at Turfspruit and Macalacaskop .................................................................... 66 

5.3.3 The Merensky Reef at Dwarsriver ....................................................................................... 69 

5.4 Samples and analytical methods .....................................................................................69 

5.5 Results ............................................................................................................................71 

5.6 Discussion .......................................................................................................................76 

5.6.1 Sulfur isotope variation .......................................................................................................... 77 

5.6.2 Model for Flatreef formation.................................................................................................. 78 

5.6.3 S isotope variation of the Merensky Reef at TRP ............................................................. 79 

5.6.4 Comparison of S isotope composition in the Flatreef and Merensky Reef at TRP ...... 80 

5.7 Conclusion .......................................................................................................................81 

5.8 Acknowledgements .........................................................................................................81 

5.9 References ......................................................................................................................82 



 
IX 

 

CHAPTER 6 : ARTICLE 3 Neodymium isotope variations in the Flatreef on Macalacaskop, 

northern limb, Bushveld Complex .............................................................................................88 

6.1 Abstract ...........................................................................................................................88 

6.2 Introduction ......................................................................................................................88 

6.3 The Rustenburg Layered Suite of the northern limb .........................................................92 

6.3.1 Marginal Zone (MZN) ............................................................................................................. 92 

6.3.2 Lower Zone (LZ) ..................................................................................................................... 92 

6.3.3 Critical Zone (CZ) ................................................................................................................... 93 

6.3.4 Main Zone (MZ) ...................................................................................................................... 95 

6.3.5 Upper Zone (UZ) .................................................................................................................... 96 

6.4 Location of samples analysed ..........................................................................................96 

6.5 Analytical methods ..........................................................................................................98 

6.6 Results .......................................................................................................................... 100 

6.7 Discussion ..................................................................................................................... 107 

6.7.1 Comparison between Platreef/Flatreef and RLS of the remainder of the BC ............ 107 

6.7.2 Sr-Nd isotopic constraints on the petrogenesis of the Flatreef .................................... 109 

6.8 Conclusions ................................................................................................................... 112 

6.9 Acknowledgements ....................................................................................................... 112 

6.10 References .................................................................................................................. 112 

CHAPTER 7 : DISCUSSION AND CONCLUSIONS................................................................ 121 

7.1 Summary of key findings ............................................................................................... 121 

7.1.1. Summary of key findings emanating from Chapter 4 .................................................. 121 

7.1.2. Summary of key findings emanating from Chapter 5 .................................................. 121 

7.1.3. Summary of key findings emanating from Chapter 6 .................................................. 122 

7.2 Correlation of Platreef/Flatreef with the UCZ in the remainder of the BC ....................... 122 

7.3 Towards a petrogenetic model for the Flatreef ............................................................... 123 

7.4 Conclusions and Recommendations .............................................................................. 128 

REFERENCES ....................................................................................................................... 129 

LIST OF APPENDICES .......................................................................................................... 143 

APPENDIX A ...................................................................................................................... 143 

For Chapter 4, Article 1. .................................................................................................... 143 

APPENDIX B ...................................................................................................................... 153 

For Chapter 5: Article 2 ..................................................................................................... 153 

APPENDIX C ...................................................................................................................... 177 

For Chapter 6: Article 3 ..................................................................................................... 177 



 
X 

 

 

  



 
XI 

 

LIST OF FIGURES 

CHAPTER 1 

Figure 1.1 a-e Global demand for Pt, Pd, Ir, Ru and Rh, respectively, over selected years. 

Modified after Thormann et al. (2017).  f Pt, Pd and Rh (referred to as ‘pgm’) demand from the 

auto catalyst industry and light vehicle output between 2016 and 2020. Modified after Cowley 

(2021). ....................................................................................................................................... 2 

Figure 1.2 a Graph showing the global supply of PGE; modified after Mudd et al. 2018. b Inset, 

PGE mines and mine projects in the RLS of the BC, modified after Thormann et al. (2017). Pt/Pd 

ratios, as calculated by Grobler et al. (2019) for Merensky Reef and Platreef from several online 

resource statements, is included in blue text .............................................................................. 3 

CHAPTER 2 

Figure 2.1 Simplified geological map of the BC indicating the different stratigraphic zones of the 

RLS. Modified after Eales and Cawthorn (1996); Yudovskaya et al. (2013). TML= Thabazimbi 

Murchison Lineament. ................................................................................................................ 8 

Figure 2.2 a Location of the Kalahari Craton in Southern Africa. b Major units and significant 

terrane boundaries (suture zones) of the Kalahari Craton, and location of the Bushveld Complex. 

Modified after Zeh et al. (2015). ................................................................................................. 8 

Figure 2.3 Schematic cross-section of the RLS and its marginal rocks. Modified after von 

Gruenewaldt et al. (1985). .........................................................................................................10 

Figure 2.4 Generalised stratigraphy of the western and eastern limbs of the RLS. Major 

subdivisions, main rock types and thickness of each zone of the RLS are indicated. Positions of 

lower (LG), middle (MG) and upper group (UG) chromitites are shown. Variations in the initial 

87Sr/86Sr ratio for whole-rocks and plagioclase (separates) as per Kruger (1994) are also shown. 

Modified after Eales and Cawthorn (1996) and Kruger (2005). ..................................................11 

Figure 2.5 SE-NW schematic section illustrating the variations in Pre-Merensky and Merensky 

pyroxenites in the southern BC. The thick reef facies is found mostly in the east adjacent to the 

Pilanesburg intrusion in the extreme NW. Structural induced irregularities in the Merensky footwall 

influences the variation of the types of reef observed laterally. Modified after Naldrett et al. (2009).

 .................................................................................................................................................16 

Figure 2.6 Geological map of the northern limb of the BC. After McDonald and Holwell (2011) 

based on van der Merwe (1976). The study area is indicated by the red rectangle. ..................19 

Figure 2.7 Schematic stratigraphic columns of the RLS, showing the putative correlation between 

stratigraphic units of the northern limb and eastern / western limbs. An inferred correlation is 



 
XII 

 

indicated between the Merensky Reef and Platreef. Modified after White (1994) and McDonald 

and Holwell (2011). ...................................................................................................................20 

Figure 2.8 Schematic section representing the relationship between Platreef, Flatreef and the 

deep downdip extension of the Flatreef at Turfspruit, and the position of its main mineralised 

zones. Large sedimentary xenoliths are metamorphosed and assimilated by intruding Bushveld 

magma. Modified after Grobler et al. (2019)………………………………………………………….23 

CHAPTER 3 

Figure 3.1 Stratigraphic log of Flatreef magmatic stratigraphy including rock types as intersected 

by drill holes a UMT-393 at Macalacaskop and b UMT-276 at Turfspruit. Drill hole UMT-393 

comprises of relatively thicker stratigraphic unit and overlies LZ whereas UMT-276 directly 

overlies a sedimentary calc-silicate assimilation zone known as the Footwall Assimilation Zone 

(FAZ). .......................................................................................................................................25 

Figure 3.2 Drill core photographs of UMT-393. a MZ norite grades into mottled anorthosite of 

HW2 that marks the top of the Flatreef, b interlayered norite-pyroxenite at various thicknesses of 

the HW1. FPX = feldspathic pyroxenite, NOR = norite. .............................................................26 

Figure 3.3 Drill core photographs of UMT-393 Bastard Reef pyroxenite indicating a net textured 

base metal sulfides present in close proximity to the chromitite stringer, b chromitite stringer with 

sharp undulating contact. Scale in mm. FPX=feldspathic pyroxenite, CR= chromitite stringer. ..27 

Figure 3.4 Drill core photograph of MD comprising of feldspathic pyroxenite (FPX) with a granitic 

vein (GRV) cross cutting the lithologies. ....................................................................................27 

Figure 3.5 Drill core photograph of a the thin upper chromitite stringer at the contact between MD 

and upper Merensky Reef (M2). The presence of a pegmatoidal pyroxenite layer (PPXT) and a 

thick granitic vein (GRV) is also observed in the M2 orthopyroxenite (OPX). b A close-up image 

of the upper chromitite stringer in (a). Base metal sulfide concentrations increase in close 

proximity to the stringer. c Pegmatoidal pyroxenite (PPXT) of the M1U. d Serpentinised 

harzburgite of the M1L. e An approximately 2 m thick mafic pegmatoidal vein (MPV) cross cutting 

pyroxenite near the base of the M2. ..........................................................................................29 

Figure 3.6 Drill core photographs of UMT-393 FCU at a depth of approximately 960 m a typical 

interlayered norite-pyroxenite-anorthosite present in the unit, b poorly developed norite cycles 

with altered hybrid lithologies, c harzburgite of LZ affinity that underlies the FCU. ....................30 

Figure 3.7 Drill core photographs of UMT-276 showing a sharp contact of overlying MZ 

gabbronorite and HW2 mottled anorthosite, b a relatively thin Merensky Reef package consisting 

of M2 feldspathic pyroxenite, M1U pegmatoidal pyroxenite and M1L feldspathic harzburgite. c 

altered and hybrid lithologies  representing calc-silicate assimilation of the FAZ. ......................32 



 
XIII 

 

CHAPTER 4 

Figure 4.1 Schematic longitudinal section through the Platreef over the entire strike length 

modified after Kinnaird et al. 2005. Note that between Witrivier and Altona a significant strike 

length is not shown ...................................................................................................................35 

Figure 4.2 a Schematic illustration of a stratigraphic comparison between Platreef/Flatreef with 

the Upper Critical Zone (UCZ) in the western BC. b Stratigraphy of the Flatreef in borehole UMT-

393, Macalacaskop. Modified after Grobler et al., 2019. Cr = chromite stringer .........................38 

Figure 4.3 Photomicrographs showing typical petrological features of the FCU, M1L and M2 in 

cross-polarized, transmitted (A-G) and plane-polarized, reflected light (H). A Bent intercumulus 

plagioclase with cumulus orthopyroxene displaying exsolution lamellae in the FCU feldspathic 

pyroxenite; B Clinopyroxene oikocryst with reactionary relationships with included 

orthopyroxenes in norite of the FCU; C Orthopyroxene crystal with anhedral plagioclase inclusions 

in a FCU norite sample; D  Bent orthopyroxene crystal from a poorly developed norite cycle in the 

FCU; E Serpentinised olivine accompanied by euhedral chromite and mesh textured magnetite in 

the Merensky Reef (M1L); F Undulose extinction of serpentinized olivine in M1L harzburgite; G 

Equilibration textures in FCU pyroxenite adjacent to M1L, note apparent 120˚ triple junctions 

indicated by red circles; H Chromite stringer found near the base of the Merensky Reef (M1L). 40 

Figure 4.4 Photomicrographs of typical rock textures of the Flatreef under cross-polarized, 

transmitted (A, B, D-H) and plane-polarized, reflected light (C). A Orthopyroxenite comprised 

mainly of impinging orthopyroxene crystals in the Merensky Reef (M2); B Deformed 

orthopyroxene with exsolution lamellae in M2 feldspathic pyroxenite; C pentlandite (Pn) and 

chalcopyrite (Cp), overgrown by actinolite in the MD. D Clinopyroxene oikocryst with plagioclase 

chadacrysts in anorthosite from HW2. Note the zonation of plagioclase. E HW feldspathic 

pyroxenite with fine grained orthopyroxene interstitial to coarser orthopyroxene. F Norite with 

cumulus orthopyroxene and plagioclase from HW1; G Clinopyroxene oikocryst with plagioclase 

and orthopyroxene inclusions in a porphyritic textured pyroxenite in HW1; H Plagioclase 

inclusions exhibiting wide twin lamellae within an orthopyroxene from a HW norite sample. .....42 

Figure 4.5 Variation in mineral modal proportions, Pt + Pd (from assay data, Ivanplats), 

plagioclase An content and initial 87Sr/86Sr of plagioclase with increase in stratigraphic height 

across the various lithologies of the Flatreef in borehole UMT-393.The 2SE of the 87Sr/86Sri is 

represented by the error bars. ...................................................................................................48 

Figure 4.6 Variation of 87Sr/86Sri in plagioclase versus An% of plagioclase for rocks of the Flatreef 

on Macalacaskop. Labelled fields show data for the Western Limb from Karykowski et al. (2017) 

and Yang et al. (2013). Data for the Northern Limb come from Mangwegape et al. (2016). UZ = 



 
XIV 

 

Upper Zone; MZ = Main Zone; UCZ = Upper Critical Zone; LCZ = Lower Critical Zone; LZ = Lower 

Zone. ........................................................................................................................................51 

 

 

CHAPTER 5 

Figure 5.1 Simplified map showing the geology of the Bushveld Complex. Modified after Eales 

and Cawthorn (1996). TML Thabazimbi-Murchison Lineament .................................................61 

Figure 5.2 Schematic diagram of a stratigraphic comparison between the Platreef/Flatreef and 

Upper Critical Zone (UCZ) in the western BC. UMCR, UCR, LCR, PNZ, LZ, BAR, MR, and PSR 

indicate the uppermost chromite seam, upper chromite seam, lowwer chromite seam, Pyroxenite-

Norite Zone, Lower Zone, Bastard Reef, Merensky Reef, and Pseudo Reef, respectively ........63 

Figure 5.3 A Geology of the Northern Lobe of the Bushveld Complex showing the locations of 

boreholes UMT-276 and UMT-393 on the farms Turfspruit and Macalacaskop. (After van der 

Merwe 1976, as modified by Ashwal et al. 2005). A detailed geology of the study area is given on 

the right as modified after Maier et al. 2021. B Geology of the Eastern Lobe of the Bushveld 

Complex showing the location of the study area at Dwarsriver 372KT (modified after Cameron & 

Abendroth, 1957; Sharpe & Chadwick, 1982; Clarke et al., 2005; Beukes et al., 2016). Key 

features such as the Merensky Reef outcrop and the main decline of the Two Rivers Platinum 

mine from which underground samples were taken, are shown on the right. .............................68 

Figure 5.4 Dip section showing borehole UMT-276. Note that the footwall comprises of a zone of 

sedimentary rock assimilation ...................................................................................................70 

Figure 5.5 Dip section showing borehole UMT-393. Note that the FCU is underlain by the LZ. 71 

Figure 5.6 Variations in a CaO/Al2O3; b loss on ignition (LOI); c S wt. % (Ivanplats Assay data); 

d Pt +Pd in ppm (Ivanplats Assay data); e δ34S; with depth in borehole UMT393. f Variation of S 

isotope composition in the pyroxenite reef (PR PXT) and harzburgite reef (PR HA) of the Platreef, 

and the ultramafic rocks of the LZ with depth in UMT-006 after Yudovskaya et al. 2017. ..........74 

Figure 5.7 Variations in a CaO/Al2O3; b loss on ignition (LOI); c S wt. % (Ivanplats Assay data); 

d Pt +Pd in ppm (Ivanplats Assay data); e δ34S; with depth in borehole UMT276. .....................75 

Figure 5.8 Variation of a CaO/Al2O3; b loss on ignition (LOI) and c δ34S with stratigraphic height 

at a mining face at TRP, eastern BC. ........................................................................................76 

Figure 5.9 Comparison of compositional data of UMT393, UMT276, UMT094 (Keir-Sage et al. 

2021) and MR at TRP (EBC). a S vs. δ34S and b CaO/Al2O3 vs δ34S. .......................................81 

CHAPTER 6 



 
XV 

 

Figure 6.1 a Regional map of the Bushveld Complex displaying the different limbs and simplified 

geology modified after Eales & Cawthorn (1996) and Yudovskaya and Kinnaird (2010). TML = 

Thabazimbi-Murchison Lineament. b Simplified geological map of the Bushveld Complex showing 

the locations of boreholes UMT-276 and UMT-393 on the farms Turfspruit and Macalacaskop. 

(After van der Merwe 1976, as modified by Ashwal et al., 2005) ...............................................90 

Figure 6.2 Schematic representation of the association between Platreef, Flatreef and the deep 

Flatreef extension at Turfspruit. Large sedimentary xenoliths are metamorphosed and assimilated 

by intruding Bushveld magma. Modified after Grobler et al. (2019). ..........................................94 

Figure 6.3 Simplified stratigraphic column of the Flatreef in borehole UMT-393, Macalacaskop. 

Modified after Beukes et al. (2021). Cr = chromitite stringer ......................................................97 

Figure 6.4 Dip section showing the various stratigraphic units of the Flatreef overlying LZ as 

intersected by borehole UMT-393. Modified after Keet et al. (2021) ..........................................98 

Figure 6.5 Chondrite-normalized REE plot of the Flatreef samples analysed in this study. The 

area in grey represents data of the UCZ (excluding UG-2) as reported by Maier et al. (2013). 

Normalizing values are taken from Anders and Grevesse (1989). HW, UR, Md, M1U, M1L, and 

FCU indicate hanging wall, Upper Reef, Middling unit, M1 Upper, M1 Lower and the Footwall 

Cyclic unit, respectively. N, FPX, OPX and HA indicate norite, feldspathic pyroxenite, 

orthopyroxenite and harzburgite, respectively. ........................................................................ 101 

Figure 6.6 Variation in mineral modes, a εNd (2.06 Ga), and b (87Sr/86Sr)i (2.06 Ga) with increase 

in stratigraphic height across Flatreef stratigraphy as intersected by borehole UMT-393. Error 

bars represent 2SE and 2σ for εNd and 87Sr/86Sri, respectively ............................................... 105 

Figure 6.7 Binary variation diagrams of a εNd vs 87Sr/86Sri, b Th/La vs 87Sr/86Sri, and c Th/La vs 

εNd in Flatreef rocks. WBC data from Maier et al. (2000) are included for comparison (grey 

symbols). Epsilon Nd and initial Sr values calculated for an age of 2.06 Ga ............................ 106 

Figure 6.8 Plot of εNd vs 87Sr/86Sri  for the a Flatreef (Abernethy 2020; this study), samples of the 

RLS of the western limb (Maier et al., 2000) and eastern limb (EBC, Lee and Butcher, 1990; 

Rains, 2014; Bourdeau et al., 2022) of the Bushveld Complex; b Flatreef samples of this study 

and adjacent farm, Turfspruit (Abernethy 2020) and Main Zone samples (Roelofse and Ashwal, 

2012; Mwenze, 2019; Abernethy, 2020; Scoon et al., 2020) of the northern limb;  c  Flatreef 

samples (this study; Abernethy, 2020) and Platreef samples (Pronost et al., 2008; Mwenze, 2019; 

Scoon et al., 2020) of the northern limb................................................................................... 108 

Figure 6.9 Plot of εNd vs 87Sr/86Sri calculated at 2.06 Ga of whole-rock data for the Transvaal 

Supergroup, Turfloop Batholith, Archaean granite gneiss, the Johannesburg Dome , the Rooiberg 

Group, the Vredefort Dome OGG and ILG, the Flatreef and a mantle melt end member (εNd =2, 



 
XVI 

 

87Sr/86Sri= 0.702, Maier et al., 2000); modified after Roelofse and Ashwal (2012). It is clear that 

the isotopic composition of the Flatreef cannot be explained through the interaction of a mantle-

derived melt and locally available contaminants as exposed along the base of the Platreef. ... 110 

Figure 6.10 Isotopic mixing models for mixtures between mantle-derived melt (εNd
i =2.0, 

87Sr/86Sri=0.702, Maier at al. 2000; Nd =14 ppm, Sr =216 ppm, Roelofse and Ashwal 2012) and a 

60:40 B1:B2 mixture (εNd
i =-5.8; 87Sr/86Sri=0.706; Nd =13 ppm; Sr =236 ppm; Curl, 2001) 

respectively, and devolatized Transvaal Supergroup carbonates (εNd
i =-6.4; 87Sr/86Sri=0.709; Nd 

=4 ppm; Sr =60 ppm; Abernethy, 2020). ................................................................................. 111 

CHAPTER 7 

Figure 7.1 An illustration showing the transition from near surface Platreef to the UCZ with depth 

on the Turfspruit locality. Modified after Stephenson (2019). ................................................... 124 

Figure 7.2 A schematic representation showing the sill-like emplacement of the mineralized 

portions of the Flatreef. A Emplacement of LZ and CZ magmas into basinal structures. B Sill-like 

emplacement of the mineralized units of the Flatreef. Where magma is brought into contact with 

country rocks, the erratic Platreef characterised by disturbed S-, Sr- and Nd-isotopic compositions 

is developed. Insets B1 and B2 show the displacement of extant CZ magma as a result of the sill 

intrusions. C The MZ is emplaced following a hiatus during which the Platreef / Flatreef solidified, 

leading to cross-cutting relationships between Platreef and MZ as seen at Sandsloot and 

Zwartfontein. ........................................................................................................................... 127 

  



 
XVII 

 

LIST OF TABLES 

CHAPTER 4 

Table 4-1 Average in situ Sr isotope compositions and anorthite contents of plagioclase crystals 

in the various rock types of the Flatreef. Initial 87Sr/86Sr calculated at 2054.89 Ma. The full 

isotopic dataset is given in Table A-1 (Appendix A). ..................................................................46 

CHAPTER 6 

Table 6-1 Selected trace elements and REE (ppm). ............................................................... 102 

Table 6-2 εNd and initial 87Sr/86Sri data calculated for an age of 2.06 Ga. Sr isotope values are 

averages of (87Sr/86Sr)i reported by Beukes et al. 2021. ........................................................... 104 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 



 
XVIII 

 

LIST OF APPENDICES 

APPENDIX A for Chapter 4 

APPENDIX B for Chapter 5 

APPENDIX C for Chapter 6



 
1 

 

CHAPTER 1 :  

INTRODUCTION 

 

1.1 Study context 

With the continuous advancement in environmentally-related technologies such as catalytic 

converters, hydrogen fuel cells, electronics, etc., the global demand for platinum group 

elements (PGE) is continuously growing in modern society (Mudd et al., 2018) (Fig. 1.1). Of 

the multiple applications that use PGE, the automotive industry has the highest demand for 

Pt, Pd and Rh (Fig. 1.1a, b and e) and the electronics industry the highest demand for Ru and 

Ir (Fig. 1.1c and d). The past decade has seen the demand for Pd overtaking that of Pt in the 

automotive industry (Fig 1.1 a, b and f) (Cowley, 2021).  South Africa is one of the leading 

producers of PGE in the world (Fig. 1.2a), with vast deposits located in the Bushveld Complex 

(BC) (Fig. 1.2b) (Thormann et al., 2017). With an area of approximately 20000 km2 and an 

average thickness of approximately 4 km north (Finn et al., 2015), and approximately 65000 

km2, and nearly 8 km thick south of the Thabazimbi-Murchison Lineament (TML) (Eales and 

Cawthorn, 1996; Cawthorn and Walraven, 1998), the Rustenburg Layered Suite of the BC is 

not only the largest layered igneous intrusion on Earth, but it is also the largest known PGE 

reserve (Cawthorn, 2010). Other metals such as V, Cr, Ni and Co are also enriched in this 

intrusion. The layered part of the complex outcrops as four distinct limbs, namely the eastern, 

western, far western and northern limb (Fig. 1.2b). The BC is formally divided into three main 

suites, namely the Rustenburg Layered Suite (RLS), consisting of mafic to ultramafic layered 

rocks; the Lebowa Granite Suite, a succession of granites; and the Rashoop Granophyre 

Suite, consisting of granophyres and granites (Eales and Cawthorn 1996, Kruger 2005). Three 

key stratigraphic units within the RLS are associated with PGE-Ni-Cu mineralization, namely, 

the Merensky Reef, the UG-2 Reef and the Platreef. Mining operations are therefore 

concentrated close to where these reefs outcrop (Fig. 1.2b). The recently discovered Flatreef, 

which is the main focus of this study, is located in the northern limb of the BC, and represents 

the down-dip, sub-horizontal extension of the Platreef, the latter which outcrops along much 

of the eastern margin of the northern limb, where it is in contact with Transvaal Supergroup 

sedimentary rocks and Archaean granite basement (van der Merwe 2008).  

Significant interaction with floor rocks is thought to have contributed to the apparent erratic 

and complex nature of the Platreef. This complexity made it difficult to correlate the Platreef 

with the Merensky Reef in the remainder of the BC even though it is located at a similar 

stratigraphic position to the latter.  The Flatreef was discovered in the course of a deep drilling 

programme conducted by Ivanplats in 2007 (Grobler et al. 2019). This discovery changed the 



 
2 

 

narrative regarding Platreef mineralization in the northern limb relative to the rest of the BC as 

previous studies in the northern limb focused on the relatively “shallow” Platreef that has been 

subjected to intense assimilation and contamination by footwall rock.  

 

Figure 1.1 a-e Global demand for Pt, Pd, Ir, Ru and Rh, respectively, over selected years. Modified 
after Thormann et al. (2017).  f Pt, Pd and Rh (referred to as ‘pgm’) demand from the auto catalyst 
industry and light vehicle output between 2016 and 2020. Modified after Cowley (2021). 



 
3 

 

 

Figure 1.2 a Graph showing the global supply of PGE; modified after Mudd et al. 2018. b Inset, PGE 
mines and mine projects in the RLS of the BC, modified after Thormann et al. (2017). Pt/Pd ratios, as 
calculated by Grobler et al. (2019) for Merensky Reef and Platreef from several online resource 
statements, is included in blue text 

1.2 Research aims and objectives 

The growing economic importance of the Platreef in the northern limb of the BC is undeniable. 

Advances in mineral exploration of the Platreef in recent years have led to the northern limb 

being the focus of intensive research, where historically the focus was mostly on the western 

and eastern limbs of the BC. The discovery of the Flatreef, i.e., the down dip, sub-horizontal 

extension of the Platreef, allows for the study of a magmatic stratigraphy where the influence 

of local footwall interaction is negligible (Grobler et al. 2019). The aim of this study is to 

establish whether the Flatreef correlates with the Merensky and Bastard Cyclic Units, which 

constitutes the transitional zone between the Upper Critical Zone and Main Zone of the 

western and eastern BC, based on its isotopic signature. 

The principal objectives are as follows: 

• To establish the Sr-, S- and Nd-isotopic stratigraphy of the Flatreef.  

• Comparing variations in the Sr-, S- and Nd-isotopic composition across the Flatreef 

with variations across the Merensky and Bastard reefs in order to establish whether 

the Flatreef may be considered a correlate of the Merensky and Bastard reefs as 

known from the eastern and western limbs of the BC.  



 
4 

 

• To constrain the petrogenesis of the Flatreef and to consider the implications thereof 

for the petrogenesis of the Merensky Reef. 

1.3 Thesis Layout and Author Contributions 

This is a publication-based thesis consisting of seven chapters, three of which are presented 

as journal-style published / submitted papers. Information such as the regional geology and 

specific explanations may be repeated in these chapters, however, each journal-style chapter 

focuses on one or more of the objectives listed above. This introductory chapter provides a 

broad overview of the global PGE demand and supply (with special reference to the BC) as 

well as the outline of this study.  

Chapter 2: This chapter provides an introduction to the regional geology of the BC as well as 

the northern limb, summarizing key aspects of available literature regarding its stratigraphy, 

mineralization and emplacement. 

Chapter 3: This chapter presents detailed descriptions of the Flatreef units intersected by 

boreholes UMT-393 and UMT-276.  

Chapter 4: Article 1 titled “Strontium isotope variations in the Flatreef on Macalacaskop, 

northern limb, Bushveld Complex: Implications for the source of platinum-group elements in 

the Merensky Reef”. This article was published in Mineralium Deposita in 2021.  

Authors: Jarlen J. Beukes*, Frederick Roelofse, Christoph D. K. Gauert, Danie F. Grobler and 

Henriette Ueckermann (*Jarlen J. Keet (née Beukes)) 

This article investigates strontium isotopic variations in plagioclase across the Flatreef and 

how these relate to the Merensky Reef in the remainder of the BC. 

Candidate’s contribution: 

The candidate was responsible for logging, sample collection, analyses, interpretation of data, 

the writing of the paper as well as figure production.  

The candidate contributed between 85 and 90% of the paper. 

Frederick Roelofse and Christoph Gauert assisted with data interpretation and discussion 

during preparation of the manuscript. 

Danie Grobler assisted in understanding the magmatic stratigraphy of the Flatreef. 

Henriette Ueckermann assisted with LA-MC-ICP-MS analyses.  



 
5 

 

Chapter 5: Article 2 titled “A comparative study of sulfur isotope variations within the Flatreef 

and Merensky Reef of the Bushveld Complex, South Africa”. This article was published in The 

Canadian Mineralogist in 2021. 

Authors: Jarlen J. Keet, Frederick Roelofse, Christoph D. K. Gauert, Danie F. Grobler and 

Mike Butler 

This article provides whole-rock S isotope data collected across the Flatreef as intersected by 

boreholes UMT-276 and UMT-393, as well as across the Merensky Reef at Two River’s 

Platinum, eastern BC. 

The candidate contributed between 75 and 85% of the paper.  

The candidate was responsible for the logging, sample collection, preparation of samples for 

analyses, data interpretation and write-up of the paper. 

The candidate was responsible for the production of all figures except for figures 4 and 5. 

Frederick Roelofse and Christoph Gauert assisted with data interpretation and discussion 

during preparation of the manuscript. 

Danie Grobler assisted in understanding the magmatic stratigraphy of the Flatreef. 

Mike Butler conducted the S isotope analyses. 

Chapter 6: Article 3 titled “Neodymium isotope variations in the Flatreef on Macalacaskop, 

northern limb, Bushveld Complex”. The present version of this article represents a revised 

version of the article originally submitted to Mineralium Deposita in early 2022. The revised 

article is currently under review with Mineralium Deposita.  

Authors: Jarlen J. Keet, Frederick Roelofse, Christoph D. K. Gauert, Linda Iaccheri, Danie F. 

Grobler and Henriette Ueckermann 

The candidate contributed between 75 and 80% of the paper.  

The candidate was responsible for the logging, sample collection, preparation of samples for 

analyses, data interpretation and write-up of the paper. 

The candidate was responsible for the production of all figures. 

Frederick Roelofse and Christoph Gauert assisted with data interpretation and discussion 

during preparation of the manuscript. 

Linda Iaccheri prepared the samples for Nd-isotopic analysis.  

Danie Grobler assisted in understanding the magmatic stratigraphy of the Flatreef. 



 
6 

 

Henriette Ueckermann assisted with the Nd-isotopic determinations. 

Chapter 7: This chapter brings together data sets and main ideas from the results chapters, 

i.e., journal style chapters and additional data, to achieve the aims set out in chapter 1 through 

a detailed discussion. 

Note that spelling (American vs. British spelling) and / or stylistic variations may be 

encountered in chapters 4-6 due to the requirements of the journals to which the work was 

submitted.  

1.4 Methodology 

Data for this thesis were collected using a range of analytical techniques, a detailed account 

of which is provided in each article chapter. 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 



 
7 

 

CHAPTER 2 :  

GEOLOGICAL BACKGROUND 

 

2.1 The Bushveld Complex 

The Bushveld Magmatic Province consists of the felsic volcanic Rooiberg Group, 

representative of the initial phase of Bushveld-related magmatism (Buchanan et al., 2002) and 

the following intrusive suites: a suite of early mafic Bushveld sills, the RLS (which many 

authors incorrectly call the Bushveld Complex), consisting of ultramafic-mafic layered rocks, 

the Rashoop Granophyre Suite and the Lebowa Granite Suite (Kruger, 2005) (Fig. 2.1). The 

BC, located in the north-eastern part of South Africa, was emplaced into the central portion of 

the Kalahari Craton, upon the Witwatersrand and Pietersburg blocks of the Kaapvaal Craton, 

at approximately 2.06 Ga (Zeh et al., 2015) (Fig 2.2). Major cratonic lineaments such as the 

TML are possible representatives of suture zones through which younger domains accreted 

during the development of the Kaapvaal Craton (Elington and Armstrong, 2004). The 

sedimentary lithologies of the Transvaal Supergroup surround the BC and these sediments 

are located within the Transvaal and Griqualand West basins, where they typically overlie 

Archaean basement (Bekker et al. 2001). The RLS south of the TML is estimated to have an 

areal extent of 65000 km2 and a thickness of about 8 km (Eales and Cawthorn, 1996; 

Cawthorn and Walraven, 1998), making it the world’s largest known layered igneous intrusion. 

Recent geophysical data on the BC north of the TML (Finn et al., 2015) increases the total 

areal extent of the BC to approximately 85000 km2.  



 
8 

 

 

Figure 2.1 Simplified geological map of the BC indicating the different stratigraphic zones of the RLS. 

Modified after Eales and Cawthorn (1996); Yudovskaya et al. (2013). TML= Thabazimbi Murchison 

Lineament. The Lebowa Granite and Rashoop Granophyre Suites are combined in the legend.  

 

Figure 2.2 a Location of the Kalahari Craton in Southern Africa. b Major units and significant terrane 

boundaries (suture zones) of the Kalahari Craton, and location of the Bushveld Complex. Modified after 

Zeh et al. (2015). TSZ – Triangle shear zone; ZPS – Zoetfontein–Palala shear zone; TML – 

Thabazimbi–Murchison-Lineament 



 
9 

 

The RLS outcrops in four different geographical areas, namely, the eastern, the northern, the 

western and far western limbs (Eales and Cawthorn, 1996, Fig. 2.1). A fifth limb, known as the 

southern (Bethal) limb, is located under younger Karoo sedimentary cover. Only the northern 

limb is located north of the TML, with the remaining four limbs located to the south. The 

recently discovered ultramafic-mafic succession known as the Waterberg Project could 

potentially extend the northern limb’s border further north (Huthmann et al., 2016, 2018; 

Kinnaird et al. 2017). Considering isostasy of the crust following the emplacement of the 

enormous BC, Cawthorn and Webb (2001) reinterpreted gravity data and suggested a 

connection at depth between the arcuate eastern and western limbs. This hypothesis is further 

supported by seismic and magnetic modelling (Webb et al., 2004; Cole et al., 2013). However, 

connectivity between the remainder of the limbs remains debatable (Cawthorn, 2015).  

2.1.1 The Rustenburg Layered Suite 

The RLS is a layered ultramafic-mafic package, approximately 9 km thick, subdivided vertically 

into distinct zones on the basis of variations in its mineralogy and petrology (Hall, 1932) or 

major shifts in isotopic composition (Kruger, 2005). These units are, from bottom to top: the 

Marginal (MZN), Lower (LZ), Critical (CZ), Main (MZ) and Upper (UZ) zones (Fig. 2.1). The 

RLS was emplaced along a regional unconformity between the Rooiberg Group volcanics and 

the Transvaal Supergroup sediments or Archaean basement granites and gneisses (von 

Gruenewaldt et al., 1985; Eales and Cawthorn, 1996; Bekker et al., 2001).  

The zones of the RLS consist of a variety of rock types and different magmas were apparently 

responsible for their formation. Earlier studies (e.g., Davies et al. 1980; Sharpe, 1981; 

Cawthorn and Davies, 1983; Harmer and Sharpe, 1985) proposed that the RLS was produced 

by three distinct magma types. The first magma type, B1, which is MgO-rich, produced the 

Lower Zone and the lower Critical Zone. The second magma type is the Al2O3-rich B3 magma, 

which produced the rocks of the Main Zone, whereas the intermediate B2 magma is thought 

to have produced the rocks of the upper Critical Zone. However, more recent studies (e.g., 

Wilson, 2012) have argued that the exact number of magmas responsible for the formation of 

the LZ and underlying units was unclear. A wide variety of pre- and syn-BC mafic sills intruded 

the country rocks surrounding the BC (Sharpe and Hulbert, 1985; von Gruenewaldt et al., 

1985) (Fig. 2.3). Whole-rock Sr isotope data recorded by Kruger (1994) were interpreted to 

suggest that the lower MZ, CZ and LZ formed during an open-system “Integration Stage” that 

was characterised by multiple injections of fresh mafic magma; whereas the upper MZ and UZ 

formed during a closed-system “Differentiation Stage”, which was dominated by large-scale 

fractional crystallization and limited injections of new magma (Fig. 2.4). 

 



 
10 

 

 

Figure 2.3 Schematic cross-section of the RLS and its marginal rocks. Modified after von Gruenewaldt 

et al. (1985). 



 
11 

 

 

Figure 2.4 Generalised stratigraphy of the western and eastern limbs of the RLS. Major subdivisions, 

main rock types and thickness of each zone of the RLS are indicated. Positions of lower (LG), middle 

(MG) and upper group (UG) chromitites are shown. Variations in the initial 87Sr/86Sr ratio for whole-rocks 

and plagioclase (separates) as per Kruger (1994) are also shown. Modified after Eales and Cawthorn 

(1996) and Kruger (2005). 

2.1.1.1 Marginal Zone (MZN) 

Due to poor exposure and limited drilling, the provenance of this lowermost succession of the 

RLS remains poorly understood (Wilson, 2015). The MZN is the most enigmatic of the RLS 

zones (Cawthorn, 2015) and in some areas it is missing from the RLS succession (Wilson, 

2012). Where present, it consists of variably fine-grained norite, pyroxenite and gabbronorite, 

with a thickness varying between 150 m and 800 m (Eales and Cawthorn, 1996). Due to the 

common occurrence of quartz and biotite associated with quartzitic and dolomitic xenoliths 

from the Transvaal Supergroup within the MZN rocks, Cawthorn et al. (2006) and Cawthorn 

(2007) proposed that the MZN is unlikely to represent the true chilled margins of the RLS, but 

more likely, contaminated magmas that rapidly crystallized before the emplacement of the 

voluminous layered intrusion. A previously unknown Basal Ultramafic Series (BUS) was 

reported in the eastern limb of the BC, occurring below the MZN (Wilson, 2015).   



 
12 

 

2.1.1.2 Lower Zone (LZ) 

The LZ cumulate rocks are dominated by orthopyroxene and lesser olivine (Cawthorn, 2007). 

The rocks of the LZ include dunite, harzburgite and pyroxenite (Cameron, 1978). The LZ lacks 

consistency in thickness and lateral continuity throughout the BC. The floor topography largely 

controls the thickness of the LZ (Eales and Cawthorn, 1996). In the Olifants River Trough in 

the eastern limb, where the LZ is best exposed, the LZ is up to 1584 m thick and is subdivided 

into a Lower Pyroxenite subzone, a Harzburgite subzone, and an Upper Pyroxenite subzone 

(Cameron, 1978). Further to the south, in the Clapham area, the LZ is approximately 900 m 

thick (Wilson, 2015). In the western limb, the LZ is about 1000 m thick and is subdivided into 

two subzones. Relative to the lower subzone that is dominated by pyroxenite and olivine- 

bearing rocks, the upper subzone consists of less pyroxenite (Teigler and Eales, 1996).   The 

distribution and thickness of the LZ in the northern limb is irregular. It occurs as seven isolated 

satellite bodies and as an apparent xenolithic inclusion along the eastern margin of the mafic 

intrusion (de Villiers 1970; van der Merwe 1976; van der Merwe, 2008). It is usually separated 

from the overlying mineralized mafic package by intervals of country rock (Yudovskaya et al. 

2013).  Economically important chromitite seams have been documented in the LZ in the 

northern limb but are not present in the LZ of the eastern and western limbs (Hulbert and von 

Gruenewaldt, 1982). Except for an approximately 90 cm thick noritic layer occurring midway 

in the LZ succession, which has been identified in both the western and eastern limbs, no 

cumulus plagioclase is present in the LZ (Lee and Tredoux, 1986; Teigler, 1990). Although 

similarity in rock types and the cyclical development of rock units is observed in the various 

LZ successions in the different limbs of the BC, Wilson (2015) concluded, upon assessing 

details of these LZ sequences, that no precise correlation exists.   

2.1.1.3 Critical Zone (CZ) 

The CZ, located stratigraphically above the LZ (Fig. 2.4), is economically significant as it hosts 

world class PGE deposits, most notably within the UG-2 Reef, Merensky Reef and Platreef 

(Naldrett et al. 2009). It is between 1300 m and 1800 m thick and is characterized by the 

appearance of cumulus chromite (Naldrett et al. 2009). Based on the first appearance of 

cumulus plagioclase, the CZ is subdivided into 2 subzones, namely, the Upper Critical Zone 

(UCZ) and the Lower Critical Zone (LCZ) (Cameron, 1980, 1982). The LCZ, which is 

approximately 800 m thick, consists mostly of pyroxenite with occasional olivine-bearing units 

and chromitite layers (Cawthorn, 2015). These chromitite layers are grouped into two 

packages, the Lower Group (LG1-LG7, numbered from the bottom upwards) and the Middle 

Group (MG1 and MG2) (Fig. 2.4). The boundary between the LCZ and UCZ occurs between 

the MG2 and MG3 chromitite layers (Eales and Cawthorn, 1996), where plagioclase becomes 

a cumulus phase.  



 
13 

 

The major chromitite layers of the UCZ are grouped into the Middle Group (MG3 and MG4) 

and Upper Group chromitites (UG1 and UG2) (Fig. 2.4). Disseminated chromite as well as 

several minor chromitite layers/stringers, such as the chromitite stringers bracketing the 

Merensky Reef pegmatoid, are also hosted by the UCZ (Maier et al. 2013). The UCZ may be 

subdivided into two parts. The succession of rocks in the lower part of the UCZ is mainly non-

cyclical and comprises of anorthosite, norite and minor orthopyroxenite (Naldrett et al. 2009). 

The upper part, however, consists of cyclic units containing part or all of chromitite, 

harzburgite, pyroxenite, norite and anorthosite, generally in that order (Naldrett et al. 2009). It 

is in this subunit of the UCZ where the stratiform PGE mineralization of the UG-2 and 

Merensky Reef is located.  

2.1.1.4 Main Zone (MZ) 

The MZ is 2200 to 3100 m thick and overlies the CZ (Ashwal et al., 2005; Fig. 2.4). The exact 

boundary between the CZ and MZ is not distinct and is currently taken as the top of the Bastard 

Cyclic Unit, where it is capped by the Giant Mottled Anorthosite (Eales and Cawthorn, 1996). 

Others have suggested that the boundary should be placed within the Merensky Cyclic Unit 

due to the presence of a distinct excursion in Sr-isotopic values (Kruger 1992, 2005, Cawthorn 

2015). The MZ mostly consists of norite, gabbro, gabbronorite and anorthosite (Mitchell, 

1990). It lacks olivine (apart from in the northern limb) and chromian spinel, and is more 

homogeneous and unlayered compared to the CZ (Eales and Cawthorn, 1996). It was 

proposed by Kruger (2005) that the MZ be subdivided into a cyclic and heterogeneous Lower 

Main Zone and differentiated Upper Main Zone (Fig. 2.4). A significant layer, known as the 

Pyroxenite Marker, is located towards the top of the MZ (Eales and Cawthorn, 1996). It is 

characterized by a major reversal in mineral compositions and a distinct excursion towards 

lower 87Sr/86Sr isotope ratios (Sharpe, 1985, Fig. 2.4), and likely represents the level at which 

the last major influx of magma into the RLS occurred. The Pyroxenite Marker hosts minor PGE 

mineralization (Maier et al., 2001). 

2.1.1.5 Upper Zone (UZ) 

The UZ is stratigraphically located above the MZ and is the most laterally extensive zone of 

the RLS. The position of the base of the UZ is taken at the level where cumulus magnetite first 

appears within a mottled anorthosite layer (SACS 1980; Eales and Cawthorn, 1996). Others 

have suggested that the boundary between the MZ and overlying UZ should be placed at the 

level of the Pyroxenite Marker, concomitant with the observed shift in 87Sr/86Sr ratios (Kruger, 

1990, 1994, 2005, Fig. 2.4). The well-layered UZ is approximately 2800 m thick and is 

composed of basal noritic rocks and subsequent gabbroic rocks inclusive of gabbronorite, 

anorthosite and magnetite gabbro (Molyneux, 1970; von Gruenewaldt, 1973). The number of 

magnetitite layers identified across the RLS differs (Maier et al. 2013). In the eastern and 



 
14 

 

western limbs up to 26 magnetitite layers have been recorded (Cawthorn and Molyneux, 1986; 

Tegner et al., 2006). These magnetitite layers usually have sharp lower contacts and 

gradational upper contacts (Eales and Cawthorn, 1996). The UZ is subdivided into 3 subzones 

based on cumulus mineralogy (SACS, 1980; Eales and Cawthorn, 1996; Maier et al. 2013). 

The rocks of subzone A consist of cumulus plagioclase, pyroxene and magnetite and 

subzones B and C additionally contains olivine and apatite, respectively (Cawthorn, 2015).  

2.1.2 Mineralization in the Rustenburg Layered Suite of the western and eastern limbs 

The presence of world-class PGE deposits in the RLS, particularly in its CZ, is one of the 

important reasons why it has been studied intensively over the past century. By virtue of their 

PGE concentrations and lateral traceability it comes as no surprise that most of the BC 

research focused on the Merensky Reef and UG-2 stratiform horizons relative to its other 

PGE-enriched layers. These layers include the Bastard and Pseudo reefs which usually 

overlies and underlies the Merensky Cyclic Unit, respectively (Viljoen and Schürmann, 1998). 

They are not laterally continuous and have non-economic PGE concentrations (Naldrett et al., 

1986). The UG-2 represents the largest PGE resource in the world (Vermaak, 1985) with 

reported PGE grades exceeding 10 ppm (Kinnaird et al., 2002). With average PGE grades 

ranging between 5 and 10 ppm and Pd/Pt ratios of ~0.6 (Barnes and Maier, 2002; Cawthorn 

et al. 2005; Naldrett et al. 2009, Wilson and Chunnett, 2006; Osbahr et al., 2013), the 

Merensky reef represents the second largest PGE resource on Earth (Godel, 2015). The PGE 

mineralization in both the Merensky Reef and the UG-2 is associated with base metal sulfides 

which include pyrrhotite, pentlandite, chalcopyrite and minor pyrite (Osbahr et al., 2014).  

However, no consensus has yet been reached regarding the origin of the Merensky Reef and 

associated PGE mineralization. Two main schools of thought about the Merensky Reef exists, 

namely “Uppers” and “Downers” (Naldrett et al. 2009). The “Uppers” school argues that the 

source of metals in the Merensky Reef are from ascending fluids from underlying footwall 

cumulates which are enriched in PGE (e.g., Boudreau and McCallum, 1992; Boudreau, 2008; 

Mathez, 1995). These authors suggest that volatile-rich fluids, carrying dissolved sulfur and 

PGEs, mixed with the resident magma, causing sulfur saturation and associated deposition of 

PGMs.  In contrast, “Downers” propose that PGEs were scavenged from an overlying magma 

column by a dense sulfide melt that percolated downwards (e.g., Naldrett, 1989; Naldrett et 

al., 2008; Cawthorn, 2005). Traditionally it was proposed that sulfur saturation was caused by 

magma mixing (e.g., Campbell et al. 1983; Naldrett et al., 1986; Naldrett and von Gruenewaldt, 

1989; Li et al. 2001) or pressure changes rather than chemical processes in the magma 

chamber (Cawthorn, 2005). Alternatively, it has been proposed that stratiform PGE 

mineralization developed through in-situ crystallization (Latypov et al., 2015, 2017) or by 



 
15 

 

slumping of semi-consolidated cumulate slurries and associated hydrodynamic sorting (Maier 

et al., 2013).   

2.1.2.1 The Merensky Reef 

Wilson and Chunnett (2006) describe the term “Reef” as the mineralized zone located within 

medium- to coarse-grained plagioclase-pyroxenite with metal content that is considered 

economic to mine. “Merensky Reef” is a mining term (Lee, 1996) that can be defined as “a 

mineralized zone within or closely associated with ultramafic cumulates at the base of the 

Merensky (cyclic) unit” (Cawthorn et al., 2002). The Merensky Cyclic Unit (MCU), which 

contains the Merensky Reef at its base (Roberts et al., 2007), is found at the transitional zone 

between the UCZ and MZ, and is overlain by the petrographically similar, though sparsely 

mineralized Bastard Cyclic Unit (BCU) (Wilson and Chunnett, 2006). The MCU generally has 

a footwall consisting of anorthosite, norite or pyroxenite overlain by a basal chromitite layer (2 

to > 4 cm thick); followed by pegmatoidal feldspathic pyroxenite or feldspathic harzburgite and 

dunite of varying thickness that is usually capped by a 2 to 10 cm thick upper chromitite layer 

overlain by an orthopyroxenite sequence (e.g., Naldrett et al., 2009, Viljoen 1999, Cawthorn 

and Boerst, 2006). The Merensky Reef dips with a slope ranging from 9˚ to 27˚ towards the 

center of the BC, however in some localized areas dips as steep as 65˚ have been recorded 

(Lee, 1996). Large-scale mining of the Merensky Reef started in 1929 near Rustenburg in the 

southern sector of the Western Limb (Viljoen, 2016) 

The Merensky Reef was traditionally thought to occur at consistent stratigraphic levels and to 

be continuous over kilometers of strike length (Cawthorn et al. 2002). However, in actuality, 

the Merensky Reef varies in thickness, composition and the location of mineralization within it 

across the RLS of both the eastern and western BC (Viljoen, 2016, Fig. 2.5). The mines use 

the term, “facies”, to describe these lateral variations of the Merensky Reef. The Merensky 

Reef can generally be subdivided into a normal reef and potholed reef, where the normal reef 

displays the least degree of transgression and conformably overlies the footwall cumulates 

(Barnes and Maier, 2002, Fig. 2.5). The normal reef typically consists of a pegmatoidal 

feldspathic pyroxenite bracketed by two chromitite seams (Barnes and Maier, 2002). Up to 

four and as little as one chromitite seam may be present in the reef.  “Potholes” refer to where 

the Merensky Reef abruptly transgresses the footwall (Lee, 1996). Potholed reefs overlie 

footwall cumulates unconformably, where potholes cut between 1 and 100 m deep into the 

footwall (Latypov, 2015).  Potholed reefs are more dominant in the northwestern sector of the 

BC, and are located at a greater depth than the normal reef by up to 30 m (Barnes and Maier, 

2002). 



 
16 

 

 

Figure 2.5 SE-NW schematic section illustrating the variations in Pre-Merensky and Merensky 
pyroxenites in the southern BC. The thick reef facies is found mostly in the east adjacent to the 
Pilanesberg intrusion in the extreme NW. Structural induced irregularities in the Merensky footwall 
influences the variation of the types of reef observed laterally. Modified after Naldrett et al. (2009). 

2.1.2.2 The UG-2 Chromitite 

The RLS hosts several chromitite layers that are enriched in PGEs relative to their host rocks, 

however, the highest PGE concentrations are recorded for the UG-2 chromitite horizon 

(Kinnaird et al. 2002). Although chromite is mined from selected chromitite layers from all 3 

chromitite groups of the CZ (LG, MG and UG) (Viljoen, 2016), most of these chromitite layers 

are not thick enough or rich enough in PGEs to be economic for the mining of PGEs with the 

exception of the UG-2 chromitite. It is interesting to note that the Merensky Reef and Platreef 

are commonly characterised by their highest PGE grades being associated with the presence 

of one or more chromitite stringers (Viljoen, 2016). The UG-2 is found 40-140 m (Cawthorn, 

1999) and up to 360 m (Barnes and Maier, 2002) below the Merensky Reef in the eastern and 

western limbs, respectively. The UG-2 unit consists of a 0.5-1.0 m thick chromitite typically 

underlain by a pegmatoidal feldspathic pyroxenite, and less commonly anorthosite (Lee, 

1996). The occurrence of two to four minor chromitites in the hanging wall feldspathic 

pyroxenite has also been noted. The UG-2 chromitite is generally composed of 60-90% 

chromite with 10-35% orthopyroxene oikocrysts and interstitial plagioclase, has a Cr2O3 

content of ~43.5%, and an average Cr/Fe ratio between 1.26 and 1.4 (Lee, 1996; Barnes and 

Maier, 2002; Viljoen, 2016). As with the Merensky Reef, potholes is a common feature of the 

UG-2 chromitite and have been observed in all mines mining this layer (Hahn and Ovendale, 

1994). Base metal sulfides generally occur at low concentrations (Vermaak, 1985). Teigler 

and Eales (1993) suggested that PGE enrichment in chromitite layers may have taken place 

without bulk sulfur saturation contrary to what is suggested for the Merensky Reef where bulk 



 
17 

 

saturation may have been responsible for base metal sulfide crystallisation and associated 

PGE scavenging. 

2.2 The Northern Limb of the RLS 

The ENE-WSW trending Thabazimbi-Murchison Lineament (TML) separates the northern limb 

from the remainder of the RLS.  The N-E striking Ysterberg-Planknek and Zebediela faults are 

the near-surface expressions of the TML (Fig. 2.1; Kinnaird and Nex, 2015; Grobler et al. 

2019). The role of the TML in the emplacement of the BC is poorly understood, but it has been 

proposed that the TML may have acted as a dyke-like feeder for the magmas that gave rise 

to the RLS (Kinnaird et al. 2005). The thickness of the N-S trending northern limb outcrop 

varies over a strike length of 110 km. Its shape is evidently influenced by several underlying 

structural features (Fig. 2.6; van der Merwe, 1976). Recent geophysical modelling of the 

shallowly buried (commonly <1500 m) RLS intrusion north of the TML have shown it to have 

an areal extent of approximately 20000 km2 (Finn et al. 2015).  

The floor rocks of the northern limb vary along strike (Fig. 2.6; Sharman-Harris et al., 2005; 

Holwell and McDonald, 2006; van der Merwe, 2008; Kinnaird and Nex, 2015). To the south, 

between the farms Tweefontein and Townlands, the footwall rocks consist of shale, banded 

ironstone, calc-silicates, mudstone and siltstone of the Duitschland Formation. On the farms 

Tweefontein to Zwartfontein, the northern limb is underlain by dolostone of the Malmani 

Subgroup. Archaean granite underlies the northern limb to the north on the farms Overysel to 

Witrivier. A distinctive characteristic of the northern limb is the prominent transgression of the 

RLS northwards from the TML, through the Transvaal Supergroup (van der Merwe, 2008). 

This transgressive relationship of the mafic magma with floor rocks is significant as the degree 

of reaction and assimilation of the floor rocks is directly associated with the type of floor rock 

(van der Merwe, 2008).  

The relationship between the RLS stratigraphy of the northern limb and the remainder of the 

BC is illustrated in Fig. 2.7. It is apparent that the mafic package in the northern limb differs 

somewhat from the conventional RLS stratigraphy in the western and eastern limbs. Some of 

the inconsistencies include: the presence of economic, well-developed chromitite layers in the 

LZ of the northern limb that are not present in the LZ, and mostly restricted to the CZ, of the 

western and eastern limbs (Hulbert and von Gruenewaldt, 1982; Eales and Cawthorn, 1996; 

Teigler and Eales, 1996); the apparent absence of the LCZ from the stratigraphy of the RLS 

in the northern limb (McDonald et al. 2005); the presence of an approximately 200 m thick 

sequence of troctolite and olivine gabbronorites in the MZ of the northern limb that is not 

reported for the MZ in the rest of the BC (van der Merwe 1976; Ashwal et al., 2005).  



 
18 

 

Ultramafic cumulates representing the LZ only crop out in the southernmost part of the 

northern limb, south of the town Mokopane. To the north of the town, they crop out as distinct 

satellite intrusions such as the Zwartfontein and Uitloop intrusions (Fig. 2.6; van der Merwe, 

1976; Hulbert and von Gruenewaldt, 1985; Yudovskaya et al., 2013). A Ni-Cu-PGE 

mineralized, layered, mafic package termed the Grasvally norite-pyroxenite-anorthosite 

(GNPA) member stratigraphically overlies the LZ sequence south of the Ysterberg-Planknek 

Fault (Hulbert, 1983). The GNPA sequence is 400-800 m thick and consists of pyroxenites, 

norites, anorthosites, gabbronorites and PGE-rich chromitite layers (van der Merwe, 2008; 

Smith et al., 2014). The base metal sulfides and PGE reside in close association in both the 

upper and lower parts of the GNPA (Maier et al. 2008).   

It was initially believed that the well-developed Pyroxenite Marker, which is a thin, continuous 

unit of orthopyroxenite (Cawthorn et al., 1991; Nex et al., 2002), occurring at the boundary 

between the lower and upper Main zones in the western and eastern limbs (Kruger, 2005), 

was not developed in the northern limb (Ashwal et al. 2005). A recent study, however, 

suggested that a layer equivalent to the Pyroxenite Marker is indeed present in the northern 

limb (Cawthorn, 2020). Although the MZ has proven to be largely barren of PGEs in the 

western and eastern BC due to inferred depletions in metal contents (Maier and Barnes, 

1999), the upper MZ of the northern limb presents intriguing possibilities for PGE mineralized 

horizons such as the Troctolite Unit (Cheshire, 2011), PGE-sulfide mineralization at Moordrift 

(Maier and Barnes, 2010; Holwell et al., 2013), the Aurora project (Maier et al., 2008; 

McDonald et al., 2017) and the Waterberg Project (Kinnaird et al., 2012; Huthmann et al., 

2016). 



 
19 

 

 

Figure 2.6 Geological map of the northern limb of the BC. After McDonald and Holwell (2011) based 
on van der Merwe (1976). The study area is indicated by the red rectangle. 



 
20 

 

 

Figure 2.7 Schematic stratigraphic columns of the RLS, showing the putative correlation between 
stratigraphic units of the northern limb and eastern / western limbs. An inferred correlation is indicated 
between the Merensky Reef and Platreef. Modified after White (1994) and McDonald and Holwell 
(2011). 

2.2.1 The Platreef 

Kinnaird and McDonald (2005) define the Platreef as: “Mafic units enriched in Ni-Cu-PGE that 

occur between the Archaean granite-gneiss basement or the Transvaal Supergroup and the 

gabbros and gabbronorites of the Main zone, north of the Planknek Fault”. Located at the base 

of the mafic succession north of the Ysterberg-Planknek fault, the Platreef is the principal ore 

horizon in the northern limb and outcrops along a 30 km strike length before it is cut out by the 

overlying MZ (Fig. 2.6; Ashwal et al., 2005; van der Merwe, 2008; Kinnaird and Nex, 2015). It 

is broadly accepted that the entire mineralized Platreef sequence intruded as multiple 

distinctive pyroxenite sills (Kinnaird et al., 2005; Manyeruke et al., 2005; Maier et al., 2008). 

The thickness of the Platreef is mainly controlled by the topography of the floor rocks and has 

been reported to be between 10-400 m thick (Kinnaird, 2005). The Platreef is therefore 

substantially thicker than the < 2 m thick Merensky and UG-2 reefs that occur at approximately 

the same stratigraphic position in the eastern and western limbs of the RLS. The Platreef is 

generally dominated by orthopyroxenite, followed by norites, gabbros, anorthosites and 

sporadic serpentinized harzburgite and xenoliths of floor rocks (Gain and Mostert, 1982; 

Kinnaird et al., 2005). The presence of interlayered metasedimentary rocks in the Platreef 

package may increase its thickness (Kinnaird and Nex, 2015). 



 
21 

 

The transgressive relationship of the Platreef with the diverse floor rocks along strike in 

addition to magmatic, metasomatic and hydrothermal processes, contributed to the complex 

and erratic nature of the deposit as well the associated PGE mineralization in specific layers 

(e.g., Harris and Chaumba, 2001; Hutchinson and Kinnaird, 2005; Kinnaird, 2005; Manyeruke 

et al., 2005; Sharman-Harris et al., 2005; Holwell and McDonald, 2006, 2007; van der Merwe, 

2008; McDonald et al. 2009). Contamination by floor rocks also resulted in the formation of 

hybrid lithologies, such as wehrlite and lherzolite, informally termed para-pyroxenites 

(McDonald and Holwell, 2011). 

The distribution of base metal sulfide and PGE mineralization in the Platreef vary with depth 

and along strike (Kinnaird et al., 2005; McDonald and Holwell, 2011). PGE mineralization 

throughout the Platreef stratigraphy is largely related to the presence of base metal sulfides 

(mainly pyrrhotite, pentlandite, chalcopyrite and minor pyrite), bismuthides, tellurides, 

antimonides and arsenides (Gain and Mostert, 1982; Kinnaird et al., 2005; Hutchinson and 

Kinnaird, 2005; Hutchinson and McDonald, 2008). The Ni-Cu-PGE mineralization is not limited 

to the pyroxenitic units of the Platreef, but is also present in xenoliths, footwall rocks and 

occasionally in the base of the overlying MZ (Holwell et al., 2006; Kinnaird and Nex, 2015). At 

Overysel, sulfide mineralization is reported to penetrate into footwall gneiss (Holwell and 

McDonald, 2006). The Pt/Pd ratio of the Platreef is around unity compared to the remainder 

of the BC where the ratio is typically > 2 (Hutchinson and Kinnaird, 2005; McDonald and 

Holwell, 2011). The magmatic origin of PGE mineralization in the upper Platreef has been 

broadly confirmed (e.g., Holwell et al. 2007; Maier et al. 2008; Penniston-Dorland et al. 2008; 

Yudovskaya et al. 2017; Junge et al. 2019; Klemd et al. 2020), however, how it formed and 

how it relates to the rest of the BC is poorly understood. 

Numerous studies have proposed various models to account for the mineralization of the 

Platreef (see McDonald and Holwell 2011, for an extensive review of these models). The 

“pudding basin” model as proposed by Naldrett et al. (2008, 2009), describes the BC as having 

been formed from multiple influxes of magma, where a new influx would cause over-

pressurization in the magma chamber resulting in partially fractionated magma being expelled, 

similarly to the space between two similar-sized nested pudding basins being filled through a 

feeder located at the base of the lower basin. In relation to the Platreef and UCZ in the 

remainder of the BC, the authors suggest that the Platreef represents PGE-enriched UCZ 

magma that escaped up the margins of the limb at different stages of the UCZ development. 

Furthermore, multiple “pudding basins” are suggested to be present in the BC, one consisting 

of the eastern and western limbs, one consisting of the northern limb and one at the far western 

limb, all exposed at different erosion levels.  



 
22 

 

Buchanan et al. (1981) proposed a model of contact style mineralization where sulfides 

separated due to country rock contamination after emplacement of Platreef magma, i.e., in 

situ contamination. Subsequent studies on the contact style Platreef have proposed that 

sulfides and associated mineralization formed in a deep staging chamber (Holwell et al., 2007; 

McDonald and Holwell, 2007; McDonald et al., 2009). Sulfur saturation and subsequent 

collection of PGE by the sulfide melt thus occurred prior to intrusion (Holwell et al., 2007). 

Sharman et al. (2013) proposed that S was assimilated in a staging chamber bounded by 

Duitschland Formation of the Transvaal Supergroup. Based on the near mantle range of S 

isotope data of the upper part of the Platreef, Holwell et al. (2007) proposed that the sulfides 

and associated mineralization of the Platreef did not form by in-situ contamination. The 

assimilation of S from country rock is considered as an ore modifying process rather than an 

initiator of primary mineralization (McDonald and Holwell, 2011), consistent with the findings 

of Holwell et al. (2007) and Hutchinson and McDonald (2008) who showed that metal tenors 

of the sulfides are diluted by high proportions of country rock input.  

2.2.2 The Flatreef 

Until recently, mining of the Platreef in the northern limb has been carried out exclusively by 

the Mogalakwena mine, the world’s largest PGE producing mine, with five open pit operations 

on the farms Overysel, Zwartfontein and Sandsloot.  The deep drilling exploration programme 

undertaken by Ivanplats Ltd. in 2007, following initial drilling on the farms Turfspruit and 

Macalacaskop to delineate mineralization suitable for open pit extraction, led to the discovery 

of the Flatreef. The Flatreef, so called due to its sub-horizontal orientation, refers to the down-

dip extension of the Platreef (Fig. 2.8). It is a thick, PGE-bearing mafic zone located at depths 

of about 750 to 850 m. Mining development of the Flatreef deposit represents the first attempt 

at underground mining of the “Platreef” in the northern limb. 

The Flatreef differs from the steep, shallow Platreef in that its layering is undisturbed, less 

contaminated and more laterally continuous (Fig. 2.8; Grobler et al. 2019). Historically, several 

studies of the Platreef in the northern limb focused on the near surface, complex, highly 

contaminated Platreef (e.g., Kruger, 2005; Kinnaird, 2005; Kinnaird et al., 2005; Sharman-

Harris et al., 2005; Hutchinson and Kinnaird, 2008; Pronost et al. 2008; Sharman et al., 2013). 

Some of the findings of these studies made it difficult to relate the Platreef stratigraphy to the 

remainder of the BC. The Platreef has been characterized as a contact-style mineralization 

PGE deposit as it is located at the base of the intrusion, is significantly thicker than reef-style 

deposits, have massive to disseminated ores and are enriched in base metal sulfides (Maier 

et al. 2008). It has, however, been shown that stratiform or reef-style mineralization is present 



 
23 

 

in the thick, undisturbed, upper section of the Platreef, i.e., Flatreef (sensu stricto; Yudovskaya 

et al. 2017). 

A stratigraphic correlation between the Platreef in the northern limb and the Merensky Reef in 

the eastern and western limbs was originally proposed by Wagner (1929) and this idea formed 

the core of interpretations of subsequent studies (e.g., White 1994).  However, the proposed 

link between the northern limb and the remainder of the BC has been challenged (McDonald 

et al. 2005). Recent lithostratigraphic and isotopic studies on the newly discovered Flatreef 

(Grobler et al., 2019; Beukes et al., 2021; Keet et al., 2021; Keir-Sage et al., 2021; Mayer et 

al., 2021) have shifted the “complicated” narrative of the Platreef. 

 

Figure 2.8 Schematic section representing the relationship between Platreef, Flatreef and the deep 

downdip extension of the Flatreef at Turfspruit, and the position of its main mineralised zones. Large 

sedimentary xenoliths are metamorphosed and assimilated by intruding Bushveld magma. Modified 

after Grobler et al. (2019). 

 

 

 

  



 
24 

 

CHAPTER 3 :  

STRATIGRAPHY OF THE FLATREEF 

 

In this chapter detailed stratigraphic descriptions are presented for the Flatreef units as 

intersected by boreholes UMT-393 and UMT-276. The igneous lithologies and presence of 

hybrid lithologies, such as parapyroxenites, as well as floor rock xenoliths is closely related to 

the interaction of the Platreef magmas with the local country rocks and underlying structures 

(McDonald and Holwell, 2011). The principal borehole selected for this research study is UMT-

393, as it intersected a relatively thick Flatreef succession and evidently contains insignificant 

hybrid rocks and dolomitic / calc silicate xenoliths, indicating that the Flatreef intersected by 

this borehole likely underwent limited localized country rock contamination. The Flatreef 

mainly consists of anorthosite, norite, feldspathic pyroxenite, orthopyroxenite and harzburgite.  

3.1. Drill Core UMT-393  

Borehole UMT-393 is located on the farm Macalacaskop (243KR) which is adjacent to the 

town, Mokopane. The borehole is 1164 m deep, intersecting approximately 743 m of MZ rocks 

and an underlying 324 m thick mafic-ultramafic Flatreef (inclusive of the immediate footwall), 

which overlies ultramafic rocks of LZ affinity (Fig. 3.1 a). Several granitic veins are observed 

cross cutting the Flatreef stratigraphy as described by Kinnaird et al. (2005). The uppermost 

unit of the Flatreef is termed the hanging wall (HW) unit, which is further divided into the HW1, 

HW2 and HW3 units, from bottom to top. The HW3 unit represents the basal part of the MZ 

which comprises of mostly norite at its base, where it is in contact with HW2. Further from its 

basal contact, the HW3 unit consists mostly of medium-grained gabbronorite. This unit is 

visually more homogeneous compared to the underlying, layered Flatreef. The MZ norite of 

HW3 has a gradational contact with the underlying HW2 (Fig. 3.2a). The HW2 unit mainly 

consists of mottled anorthosite in sharp contact with an underlying succession of interlayered 

norite-pyroxenite-anorthosite, termed “norite cycles” (Kinnaird, 2005), of the HW1 (Fig.3.2b).  



 
25 

 

 

Figure 3.1 Stratigraphic log of Flatreef magmatic stratigraphy including rock types as intersected by 
drill holes a UMT-393 at Macalacaskop and b UMT-276 at Turfspruit. Drill hole UMT-393 comprises of 
relatively thicker stratigraphic unit and overlies LZ whereas UMT-276 directly overlies a sedimentary 
calc-silicate assimilation zone known as the Footwall Assimilation Zone (FAZ). 



 
26 

 

 

Figure 3.2 Drill core photographs of UMT-393. a MZ norite grades into mottled anorthosite of HW2 that 
marks the top of the Flatreef, b interlayered norite-pyroxenite at various thicknesses of the HW1. FPX 
= feldspathic pyroxenite, NOR = norite.  

Below the HW unit, at a depth of about 846 m to 852 m, is a mineralized feldspathic pyroxenite 

horizon equated with the Bastard Reef in the remainder of the BC, which is in sharp contact 

with a basal chromitite stringer (Fig. 3.3 a and b; Grobler et al., 2019). Yudovskaya et al. 

(2017) termed this layer the “Upper Reef”. The Upper Reef has elevated PGE grade with a 

3PGE grade of about 1.6 ppm. The HW together with the Upper Reef represents the 

stratigraphic equivalent of the Bastard Cyclic Unit in the western and eastern limbs of the BC 

(BCU; Fig. 3.1a).  



 
27 

 

 

Figure 3.3 Drill core photographs of UMT-393 Bastard Reef pyroxenite indicating a net textured base 
metal sulfides present in close proximity to the chromitite stringer, b chromitite stringer with sharp 
undulating contact. Scale in mm. FPX=feldspathic pyroxenite, CR= chromitite stringer. 

Underlying the BCU is an ~35 m thick Middling Unit (MD) consisting of unmineralized 

feldspathic pyroxenite (Fig. 3.4) and orthopyroxenite with a basal chromitite stringer of < 1 cm 

thick (Fig. 3.5b). The MD may also be characterised by interlayered pyroxenite-norite-

anorthosite and feldspathic orthopyroxenite in other intersections of the Flatreef on the mine 

property (Grobler et al. 2019).  

 

Figure 3.4 Drill core photograph of MD comprising of feldspathic pyroxenite (FPX) with a granitic vein 
(GRV) cross cutting the lithologies. 

 



 
28 

 

Below the MD, at a depth of about 885 m, is a mineralized package of orthopyroxenite, 

feldspathic pyroxenite, pegmatoidal orthopyroxenite, harzburgite and chromitite stringers. This 

unit has the highest PGE content in this Flatreef intersection (total Pt+Pd of up to 5 ppm). This 

unit is correlated with the Merensky Reef in the western and eastern limbs (Grobler et al. 

2019). Yudovskaya et al. (2017) termed this unit the “Main Reef”. The term “Merensky Reef” 

is used for this mineralized package in this study, but is, however, referred to as the “Main 

Reef” henceforth and in some of the subsequent chapters for simplicity in comparison to the 

remainder of the BC. The MD and Main Reef is grouped into the Merensky Cyclic Unit (MCU) 

of the Flatreef (Fig. 3.1a). The upper part of the Main Reef (Fig. 3.5a), referred to as the M2, 

consists of orthopyroxenite and feldspathic pyroxenite and is associated with a well-developed 

thin (<1 cm) upper chromitite stringer (Fig. 3.5b) and a second, poorly developed, chromitite 

stringer near the base of the unit. The lower part of the Main Reef, referred to as the M1, is 

intersected at a depth of about 927 m. It is subdivided into an M1U consisting of pegmatoidal 

orthopyroxenite (Fig. 3.5c) and medium-grained orthopyroxenite, and an M1L consisting 

mainly of harzburgite (Fig. 3.5d). An internal chromitite stringer is observed near the base of 

the M1L. Base metal sulfides occur disseminated throughout the Merensky Reef unit. Mafic 

pegmatitic veins occur closely associated with the Merensky Reef and are absent from the 

overlying units (Fig. 3.5e).  



 
29 

 

 

Figure 3.5 Drill core photograph of a the thin upper chromitite stringer at the contact between MD and 
upper Merensky Reef (M2). The presence of a pegmatoidal pyroxenite layer (PPXT) and a thick granitic 
vein (GRV) is also observed in the M2 orthopyroxenite (OPX). b A close-up image of the upper 
chromitite stringer in (a). Base metal sulfide concentrations increase in close proximity to the stringer. 
c Pegmatoidal pyroxenite (PPXT) of the M1U. d Serpentinised harzburgite of the M1L. e An 
approximately 2 m thick mafic pegmatoidal vein (MPV) cross cutting pyroxenite near the base of the 
M2. 



 
30 

 

The Main Reef overlies the 106 m thick Footwall Cyclic Unit (FCU). This unit varies 

considerably along strike and depth in the study area. Similarly to the HW1, interlayered norite-

pyroxenite-anorthosite (norite cycles) is a common feature in the FCU (Fig. 3.6a). Certain 

packages of norite cycles are poorly developed and relatively more altered as hybrid 

lithologies are also present, commonly near the base of the unit (Fig. 3.6b). The FCU in UMT-

393 is underlain by ultramafic lithologies of LZ affinity (Fig. 3.6c).  

 

Figure 3.6 Drill core photographs of UMT-393 FCU at a depth of approximately 960 m a typical 
interlayered norite-pyroxenite-anorthosite present in the unit, b poorly developed norite cycles with 
altered hybrid lithologies, c harzburgite of LZ affinity that underlies the FCU. 

 

 



 
31 

 

3.2. Drill Core UMT-276  

Drill core UMT-276 was drilled about 3 km north-west of UMT-393 on the adjacent farm, 

Tursfpruit. The Flatreef stratigraphy in UMT-276 is broadly similar to that of UMT-393 in that 

HW3 (basal part of the MZ), HW2, HW1, BCU, MCU and FCU are intersected and granitic 

veins of various thicknesses cross cut the lithologies. A significant difference between these 

two boreholes is that the Flatreef is significantly thinner in UMT-276.  Furthermore, in UMT-

276, a zone of sedimentary calc-silicate assimilation, referred to as the Footwall Assimilation 

Zone (FAZ), occurs below the Flatreef. The difference in the thickness of the Flatreef between 

UMT-393 and UMT-276 may be attributed to the control of floor rock topography, where thicker 

successions tend to develop in sub-basins (Kinnaird et al., 2005). In borehole UMT-276, the 

MZ gabbronorite is underlain by mottled anorthosite of the HW2 (Fig. 3.7a). Mineralized 

feldspathic pyroxenite of the Upper Reef of about 3 m thick is intersected at a depth of 

approximately 772 m. The MD consisting of feldspathic pyroxenite underlies the Bastard Reef 

and is capped by a chromitite stringer.   The Main Reef is located below the MD at a depth of 

approximately 780 m and consists mainly of mineralized medium- to coarse-grained 

feldspathic pyroxenite of the M2, pegmatoidal orthopyroxenite of the M1U and medium- to 

coarse-grained feldspathic harzburgite of the M1L (Fig. 3.7b).  Below the FCU at a depth of 

865 m, lithologies are more altered and hybrid lithologies predominate (Fig. 3.7c). 



 
32 

 

 

Figure 3.7 Drill core photographs of UMT-276 showing a sharp contact of overlying MZ gabbronorite 
and HW2 mottled anorthosite, b a relatively thin Merensky Reef package consisting of M2 feldspathic 
pyroxenite, M1U pegmatoidal pyroxenite and M1L feldspathic harzburgite. c altered and hybrid 
lithologies representing calc-silicate assimilation of the FAZ. 

 

 

 

 

 

 

 

 

 



 
33 

 

CHAPTER 4 :  

ARTICLE 1 

Strontium isotope variations in the Flatreef on Macalacaskop, 

northern limb, Bushveld Complex: Implications for the source of 

platinum-group elements in the Merensky Reef 

 

4.1 Abstract 

The Platreef of the Bushveld Complex is one of three key stratigraphic units, along with the 

Merensky and UG-2 reefs, in which platinum-group elements are concentrated. The 

correlation between the Platreef and the stratigraphy from the western and eastern limbs of 

the BC has been challenging due to the heterogeneous and variable nature of the Platreef 

along strike. However, the discovery of the Flatreef, interpreted as the down dip, sub-

horizontal extension of the Platreef, on the farms Turfspruit and Macalacaskop, opened new 

avenues for enquiry that allow the study of a magmatic stratigraphy less affected by footwall 

interaction. In this study, we report on the Sr isotope composition of plagioclase through a 

~ 280-m-thick intersection of the Flatreef and its foot- and hanging walls as intersected by 

borehole UMT-393 drilled on the farm Macalacaskop. Comparison of the Sr-isotopic 

composition and the anorthite content of plagioclase across this intersection with data from 

the Merensky and Bastard cyclic units elsewhere in the Bushveld Complex, supports the 

contention that the Flatreef is a correlative of the Upper Critical Zone – Main Zone transition, 

including the Merensky and Bastard cyclic units. The available data suggest that the sulfides 

of the Flatreef and its correlatives in the remainder of the Bushveld Complex originally formed 

in a deeper staging chamber. During continued magma ascent some of the sulfides were 

entrained and deposited in the Flatreef. 

4.2 Introduction 

The Bushveld Complex (BC) (Fig. 1, Maier et al. 2020) is not only the largest layered igneous 

intrusion on Earth but also hosts the largest resources of platinum group elements (PGE), 

chromium and vanadium. Although the complex is well studied, many questions remain 

unanswered and extensions to the known mineral resources are still being made on a regular 

basis. An assessment of 2015 total PGE resources versus 2010 data shows that the total PGE 

resources for the Platreef nearly doubled (Mudd et al. 2018). It is only in recent years, with 

advances in mineral exploration of the Platreef, that the northern limb of the BC has become 

the focus of more intensive research relative to the western and eastern limbs.  

The study of whole rock Sr isotopic variations across the stratigraphy of the BC provided 

valuable insight into the origin of the PGE-rich Merensky Reef, in the western and eastern 



 
34 

 

limbs (Kruger and Marsh 1982; Sharpe 1985; Lee and Butcher 1990; Kruger 1994). These 

authors concluded that the Merensky Reef formed in response to mixing of compositionally 

and isotopically distinct magmas that gave rise to a distinct inflection in the Sr isotope 

composition close to the level of the Merensky Reef. It has subsequently been argued, based 

on plagioclase separates from the Merensky and Bastard cyclic units showing Sr isotope 

compositions characteristic of the Main Zone, and orthopyroxene separates showing Cr/MgO 

ratios characteristic of the Critical Zone, that the mixing of cumulus minerals rather than 

magmas was instrumental in the gene