Plenary Sessions 25

Onset of Crustal Gold Cycle Triggered by First Oxygenic Photosynthesis

Hartwig E Frimmel Bavarian Georesources Centre, Institute of Geography and Geology, University of Würzburg, Am Hubland, D-97074 Würzburg, Germany

Abstract. First large-scale concentration of gold in the continental crust resulted from Mesoarchaean atmo-spheric and biological evolution. Chemistry of the Mes-oarchaean atmosphere and hydrosphere enabled a high Au flux off the Archaean land surface. When early life gradually changed from anaerobic anoxygenic to oxygenic photosynthesizers at around 3.0 Ga, the first “whiffs” of free oxygen produced by emerging oxygenic photosynthesizers under an overall reducing atmosphere provided a trap for Au dissolved in the huge reservoir of meteoric and shallow sea waters. Remnants of this microbially mediated gold are preserved in thin kerogen layers in some Witwatersrand reefs, which developed on erosional unconformities, scour surfaces and bedding planes in near-shore environments at around 2.9 Ga. This gold provided the principal source for the very rich Meso- to Neoarchaean placer deposits that formed by subsequent sedimentary reworking of the delicate microbial mats on aeolian deflation surfaces, into fluvial channels and delta deposits. Later tectonic reworking of these ancient gold-rich sediments by orogenic processes explains the temporal peak of orogenic-type gold formation at around 2.7 to 2.4 Ga. With time the strongly gold-enriched Archaean sediments became progressively eroded, covered or tectonically reworked, and their role as potential source of younger placer deposits diminished.

Keywords: oxygenic photosynthesis, gold, palaeoplacer, Mesoarchaean, Witwatersrand  

1 Background

Gold has been concentrated in a huge variety of deposit types through a diverse range of magmatic, hydrothermal and sedimentary processes. By far, the two most important gold deposit types are conglomerate-hosted (Witwatersrand-type) and orogenic-type deposits. Temporal distribution of orogenic gold deposits is closely linked with periods of enhanced crustal growth between 2.7 and 1.9 Ga as well as lesser peaks at around 2.9, 0.6, 0.28 and 0.12 Ga (Groves et al. 2005). Mode of deposition and timing of Witwatersrand gold mineralization has been a source of extreme controversy with disparate depositional models, syngenetic (palaeoplacer) and epigenetic, having been proposed (e.g. Robb and Meyer 1991; Minter et al. 1993; Barnicoat et al. 1997; Phillips and Law 2000). Data gathered over the past two decades clearly indicate that a purely epigenetic model is untenable, however, and that the principle stage of mineralization was syn-sedimentary. Limited local remobilization of gold and other ore constituents by post-depositional fluids has occurred (modified palaeoplacer model of Robb and Meyer 1991); for a review and arguments for/against the various models see Frimmel et al. (2005).

It has been estimated that more than 80% of all known gold was initially deposited in continental crust of Mesoarchaean age and this gold was likely mantle derived (Frimmel 2008). Melts and hydrothermal fluids in the course of tectonic recycling of Archaean crust undoubtedly remobilized much of this early gold as indicated by order of magnitude less Os in (hydrothermally) recycled gold (Frimmel and Hennigh 2015). A syngenetic model for Witwatersrand deposits requires a mega-gold depositional event during the Mesoarchaean. The ultimate source of this gold has remained the biggest challenge for the palaeoplacer model and has remained highly controversial for many decades. Recent research has shown the possibility of microbial mediation in the concentration of gold and of high Au-solubility in syn-depositional meteoric waters and seawater, thus offering a potential solution to the source problem (Horscroft et al. 2011; Frimmel 2014; Frimmel and Hennigh 2015; Heinrich 2015).

2 Witwatersrand-type deposits

Conglomerate-hosted gold mineralization is by no means restricted to the Mesoarchaean Witwatersrand Basin but is a style of mineralization observed on almost all cratons and is common to many siliciclastic successions that developed on Archaean to Palaeoproterozoic basement. The overall ore mineralogy of the Witwatersrand-type deposits/occurrences is strikingly similar, with differences being related primarily to age. Deposits that are older than the Great Oxidation Event (GOE) at c. 2.4 Ga are rich in both detrital and synsedimentary pyrite as well as detrital uraninite, whereas those younger than the GOE contain detrital Fe-oxides (haematite, magnetite) and lack detrital uraninite. Moreover, many of the older Archaean examples, especially those that are particularly well endowed in gold, contain more or less abundant stratiform kerogen layers (“carbon seams”) that represent former microbial mats (Frimmel and Hennigh 2015). Such carbon seams are conspicuously missing in all the younger, Palaeoproterozoic examples.

A further significant difference lies in the shape and size of gold particles in the various deposits. The presence of detrital gold particles has been ascertained in almost all examples but much of that gold had been remobilized into texturally late positions in those deposits that had experienced low- to medium-grade metamorphic overprint. This notwithstanding, many of these deposits, except for those in the Witwatersrand Basin, contain well-preserved gold nuggets that display variable degrees of mechanical abrasion and defor-

Acknowledgements

I wish to acknowledge at least 200 mine geologists whoprovided mine tours and encouraged research projects.Numerous mining and exploration companies, both the majors and juniors continue to provide funding andlogistic support for research and project access throughgood and bad financial times. Without their support,advancing our science would be less possible.Innumerable graduate students have challenged me andgenerated new knowledge, some of it proving me wrong.Researchers in both university and government departments continue to advance the frontiers ofknowledge and provide guidance to students and industry.Finally, visionary managers and research funding agencies that encourage both the underpinning basic research andtake a risk to support frontier work in the oceans and in remote areas are absolutely required if we are to advance.The results of all of the above enhance both the social andeconomic well-being of virtually every nation.

References

Barrie CT et al. (1999). Heat and fluid flow in volcanic associatedmassive sulfide-forming hydrothermal systems. In: Barrie CT,Hannington MD (eds) Volcanic-associated massive sulfide deposits; processes and examples in modern and ancientsettings. Socorro, NM, United States, Society of EconomicGeologists. 8:201-219

Crerar D et al. (1985). Chemical controls on solubility of oreforming minerals in hydrothermal solutions. CanadianMineralogist 23:333-352

Cruse AM, Seewald (2001) Metal mobility in sediment coveredridge-crest hydrothermal systems: Experimental and theoreticalconstraints. Geochimica et Cosmochimica Acta 65(19):3233-3247

Embley RW et al. (1988) Submersible investigation of an extincthydrothermal system on the Galapagos Ridge: sulphide mounds,stockwork zone and differentiated lavas. In: Marti RF (ed)Seafloor hydrothermal mineralization. Canadian Mineralogist26:517-540

Franklin JM et al. (2005) Volcanogenic Massive Sulfide Deposits.Economic Geology 100th Anniversary Volume: 523559

Galley AG et al. (2007) Volcanogenic massive sulphide deposits.Mineral Deposits of Canada. In: Goodfellow WD (ed). St Johns,Geological Association of Canada. Special PublicationN°5:141-162

Gibson HL, Watkinson DH, Comba CDA (1983) Silicification:Hydrothermal Alteration in an Archean Geothermal Systemwithin the Amulet Rhyolite Formation. Econ Geol 78:954-971

Hannington MD et al. (2005) Hydrothermal Ore Deposits in theSubmarine Environment. Economic Geology 100th Anniversary Volume 111-142

Hannington MD et al. (1988) Gold and native copper in supergene sulfides from the Mid-Atlantic Ridge. Nature 333(6168): 64-66

Langseth MG, Becker K (1994) Structure of igneous basement atsites 857 and 858 based on Leg 139 downhole logging.Proceedings of the Ocean Drilling Program, scientific results;Middle Valley, Juan de Fuca Ridge; covering Leg 139 of thecruises of the drilling vessel JOIDES Resolution, San Diego,California, to Victoria, British Columbia, Canada, sites 855-858,4 July-11 September, 1991. J Mottl Michael, E Davis Earl, TFisher Andrew et al. College Station, TX, United States, TexasA & M University, Ocean Drilling Program 139:73-583

Scott SD, Kissin SA (1973) Sphalerite Composition in the Zn-Fe-S System Below 300°C. Econ Geol 68(4):475-479

Seward TM, Barnes HL (1997) Metal transport by hydrothermalfluids. Geochemistry of Hydrothermal Ore Deposits. Barnes HL(ed), John Wiley and Sons Inc. 235-285

Seyfried WE Jr et al. (1999) Experimental and theoretical controlson the composition of mid-ocean ridge hydrothermal fluids.Volcanic-associated massive sulfide deposits; processes andexamples in modern and ancient settings. Barrie CT, HanningtonMD. Socorro, NM, United States, Society of EconomicGeologists 8:181-200

Seyfried, WEB, Seewald JS (1988) Hydrothermal alterationprocesses at mid-ocean ridges: constraints from diabasealteration experiments, hot-spring fluids and composition of the oceanic crust." In: Marti RF (ed) Seafloor hydrothermalmineralization. Canadian Mineralogist 26:787-804

Skirrow RG, Franklin JM (1994) Silicification and metal leaching insemi conformable alteration beneath the Chisel Lake massivesulfide deposit, Snow Lake, Manitoba. Economic Geology andthe Bulletin of the Society of Economic Geologists 89(1):31-50

Spooner ETC, Fyfe WS (1975) Sub-seafloor metamorphism, heatand mass transfer. Contributions to Mineralogy and Petrology 42:287-304

Williams-Jones AE, Migdisov A (2014) Experimental contraints onthe transport and deposition of metals in ore-forming hydrothermal systems. Special Publication 18, Society ofEconomic Geologists 18:77-95

Winter LS et al. (2004) A Reconstructed Cretaceous DepositionalSetting for Giant Volcanogenic Massive Sulfide Deposits atTambogrande, Northwestern Peru. Andean Metallogeny: NewDiscoveries, Concepts, and Updates: Special Publication No. 11Sillitoe RH, Perelló J, Vidal Cobián CE (eds). Society ofEconomic Geology Special Publication n°11:319-340

Yergeau D (2015) Géologie du gisement synvolcanique aurifèreatypique Westwood, Abitibi, Québec. Georessources. Quebec City, Canada, INRS, PhD

MINERAL RESOURCES IN A SUSTAINABLE WORLD • 13th SGA Biennial Meeting 2015. Proceedings, Volume 126

5 The Precambrian crustal gold cycle

Similar to the secular changes in the concentration of other metals, such as Fe and Mn, the interplay between Earth’s evolving mantle, atmosphere and biology alsodictated the concentration of gold into Earth’s crust (Fig. 2). For a highly siderophile element (HSE), such as Au, sequestration into juvenile continental crust would havereached a maximum in Palaeo- to early Mesoarchaeantimes and been a function of crustal growth rate, degreeof melting in the mantle (and thus mantle temperature), and HSE-availability in the source region within the mantle. Thus elevated background concentrations of Auare to be expected in Archaean crustal rocks, with likely differences between better and less well-endowedcratons dependent on the HSE-pre-enrichment in their respective source regions in the mantle, be it due toendogeneous processes (e.g. incomplete core-mantleseparation) or exogeneous processes (Hadaean-Palaeoarchaean heavy meteoroid bombardement). Considering the exceptionally high amounts of Au and PGE in the Kaapvaal Craton in the Witwatersrand Basin and the Bushveld Complex, respectively, that cratonappears to represent crust that differentiated from aparticularly well-endowed mantle domain.

Figure 1. Schematic block diagram illustratingthe likely depositional environments of a braiddelta, using the Steyn and Basal reefpalaeoplacers in the Welkom goldfield as example (after Minter et al. 1993), alsohighlighting the four principal stages involvedin the formation of the gold deposits; b – dPhotographs illustrating different stages of reworking of microbial mat-hosted gold toplacer gold: b Gold platelets/filaments onsurface of microbial columnar structures, VaalReef ; c Gold toroid with overfolded rim inconglomerate next to kerogen seam shown inFig. 1b; d SEM image of gold toroid withoverfolded rim, Basal Reef; e In-situ gold toroidwith overfolded rim (arrow) between detritalpyrite grains (Py) in Basal Reef, under combined transmitted and reflected light; f Goldconcentrated along crossbeds in pebbly quartzarenite at base of Basal Reef, Free State Geduldmine, Welkom goldfield (from Frimmel andHennigh 2015).

By about 3 Ga, maximum Au flux off the landsurface should have been reached because of maximum Au background values in the young continental crust and most suitable environmental conditions. Only by about2.90 Ga had sufficiently large oxygenic photo-synthesizing microbial communities developed that could scavenge large amounts of gold from the surrounding waters. This first concentration of gold to, in places, ore grade had a profound effect as it triggered theonset of the crustal gold cycle. Much of the microbiallyfixed gold soon became reworked by sedimentaryprocesses into placer deposits that are marked by small grain size of the detrital particles. With a delay of some200-400 myr, tectonic recycling of the Mesoarchaean sediment-hosted gold by orogenic processes led to thesecular peak in orogenic gold deposit formation in theNeoarchaean. Subsequently, these orogenic gold depositsbecame more and more important sources for younger, Neoarchaean to Palaeoproterozoic gold placer accumulations, whereas older palaeoplacersprogressively lost significance as source rocks over time.Concentration of gold in deeper marine environmentswas probably slowed down from Meso- to Neoarchaeantimes because of a change in seawater chemistry due to the transition from predominantly submarine to predominantly subaerial volcanism.

mation, thus reflecting a range of transport distances from their respective sources in a former hinterland. Specific sources in the form of older, eroded gold deposits, be it greenstone-hosted gold-quartz veins or older conglomerate-hosted placer deposits, is indicated for many of the Witwatersrand-type deposits that are younger than 2.8 Ga, such as those in the Fortescue, Caraça, Jacobina, and Tarkwa groups, the Roraima Supergroup, as well as in the younger deposits in the Kaapvaal Craton, specifically the Elsberg Reefs, Ventersdorp Contact Reef and the Black Reef. In contrast, no specific sources are known for the older examples in the Witwatersrand Supergroup, which has posed a major problem for the palaeoplacer model and has been used as critical argument by those favouring an epigenetic gold mineralization (e.g. Phillips and Law 2000).

3 Hydrocarbons and gold

Hydrocarbons are, though usually very rare, a ubiquitous feature in many Witwatersrand-type deposits, especially in Archaean ones. Two principal types of textural appearance of such hydrocarbons can be distinguished: (i) bituminous globules and (ii) “carbon seams”. Carbonisotope data on both types attest to a biogenic origin ofthe C (Spangenberg and Frimmel 2001). The texturalposition of pyrobitumen globules leaves little doubt thatthey are related to migrating oils, which is furthersupported by the presence of oil-bearing fluid inclusions(Drennan and Robb 2006). In contrast, the “carbonseams” occur as stratiform, mm- to cm-thick layers, or asthin films on erosional unconformities, scour surfacesand bedding planes at the bottom of the auriferousconglomerate beds, where they display stromatolite-likecolumnar microstructures perpendicular to bedding.Their microstructural and textural characteristics as wellas very little variation in δ13C between individual n-alkanes suggest that they represent in-situ microbial matsand thus kerogen (Frimmel and Hennigh 2015). Inplaces, as much as 70 % of all the gold has been reportedas attached to such kerogen seams in the form of thinplatelets on the surface of the stromatolite-like surfacestructures. This raises the question of the possibility ofmicrobially mediated fixation of gold. Considering theage of the kerogen seams (mainly at around 2.9 Ga) andthe growing recognition that the change-over fromanoxygenic to oxygenic photosynthesis took place ataround 3 Ga (Lyons et al. 2014), it is most likely thatthese kerogen seams represent remnants of first oxygenicphotosynthesizers, probably cyanobacteria.

4 From microbially mediated to placer gold

In an essentially O2-free Archaean atmosphere, high CO2 as well as SO2 and H2S concentrations were invariably a consequence of volcanic degassing, with the SO2/H2S ratio being predominantly controlled by the proportion of submarine to subaerial volcanism. Contemporaneous rainwater was acidic (pH ≈ 4) and caused deep chemical weathering of the Archaean land surface as indicated by

geochemical data. Widespread kaolinitization of feldspar-bearing rocks buffered the pH of local meteoric waters to c. 6. This, together with elevated H2S fugacity, as indicated by the omnipresence of detrital and synsedimentary pyrite enabled orders of magnitude higher solubility of Au as AuHS(aq) or as Au(HS)2

-, thus causing a high fluvial Au flux from the Mesoarchaean land surface (Frimmel 2014; Heinrich 2015). Wherever and whenever river water with a high dissolved Au-load came into contact with the surface of the postulated cyanobacterial mats, gold would have precipitated according to an oxidation reaction, such as: 4 Au(HS)2

- + 15 O2 + 2 H2O = 4 Au + 8 SO42- + 12 H+.

The microbial mats, which had grown in low-energy environments, such as temporarily flooded river banks in braided river systems, estuaries, lagoons or shallow shelves, are delicate structures with little preservation potential. Any regressive stage in the littoral region or flooding event in the braided river system would invariably have led to fluvial reworking of the freshly deposited microbial mats into coarser-grained fluvial deposits, such as the auriferous conglomerates (Fig. 1). This is beautifully evidenced by the truncation of carbon seams that developed on very gently dipping unconformities or disconformities, by fluvial erosion channels. Platy and filamentous gold particles released from the carbon seams by fluvial reworking would have been rolled into micro-nuggets and re-deposited downstream as heavy mineral concentrate. Many of the rounded micro-nuggets described from several Witwatersrand reefs (e.g. Minter et al. 1993) can be thus explained. Elsewhere, on aeolian deflation surfaces, microbial mats would have dried up and become removed by wind. The delicate thin gold platelets on the growth surfaces of the microbial columns would have been released and blown away as flakes. A typical morphological expression of wind-blown gold particles is toroids with double-sided over-folded rims that result from peening by saltating sand grains (Fig. 1e) – a morphological feature common in some Witwatersrand reefs (e.g. Basal Reef, Minter et al. 1993). Although gold-enrichment processes may have acted in the shallow marine environment, evidence of gold enrichment in deep marine siliciclastic sedimentary rocks from the Mesoarchaean is unclear. The lack of gold-enrichment in deeper marine sediments of the Witwatersrand may be explained by a lack of planktonic photosynthesizing microbes at that time, thus no gold-precipitation mechanism was available to contribute gold to the sediment column, or perhaps gold grades were diluted where sedimentation rates were higher. Mesoarchaean marine turbidite sequences, the material with which the latter hypothesis could be tested, are only poorly represented in the geological record. One of the few well-preserved turbidite sequences from this time is the 2.93 Ga Mosquito Creek Formation near Nullagine in the Pilbara Craton and it displays high background Au concentrations of up to 1.4 ppm in conglomerate facies within the fan apex (Frimmel and Hennigh 2015).

Plenary Sessions 27

5 The Precambrian crustal gold cycle

Similar to the secular changes in the concentration of other metals, such as Fe and Mn, the interplay between Earth’s evolving mantle, atmosphere and biology also dictated the concentration of gold into Earth’s crust (Fig. 2). For a highly siderophile element (HSE), such as Au, sequestration into juvenile continental crust would have reached a maximum in Palaeo- to early Mesoarchaean times and been a function of crustal growth rate, degree of melting in the mantle (and thus mantle temperature), and HSE-availability in the source region within the mantle. Thus elevated background concentrations of Au are to be expected in Archaean crustal rocks, with likely differences between better and less well-endowed cratons dependent on the HSE-pre-enrichment in their respective source regions in the mantle, be it due to endogeneous processes (e.g. incomplete core-mantle separation) or exogeneous processes (Hadaean-Palaeoarchaean heavy meteoroid bombardement). Considering the exceptionally high amounts of Au and PGE in the Kaapvaal Craton in the Witwatersrand Basin and the Bushveld Complex, respectively, that craton appears to represent crust that differentiated from a particularly well-endowed mantle domain.

Figure 1. Schematic block diagram illustrating the likely depositional environments of a braid delta, using the Steyn and Basal reef palaeoplacers in the Welkom goldfield as example (after Minter et al. 1993), also highlighting the four principal stages involved in the formation of the gold deposits; b – d Photographs illustrating different stages of reworking of microbial mat-hosted gold to placer gold: b Gold platelets/filaments on surface of microbial columnar structures, Vaal Reef ; c Gold toroid with overfolded rim in conglomerate next to kerogen seam shown in Fig. 1b; d SEM image of gold toroid with overfolded rim, Basal Reef; e In-situ gold toroid with overfolded rim (arrow) between detrital pyrite grains (Py) in Basal Reef, under combined transmitted and reflected light; f Gold concentrated along crossbeds in pebbly quartz arenite at base of Basal Reef, Free State Geduld mine, Welkom goldfield (from Frimmel and Hennigh 2015).

By about 3 Ga, maximum Au flux off the land surface should have been reached because of maximum Au background values in the young continental crust and most suitable environmental conditions. Only by about 2.90 Ga had sufficiently large oxygenic photo-synthesizing microbial communities developed that could scavenge large amounts of gold from the surrounding waters. This first concentration of gold to, in places, ore grade had a profound effect as it triggered the onset of the crustal gold cycle. Much of the microbially fixed gold soon became reworked by sedimentary processes into placer deposits that are marked by small grain size of the detrital particles. With a delay of some 200-400 myr, tectonic recycling of the Mesoarchaeansediment-hosted gold by orogenic processes led to thesecular peak in orogenic gold deposit formation in theNeoarchaean. Subsequently, these orogenic gold depositsbecame more and more important sources for younger,Neoarchaean to Palaeoproterozoic gold placeraccumulations, whereas older palaeoplacersprogressively lost significance as source rocks over time.Concentration of gold in deeper marine environmentswas probably slowed down from Meso- to Neoarchaeantimes because of a change in seawater chemistry due tothe transition from predominantly submarine topredominantly subaerial volcanism.

mation, thus reflecting a range of transport distances from their respective sources in a former hinterland. Specific sources in the form of older, eroded golddeposits, be it greenstone-hosted gold-quartz veins orolder conglomerate-hosted placer deposits, is indicated for many of the Witwatersrand-type deposits that are younger than 2.8 Ga, such as those in the Fortescue,Caraça, Jacobina, and Tarkwa groups, the Roraima Supergroup, as well as in the younger deposits in the Kaapvaal Craton, specifically the Elsberg Reefs,Ventersdorp Contact Reef and the Black Reef. In contrast, no specific sources are known for the older examples in the Witwatersrand Supergroup, which has posed a major problem for the palaeoplacer model andhas been used as critical argument by those favouring anepigenetic gold mineralization (e.g. Phillips and Law 2000).

3 Hydrocarbons and gold

Hydrocarbons are, though usually very rare, a ubiquitousfeature in many Witwatersrand-type deposits, especiallyin Archaean ones. Two principal types of texturalappearance of such hydrocarbons can be distinguished:(i) bituminous globules and (ii) “carbon seams”. Carbon isotope data on both types attest to a biogenic origin ofthe C (Spangenberg and Frimmel 2001). The textural position of pyrobitumen globules leaves little doubt that they are related to migrating oils, which is further supported by the presence of oil-bearing fluid inclusions(Drennan and Robb 2006). In contrast, the “carbon seams” occur as stratiform, mm- to cm-thick layers, or as thin films on erosional unconformities, scour surfaces and bedding planes at the bottom of the auriferousconglomerate beds, where they display stromatolite-like columnar microstructures perpendicular to bedding.Their microstructural and textural characteristics as well as very little variation in δ13C between individual n-alkanes suggest that they represent in-situ microbial matsand thus kerogen (Frimmel and Hennigh 2015). Inplaces, as much as 70 % of all the gold has been reportedas attached to such kerogen seams in the form of thinplatelets on the surface of the stromatolite-like surfacestructures. This raises the question of the possibility ofmicrobially mediated fixation of gold. Considering the age of the kerogen seams (mainly at around 2.9 Ga) and the growing recognition that the change-over from anoxygenic to oxygenic photosynthesis took place at around 3 Ga (Lyons et al. 2014), it is most likely that these kerogen seams represent remnants of first oxygenic photosynthesizers, probably cyanobacteria.

4 From microbially mediated to placer gold

In an essentially O2-free Archaean atmosphere, high CO2as well as SO2 and H2S concentrations were invariably aconsequence of volcanic degassing, with the SO2/H2S ratio being predominantly controlled by the proportion of submarine to subaerial volcanism. Contemporaneousrainwater was acidic (pH ≈ 4) and caused deep chemicalweathering of the Archaean land surface as indicated by

geochemical data. Widespread kaolinitization offeldspar-bearing rocks buffered the pH of local meteoricwaters to c. 6. This, together with elevated H2S fugacity, as indicated by the omnipresence of detrital and synsedimentary pyrite enabled orders of magnitudehigher solubility of Au as AuHS(aq) or as Au(HS)2

-, thus causing a high fluvial Au flux from the Mesoarchaean land surface (Frimmel 2014; Heinrich 2015).

Wherever and whenever river water with a high dissolved Au-load came into contact with the surface of the postulated cyanobacterial mats, gold would haveprecipitated according to an oxidation reaction, such as:4 Au(HS)2

- + 15 O2 + 2 H2O = 4 Au + 8 SO42- + 12 H+.

The microbial mats, which had grown in low-energyenvironments, such as temporarily flooded river banks inbraided river systems, estuaries, lagoons or shallowshelves, are delicate structures with little preservation potential. Any regressive stage in the littoral region or flooding event in the braided river system would invariably have led to fluvial reworking of the freshlydeposited microbial mats into coarser-grained fluvial deposits, such as the auriferous conglomerates (Fig. 1). This is beautifully evidenced by the truncation of carbonseams that developed on very gently dippingunconformities or disconformities, by fluvial erosion channels. Platy and filamentous gold particles releasedfrom the carbon seams by fluvial reworking would havebeen rolled into micro-nuggets and re-depositeddownstream as heavy mineral concentrate. Many of therounded micro-nuggets described from severalWitwatersrand reefs (e.g. Minter et al. 1993) can be thusexplained.

Elsewhere, on aeolian deflation surfaces, microbial mats would have dried up and become removed bywind. The delicate thin gold platelets on the growthsurfaces of the microbial columns would have beenreleased and blown away as flakes. A typical morphological expression of wind-blown gold particles is toroids with double-sided over-folded rims that resultfrom peening by saltating sand grains (Fig. 1e) – amorphological feature common in some Witwatersrand reefs (e.g. Basal Reef, Minter et al. 1993).

Although gold-enrichment processes may have acted in the shallow marine environment, evidence of gold enrichment in deep marine siliciclastic sedimentary rocks from the Mesoarchaean is unclear. The lack of gold-enrichment in deeper marine sediments of theWitwatersrand may be explained by a lack of planktonicphotosynthesizing microbes at that time, thus no gold-precipitation mechanism was available to contributegold to the sediment column, or perhaps gold gradeswere diluted where sedimentation rates were higher.Mesoarchaean marine turbidite sequences, the materialwith which the latter hypothesis could be tested, are onlypoorly represented in the geological record. One of thefew well-preserved turbidite sequences from this time isthe 2.93 Ga Mosquito Creek Formation near Nullagine in the Pilbara Craton and it displays high background Au concentrations of up to 1.4 ppm in conglomerate facies within the fan apex (Frimmel and Hennigh 2015).

MINERAL RESOURCES IN A SUSTAINABLE WORLD • 13th SGA Biennial Meeting 2015. Proceedings, Volume 128

Mathematical Geology and the GeoChron Model

Jean-Laurent MalletParadigm, Houston, US - Jean-Laurent.Mallet@PDGM.com

Abstract. As a conclusion of a whole life dedicated tomathematical geology, in this talk I would like to present myvision of this branch of science which, similarly togeophysics and geochemistry must be considered first, andabove all, as belonging to the geological realm. As anexample of what mathematical geology consists of, I am also pleased to introduce my final contribution that I called “GeoChron” which is fully described in my latest andprobably last book (Mallet 2014).

Keywords. Geomodeling, numerical geology, naturalresources SKUA, GOCAD©

1 Mathematical Geology

“The grand Book of Nature stands continually open to ourgaze, but it cannot be understood unless one first learns tocomprehend the language and interpret the characters inwhich it is written. It is written in the language of mathematics […] without which it is humanly impossible tounderstand a single word of it; without mathematics, one iswandering around in a dark labyrinth.” (Galileo Galilei: IlSaggiatore 1623)

The Earth is part of the physical realm and the goal of mathematical geology is to use mathematics to explain the laws of Nature as Galileo suggested in is famous opus "IlSaggiatore". Since Pythagoras, Archimedes, Kepler and Galileo, we know that the physical realm is ruled bymathematical equations and, since Newton, we know thatmost of these equations are differential equations:therefore, one should not be surprised that Mathematical Geology must primarily consist of partial differentialequations aimed at describing a unied mathematicalframework for modeling the geometry and the properties of geological structures.

Comparing mathematical physics and Mathematical Geology is relevant on more than one count. First andforemost, in physics two types of physicists have to bedistinguished:

experimenting physicists who seek to bring physicalphenomena to light and to measure their effects,possibly inventing their own devices to do so;

modeling physicists who seek to build mathematicalmodels which enable to predict observed phenomenawith the highest possible precision.

Both these classes of physicists are not one above theother on the values scale and indeed they are intimately linked.

1 Albeit it is well known that he did famous "thought"experiments, it is well known that Einstein never performed any"real" experiment!

Figure 1. Example of level set surfaces which correspond toconstant values of the functions u(x; y; z), v(x; y; z) and t(x; y; z) in the G-space. Note that these functions are continuouseverywhere but across the faults. Note also that the intersectionsof these level set surfaces induce a(u; v; t) curvilinear coordinatesystem associated with the local frame {ru; rv; rt}r..

Kepler for example would not have achieved anythingwithout Tycho Brahe's observations and Einstein1 would probably not have formulated the "special relativity" theorywithout the lack of evidence for the "luminiferous ether" demonstrated by Michelson and Morley's famousexperiment. Reciprocally anyone's observations wouldonly be a pleasant way to pass the time without othersputting things into equations. In his essay "La Science etl'hypothèse (1908)" Henri Poincaré famously sums this up as follows:

“Science is built from facts as a house is built from stones but a jumble of facts is no more science than a jumble ofstones is a house”

Since the end of the 18th century, "naturalist geologists"have accumulated a considerable base of observations andhave deduced thereof mainly qualitative phenomenologicalmodels. The advent of computers combined with thegrowing need for quantitative answers to the questionsraised by water resources and oil and mining industries atthe end of the 1960s led to the rise of a new class of geologists who can also be termed "modelers" as somephysicists.

"Pure" mathematics is characterized by a very conciseliterary style. The reason for this is that in the mathematicalrealm everything is perfect and does not warrant ajustification as to the physical realm. By contrast, inphysics2 and in geology the constant need to justifymathematical models as to observations leads to much more verbose discourses. This explains the generallyverbose style of Mathematical Geology.

2 e.g., have a look to the wonderful book by Hawking (2002)and the famous article by Einstein (1905).

Figure 2. Main stages of gold concentration styles in relation to the evolution of Earth’s atmosphere and secular distribution of iron formations from Archaean to Palaeoproterozoic time: a Evolution of atmospheric partial pressure of oxygen (pO2) relative to present atmospheric level (PAL); blue arrows denote first “whiffs” of O2 emanating from first emerging oxygenic photosynthesizing microbes; b Evolution of atmospheric H2S/SO2 ratio and pCO2; c Secular distribution of marine Fe deposits (banded and granular iron formations) compared with that of mantle plumes (red), also shown is the secular variation in mantle temperature (stippled curve); d Secular variation in orogenic gold formation and the main stages of surficial gold concentration on microbial mats, as uraninite-pyrite placers and as Fe-oxide-bearing placers; also shown (brown stippled curve) is the estimated secular variation in the Au-flux off the ancient land surface (from Frimmel and Hennigh 2015).

As time progressed, more and more of the gold-rich Archaean sediments became recycled by orogenic or other tectonic overprints into metamorphic crustal terrains or via subducted slabs back into the upper mantle. Some of them melted partially to form potential reservoirs for later gold-mineralizing magmatic systems. Thus a variety of gold sources for a range of magmatic and hydrothermal mineralizing systems evolved, giving rise to the variety of gold deposit types known. The role of palaeoplacers as sites for large-scale gold concentration, however, began to diminish, as their most fertile sources, the Mesoarchaean microbial mats and resulting placers, had been either eroded or covered by younger sediments or tectonically reworked. Palaeoproterozoic placers are, therefore, mainly derived from point sources, i.e. discrete deposits in their former hinterland. With the drastic changes in atmospheric composition at around 2.4 Ga (GOE), the capacity of

surface waters to take up Au from chemically weathered rocks diminished by orders of magnitude. The Au flux from the land surface towards the ocean became effectively insignificant and the Au concentration in Palaeoproterozoic seawater must have decreased dramatically. Further large-scale microbially mediated fixation of gold as postulated for Archaean times, be it in fluvial or shallow marine environments, became practically impossible. As the preservation potential of the Archaean and possibly earliest, pre-GOE Palaeoproterozoic gold-rich microbial mats and placer deposits on the surface quickly decreased with time, they became less and less likely source candidates for younger placer gold accumulation. This might explain the lack of Witwatersrand-type deposits younger than 1.8 Ga. Instead, much of the crustal gold cycle took place by magmatic, hydrothermal and metamorphic processes within the lithosphere.

References

Barnicoat AC, Henderson IHC, Knipe RJ, Yardley BWD, Napier RW, Fox NPC, Kenyon AK, Muntingh DJ, Strydom D, Winkler KS, Lawrence SR, Cornford C (1997) Hydrothermal gold mineralization in the Witwatersrand basin. Nature 386:820-824

Drennan GR, Robb LJ (2006) The nature of hydrocarbons and related fluids in the Witwatersrand Basin, South Africa: Their role in metal redistribution Geol Soc Am Spec Pap 405:353-385

Frimmel HE (2014) A giant Mesoarchean crustal gold-enrichment episode: Possible causes and consequences for exploration In: Kelley K, Golden HC (eds) Society of Economic Geologists, Spec Publ 18:209-234

Frimmel HE, Hennigh Q (2015) First whiffs of atmospheric oxygen triggered onset of crustal gold cycle. Miner Deposita 50: 5-23

Frimmel HE, Groves DI, Kirk J, Ruiz J, Chesley J, Minter WEL (2005) The formation and preservation of the Witwatersrand goldfields, the largest gold province in the world In: Hedenquist JW, Thompson JFH, Goldfarb RJ, Richards JP (eds) Economic Geology One Hundredth Anniversary Volume. Society of Economic Geologists, Littleton, Colorado, pp 769-797

Groves DI, Condie KC, Goldfarb RJ, Hronsky JMA, Vielreicher RM (2005) Secular changes in global tectonic processes and their influence on the temporal distribution of gold-bearing mineral deposits. Econ Geol 100:203-224

Heinrich CA (2015) Witwatersrand gold deposits formed by volcanic rain, anoxic rivers and Archaean life. Nature Geosci 8: 06-209

Horscroft FDM, Mossman DJ, Reimer TO, Hennigh Q (2011) Witwatersrand metallogenesis: The case for (modified) syngenesis. SEPM Spec Publ 101:75-95

Lyons TW, Reinhard CT, Planavsky NJ (2014) The rise of oxygen in Earth's early ocean and atmosphere. Nature 506:307-315.

Minter WEL, Goedhart ML, Knight J, Frimmel HE (1993) Morphology of Witwatersrand gold grains from the Basal Reef: Evidence for their detrital origin. Econ Geol 88:237-248

Robb LJ, Meyer FM (1991) A contribution to recent debate concerning epigenetic versus syngenetic mineralization processes in the Witwatersrand Basin. Econ Geol 86: 396-401

Phillips GN, Law JDM (2000) Witwatersrand gold fields: Geology, genesis and exploration. SEG Rev 13:439-500

Spangenberg J, Frimmel HE (2001) Basin-internal derivation of hydrocarbons in the Witwatersrand Basin, South Africa: evidence from bulk and molecular δ13C data. Chem Geol 173:339-355