Abstract
The lower cap carbonate (Rasthof Formation) overlies Neoproterozoic glacial deposits (Chuos Formation) and is exposed in the Khowarib-Warmquelle area in Northern Namibia. The basal 14.2 m part of the Rasthof Formation (total about 220 m) consists of the carbonate rhythmite. The rhythmite part of the Rasthof Formation contains 1 m cycles of dark- and light-coloured rhythmites. Alizarin-red staining of thin sections and elemental mapping of polished samples indicate that the dark-coloured parts are rich in calcite, whereas the light-coloured parts are dolomite-rich. On a 1 cm scale, a reddish clay layer is intercalated in each calcite rich dark-coloured rhythmite part. These cycles of reddish clays as well as some associated major turbidites can be well correlated between columns up to about 20 km distance. Furthermore, at one locality (K4), rip-up clasts occur in a turbidite bed. Their lithology consists of dark- and light-coloured rhythmite and a reddish clay layer and can be judged to have been derived from underlying horizons. Because the clasts are elastically deformed, it is strongly suggested that the difference in lithology observed within the basal part of the Rasthof Formation existed when clasts were ripped-up shortly after sedimentation. This suggests that the cycle involving dolomite is synsedimentary, and not a diagenetic feature. Direct precipitation of dolomite does not occur in present day open marine seawater. Hurtgen et al. (2002) suggest that seawater was depleted in sulphate after Neoproterozoic glaciation. It is proposed here that some possible depositional models of cycles that consist of calcite-rich and dolomite-rich parts as well as reddish clay beds in rhythmites of the Rasthof cap carbonate.
The Neoproterozoic is a remarkable era in the history of the Earth because of low latitude glaciations and the emergence of the Metazoa. The era is characterized by geological phenomena of controversial nature: global distribution of glacial deposits, negative excursion of carbon isotopes, deposition of banded iron formations and thick carbonates that cover the glacial deposits. The ‘Snowball Earth’ hypothesis emerged as a solution for the enigma of these Neoproterozoic geological phenomena (Kirschvink 1992; Hoffman et al. 1998). Associated with the ‘Snowball Earth’ hypothesis, some scenarios have been suggested to explain the emergence of the Metazoa at this time, for example, increasing pO2 after the glaciation, or increasing evolutionary rate within small communities in isolated refugees during the global glaciation. Researchers dealing with the Neoproterozoic glaciations expect that the deeper insights into the glaciation events of the time may provide a sounder basis for understanding the emergence of the Metazoa.
Neoproterozoic successions bearing glacial deposits are distributed globally (Hambrey & Harland 1981; Evans 2000). The Otavi Group is one of the most famous of these Neoproterozoic successions. It is rich in carbonates, and Hoffman et al. (1998) revived the ‘Snowball Earth’ hypothesis with carbon isotope studies of the carbonates of this group. Following Hoffman et al. (1998), many researchers turned their attention to the Otavi Group in an attempt to understand the Neoproterozoic glacial episodes. Carbonates that cover these glacial deposits are called ‘cap carbonates’ and are regarded as the important keys to understanding the Neoproterozoic glaciations. In Namibia, glacial deposits are known at two different stratigraphic horizons (Hoffmann & Prave 1996). This paper examines the cyclic fluctuations of calcite–dolomite composition in the lower cap carbonate sequence (Rasthof Formation), which covers the lowermost glacial deposits (Chuos Formation), correlated with the Sturtian glaciation. Here it is reported that the calcite–dolomite cycles in the Rasthof Formation are primary, and not diagenetic features, based on field, petrological and chemical analysis of the rocks reflecting these cycles.
In the modern sea, dolomite is rarely precipitated, although it is far more common than calcite in the Precambrian strata. Today dolomite is formed as a product of diagenesis. The lack of good modern analogues for understanding ancient dolomite formation led to the notion of the ‘dolomite problem’ (e.g. Zenger 1972; Tucker 1990). In the modern sea, sulphate impedes dolomite precipitation (Tucker 1990). Seawater depleted in sulphate (brought about by the reducing environment that would have existed under a global ice cap) may have promoted dolomite precipitation (Hurtgen et al. 2002). The calcite–dolomite cycle in the Rasthof Formation is, therefore, an interesting sequence to study in order to understand ocean chemistry of the Neoproterozoic and its fluctuations.
Neoproterozoic glacial deposits
One of the most important problems concerning the Neoproterozoic glaciations is just how many glaciation events occurred. It is generally assumed there are at least two groups of Neoproterozoic glacial deposits that exhibit similar geological and geochemical characteristics (Kennedy et al. 1998). The younger group is often termed the Marinoan glaciation, the older the Sturtian. Geochemical studies suggest there are at least four δ13C excursions in the Neoproterozoic (Kaufman et al. 1997), which indicate at least four glacial events. In addition, there are some reliable age constraints for Neoproterozoic glacial deposits, for example, the 713 Ma Gubrah Member (Oman) (Allen et al. 2002), the 635 Ma Ghaub Formation (Namibia) (Hoffmann et al. 2004), and the 580 Ma Gaskiers Formation (Newfoundland) (Bowring et al. 2003).
In northern Namibia, the Otavi Group, which was deposited on the margin of the Congo Craton, contains two glacial horizons, each overlain by a cap carbonate. The lower glacial deposit is the Chuos Formation. A reliable maximum age constraint for the Chuos Formation is 746 ± 2 Ma (U–Pb zircon), based on studies of felsic volcanic rocks of the Naauwpoort Formation (Hoffman et al. 1996). The lower cap carbonate, which lies atop the Chuos Formation, is the Rasthof Formation. The upper glacial deposit is the Ghaub Formation. Mafic lava flows and thin felsic ash beds are interbedded within the Ghaub Formation and provide a date of 635.5 ± 1.2 Ma (U–Pb zircon) (Hoffmann et al. 2004). The upper cap carbonate is the Maieberg Formation. Both cap carbonates contain, at their bases and just above the glacial sediments, a considerable amount of dolomite associated with negative carbon isotope excursions (Hoffman et al. 1998; Halverson et al. 2002; Yoshioka et al. 2003).
Rasthof Formation
Rasthof Cap carbonate sequence
The Rasthof Formation overlies the Chuos diamictite (the Chuos Formation) and is typically a light grey-coloured carbonate with a thickness of 200–240 m, in the Khowarib-Warmquelle area of northern Namibia (Fig. 1). The Otavi Group is well exposed around this region. At several outcrops we measured columnar sections and collected samples for chemical profiling (Fig. 2).
Area studied in Northern Namibia. W1 to W3: localities of measured columnar sections near Warmquelle in the north. K1 to K5: localities of measured columnar sections near Khowarib. Columnar sections are shown in Figure 2.
Columnar sections. Correlation is indicated by thin dotted tie lines. The cycle of dark- and light-coloured rhythmites in the lower part of the columns constitute especially good marker beds. The clay layers are always associated with dark-coloured rhythmite in interval II.
The basal 14.2 m thick part of the formation consists of thinly laminated carbonate rhythmites, overlain by stromatolite-rich sediments. In the rhythmite-bearing part of the formation, the thickness of the lamination varies from a scale of mm down to 100 µm. A number of thin, massive carbonate layers occur among the rhythmites, some of which show grading in grain size or/and colour, the thickness of these layers ranging from 1–30 cm. Some thick layers show grading and contain rip-up clasts. These features indicate that these layers are turbidites. The rhythmites are overlain by a stromatolitic part. In general, stromatolites are dome-like in structure, with thick layers at the top and thinner layers in the steep furrows on both sides of the crest. To the contrary, stromatolites in the Rasthof Formation are constructed of nearly parallel lamina, and there are numerous brecciated horizons.
δ13C values
The rhythmite section located at the base of the Rasthof Formation, with about 14.2 m thickness, is the main target of this study. Several columnar sections of this part of the Formation around Khowarib-Warmquelle region were measured and recorded in detail (Fig. 3). The K1 section (Fig. 3a) exhibits the best exposure and was sampled for the geochemical studies by Yoshioka et al. (2003) as well as in this present study. Yoshioka et al. (2003) divided this section into three intervals, based on stable isotope signature. In interval I (base of the rhythmites to 1.5 m) δ13C values [PDB] increased from −5 to −2 per mm upwards. In the interval II (1.5 m to 12 m) the value remained nearly constant at −2 to −1 per mm. In interval III (12 m to the top of the rhythmites) the value increased from −1 to +5 per mm. Each interval exhibits different sedimentary features.
Photographs of the outcrops. (a) Locality K1 (Khowarib area), where a 14.2 m long set of continuous section was recovered and used for a chemical profiling study. Beds are tilted almost vertically (older to the left). Dark bands are clearly evident; (b) Locality K4 (Khowarib area). Photograph depicts the lower part of the interval II with conspicuous cycles of dark-coloured, calcitic bands alternating with light-coloured dolomitic bands; (c) Locality W1 (Warmquelle area). The basal part of the interval II. Also in the area near Warmquelle, rhythmites contain dark-coloured calcitic bands; (d) Locality W1 (Warmquelle area). The middle part of the interval II exposes the dark-coloured calcitic band.
Field observations
In interval I, the basal part of the section contains three massive layers: a calcite-rich layer, a dolomite-rich layer, and a clay-rich layer, in ascending order. Each is about 5 cm thick. Above these layers, the interval I consists of light grey, finely laminated rhythmites; the thickness of the lamination ranges from 100 µm to 1 mm. The rhythmites consist of dolomite and this part of the section is almost free of turbidites. The δ13C value is about –5 per mil at the base of the massive layers, gradually increasing to about −2 per mil in the rhythmite part of the interval I.
In interval II, the colour of the rhythmite changes cyclically from dark to light colours (Fig. 3). The rhythmite contains intercalations of thin turbidites. The dark part of interval II consists of calcite-rich carbonate, and the light-coloured is made up of rhythmites rich in dolomite. The rhythmites in interval I consist mostly of dolomite, and their colour is slightly green, but the light-colour dolomitic rhythmites in interval II contain some calcite and are yellow to brown. Each dark part contains a reddish clay layer of about 1 cm thickness (Fig. 4e). In the lower part of interval II, clay layers always accompany the dark-coloured rhythmites (Fig. 2). In interval III, clay layers are thin and irregular.
Thin sections and polished slabs. Specimens collected at K1 locality. Lower left scale bar is 1 mm long and for Figure 4a–d. Lower right scale bar is 3 cm long and for Figure 4e–f. (a) Thin section of dark-coloured rhythmite from the interval II (under crossed nicols). Brick-shaped calcite grains are arranged with their long axes parallel to the bedding (horizontal in the photomicrograph), forming fine lamination. Dark and very fine clay particles are arranged into lamina. One such lamina is sub-horizontal and 2/3 of the height from the base of the micrograph on the right hand side. (b) Thin section of the specimen from Interval I (open nicol). Micritic dolomite (white to grey) matrix and dark-coloured clay particles arranged in lamina are visible. Larger dark spots along the clay lamina are spherical clay particles. (c) Thin section of a light-coloured rhythmite in the interval II (crossed nicols). Within the large ‘bricks’ of calcite crystals, dark, micritic dolomite grains are scattered, most easily discernible near the upper left-hand corner. Clay particles form thin dark laminations. (d) Thin section of the dolomitic part from the interval III. Coarse dolomite crystals and clay laminations, with the most conspicuous one running horizontally near the middle of the height of the photographs, are visible. (e) Clay layer in the interval II, about 1 cm thick and intercalated in dark-coloured calcitic rhythmite. Such clay layers intercalate in each dark rhythmite in the interval II. In the lower part of the photograph, fine laminae of dolomite are visible. (f) Turbidite layer in carbonate rhythmites in the interval II, grading from light to dark colour. The upper 1/3 shows an apparent secondary grading from light to dark colour.
The thickness of the dark band, as well as the colour contrast of the rhythmite cycles, decreases upwards. The thin turbidite layers are only a few centimetres in thickness. The thick turbidites are about 10 cm in thickness and exhibit colour grading (Fig. 4f). Turbidites of c. 0.5 to 1 m thickness contain rip-up clasts occasionally. In this interval, δ13C values are nearly constant around −2 to −1 per mil.
In interval III, the rhythmites exhibit rather thick and undulating lamination, and consist of pure dolomite. Their colour is light grey and blue. Laminations contain clay and hematite. This interval contains massive turbidites, but their thickness is generally thinner than those in interval II. In this interval, δ13C values increase from −1 to +5 per mil upsection.
Columnar sections were measured in the Khowarib-Warmquelle area, with particular emphasis on the rhythmite part of the Rasthof Formation (Fig. 2). The same set of dark- and light-coloured carbonate rhythmites and thin reddish clay layers, as well as some turbidite layers, can be correlated with the measured columnar sections up to a distance of 20 km away from the K1.
Thin section data
In thin sections, calcite and dolomite were distinguished by using alizarin red. The rhythmites of interval I consist of micritic dolomite layers and fine clay layers (Fig. 4b). The thickness of the lamina is about 100 µm to 1 mm. Clay galls occur as spherical clusters, containing micritic dolomite grains in some horizons. These clusters could have been flocculated by the mixing of freshwater and seawater.
Carbonates in the dark bands of interval II consist of brick-shaped calcite crystals (Fig. 4a). Lamination of these rhythmites reflects the structure of these crystals. Thickness of the lamination is about 100 µm. The light-coloured carbonates in interval II consist of, in addition to brick shaped calcites, micritic dolomite inclusions (Fig. 4c). The thickness of the lamination is thicker than that of the dark-coloured rhythmites.
Carbonate rhythmites in interval III consist of dolomite and clay (Fig. 4d). Dolomite crystals are coarser than those in intervals I and II. Laminations are also thicker than those in intervals I, II, and undulating. Goethite occurs in the dark clay lamina.
Chemical profiles
Element mapping using XRF was carried out of the rhythmite part of the Rasthof Formation. At the K1 locality, a continuous series of samples from the entire rhythmite unit of 14.2 m were collected and embedded in epoxy resin in order to produce slabs that could be polished. These were analysed using a scanning XRF analytical microscope (SXAM). The element mapping images were processed by ‘lamination tracer algorithm’ (Katsuta et al. 2005) to obtain chemical profiles of the section. Furthermore at the K2-4, and W1 localities a part of the rhythmite unit was chemically profiled using the same method (Fig. 5).
Chemical profiles. Element mapping using XRF was carried out for the rhythmite part of Rasthof Formation at K1 locality with the spatial resolution of about 0.8 mm. The obtained element mapping images were processed by the ‘lamination tracer algorithm’ (Katsuta et al. 2005) to obtain this chemical profiles. Ca intensity fluctuates in interval II related to dark- and light-coloured cycles.
Characteristics of the chemical profiles obtained for the Rasthof rhythmites mirror the lithology. The moderate and constant Ca intensity, of about 100 XRF counts, in intervals I and III, correlates well with the relative dominance of dolomite over calcite. In interval II, the Ca profile exhibits a metre-scale cyclic fluctuation. Higher Ca intensity corresponds to the dark bands with high calcite content, and a lower intensity relates to the light-coloured bands, with relatively high dolomite content. Fe and Si peaks correspond to the reddish clay layers, defining the boundaries of cycles.
Mn and Sr profiles exhibit unique fluctuations in interval II. Their content is low in the clay layer, whereas higher in the carbonate-rich layer. The metre-scale cyclic fluctuations in the Mn profile correlate with the calcite–dolomite cycles in interval II. The Mn profile has a negative correlation with the Ca profile, and Mn seems associated with the dolomite. On the other hand, the Sr profile has some correlation with the Ca profile. In addition to metre-scale fluctuation, the Sr profile shows a 5 m-order fluctuation.
Rip-up clasts in turbidites
During the field survey, we made an interesting observation on outcrop K4. A 30 cm thick turbidite layer (T1) contains dark-coloured rip-up clasts as well as red and light-coloured clasts (Fig. 6). These clasts show elastic deformation, which indicates that they are derived from unconsolidated layers when they were ripped up.
Rip-up clasts within a 30 cm thick turbidite layer (T1). Dark-coloured rip-up clasts, as well as red and light-coloured clasts, in outcrop. (a) T1 turbidite layer intercalates within the rhythmites. The massive turbidite lies nearly horizontal and occupies the middle part of the composite photographs. (b) Enlarged photograph (b) of the part marked with a frame in (a) and (c) outline drawing of the rip up clasts (c) visible on the enlarged photograph (b). Scale is 5 cm in diameter for (b) and (c). Elastic deformation is easily to be observed, which is reproduced on (c). Large fragment of an elastically deformed clast with clear deformation of calcitic lamina. An intercalated reddish clay layer also demonstrates strong elastic deformation and acute folded, is denoted with grey tone in (c). Thin, light-coloured, dolomitic fragments are apparent near the lower left corner of parts (b) and (c).
Figure 6 shows typical rip-up clasts in the upper part of the bed. The elastically deformed dark block has fine laminations and intercalation with a thin, reddish clay layer. Thin fragments of light-coloured layers were also observed. Therefore, the clasts seem to have not fully been consolidated when they were incorporated into the turbidite. The dark and light layers in the rip-up clasts correspond to the calcitic and dolomitic parts of interval II, respectively and suggest that the calcite–dolomite cycles in interval II were already formed before the later deposition of the turbidite (T1).
Discussion
Synsedimentary dolomites
In interval II, the dark- and light-coloured carbonates form cycles, which can be correlated between localities over at least 20 km distance in the Khowarib-Barmquelle area. Observation in the field, as well as on stained thin sections, and chemical profiling of the rhythmites demonstrate that the rhythmic change in colour reflects the cyclic change of carbonate mineral compositions.
The basal rhythmite of the Rasthof Formation contains turbidite layers of about 1 to 30 cm in thickness. Some of them exhibit colour grading. The rip-up clasts within a 30 cm thick turbidite at the K4 locality exhibit elastic deformation and consist of dark-coloured and light-coloured rhythmite and reddish clay layers (typically observed in the sequence above and below the clasts horizon). The elastic deformation of the rip-up clasts suggests that the characteristic features of the calcite–dolomite cycles observed in interval II existed as primary, synsedimentary structures when the carbonate was unconsolidated and before the formation of rip-up clasts.
Model for synsedimentary dolomite formation
The existence of dolomite in sediment which is still elastic suggests that the dolomite has been formed either through sea floor dolomitization in the very early stage of diagenesis, or, has been formed as a primary precipitate from the seawater. In recent environments, the former can occur the interface where freshwater and seawater intermix. However, the depth of deposition of the rhythmites in the Rasthof Formation is estimated to have formed below storm wave base. Sea-level rise that occurred during deglaciation in the Neoproterozoic is estimated to be of a significant magnitude (Hoffman et al. 1998; Hoffman & Schrag 2002). Deep marine rhythmite formation is not concordant with the freshwater mixing required for primary dolomitization and thus, because the dolomite at the base of the Rasthof Formation appears to be a primary, synsedimentary precipitate, there must be another explanation for its presence.
Direct precipitation of dolomite does not occur in the present open oceans, because of the inhibiting effect of sulphate in the seawater. If sulphate is depleted up to 5% of the present concentration, dolomite precipitation may occur directly from seawater (Tucker 1990). Hurtgen et al. (2002) suggested that Neoproterozoic seawater was depleted in sulphate nearly 5% of the present concentration during meltdown of the ‘Snowball Earth’ condition, based on the sulphur isotope data from the Otavi Group. Dolomite in interval I in the basal part of the Rasthof cap carbonate may have been directly deposited within such sulphate depleted seawater. However, the residence time of Mg2+ is 1.3×107 years (Broecker & Peng 1982). The ‘Snowball Earth’ hypothesis implies about the same time scale for the duration of the glaciation. Thus seawater immediately after the glaciation should have contained only a small amount of dissolved Mg2+. The dolomite precipitation, therefore, likely ceased when the dissolved Mg2+ was rapidly exhausted. The end of dolomite precipitation coincides with the boundary between intervals I and II. But this will not immediately lead to the precipitation of calcite because sulphate inhibition.
After the end of the glacial events, the supply of Mg2+ may have resumed with the recovery of freshwater inflow from the land. River water would have been the main source of Mg2+, supplying about 5.3×1012 moles of Mg2+ per year to seawater in today's oceans (Edmond et al. 1979). This may have contributed to formation of the dolomite in interval II. However, it seems that there was a fluctuation in Mg2+ supply, indicated by the alternating precipitation of dolomite and calcite in this interval. Because the ratio of Mg2+/Ca2+ contained in river water is <0.4, this will have been rapidly consumed by dolomite precipitation in ocean water. If the flux of Mg2+ was insufficient, then dolomite precipitation would have ceased and been followed by calcite precipitation. The fluctuation of Mg2+ supply may be brought about in two ways: fluctuation in volume and rate of river inflow or fluctuation of sea level. During the time of high river water influx, the rich supply of Mg2+ will promote dolomite precipitation. To the contrary, during times of low river water influx, Mg2+ will be consumed prior to reaching a site of precipitation, so resulting in the calcite formation. Sea level change will influence the distance from land where dolomite can be formed. During a sea level high stand, the distance between the site of precipitation and the shoreline may increase and the inflowing Mg2+ may be consumed before reaching precipitation site; then calcite instead of dolomite will be deposited. On the other hand, when sea level is low, the site of precipitation will be close to the shoreline and increased Mg2+ supply promotes the precipitation of dolomite. In both cases, the fluctuation of Mg2+ supply apparently caused the cyclic formation of dolomite in interval II. It is concordant with the fact that calcite-rich rhythmites, which should be precipitated at the time of low river inflow or when the distance of the site from the shoreline is greater, contain less abundant clastic material.
Reddish clay layers in the calcite-rich rhythmites of interval II can be widely correlated between the many localities in our study area. Similar clay layers intercalated within carbonates are also known from sediments deposited during the Paleocene–Eocene Thermal Maximum (PETM) event. At that time, the postulated rapid release of a large amount of carbon in the form of methane appears to have brought about deep-sea acidification and lead to the dissolution of carbonates, leaving clay layers behind as residuals. At the end of Snowball Earth conditions, Hoffman et al. (1998) estimated that it is likely that 250 000 gigatons of carbon were released in the form of CO2. Such huge amounts of released CO2 could have brought about rapid acidification of the oceans, similar to the Phanerozoic PETM. It is, therefore, an interesting theme for future investigation; if the reddish clay layers of interval II are truly linked with the release of CO2 at the end of the ‘Snowball Earth’ condition. If the clay layer is evidence of ocean acidification after the Neoproterozoic glaciations, this most certainly would have influenced the marine biosphere. Neoproterozoic glaciation influenced evolution of life not only physically (temperature), but also chemically (pH).
Conclusions
The basal rhythmite of Rasthof Formation contains cycles of dark- and light-coloured rhythmites. Alizarin red staining of thin sections and elemental mapping of polished samples indicate that the dark-coloured part of these sediments is rich in calcite, whereas the light-coloured parts are dolomite-rich with a weak calcite signal. Reddish clay layers are intercalated in each calcite-rich, dark-coloured rhythmite layer. These cycles, as major turbidites, can be well correlated between sections up to 20 km apart. Rip-up clasts of dark- and light-coloured rhythmites and reddish clay layers exhibiting elastic deformation preserved in a turbidite at K4 suggest a synsedimentary, not a diagenetic origin.
Acknowledgments
The authors greatly thank P. Hoffman and G. P. Halverson for field guidance and help for sampling. The authors also express our cordial thanks to Gabi Schneider, Director of the Geological Survey of Namibia, for her kind permission and support for our research activities in Namibia. We thank M. Kumazawa, H. Yoshioka and T. Okaniwa, for their support in the research and sampling. This work was supported by the ‘Decoding the Earth Evolution’ Program.
- © The Geological Society of London 2007
References
Calcite–dolomite cycles in the Neoproterozoic Cap carbonates, Otavi Group, Namibia
