Archaean Hydrothermal Systems in the Barberton Greenstone Belt and Their Significance as a Habitat for Early Life
Hydrothermal systems have played an important role in shaping the 3.54–3.23 Ga volcano-sedimentary succession of the Barberton greenstone belt. Evidence for relatively low-temperature (≤150°C) seafloor hydrothermal activity is widely recorded in extensive silicification of volcanic and sedimentary rocks, leaching of elements commonly mobile during water-rock interaction, and extensive hydraulic fracturing. Evidence for the presence of high-temperature hydrothermal vents is scarce and restricted to a massive sulphide deposit near the top of the succession. Many of the zones affected by seafloor alteration are spatially associated with traces of early life, such as carbonaceous matter and bioalteration features. Diffuse venting of low-temperature hydrothermal fluids was a widespread phenomenon on the Palaeoarchaean seafloor, making it an ideal habitat for hyperthermophiles and the possible birthplace of life during earlier times.
KeywordsBarberton greenstone belt Archaean Hydrothermal activity Seafloor alteration Silicification
The Barberton greenstone belt contains a well preserved Palaeoarchaean volcano-sedimentary succession that provides abundant evidence for the interaction of submarine volcanic rocks and seafloor sediments with the hydrosphere and biosphere. Petrological and geochemical studies have provided evidence for the existence, as early as 3.5 Ga, of a rich microbial ecosystem. This is preserved as filamentous and spherical structures of carbonaceous matter and microbial mats in cherts (Walsh 1992; Westall et al. 2001, 2006; Tice and Lowe 2006; Glikson et al. 2008) as well as possible bioalteration features in pillow lavas and hyaloclastites (Furnes et al. 2004; Banerjee et al. 2006). Many of these traces of life are associated with rocks that show evidence for hydrothermal activity on the Archaean ocean floor. Submarine hydrothermal systems, driven by circulation of seawater through volcanic rocks, constitute a possible site where life emerged and evolved. This paper provides a summary of features preserved in the Barberton greenstone belt that may be directly related to the presence of hydrothermal systems. The importance of these systems as a habitat of ancient life is briefly evaluated.
2 Geological Setting
2.1 Onverwacht Group
The Onverwacht Group formed between 3.54 and 3.30 Ga (Kröner et al. 1996; Byerly et al. 1996) and consists of komatiites, komatiitic basalts, basalts and minor dacites. Sedimentary rocks make up less than 10% of the succession and formed in a deep to shallow marine environment. South of the Inyoka Fault, the Onverwacht Group has been subdivided into six formations, the Sandspruit, Theespruit, Komati, Hooggenoeg, Kromberg and Mendon Formations (Viljoen and Viljoen 1969; Lowe and Byerly 2007a). The Sandspruit and Theespruit Formations are separated from the Komati Formation by a fault zone, but the younger units are stratigraphically conformable. The Onverwacht Group is best developed and least deformed in the southwestern part of the belt northeast of Tjakastad (Fig. 1). Metamorphic grade is mainly greenschist facies, but locally reaches amphibolite facies, in particular in the Sandspruit and Theespruit Formations. Onverwacht Group rocks north of the Inyoka Fault have been grouped together as the Weltevreden Formation (Lowe and Byerly 2007a).
2.2 Fig Tree Group
The 3.26–3.23 Ga Fig Tree Group consists of a several kilometres thick siliciclastic-volcaniclastic sedimentary sequence that is capped by felsic volcanic rocks (Condie et al. 1970; Heinrichs 1980; Hofmann 2005; Lowe and Byerly 2007a). South of the Inyoka Fault, a variety of lithofacies are present that formed in deep- to shallow-water to alluvial environments (southern facies). The Loenen Formation is a sequence of sandstones and shales that contain abundant felsic detritus. It is overlain by ferruginous shales and jaspilitic banded iron formations of the Ngwenya Formation. The Mapepe Formation consists of shale, tuffaceous shale, laminated felsic tuff, chert-clast conglomerates and, locally, baryte beds. Zircon dates from the Mapepe Formation vary from 3.26 to 3.23 Ma (Kröner et al. 1991; Byerly et al. 1996). The ca. 3.25 Ga Auber Villiers Formation is approximately 1.5 km thick and consists of dacitic, plagioclase-phyric volcaniclastic rocks and terrigeneous chert-clast conglomerates and sandstones.
In the northern part of the belt, the Fig Tree Group is mainly characterized by turbiditic sandstones and shales of the Ulundi, Sheba and Belvue Road Formations that formed in a relatively deep-water environment (northern facies). Kohler and Anhaeusser (2002) defined the Bien Venue Formation, which consists of quartz-muscovite schist derived from dacitic to rhyodacitic volcaniclastic protoliths dated at ca. 3.25 Ma. Subordinate rock types include banded chert, phyllite, and biotite-plagioclase and chlorite schists, derived, in turn, from dacitic and basaltic precursors. Massive sulphide and baryte deposits are also present. The Schoongezicht Formation overlies the Belvue Road Formation in the central and northwestern part of the greenstone belt. It consists of plagioclase-rich turbidites intercalated with shale at the base and cross-bedded volcaniclastic sandstones and dacite clast conglomerates at the top. Intercalated felsic volcanic rocks have been dated at 3,226 ± 1 Ma (Kamo and Davis 1994).
2.3 Moodies Group
The Moodies Group was deposited close to 3.227 Ga, but not later than 3.11 Ga (Kamo and Davis 1994). It consists of shallow-marine to fluvial sandstone and conglomerate with minor shale and banded iron-formation. The Moodies Group is approximately 3 km thick and consists of several fining-upward sequences ranging from conglomerate or pebbly quartzose sandstone at the base, to a thick sandstone unit, to capping siltstone and shale (Anhaeusser 1976).
3 Hydrothermal Systems in the Onverwacht Group
The Onverwacht Group is a thick pile of submarine lava flows that erupted intermittently over a time period of ca. 250 Ma. Breaks in volcanic activity resulted in the deposition of thin interflow sedimentary units characterized by predominantly tuffaceous volcaniclastic material. At the same time, hydrothermal processes were operating on, and immediately below, the seafloor, leaving behind extensive zones of alteration that are mainly characterized by the presence of silicified volcanic rocks, bedded cherts, and chert veins.
3.1 Silica Alteration Zones
Silicification of volcanic rocks immediately below sedimentary chert horizons is a common phenomenon of Palaeoarchaean volcano-sedimentary successions in general and the Barberton greenstone belt in particular (Kitajima et al. 2001; Terabayashi et al. 2003; Van Kranendonk and Pirajno 2004; Van Kranendonk 2006; Hofmann and Wilson 2007; Hofmann and Harris 2008). Because of the high silica content of the alteration zones, early workers in the Barberton greenstone belt regarded some of these zones to consist of felsic volcanic rocks, which were later shown to be silicified mafic and ultramafic rocks (Paris et al. 1985; Lowe and Byerly 1986; Duchac and Hanor 1987). The origin of the silicification has been the subject of debate, with interpretations ranging from weathering of volcanic flow tops (Lowe and Byerly 1986) to various processes of hydrothermal alteration (de Wit et al. 1982; Paris et al. 1985; Duchac and Hanor 1987; Hanor and Duchac 1990; de Wit and Hart 1993).
Hofmann and Harris (2008) observed that δ18O values of silicified volcanic rocks show a positive linear relationship with silica content and plot between the mantle value (ca. 5.7‰) and a δ18O value of ca. 18‰ extrapolated for a composition of 100% SiO2. A Si isotope study was reported by Abraham et al. (2007) from the same alteration zone shown in Fig. 3. Unsilicified basalts display negative δ30Si values (−0.15‰ to −0.40‰), similar to average mafic magmatic rocks, whereas δ30Si for silicified basalts shifts towards more positive values (up to 0.8‰) with increasing grade of silicification, with overlying chert displaying the highest Si isotope values (up to 1.1‰).
Hofmann and Harris (2008) concluded that the element depletion–enrichment patterns and oxygen isotope data indicate low-temperature (ca. 100–150°C) hydrothermal processes for the origin of the alteration zones. Diffuse venting over large tracts of ocean floor was envisaged to have given rise to silicification of the volcanic rocks and of seafloor sediments to form cherts. The Si isotope variation may be attributed to simple mixing of a mafic volcanic and seawater endmember composition, both of which had values similar to the modern day. Apparent zonations of alteration, with silicification at the top and carbonate alteration at depth may be a result of differing fluid chemistries with depth. In the deeper part of the alteration zones, hydrothermal fluids may have been more alkaline, becoming more acidic towards the contact with the likely acidic Archaean ocean due to mixing with seawater.
3.2 Bedded Cherts
Bedded chert horizons are typically 1–20 m thick and were deposited on the seafloor in between phases of extrusive submarine volcanic activity. With the exception of the Komati Formation, bedded cherts are common throughout the Onverwacht Group and the lower part of the Fig Tree Group, but are rare to absent in the upper Fig Tree and Moodies Groups. They typically consist of microcrystalline quartz with minor amounts of phyllosilicates, Fe-Mg-Ca-carbonates, Fe-Mn and Ti-oxide minerals, disseminated sulphides and carbonaceous matter. Cherts are highly siliceous rocks (SiO2 commonly >95 wt%) that contain clastic or tuffaceous material of ultramafic to felsic composition replaced by microquartz. Despite the fact that addition of silica has resulted in dilution of most other elements, they have relatively high alkali contents and are enriched in many metals, particularly in Co, Ni, and Cu. The δ18O values of cherts are somewhat higher than for underlying silicified volcanic rocks and are generally in the range of 15–20‰. The highest δ18O value yet obtained for cherts of the Barberton greenstone belt is 22.1‰ (Knauth and Lowe 2003).
It is generally agreed that silicification took place shortly after deposition of the precursor sediment on the seafloor (Lowe 1999), but the mechanism is a matter of debate. Many researchers regard the silicification as a direct result of hydrothermal processes on the seafloor (de Wit et al. 1982; Paris et al. 1985; Duchac and Hanor 1987). Hofmann and Harris (2008) suggested silicification being the result of diffuse flow of hydrothermal fluids that were also responsible for the silicification of the underlying rocks. Others refute the evidence for hydrothermal activity, and ascribe the formation of these rocks to marine processes that acted in a hot Archean ocean (e.g. Knauth and Lowe 2003; Tice and Lowe 2006). Cherts of the Barberton greenstone belt have much lower δ18O values relative to Phanerozoic cherts (difference of ≥10‰ on average; Perry 1967; Knauth and Lowe 2003). These signatures have been interpreted in various ways, but mainly being the result of one or more of the following: (1) lower δ18O values of Archaean seawater (Perry 1967; Kasting et al. 2006; Jaffres et al. 2007), (2) higher temperatures of the Archaean ocean (ca. 70°C; Perry 1967; Knauth and Lowe 2003), (3) temperatures of hydrothermal fluid–seawater mixtures and not the temperature of the Archaean ocean (Hofmann 2005; Hofmann and Harris 2008).
If seawater had δ18O values as low as −13‰ at 3.4 Ga (Jaffres et al. 2007), this would imply that the alteration took place at temperatures of around 20°C. Such low temperatures are inconsistent with the geochemical and mineralogical changes in the alteration zones that need elevated temperatures to form. The model of a hot Archaean ocean, potentially resulting in downward infiltration of hot fluids from the seafloor and leading to top down alteration of volcanic rocks, is inconsistent with a number of observations that indicate the presence of upwelling hydrothermal fluids, such as depletion of metals (Co, Ni and Cu) in subsurface volcanic rocks and their relative enrichment in overlying chert (Hofmann 2005; Hofmann and Harris 2008), and the development of chert veins, regarded as hydraulic fractures beneath chert beds necessitating high fluid pressures at times (Hofmann and Bolhar 2007).
3.3 Chert Veins
Vein and bedded cherts of a single stratigraphic section are petrographically, isotopically as well as geochemically very similar. They are all characterized by sand-sized carbonaceous grains of apparent detrital origin. The differences between several pairs of vein chert and directly overlying bedded chert is <2‰ for C isotopes and <1.1‰ for O isotopes. The same applies to REE patterns and trace element ratios (Hofmann and Bolhar 2007; Hofmann and Harris 2008). All these observations suggest a genetic link between the veins and the sedimentary horizons.
While some workers regard chert veins as feeder channels for silicified sedimentary deposits in the Pilbara craton (e.g. Brasier et al. 2002; Van Kranendonk 2006), in the Barberton belt they have been regarded to represent fractures that were filled with sediment from above (Lowe and Byerly 1986; Hofmann and Bolhar 2007). In the model presented by Hofmann and Bolhar (2007), diffuse upflow of hydrothermal fluids in a shallow, but laterally very extensive hydrothermal system that was initially open to the ocean is envisaged during deposition of the sedimentary chert precursor. This system became temporarily closed during progressive sedimentation and hydrothermal silicification of the seafloor sediments that resulted in the bedded chert horizons. The cherts acted as a poorly permeable barrier for ascending fluids, resulting in the build-up of fluid overpressure and hydraulic fracturing of the volcanic rocks and the cap rocks at times. The hydraulic fractures were filled with sedimentary material from as yet unsilicified sediments overlying the chert for some time before the hydraulic system was sealed again and the next cycle of pressure build-up and fracturing took place. This eventually resulted in multiple injections of fluidized sedimentary material and hydrothermal fluids in cross-cutting veins, partly brecciating the host rock and incorporating host rock fragments into the veins.
3.4 Ironstone Pods: Archaean Hydrothermal Systems or Products of Recent Weathering
Lowe and Byerly (2003) studied Fe-oxide-rich deposits at different localities in the Barberton greenstone belt, including the ironstone pod described by de Ronde et al. (1994). They concluded that the ironstones were deposited on and directly below the modern ground surface by active groundwater and spring systems during relatively recent (Cenozoic) times. This reinterpretation resulted in some debate in the literature (de Ronde et al. 2004; Lowe and Byerly 2004), which has since cooled down with additional work published by Lowe and co-workers (Hren et al. 2006; Lowe and Byerly 2007b).
Debate has centred on the pod originally described by de Ronde et al. (1994) to be of Archaean age (Fig. 7). De Ronde et al. (2004) cited the following main arguments for the ancient age of this pod: (1) The pod is regarded to be part of the Archaean stratigraphy; it is cut by a porphyry intrusion and clasts derived from ironstone pods are present in overlying sedimentary units. (2) It contains rounded chert boulders interpreted as talus derived from scarps in the vicinity of the mound. (3) Quartz-haematite-goethite stockwork veins are present in the pod. (4) Its haematitic composition is unlike low-T spring deposits. Lowe and Byerly (2004, 2007b) argued for a recent origin of the ironstone pod, through partial replacement of Archaean chert host rocks by Fe-oxide-rich material. These authors based their interpretation on the following main observations: (1) No evidence was found that the pods are part of the stratigraphy, as they are not cut by Archaean porphyry or provided detritus upon erosion during Archaean times as claimed by de Ronde et al. (1994). Instead, they locally observed layering in the pod that dips parallel to the present slope. (2) Internal chert blocks are not rotated and represent in situ remnants not affected by replacement of chert by haematite. (3) The pods do not show evidence for deformation that characterizes the surrounding rocks and contain common open cavities, which would not have been preserved up to the present. Lowe and Byerly (2004, 2007b) further argue that the quartz stockworks represent remnants of Archaean quartz veins that were only partially replaced by ironstone.
Our own observations of ironstone pods in the Barberton greenstone belt tend to agree with a relatively recent origin of these features. Nevertheless, a large part of the work by de Ronde and co-workers was done on fluid inclusions of quartz veins that are of undoubted Archaean age. These quartz veins are commonly associated with chert veins interpreted to have formed during seafloor hydrothermal activity (Hofmann and Bolhar 2007) and may thus still provide evidence for hydrothermal fluid composition and temperature of an Archaean hydrothermal system. This is in contrast to the claim by Lowe and Byerly (2007b) that these features are a result of late-stage hydrothermal alteration of the belt (as late as 2.7 Ga). If the haematite-rich ironstone pod is a result of relatively recent spring activity, two questions still remain. One issue is the predominance of haematite in the pod, which generally only precipitates from relatively high-temperature fluids. This led Lowe and Byerly (2007b) to suggest a hot spring origin for the ironstone pod. No hot spring has so far been reported from the greenstone succession of the Barberton belt. Another major issue is the source of the iron. Lowe and Byerly (2007b) suggest that the iron was likely derived from the weathering of siderite-rich sedimentary units at depth, such as banded iron-formation. This is far from proven, and will require further investigations.
3.5 Heat Source for Onverwacht Hydrothermal Activity
Hydrothermal systems consist of two essential components, a heat source and a fluid phase (Pirajno 1992). While the fluid phase is most readily attributed to seawater circulating through the volcanic pile of the Onverwacht Group, the heat source is less easy to define. Most hydrothermal systems operating on the seafloor today are driven by heat generated from shallow-level magma chambers, especially those developed close to the mid-ocean ridges (Alt 1995). Other heat sources may include shallow-level magma chambers, intrusive igneous bodies or volcanic complexes associated with volcanic islands, oceanic plateaus and island arcs. Crustal sections close to these heat sources may experience extensive alteration related to the interaction with actively circulating >350°C hot hydrothermal fluids, which may locally discharge on the seafloor through black smokers. Further away from these zones fluid temperatures will be much lower, and hydrothermal fluid discharge is much more diffuse (Alt 1995; German and von Damm 2003).
Although the Onverwacht Group was deposited in a submarine setting and formed mainly as a result of subaquatic deposition of mafic-ultramafic lavas, there is no evidence that the volcanic rocks formed close to an Archaean mid-oceanic ridge. The lavas accumulated during an extensive period of time far longer than the average life span of modern oceanic crust. The alteration zones are laterally extensive and tabular, bearing no resemblance to the heavily faulted spreading centres along modern mid-ocean ridges. Instead, the Onverwacht Group is characterized by laterally continuous submarine lava flows more akin to large submarine shield volcanoes or oceanic plateaus (Cloete 1999; Hofmann and Harris 2008). Focused hydrothermal upflow zones in the form of epidosites and black smoker-type massive sulphides are lacking throughout the Onverwacht Group, suggesting diffuse upflow of hydrothermal fluids through the volcanic rocks.
Although there are several sub-volcanic intrusions in the Barberton sequence, ranging from ultramafic to felsic in composition, that could have been the heat source for the hydrothermal activity, there are many more examples of stratiform alteration zones in the Barberton greenstone belt that cannot be directly linked to such intrusions. Each volcanic sequence capped by chert is altered at the top, indicating that the alteration took place during a time interval of relative volcanic quiescence, when there was enough time for the deposition of finely laminated sediments before the onset of the next volcanic episode. Hofmann and Harris (2008) attributed the hydrothermal activity to heat derived from cooling of volcanic rocks and a high regional heat flow in an oceanic plateau-like setting in the vicinity of a long-lived source of hot, upwelling mantle that resulted in the establishment of shallow subseafloor convection cells. Shallow-level igneous intrusions may have provided additional heat, but they were not the main factor for elevated heat flow. Extensive silicification was the result of diffuse venting of hydrothermal fluids over broad areas of the ancient ocean floor, likely aided by seawater chemistry close to silica saturation and thus more conducive for widespread silicification (Hofmann and Harris 2008).
3.6 Buck Reef Chert: Deposition During Hydrothermal or Normal Marine Conditions?
Debate has centred on the interpretation of the palaeoenvironment. De Vries (2004) and de Vries et al. (2006) consider the BRC to be the uppermost silicified part of the Buck Ridge volcano-sedimentary complex (BR-vsc), a series of arc-like basaltic and felsic volcanic rocks, capped by pervasively silicified sedimentary rocks. This complex is believed to be characterized by syndepositional normal faulting, which widely controlled felsic magma upwelling and sedimentation. De Vries et al. (2006) advocate hydrothermal venting as a common growth-fault related phenomenon directly related to the depositional history of the unit, giving rise to features such as chert veins, hydraulic breccia bodies and quartz-filled cavities.
In contrast, Lowe and Byerly (2007a) regard the BRC to have been deposited on a subsiding volcanic platform representing the eroded felsic volcanic edifice of H6 that was emplaced well before deposition of the BRC. Syn-depositional normal faulting was observed to have taken place during deposition of the lower half of the BRC (Lowe and Fisher Worrel 1999). Tice and Lowe (2004, 2006) suggested that the BRC consists of normal marine sediments that were silicified through interaction with hot marine waters supersaturated with amorphous silica and related abundant carbonaceous matter to microbial mats in a non-hydrothermal environment.
Hofmann and Bolhar (2007) and Hofmann and Harris (2008) suggested hydrothermal activity during deposition of at least the lower part of the BRC (the evaporitic facies). Dacite and dacite-derived clastic sedimentary rocks below and interbedded with the evaporitic cherts are highly silicified and show the same chemical variations with depth as observed in all the other alteration zone below chert horizons. Multiple generations of cross-cutting veins of chert are common in the alteration zone. These are both stratiform as well as at a high angle to bedding, the latter extending several 100 m below the BRC.
If the evaporitic facies was indeed affected by syn-depositional hydrothermal activity, then the chert clast breccias commonly observed in this zone may be attributed to subsurface dissolution of evaporites by hydrothermal fluids rather than to evaporite dissolution in the phreatic zone (Lowe and Fisher Worrel 1999). A study of fluid inclusions in quartz veins and quartz-filled cavities by de Vries and Touret (2007) from the BRC gave conflicting results, although it pointed towards relatively low pressure of ∼100 bars for quartz precipitation, in support of a shallow water environment and synsedimentary hydrothermal activity.
In addition, the evaporites may not be precipitates from purely marine waters. Elongate crystal pseudomorphs interpreted to represent pseudomorphs after nahcolite (NaHCO3; Lowe and Fisher Worrel 1999) are associated with finely laminated and rippled cherts and grew concurrently with deposition of the host sediment. The nahcolite crystals have been interpreted as primary evaporitic minerals that grew in a hypersaline marine coastal environment with high bicarbonate and sodium contents. If the pseudomorphs are indeed former nahcolite and if the BRC was deposited during hydrothermal activity, there is a possibility that elevated concentrations of bicarbonate and/or sodium are derived from hydrothermal fluids. Alteration in the underlying dacite and dacite-derived sediments has resulted in the almost complete removal of Na from at least the uppermost 100 m of rock succession (<< 0.1 wt% Na2O, Hofmann and Harris 2008), as the Na content of the unaltered magmatic rocks was likely similar in composition to the genetically related Theespruit Pluton (ca. 7 wt% Na2O, Kleinhanns et al. 2003). It is thus possible that Na was derived from syn-depositional hydrothermal fluids. It needs to be emphasized that nahcolite has only been observed at the base of the BRC where it overlies altered felsic igneous rocks.
Chert veins were not observed cutting across the basal evaporitic facies. This may indicate that hydrothermal activity ceased before deposition of the overlying cherts or that the evaporitic facies acted as a barrier for ascending hydrothermal fluids. Cessation of hydrothermal activity can be attributed to reduced convection of seawater as the basal chert beds would have prevented continuous recharge of the hydrothermal system. If this was the case, deposition of cherts higher up in the BRC may have been less affected by hydrothermal activity. This may explain the abundance of soft-sediment deformation features in black-and-white banded cherts above the evaporitic facies (Tice and Lowe 2004), suggesting silicification acted more slowly. δ18O values for the BRC are amongst the highest values obtained by Knauth and Lowe (2003) from the Barberton greenstone belt and range from 15.7‰ to 21.4‰, which is in line with subdued hydrothermal activity.
If BRC deposition was indeed influenced by hydrothermal activity, then several potential heat sources need to be considered. The sill-like dacite intrusion underlying the BRC could have been the heat source, but this is unlikely for several reasons. (1) The dacite is erosively overlain by dacite-derived sandstones that are silicified at the top and may have had sufficiently cooled down prior to subaerial exposure and erosion not to have been responsible for the hydrothermal alteration as seen at the top. (2) Available age data suggest that the dacite formed long before the BRC was deposited (ca. 3.45 versus 3.42 Ga, see above). The heat source could thus have been an elevated regional geotherm or renewed volcanic activity that gave rise to the overlying volcanic rocks. The BRC is overlain by undated mafic/ultramafic lapillistones and tuffs of the Kromberg Formation, it is cut by abundant ultramafic dykes and sills that may be genetically related to the overlying unit, it contains rare layers of ultramafic accretionary lapilli, and has been regarded as interleaved laterally with mafic and ultramafic rocks (Ransom et al. 1999; Lowe and Byerly 2007a). All these observations suggest that mafic/ultramafic magmatism was ongoing during deposition of the middle to upper part of the BRC. It is very well possible that extensional deformation of the sequence was related to this magmatic event.
4 Hydrothermal Systems of the Fig Tree Group
The 3.26–3.23 Ga Fig Tree Group is a volcano-sedimentary sequence that is dominated by relatively fine-grained siliciclastic sediments and felsic volcanic and volcaniclastic rocks. Carbonaceous cherts are uncommon, but many of the shale horizons are rich in carbonaceous matter (Reimer 1975), although reports of microfossil finds have been met with scepticism (Schopf and Walter 1983). Evidence for hydrothermal activity during Fig Tree deposition is mainly preserved in the presence of baryte horizons, a massive sulphide deposit and localized hydrothermal alteration of the sedimentary rocks.
Detrital baryte was interpreted by Heinrichs and Reimer (1977) and Reimer (1980) to represent reworked baryte deposits of hydrothermal-exhalative origin. Multiple sulphur isotope values, in particular negative Δ33S values, suggest that the sulphate was derived from seawater, and ultimately from photolysis reactions in the atmosphere (Bao et al. 2007). Hydrothermal activity may have been roughly contemporaneous with baryte reworking and deposition. On the western limb of the Baryte Valley Syncline (Fig. 10), baryte beds directly overlie a succession of highly silicified shales, now chert, that are transected by numerous chert veins. Baryte values are high (>1,000 ppm) in several samples of these rocks (Hofmann 2005; Hofmann and Bolhar 2007). At a disused baryte mine (Main Workings, Fig. 10), chert and baryte veins and baryte-cemented breccias have been observed in cherts underlying the baryte beds (Pearton 1986). This suggests a genetic link between chert veins and reworked baryte, i.e. syndepositional hydrothermal activity. Horizons consisting of interbedded chert and baryte also occur in the Pilbara craton, where they are associated with swarms of zoned veins of chert and baryte, indicative of a hydrothermal origin (Nijman et al. 1999; Van Kranendonk and Pirajno 2004; Van Kranendonk 2006). The timing of baryte deposition in the Fig Tree Group is unclear, as dating of the Mapepe Formation in different structural domains has yielded conflicting results (Kröner et al. 1991; Byerly et al. 1996).
4.2 Hydrothermally Altered Shales and Sandstones
Because sedimentary baryte deposits are restricted to the central part of the greenstone belt where Fig Tree Group sedimentary rocks are highly altered, it is possible that this area was a geothermally active zone during Fig Tree times. Ba-rich hydrothermal fluids may have ascended mainly during lower Mapepe times and resulted, firstly, in local deposits of hydrothermal baryte that were reworked to form detrital baryte beds and, secondly, in the metasomatic alteration of Fig Tree strata in hydrothermal upflow zones. Syndepositional alteration may be related to the vicinity of the “baryte belt” to a tectonically active zone underlying, and possibly controlling parts of, the Fig Tree sedimentary basin (e.g. Proto-Inyoka zone of Heinrichs and Reimer 1977).
4.3 Bien Venue Massive Sulphide Deposit
4.4 Heat Source for Fig Tree Hydrothermal Activity
Felsic volcanic rocks of the Bien Venue Formation have been dated at 3,256 ± 1 Ma (Kohler 2003), which is the likely age of the Bien Venue VMS deposit. This is the same age as reported for the felsic volcanic Auber Villiers Formation (3,256 ± 4 Ma, Kröner et al. 1991; 3,253 ± 2 Ma, Byerly et al. 1996), suggesting a belt-wide event of felsic volcanic activity. A similar age has also been reported from dacitic tuffs at the base of the Mapepe Fm at one locality (3,258 ± 3 Ma, Byerly et al. 1996), which could suggest that at least some of the Mapepe baryte deposits formed at that time. The presence of baryte associated with the Bien Venue deposit would fit the interpretation that the Fig Tree Group both north and south of the Inyoka Fault was affected by a major hydrothermal event at ca. 3.255 Ga associated with felsic volcanic activity. No TTG plutons of this age have so far been identified in the Barberton Greenstone belt, although many of the highly sheared felsic rocks in the Bien Venue Formation may have originally formed as sub-volcanic sills. Instead there are a number of plutons with an age of 3.23 Ga that formed at the same time as the felsic volcanic rocks of the Schoongezicht Formation that forms the uppermost unit of the Fig Tree Group. Most authors (e.g. de Ronde and Kamo 2000; Kohler and Anhaeusser 2002; Lowe and Byerly 2007a) have interpreted the felsic volcanic activity in the Fig Tree Group to be associated with subduction-related magmatism in a continental arc or back-arc setting, environments conducive for hydrothermal activity.
5 Implications of Hydrothermal Activity for Early Life in the Barberton Greenstone Belt
Evidence for hydrothermal activity is widespread in the Barberton greenstone belt. This is not surprising, as the Swaziland Supergroup, with the exception of the Moodies Group, consists largely of volcanic and pyroclastic rocks that were intruded by TTG plutons and ultramafic layered complexes at several stages in the history of the succession. What remains to be answered more fully is how hydrothermal activity did affect the life forms that were present at that time and to what extent hydrothermal systems and sub-seafloor environments were occupied by microbial life. Furthermore, it is important to determine how hydrothermal alteration affected the morphology and composition of remnants of life as preserved in Barberton rocks. Research to answer these questions is still in its infancy.
Carbonaceous matter of likely organic origin (Walsh 1992; Westall et al. 2001; Tice and Lowe 2006; Glikson et al. 2008) is present in bedded cherts and in chert veins of the Onverwacht Group. In which environment this material formed is unresolved. If the carbonaceous matter formed in shallow parts of the ocean and represents background suspension sediments that settled out of the water column, its origin would be unrelated to seafloor hydrothermal activity. It is also possible that organic matter was produced by photo-/chemosynthetic bacteria close to the seafloor. Mat-like accumulations of carbonaceous matter have been described from bedded cherts (Westall et al. 2001; Tice and Lowe 2006). If biogenic and of chemosynthetic origin, low-temperature hydrothermal activity may thus have been conducive to life on the Archaean seafloor.
Alternatively, carbonaceous matter may have formed by subsurface biota that inhabited the hydrothermal systems and were deposited on the seafloor during hydrothermal effusions. Supporting the idea of a subsurface habitat are tubular features described from the rims of pillow basalts from the 3.3 Ga Kromberg Formation and interpreted to represent corrosion features by chemolithoautotrophs that colonized the originally glassy pillow rims (Furnes et al. 2004). If these structures are truly biogenic in origin, the microbes must have thrived while hydrothermal fluids were circulating the basaltic subseafloor environment.
Although thermophilic organisms may well have occupied the low-temperature hydrothermal systems at that time (e.g. Ueno et al. 2004), most of the carbonaceous matter present in the chert veins likely originated from the downward flow of sediment-laden seawater from the seafloor and was not produced at depth. This is indicated by its ubiquitous presence in subvertical veins extending to the palaeoseafloor, but its absence in stratiform chert veins that are commonly filled with botryoidal chert precipitate. The latter is interpreted to have formed when the hydrothermal system was closed off from the seafloor due to the impermeable chert cap (Hofmann and Bolhar 2007). A possible abiological carbon source includes the reduction of CO2 during serpentinisation of ultramafic rocks at depth (e.g. Horita and Berndt 1999) and redistribution of carbonaceous matter during hydrothermal venting, although these processes have been discounted to have played a major role for the origin of carbonaceous matter in the Barberton greenstone belt (Hofmann and Bolhar 2007). However, remobilisation of carbonaceous matter did take place, as indicated by its presence as secondary material in pore space and in veins (van Zuilen et al. 2007), indicating that morphological interpretations of carbonaceous structures need to be conducted with care.
Shales of the Fig Tree Group are rich in carbonaceous material (0.16–1.34 wt% TOC; Reimer 1975), which occurs as compacted carbonaceous grains. This material has very similar isotopic signatures compared to Onverwacht carbonaceous cherts (δ13C = −29‰ on average; Hofmann and Bolhar 2007), suggesting that most of the carbon must represent material that either formed and was reworked on the seafloor or settled out of the water column. A biogenic origin appears most plausible in the absence of a tenable model for the abiogenic origin of carbonaceous matter in thick and spatially widespread black shale sequences with light carbon isotope signatures comparable to those in younger settings. Carbonaceous matter is both present in hydrothermally altered and unaltered shales and throughout the Fig Tree Group, suggesting no direct link between life and hydrothermal activity, at least during Fig Tree times.
Research in the Barberton greenstone belt was supported by Deutsche Forschungsgemeinschaft (Ho 2507/1–1/2), University of the Witwatersrand Research Committee and National Research Foundation of South Africa (FA2005040400027). Numerous people provided access, support, and hospitality, including Johan Eksteen and Property Mokoena (Mpumalanga Parks Board), Colin Wille (Taurus Estate), Jan Maarten and Wilma van Rensburg (Sappi Forests) and Roelf le Roux and Chris Rippon (Barberton Mines).
- Abraham K, Opfergelt S, Cardinal D, Hofmann A, Foley S, André L (2007) Si isotopes as a clue for understanding Eoarchaean silicification. Geochim Cosmochim Acta 71(Suppl 1):A3Google Scholar
- Alt JC (1995) Subseafloor processes in mid-ocean ridge hydrothermal systems. In: Humphris SE, Zierenberg RA, Mullineaux LS, Thompson RE (eds) Seafloor hydrothermal systems: physical, chemical, biological, and geological interactions. Geophys Monogr 91:85–114Google Scholar
- Anhaeusser CR (1976) The geology of the Sheba Hills area of the Barberton Mountain Land, South Africa, with particular reference to the Eureka syncline. Trans Geol Soc S Afr 79:253–280Google Scholar
- Channer DMD, de Ronde CEJ, Spooner ETC (1997) The Cl--Br--I- composition of ∼3.23 Ga modified seawater: implications for the geological evolution of ocean halide chemistry. Earth Planet Sci Lett 150:325–335Google Scholar
- Cloete M (1999) Aspects of volcanism and metamorphism of the Onverwacht Group lavas in the southwestern portion of the Barberton greenstone belt. Mem Geol Surv S Afr 84:232 ppGoogle Scholar
- de Ronde CEJ, de Wit MJ, Spooner ETC, Channer DMdeR, Christenson BW, Bray CJ, Faure K (2004) Ironstone pods in the Archean Barberton greenstone belt, South Africa: Earth’s oldest seafloor hydrothermal vents reinterpreted as Quaternary subaerial springs. Geology, Online Forum, p e68Google Scholar
- de Vries ST (2004) Early Archaean sedimentary basins: depositional environment and hydrothermal systems. Examples from the Barberton and Coppin Gap Greenstone Belts. Geologica Ultraiectina 244:160 ppGoogle Scholar
- Heinrichs T (1980) Lithostratigraphische Untersuchungen in der Fig Tree Gruppe des Barberton Greenstone Belt zwischen Umsoli und Lomati (Südafrika). Göttinger Arbeiten zur Geologie und Paläontologie 22:118 ppGoogle Scholar
- Hofmann A, Wilson AH (2007) Silicified basalts, bedded cherts and other sea floor alteration phenomena of the 3.4 Ga Nondweni greenstone belt, South Africa. In: Van Kranendonk MJ, Smithies RH, Bennett V (eds) Earth’s oldest rocks. Dev Precambr Geol 15:571–605Google Scholar
- Hofmann A, Reimold UW, Koeberl C (2006) Archean spherule layers in the Barberton Greenstone Belt, South Africa: a discussion of problems related to the impact interpretation. In: Reimold WU, Gibson R (eds) Processes on the early Earth. Geol Soc Am Spec Pap 405:33–56Google Scholar
- Kohler EA (2003) The geology of the Archaean granitoid-greenstone terrain in the vicinity of Three Sisters, Barberton Greenstone Belt. Geol Surv S Afr Bull 133:150 ppGoogle Scholar
- Lowe DR (1999) Petrology and sedimentology of cherts and related silicified sedimentary rocks in the Swaziland Supergroup. In: Lowe DR, Byerly GR (eds) Geologic evolution of the Barberton Greenstone Belt, South Africa. Geol Soc Am Spec Pap 329:83–114Google Scholar
- Lowe DR, Byerly GR (2004) Ironstone pods in the Archean Barberton greenstone belt, South Africa: Earth’s oldest seafloor hydrothermal vents reinterpreted as Quaternary subaerial springs. Reply, Geology, p e69Google Scholar
- Lowe DR, Byerly GR (2007a) An overview of the geology of the Barberton greenstone belt and vicinity: implications for early crustal development. In: Van Kranendonk MJ, Smithies RH, Bennett V (eds) Earth’s oldest rocks. Elsevier, Amsterdam. Dev Precambr Geol 15:481–526Google Scholar
- Lowe DR, Fisher Worrel G (1999) Sedimentology, mineralogy, and implications of silicified evaporites in the Kromberg Formation, Barberton Greenstone Belt, South Africa. In: Lowe DR, Byerly GR (eds) Geologic evolution of the Barberton Greenstone Belt, South Africa. Geol Soc Am Spec Pap 329:167–188Google Scholar
- Lowe DR, Nocita BW (1999) Foreland basin sedimentation in the Mapepe Formation, southern-facies Fig Tree Group. In: Lowe DR, Byerly GR (eds) Geologic evolution of the Barberton Greenstone Belt, South Africa. Geol Soc Am Spec Pap 329:233–258Google Scholar
- Murphy PW (1990) The petrology, geochemistry and classification of the Bien Venue massive sulphide deposit, Baberton Mountainland. MSc Thesis, University of Natal, DurbanGoogle Scholar
- Pearton TN (ed) (1986) Excursion guidebook for Barberton Mountain Land, Geocongress ’86. Geol Soc S Afr, 82 ppGoogle Scholar
- Ransom B, Byerly GR, Lowe DR (1999) Subaqueous to subaerial Archean ultramafic phreatomagmatic volcanism, Kromberg Formation, Barberton Greenstone Belt, South Africa. In: Lowe DR, Byerly GR (eds) Geologic evolution of the Barberton Greenstone Belt, South Africa. Geol Soc Am Spec Pap 329:151–166Google Scholar
- Reimer TO (1975) Untersuchungen über Abtragung, sedimentation und diagenese im frühen Präkambrium am Beispiel der Sheba Formation (Südafrika). Geol Jahrb B 17:108 ppGoogle Scholar
- Rouchon V, Orberger B, Hofmann A, Pinti D (2009) Late diagenetic Fe-carbonates in Early Archean felsic sediments (Hooggenoeg Formation, Barberton Greenstone Belt, South Africa): implications for CO2 sequestration and the chemical budget of seawater. Precambr Res. doi:10.1016/j.precamres.2009.04.010Google Scholar
- Schopf JW, Walter MR (1983) Archean microfossils: new evidence of ancient microbes. In: Schopf JW (ed) Earth’s earliest biosphere: its origin and evolution. Princeton University Press, Princeton, NJ, pp 214–239Google Scholar
- Viljoen MJ, Viljoen RP (1969) An introduction to the geology of the Barberton granite-greenstone terrain. Geol Soc S Afr Spec Publ 2:9–28Google Scholar
- Ward JHW (1999) The metallogeny of the Barberton greenstone belts, South Africa and Swaziland. Mem Geol Surv S Afr 86:108 ppGoogle Scholar
- Westall F, de Ronde CEJ, Southam G, Grassineau N, Colas M, Cockell C, Lammer H (2006) Implications of a 3.472–3.333 Gyr-old subaerial microbial mat from the Barberton greenstone belt, South Africa for the UV environmental conditions on the early Earth. Phil Trans R Soc B 361:1857–1875CrossRefGoogle Scholar