Abstract
In view of the difficulties of stratigraphic correlation, the absence of being able to position continents reliably, the scarcity of climate indicators, and the fact that Vendian sequence crops out over a wide area and can be far apart, it has only so far been possible to identify the most general climatic features for three Vendian units. Late Vendian (Nemakit–Daldyn), and possibly Early Tommotian (Early Cambrian), time was characterized by a glacial climate with pronounced climatic zonation. It appears that there were northern and southern cold zones and a low-latitude, possibly mid-latitude warmer zones. Throughout the time the Middle Vendian sequence was being deposited (Redkino–Kotlin), a warm, possibly arid zone was characteristic of low and mid- palaeolatitudes. In southern high latitudes, there is evidence of a warm humid climate at some periods during this time. The Early Vendian (Laplandian or Varangian) time was characterized by a glacial climate. Some palaeocontinental reconstructions based on magnetism suggest that these glaciations affected high, middle and most of the low latitudes, extending to the palaeoequator. The timing of glacial events in the late Neoproterozoic and their extent across the globe and severity is a topic of great debate at the moment—some researchers suggesting glaciations that lasted for millions of years that locked the entire Earth in solid ice (‘Snowball Earth’). Others suggest that some ice-free oceans were present even during the most severe of glacial events and certainly some geological evidence seem to fit this latter hypothesis well.
Direct application of palaeoclimatic reconstruction methods developed for the Phanerozoic (Semikhatov & Chumakov 2004), in the Vendian is hampered by several difficulties. The first arises as a result of the lack of resolution of chronostratigraphy when compared with the precision of the Phanerozoic schemes. In some of the best understood sequences the late Neoproterozoic has been divided into up to five widely correlative units based on trace fossils, microfossils, metazoans, radiometric dates, stable isotope signatures and characteristic glacial horizons (Knoll 2000; Walter et al. 2000). But in many Vendian sections world-wide, the set of stratigraphic data is insufficient to allow recognition and broad correlation. On the global scale, only three units in time equivalent sequences can currently be distinguished and correlated.
The second difficulty in studying Vendian palaeoclimates arises from the lack of general agreement on continental reconstructions, with available reconstructions differing greatly. It would be of great value, as has been suggested by many researchers, to cross-check reconstructions arrived at by use of palaeomagnetic evidence, with palaeoclimatic indicators (Khain & Yasamanov 1987; Smith 2001; amongst many others).
The third difficulty affecting the reconstruction of Precambrian climatic zones is that, with increased age, the number of lithological and palaeontological climate indicators vanish and geochemical indicators decrease, through alteration, recycling and thus resetting of the ‘climate-signal’. The principle lithological indicators useful for Vendian/Ediacaran palaeoclimate reconstruction are tillites, evaporites (gypsum and salts) and carbonate platform sediments. Red beds, especially those lacking carbonate cement, are more ambiguous in their climate signal (unless, of course, there are additional indicators: mineralogical, structural, etc.). Indicators of humidity levels are particularly scarce in the Vendian. Bauxites occur only in exceptional cases and their origin is frequently unclear. Grey, terrigenous rocks are generally not reliable as indicators of humid conditions, though some researchers have invoked such. Grey colour is often of secondary origin in these ancient rocks. So, in practice, palaeoclimatic reconstructions are mainly based on indicators such as tillites, evaporites and carbonate platforms. This can lead to an exaggerated estimate of glacial and arid zone distribution at the expense of humid zones. Bio-lithological indicators are primarily limited to the occurrence of stromatolitic reefs, which occur occasionally as thick sequences, indicative of a warmer climate.
Finally, present-day outcrop areas of Late Precambrian deposits are separated by considerable distance. This also limits the accuracy of reconstructions.
These restrictions allow only the most general conclusions on global climates of the Vendian to be drawn.
Stratigraphic and palaeogeographic framework
The stratigraphic distribution of climatic indicators in some of the key sections of the Vendian System on five continents is summarized in Figure 1. Suggested correlations between Vendian and Ediacaran Systems are shown in Table 1. Based on the distribution of Small Shelley Fossils (SSF) and other metazoans, multicellular algae, trace fossils and microfossils, glacial horizons and radiometric dating, it is possible to define three divisions of the Vendian, which are loosely referred to here as ‘Early,’ ‘Middle,’ and ‘Late’ Vendian and are designated VI, VII, and VIII in Figure 1. The Early Vendian contains one or, in some regions, two glacial horizons (one at the base—Lower Blon’ Subformation, Smalfjord Formation etc.; another at the top—Glusk and Mortensnes formations, etc.), and is characterized by a significant negative δ13C at the base and the top. New data provide some evidence of correlation of the upper glacial formations of the Lower Vendian with the Gaskiers Formation of Newfoundland, dated at c. 580 Ma (Knoll et al. 2004). The base of lower glacial formations of the Lower Vendian may be correlated to the base of the glacial Nantuo Formation sensu stricto (age between 635 and 663 Ma; Zhou et al. 2004; Condon et al. 2005) or to the base of Nantuo Formation sensu lato (including Gucheng Formation; Evans 2000) of age >663 Ma (Zhou et al. 2004). The Middle Vendian contains primitive ichnofossils, specific assemblages of acanthomorphic acritarch, metaphyts and Ediacaran metazoans and is radiometrically dated at between 580 Ma and c. 545 Ma (Semikhatov 2000). The Late Vendian is characterized by the Nemakit–Daldyn assemblage of small shell fossils, complex ichnofossils, at the top by a significant negative δ13C and glacial horizons in some regions, along with radiometric dates between 545 and 535 Ma (Semikhatov 2000).
Attempt at correlation of glacial horizons and other climatic indicators in key sections of the Vendian. The names of supergroups, groups, and subgroups are in capital letters, and those of formations and smaller stratigraphic units are in lower case. VI, VII, and VIII denote the ‘Early,’ ‘Middle,’ and ‘Late Vendian’—informal units used in this paper. References: 1, Narbonne & Aitken 1995; 2, Walter et al. 2000; 3, Myrow 1995; 4, Myrow & Kaufman 1999; 5, Narbonne & Gehling 2003; 6, Chumakov 1978b; 7, Harland et al. 1993; 8, Vidal 1981; 9, Gorokhov et al. 2001; 10, Aksenov 1990; 11, Vidal & Moczydlowska 1995; Martin et al. 2000; 12, Bekker 1992; 13, Chumakov 1998; 14, Kiselev & Korolev 1981; 15, Hambrey et al. 1981; 16, Bobrov 1965; Chumakov 1993; Pokrovsky et al. 2005. 17, Wanget al. 1998; 18, Zhang et al. 1998; Barfod et al. 2002; Condon et al. 2005; Yin et al. 2005; 19, Brasier et al. 2000; 20, Leather et al. 2002; 21, Tiwari & Knoll 1994; Tiwari 1999; 22, Tiwari 1999; 23, Preiss 2000; 24, Walter et al. 2000; Calver 2000; 25, Germs 1995; 26, Walter 2000; 27, Trompette 1994, 1996; 28, Alvarenga &Trompette 1998, 1992.
Accepted correlation between Vendian and Ediacaran Systems
Early, Middle and Late Vendian correspond, on the Russian Platform, to: (a) the Laplandian glacial horizon (comprising the Glusk and Blon' formations); (b) the Redkino and Kotlin horizons; and (c) the Rovno horizon of B. S. Sokolov (1998). In Siberia, a likely stratigraphic correlative of the Rovno horizon is the Nemakit–Daldyn stage. Outside Russia, these three parts of the Vendian are usually referred to as Varangerian (comprising Smalfjord, Nyborg and Mortenses formations), Ediacaran, and pre-Tommotian Cambrian.
Thus, the intervals for which Vendian palaeoclimatic reconstructions might be possible are comparable in duration to long stages in the Phanerozoic (VIII) or to the longest Phanerozoic periods (VI).
The use of palaeogeography as a basis for palaeoclimatic reconstruction is more complex than it might seem (Chumakov 1992; Kirschvink 1992a; Scotese & McKerrow 1990; Dalziel et al. 1994; Young 1995; Mossakovsky et al. 1996; Piper 2000; Scotese 2000; Dalziel & Soper 2001; Smith 2001; Kheraskova et al. 2003; Smith & Pickering 2003; and many others). Many of these reconstructions are similar with regard the positioning of the main continents, but differ significantly with respect to location of microcontinents. This situation reflects the vigorous effort being made to refine such reconstructions—and, indeed, many of these are beginning to converge. In this paper, the palaeoclimatic schemes are based on the palaeogeographic reconstructions proposed by Smith & Pickering (2003) for the Early Vendian (600 Ma), and on tentative ones proposed by Smith (2001) for the Middle Vendian (580 Ma) and the earliest Cambrian Period (540 Ma). Following recommendations of the International Stratigraphic Commission, Smith assigned the Nemakit–Daldyn stage to the earliest Cambrian. In this paper and in Russia generally, this stage is regarded as Late Vendian. Smith's reconstructions were based on the most reliable palaeomagnetic and palaeotectonic data available at the time of publication, placing early Vendian tillites in high and middle latitudes. This may not now always be the case. In this paper, Smith's reconstructions are used with insignificant changes, as explained below.
Climatic zonation of the Late Vendian (Nemakit–Daldyn, ‘VIII’)
Climatic indicators of the VIII (Fig. 1) were plotted on Smith's (2001) palaeogeographical reconstruction for 540 Ma. These indicators suggest the existence of three roughly east-westerly trending climatic zones during the Late Vendian (possibly earliest Cambrian): two zones of cold climate (northern and southern), delimited by glacial deposits, and a third zone with evidence of a warm, locally arid zone (Fig. 2c). The northern zone includes glacial deposits deposited near end of Vendian time on the Tarim and North China blocks (Hambrey et al. 1981; Yin & Guan 1999). A glacially-striated basement occurs locally in sequences on the North China block beneath glacial deposits (Guan et al. 1986). According to Smith's (2001) reconstruction, at the peak of this glaciation, terrestrial ice sheets in the northern hemisphere could have reached a palaeolatitude of 45° north. The southern cold zone, developed during Early Cambrian or Late Vendian times, is exemplified by glacial deposits in north and west Africa (West African Glacial Horizon—Chumakov 1993; Bertrand–Sarfati et al. 1995; Trompette 1997), which may have been limited to the high southern palaeolatitudes (approximately 75° S). The original geographical positions of Late Vendian glacial deposits in Kazakhstan and Tien Shan (Baikonur Glacial Horizon—Chumakov 1978) are not well defined. Smith (2001) associated these tectonic blocks with the Siberian block, all located in low latitudes of the southern hemisphere. Considering the similarity of Vendian to Cambrian sequences (in particular their glacial deposits) of the Tien Shan, Kazakh blocks, Tarim blocks, and taking into account minimal subsequent movement, it seems most logical that the Kazakh, Tien Shan and Tarim blocks were relatively close to each other in the northern hemisphere. This positioning of the Kazakh, Tien Shan and Tarim blocks is more consistent with the tectonic ideas about make-up of the Central Asian fold belt (Mossakovsky et al. 1996; Kheraskova et al. 2003). Separating the cold zones during Nemakit–Daldyn time, was a warmer belt in which predominantly stromatolitic and oncolitic carbonate platforms were forming (northwestern North America, Siberia, Arabia and South China). Much of this zone was arid, as indicated by extensive deposits of salt, gypsum and anhydrite in Siberia (Zharkov 1981; Khomentovsky 1990), Arabia (Brasier et al. 2000; Walter et al. 2000), Iran and northwestern India (Strauss et al. 2001). Sabkha-type gypsum-bearing deposits and imprints of salt and gypsum crystals and anhydrite nodules are known outside these basins as well. All attest to a warm to hot arid climate in low and some middle latitudes in Nemakit–Daldyn time. Carbonate red beds of this age are known from South Africa (Germs 1995) and Australia (Walter et al. 1995). Non-carbonate red beds, showing evidence of aridity gypsification, halite films and desiccation cracks (Aksenov 1990) occur in the Reshma Formation in the East European Craton and in north India (Tiwari 1999; Kumar et al. 2000). These terrigenous red beds, including carbonate-bearing sediments, were deposited at midlatitudes between cold and warm zones and may have formed in semi-arid climate. According to the reconstruction used in this paper, the southernmost occurrence of polymict red bed lay at palaeolatitude of 60° south (Suvorovo Formation in the southwestern Ukraine). All of the above data give some support for the cold zone of the southern hemisphere being restricted to high latitudes, with the midlatitudes hosting a semi-arid or arid belt. Of course, all of these reconstructions depend heavily on precise dating of the sediments so that climatic events are contemporaneous and at the moment there is lively discussion and disagreement that may lead to quite different conclusions in the future.
Schemes of the Vendian zonality (palaeogeographic basis is modified after Smith (2001) and Smith & Pickering (2003): (a) Early Vendian (Laplandian or Varangian time); (b) Middle Vendian (Redkino–Kotlin time); (c) End of the Vendian–? beginning of the Early Cambrian (Nemakit–Daldyn, ? early Tommotian time). Au, Australia; An, Antarctica; Ar, Arabia; Af, Africa; B, Burma; G, Greenland; E, North-eastern Europe; I, India; IC, Indo-China; Ir, Iran; K, Kazakhstan blocks; N, Newfoundland; S, Central Siberia; NA, North America; NC, North China; T, Tibet; Tr, Tarim; Sp, Spitsbergen; SA, South America; SC, South China.
Analysis of the palaeoclimatic reconstruction for the Nemakit–Daldyn stage leads to three essential conclusions. First, the latitudinal positions of two cold and warm climatic belts are consistent with the principal climatic trends on Earth and, hence, Smith's (2001) palaeoclimatic reconstruction at the moment appears consistent with observed palaeoclimatic indicator distribution. The second conclusion is no less important: the climate during the Late Vendian times was glacial. The existence of an extensive Late Vendian glaciation is consistant with a large negative anomaly of δ13C at the Vendian–Cambrian boundary (Knoll 2000) and deposition of cap carbonates signalling the end of the glaciation (Kazakhstan and Tien Shan: Chumakov 1978, 1992; West Africa: Trompette 1997). Lastly, the climatic zoning at the end of the Vendian was asymmetrical, a pattern similar to such zonation during Phanerozoic glacial episodes.
Climatic zonation of the Middle Vendian (Kotlin–Redkino; Ediacaran; ‘VII’)
Smith's reconstructions for the Middle Vendian (Fig. 2b) are consistent with the distribution of carbonate platforms at low and middle latitudes (deposits currently found in Australia, Kazakhstan, North India, Arabia, South Africa, South America, Siberia and South China). Deposition of salts and gypsum in North India, Siberia, South China and in the region of the present-day Persian Gulf testifies to the wide distribution of arid environments within this climatic zone. The evaporite basins of the south Siberian and Iranian–Arabian regions are of enormous proportions. The latter probably included northwestern Hindustan. To the south of this warm zone, in high latitudes in the East European craton and in the Cis-Ural region, the sediments that accumulated in the seas invading the shallow Valdai Basin were dominated by bluish-grey, fine-grained, thin-bedded clays and siltstones, some enriched in organic matter, which preserved the characteristic Ediacaran metazoans along with metaphytes. In the marginal parts of the Valdai Basin, fine-grained and thin-bedded rocks graded into coarser grained non-carbonate, variegated red beds containing an admixture of kaolinite; indicative of erosion under humid conditions. This suggests that the land adjacent to the Valdai Basin was under the regime of a warm and humid climate. Thus, the climate in the high latitudes during the Middle Vendian times was humid and relatively warm, consistent with the lack of glacial indicators preserved in Middle Vendian deposits.
Climatic zonation of the Early Vendian (Laplandian or Varangian, ‘VI’)
Evidence of glaciation is widespread in Lower Vendian deposits. There are two glacial units in the Lower Vendian: one in the lower part and one in the upper part (Fig. 1). These lower glacial unit occurs on many large continents, and on some microcontinents as well. These units are located on the northeastern European craton with adjacent orogenic zones (Chumakov 1990, 1992; Harland et al. 1993), South Africa (Germs 1995), the Lesser Himalayas (Tiwari & Knoll 1994; Kumar et al. 2000), Australia (Preiss 1987, 2000; Walter 2000), and South China (Wang et al. 1998), as well as some other regions (Figs 1, 2a). In addition to those mentioned above, there are numerous occurrences of Lower Vendian diamictites whose age or origin is not well constrained (for a review see Evans 2000). The diversity of published continental reconstructions suggest that many of these glacial deposits were formed at low latitudes, some on the Equator. Such a wide latitudinal distribution indicates a profound glacial climate at beginning of the Early Vendian. Non-glacial facies of this time have not been identified. They may have been removed by the glacio-eustatic subaerial exposure of the continental shelves and inland basins.
This rock-record points to a severe glacial event on Earth at the beginning of the Early Vendian. Some authors supposed that this glaciation was global (Harland 1964; ‘Snowball Earth’ of Kirschvink 1992b; Hoffman & Schrag 2002 and others). Although there are currently no compelling counter-arguments against the ‘Snowball Earth’ hypothesis, which suggests that the entire Earth was covered by ice—serious doubts as to its validity do exist. There are a number of reasons for these doubts.
There are numerous indications of frequent glacial oscillations within the Lower Vendian and Upper Riphean sequences (Chumakov 1992, 1998, 2004; Condon et al. 2002; Hambrey et al. 1981; Leather et al. 2002 etc.). Typical examples are the glacial sequences of Belarus (Chumakov 1992) and of the Middle and South Urals (Chumakov 1998). The Vil'chitsy Group of Belarus consists of two units; the lower Blon' Formation and the unconformably overlying Glussk Formation (Fig. 3). The Blon' Formation is composed of glacial sediments at the base and interglacial sandstones and sandy dolomites in its upper part. In the most complete sections, the Glussk Formation includes three till members separated by varved clays with dropstones, and by well sorted, likely fluvial and aeolian sands with thin clay intercalations, rare ripple marks, and desiccation cracks (Fig. 3b). The uppermost layers of interglacial members reveal glacial deformations. Each till member has an intricate structure and consists of several layers of differing composition and facies, usually separated from each other by thin-bedded clay with glacial deformations (Fig. 3a). This complicated structure of the Vil'chitsy Group is related to the number glacial oscillations of different scale. The Vil'chitsy Group is much thicker than all Pleistocene glacial sequences at platforms and is similar to Permian glacial sequences. It is difficult therefore to suggest that the Vil'chitsy Group was deposited during deglaciation of ice sheet only, as suggested by some authors. Another typical example of glacial oscillations occurs in the Lower Member of the Kurgashlya Formation. This member consists of an alternation of tillites, cap dolomites and laminated siltstones containing dropstones (Fig. 4). The ‘Snowball Earth’ requires the stable existence of each global ice cover for several million years. A marked increase in the heat balance of the planet's surface would be required for the termination of each glaciation. Mechanisms for such dramatic, multiple and rapid warming are unknown and unlikely.
Numerical simulations of glacial events carried out for the Late Proterozoic (Jenkins & Frakes 1998; Poulsen et al. 2002; Poulsen 2003 and others) do not support long, uninterrupted total glaciations, such as that required by the ‘Snowball Earth’ hypothesis.
Fossils and biomarkers of phototrophism in glacial deposits provide clear evidence for the continuous existence of phytoplankton and, hence, photosynthesis in the late Precambrian (Fedonkin 2003, and refererences therein), suggesting that the world ocean was not completely covered by thick ice as required by the ‘Snowball Earth’ hypothesis.
New data suggest the possibility of underestimating palaeomagnetic latitudes determined from sedimentary rocks, particularly red beds. Original palaeomagnetic directions on Palaeogene red beds of Central Asia were underestimated by 20°–30° due to secondary flattening of the natural remanent magnetism (NRM) vector (Bazhenov & Mikolaichuk 2002). The possibility of secondary flattening of the NRM in the much older Upper Precambrian sedimentary rocks (in Yrelina Subgroup of Australia, for example) is a real concern due to solution, loss and redistribution of material in the rocks and the reorientation of some clastic grains through long exposure to static load metamorphism (Kholodov 1994).
A necessary condition for a ‘Snowball Earth’ is the tropical position of most continents (Schrag et al. 2002). At the same time, according to most reconstructions of continental positions published in recent years, significant landmasses were located in high and middle latitudes during the Vendian (e.g. Scotese & McKerrow 1990; Dalziel 1997; Piper 2000; Scotese 2000; Powell et al. 2001—used by Hoffman & Schrag 2002; Dalziel & Soper 2001; Smith 2001; Kheraskova et al. 2003; Smith & Pickering 2003).
The presence of cap carbonates atop glacial deposits and the occurrence of banded iron formations associated with these sequences is viewed in the ‘Snowball Earth’ hypothesis as a result of global events related to rapid deglaciation (Hoffman & Schrag 2002). However, even though cap carbonates occurred widely, they do not necessarily occur globally. Cap carbonates are shallow-water, post-glacial sediments demonstrating regional facies changes. The South Australian Nuccaleena Formation, a typical cap carbonate that occurs over a large area, consists of paralic and shallow-water, nearshore dolomites, which grade laterally into shales and sandstones that were deposited in a more open basin (Preiss 1987). Rapid replacement of cap carbonates by sandstones along strike is observable in the Kurgashly Formation of the South Urals (Fig. 4) and in some other regions. In the South (Fig. 4) and Middle Urals, Spitsbergen and in some glacial sequences elsewhere, thin carbonate beds are found within marine glacial deposits. The regional nature of Late Precambrian banded iron formations and manganese occurrences associated with glacial deposits is even more obvious (Chumakov 1992; Young 2002). Almost all occurrences of these ores are confined to volcanic areas, partly those of rift origin: Tien Shan (Dzholdoshev 1964), SW and South Africa (Martin 1965), northwestern Canada (Yeo 1981), Brazil (Ferran 1982), Middle Urals (Chumakov 1992). In some cases, iron-rich rocks and iron ores occur between marine glacial strata (the Tien Shan and Urals) or form the matrix of these strata: the Tien Shan and Canada (Dzholdoshev 1964; Eisbacher 1981; Yeo 1981). These data have led researchers studying Fe and Mn occurrences in glacial sequences to the conclusion that such metals precipitated locally from volcanic fluids in a zone containing a mixture of fresh glacial water and seawater.
Cyclic structure of the glacial Vil'chitsy Group in the Vil'chitsy drill hole, Lower Vendian sequence, Belarus: (a) middle member of the Glussk Formation; (b) Glussk Formation; (c) Vil'chitsy Group.
Cyclic structure of the glacial Lower Member of Kurgashly Formation, Lower Vendian sequence, Belaya River, South Ural. Sections: (a) Krivaya Luka, ridge; (a1) Krivaya Luka river cuts; (b) 2 km up-stream of Krivaya Luka; (c) 1.5 km NE of Muradymovo village; (d) left side of Belaya River 1.5 km down-stream of Apshak brook; (e) at mouth of Apshak brook.
Conclusion
Palaeoclimatic reconstructions indicate that, during Vendian time, there were at least two main intervals of glacial climate (VI and VIII), with a non-glacial interval between (VII). This alternation was responsible for the reorganization of climatic zonality, and it changed environments from warm to cold and back, which most certainly would have had significant effects on the biosphere. The Early Vendian ‘Snowball Earth’ hypothesis needs further critical consideration. At the moment, geological data favours a ‘Slushball Earth’ hypothesis for this period (Poulsen 2003; Poulsen et al. 2002).
Acknowledgments
The investigations were supported by Project No 05-05-64949 of the Russian Fund of Base Investigations and Programme 18 of Presidium RAS. I appreciate the suggestions by L. A. Frakes and the second anonymous reviewer. This paper was written as a part of UNESCO IGCP493.
- © The Geological Society of London 2007
