It is widely known that glacial ice contains a record of Earth’s changing atmospheric composition in the form of bubbles trapped when the ice formed. That is fine for investigations going back about a million years, in particular those that deal with past climate change. Obviously going back to the composition of air tens or hundreds of million years ago cannot use such a handy, direct source of data, but has relied on a range of indirect proxies. These include the number of pores or stomata on fossil plant leaves for CO2, variations in sulfur isotopes for oxygen content and so on. Variation over time of the atmosphere’s content of oxygen has vexed geoscientists a great deal, partly because it has probably been tied to biological evolution: forming by some kind of oxygenic photosynthesis and being essential for the rise to dominance of eukaryotic animals such as ourselves. Its presence or absence also has had a large bearing on weathering and the associated dissolution or precipitation of a variety of elements, predominantly iron. Despite progressively more clever proxies to indicate the presence of oxygen, and intricate geochemical theory through which its former concentration can be modelled, the lack of an opportunity to calibrate any of the models has been a source of deep frustration and acrimony among researchers.
Yet as is often said, there are more ways of getting rid of cats than drowning them in butter. The search has been on for materials that trap air in much the same way as does ice, and one popular, if elusive target has been the bubbles in crystals of evaporite minerals. The trouble is that most halite deposits formed by precipitation of NaCl from highly concentrated brines in evaporating lakes or restricted marine inlets. As a result the bubbles contain liquids that do a grand job of preserving aqueous geochemistry but leave a lot of doubt as regards the provenance of gases trapped within them. For that to be a sample of air rather than gases once dissolved in trapped liquid, the salt needs to have crystallized above the water surface. That may be possible if salt forms from brines so dense that crystals are able to float, or perhaps where minerals such as gypsum form as soil moisture is drawn upwards by capillary action to form ‘desert roses’. A multinational team, led by Nigel Blamey of Brock University in Canada, has published results from Neoproterozoic halite whose chevron-like crystals suggest subaerial formation (Blamey, N.J.F. and 7 others, 2016. Paradigm shift in determining Neoproterozoic atmospheric oxygen. Geology, v. 44, p. 651-654). Multiple analyses of five halite samples from an ~815 Ma-old horizon in a drill core from the Neoproterozoic Canning Basin of Western Australia contained about 11% by volume of oxygen, compared with 25% from Cretaceous salt from China, 20% of late-Miocene age from Italy, and 19 to 22% from samples modern salt of the same type.
Although the Neoproterozoic result is only about half that present in modern air, it contradicts results that stem from proxy approaches, which suggest a significant rise in atmospheric oxygenation from 2 to about 18% during the younger Cryogenian and Ediacaran Periods of the Neoproterozoic, when marine animal life made explosive developments at the time of repeated Snowball Earth events. Whether or not this approach can be extended back to the Great Oxygenation Event at around 2.3 Ga ago and before depends on finding evaporite minerals that fit stringent criteria for having formed at the surface: older deposits are known even from the Archaean.
About 635 Ma ago fossils of large-bodied organisms first appeared in the geological record: some quilt like, others with a crude bilateral symmetry, more looking like ‘mud-filled bags’ and ribbed discs but none that can easily be distinguished as animals, plants or colonial microorganisms. First found abundantly in the Ediacara Hills of South Australia, hence their sack-name the Ediacaran biota, it now seems that they were distributed globally in the late Neoproterozoic Era. Interpreting their metabolism is risky enough – some are reckoned to be animals that absorbed nutrients through their skin, others said to be dependent on photosynthesis – but a controversy has raged for many years over the kind of environment in which they thrived. In a detailed 2012 study of sedimentary structures petrography in the South Australian sandstones from which they were first described, Gregory Retallack of the University of Oregon inferred that some lived on land and are now found in palaeosols: they include Spriggina, Dickinsonia and Charnia that are among the most favoured candidates for being animals or some kind. Others inhabited shallow water. Anticipating fiery disputes a Nature editorial appeared in same issue in which Retallack published his paper .
Retallack has now moved on to the even more fossil-rich Ediacaran sediments of Newfoundland (Retallack, G.J 2016. Ediacaran sedimentology and paleoecology of Newfoundland reconsidered. Sedimentary Geology, v. 333, p. 15-31). Eye-wateringly detailed sequence stratigraphy of the now famous Mistaken Point locality and others suggests that the ecosystem there was an intertidal salt marsh. In detail it contains evidence for shallow-water graded bedding, signs of regular storms and perhaps tsunamis together with interbedded palaeosols and subaerial volcanic crystal tuffs whose feldspars survive intact. The palaeosols can be subdivided into several pedogenic types akin to those used to classify modern soils. Unlike the arid setting of the South Australian Ediacaran sediments, whose palaeosols show signs of freezing, the Newfoundland package indicates humid, cool-temperature climes
As in Australia, the palaeosols are rich in Ediacaran fossils, including the best known; the leaf-like Charnia and its discoidal support structure that appears in Retallack’s reconstruction of the environment in an analogous way to salt-tolerant shrubs in modern tidal flats. They occur together with encrusting fossils that bear some resemblance to modern foliose fungi or lichens. Further chuntering in the palaeontological community seems inevitable, but the sedimentological observations alone knock one hypothesis on the head: it has been said that the graded bedding common to both major Ediacaran assemblages constitutes evidence for deep marine origins from turbidity currents. But there is further compost in which controversy may thrive, in that Retallack ascribes the repeated palaeosols to glacially controlled sea-level fluctuations: the Newfoundland sequence contains two diamictites interpreted as tillite, one dated at ~583 Ma the other undated but at the top of the sequence.
The Earth’s earliest atmosphere undoubtedly had a chemistry dominated by carbon dioxide and nitrogen, together with transient water vapour, outgassed from volcanoes giving pervasive reducing conditions at the surface and in the oceans. Until the last couple of decades the only clear evidence of a switch to oxidising conditions and presumably significant atmospheric oxygen was direct, mineralogical evidence. The most obvious signs are ancient, reddened soils formed when soluble Fe2+ lost electrons to molecular oxygen to form the distinct red, orange and brown oxides and hydroxides of insoluble Fe3+ that impart a deep staining in even small quantities. Others include the disappearance from river-transported sediments of clearly transported grains of metal sulfides and uranium oxide that remain stable under reducing conditions but quickly break down in the presence of oxygen.
Widespread observations in Precambrian sediments, eventually linked with reliable radiometric ages, strongly suggested a fundamental environmental change at around 2.3 billion years ago: the Great Oxidation Event. A few such signs emerge from somewhat older rocks back to 2.7 Ga, but only the 2.3 Ga event created a permanent feature of our home world; at first toxic to many of the prokaryote life forms of earlier times but eventually a prime condition for the rise of the Eukarya and eventually metazoan animals. Isotopic analysis of sulfur from Precambrian sediments also gave hints of a more complex but much debated transition because of the way S-isotopes fractionate under different environmental conditions. Now other indirect, isotopic approaches to redox conditions have become feasible, with a surprising result: powerful evidence that about 3 billion years ago there was appreciable atmospheric oxygen (Crowe, S.A. et al. 2013. Atmospheric oxygenation three billion years ago. Nature, v. 501, p. 535-538).
The Danish-South African-German-Canadian group relied on a fractionation process among the isotopes of chromium, which can exist in several oxidation states. When minerals that contain Cr3+ are weathered under oxidising conditions to release soluble Cr6+ the loss in solution preferentially removes the 53Cr isotope from residual soil. If the isotope enters groundwater with reducing conditions to precipitate some Cr3+ -rich material yet more 53Cr remains in solution. Eventually such enriched water may enter the oceans, where along with iron and other transition-group metal ions chromium can end up in banded iron formations (BIFs) to preserve isotopic evidence for oxidising conditions along it route from land to sea.
The team analysed both a palaeosol and a BIF unit from a stratigraphic sequence in the Achaean of NE South Africa that is between 2980 and 2924 Ma old. A substantial proportion of the palaeosol is depleted in 53Cr whereas the lower part of the slightly younger BIF is significantly enriched. Changes in the concentration of redox sensitive elements, such as chromium itself, uranium and iron, in the two lithologies helps confirm the isotopic evidence for a major ~3 Ga oxidation event. It is possible to use the data to estimate what the atmospheric oxygen content might have been at that time: not enough to breathe, but significant at between 6 x 10-5 to 3 x 10-3 the atmospheric level at present. Oxygen can be produced abiogenically through irradiation of water vapour in the atmosphere as well as by organic photosynthesis. However, the first route seems incapable of yield more than a billionth of present atmospheric concentrations, so the spotlight inevitably falls on a ‘much deep history’ of the action of blue-green bacteria (cyanobacteria) than hitherto suspected.