The idea that the Earth was like a giant snowball during the Neoproterozoic Era arose from the discovery of rocks of that age that could only have formed as a result of glaciation. However, unlike the Pleistocene ice ages, evidence for these much older glacial conditions occurs on all continents. In some locations remanent magnetism in sedimentary rocks of that age is almost horizontal; i.e. they had been deposited at low magnetic latitudes, equivalent to the tropics of the present day. Frigid as it then was, the Earth still received solar heating and magmatic activity would have been slowly adding CO₂ to the atmosphere so that less heat escaped – a greenhouse effect must have been functioning. Moreover, an iced-over world would not have been supporting much photosynthetic life to draw down the greenhouse gas into solid carbohydrates and carbonates to be buried on the ocean floor. As far as we know the Solar System’s geometry during the Neoproterozoic was much as it is today. So changes in the gravitational fields induced by the orbiting Giant Planets would have been affecting the shape (eccentricity) of Earth’s orbit, the tilt (obliquity) of its rotational axis and the precession (wobble) of its rotation as they do at present through the Milankovich effect. These astronomical forcings vary the amount of solar energy reaching the Earth’s surface. It has been suggested that a Snowball Earth’s climate system would have been just as sensitive to astronomical forcing as it has been during the last 2 million years or more. Proof of that hypothesis has recently been achieved, at least for one of the Snowball events (Mitchell, R.N. and 8 others 2021. Orbital forcing of ice sheets during snowball Earth. Nature Communications, v. 12, article 4187; DOI: 10.1038/s41467-021-24439-4).

Another of the enigmas of the Neoproterozoic is that after and absence of more than a billion years banded iron formations (see: Banded iron formations (BIFs) reviewed, December 2017) began to form again. BIFs are part of the suite of sedimentary rocks that characterise Snowball Earth events, often alternating with glaciogenic sediments. Throughout each cold cycle – the Sturtian (717 to 663 Ma) and Marinoan (650 to 632 Ma) glacial periods – conditions of sediment deposition varied a great deal from place to place and over time. Some sort of cyclicity is hinted at, but the pace of alternations has proved impossible to check, partly because the dominant rocks (glacial conglomerates or diamictites) show little stratification and others that are bedded (various non-glacial sandstones) vary from place to place and give no sign of rates of deposition, having been deposited under high-energy conditions. BIFs, on the other hand are made up of enormous numbers of parallel layers on scales from millimetres to centimetres. Bundles of bands can be traced over large areas, and they may represent repeated cycles of deposition.

How BIFs formed is crucial. They were precipitated from water rich in dissolved iron in its reduced Fe²⁺ (ferrous) form, which originated from sea-floor hydrothermal vents. Precipitation occurred when the amount of oxygen in the water increased the chance of electrons being removed from iron ions to transform them from ferrous to ferric (Fe³⁺). Their combination with oxygen yields insoluble iron oxides. Cyclical changes in the availability of oxygen and the balance between reducing and oxidising conditions result in the banding. In fact several rhythms of alternation are witnessed by repeated packages at deci-, centi- and millimetre scales within each BIF deposit. Overall the packages suggest a constant rate of deposition: a ‘must-have’ for precise time-analysis of the cycles. BIFs contain both weakly magnetic hematite (Fe₂O₃) and strongly magnetic magnetite (Fe₃O₄), their ratio depending on varying geochemical conditions during deposition. Ross Mitchell of Curtin University, Western Australia and his Chinese, Australian and Dutch colleagues measured magnetic susceptibility at closely spaced intervals (1 and 0.25 m) in two section of BIFs from the Sturtian glaciation in the Flinders Ranges of South Australia. Visually both sections show clear signs of two high-frequency and three lower frequency kinds of cycles, expressed in thickness.

The tricky step was converting the magneto-stratigraphic data to a time series. High-precision zircon U-Pb dating of volcanic rocks in the sequence suggested a mean BIF deposition rate of 3.7 to 4.4 cm per thousand years. This allowed the thickness of individual bands and packages to be expressed in years, the prerequisite for time-series analysis of the BIF magneto-stratigraphic sequence. This involves a mathematical process known as the Fast-Fourier Transform, which expresses the actual data as a spectral curve. Peaks in the curve represent specific frequencies expressed as cycles per metre. The rate of deposition of the BIF allows each peak to be assigned a frequency in years, which can then be compared with the hypothetical spectrum associated with the Milankovich effect. One of the BIF sequences yielded peaks corresponding to 23, 97 and 106 ka. These match the effects of variation in precession (23 ka) and ‘short’ orbital eccentricity (97 and 106 ka) found in Cenozoic sea-floor sediments and ice cores. The other showed peaks at 405, 754 and 1.2 Ma corresponding to ‘long’ orbital eccentricity and long-term features of both obliquity and precession. Quite a result! But how does this bear on Snowball Earth events? Cyclical changes in solar heating would have affected the extent of ice sheets and sea ice at all latitudes, forcing episodes of expansion and contraction and thus changes in sediment supply to the sea floor. That helps explain the many observed variations in sedimentation other than that of BIFs. Rather than supporting a ‘hard’ Snowball model of total marine ice cover for millions of years, it suggests that such an extreme was relieved by period of extensive open water, much as affected the modern Arctic Ocean for the last 2 million years or so. There could have been global equivalents of ice ages and interglacials during the Sturtian and Marinoan. ‘Hard’ conditions would have shut down much of the oceans’ biological productivity, periodically to have been reprieved by more open conditions: a mechanism that would have promoted both extinctions and evolutionary radiations. Snowball events may have driven the takeover of prokaryote (bacteria) dominance by that of the multicelled eukaryotes that is signalled by the Ediacaran faunas that swiftly followed glacial epochs and the explosion of multicelled life during the Cambrian. As eukaryotes, we may well owe our existence to Snowball.