by Ted Nield
Supercontinents are arid because moisture cannot reach their interiors; but on smaller continental blocks this situation is reversed. After supercontinent fragmentation, more rain tends to fall on more land, and rock weathering speeds up. Because the continental fragments were then sitting entirely within the tropics, weathering rates were particularly high. What is more, the newly erupted basalt provinces were especially susceptible to chemical weathering.
So, as more rocks were weathered, even more carbon dioxide was removed from the atmosphere and delivered to the seas. The length of coastline increased, as did the area of shallow shelf sea, providing even more habitat for stromatolite-forming algae to colonize.
As these progressive effects took hold, they drove the climate into colder and colder territory. Icecaps began to form and expand. Normally, when icecaps expand over continents lying at high latitude, they exert a negative feedback on the process because they cover up more rock, preventing weathering and leaving more carbon dioxide in the air to keep the planet warm. But that didn’t happen. With the continents far away from the poles, no brake was applied. As the icecaps crept Equatorwards to within about 25–30 degrees of latitude, they passed a point of no return. Earth was doomed to ten million years of icy stillness.
This point came because, at a certain coverage of brilliant ice and snow, the amount of heat reflected back into space became so high that the cooling process was unstoppable. The Earth system had no choice, the cooling effect had nowhere else to go but completion, encasing the whole surface of the planet in ice. This was how Lasagne World gave rise to Snowball Earth.
Iceworld
Once the planet was encased from pole to pole, the Earth system was frozen, literally and figuratively. Apart from rare nunataks of rock, the tallest mountaintops poking above the endless ice plain, the whole sunlit globe shone in only two colours: blue above and white below. On Iceworld there was no evaporation and no clouds anywhere. The hydrological cycle, in which water evaporates from the sea to be deposited as rain and snow on land and so returns in rivers to the sea once more, became restricted to the precipitation of the small amount of water that would ‘sublime’ from the ice surface: in other words, go straight from ice to vapour. And beneath their icy carapace, the seas became stagnant as interchange between them and the atmosphere ceased.
Yet deep below, far underneath ice, oceans and rocky crust, churning away irrespective of the catastrophe at surface, Earth’s planetary heat engine rolled on, driven by the imperative to dissipate its radiogenic heat. Beneath the sealed seas, volcanically active spreading ridges pumped their acidic, superheated, mineral-rich waters into the frigid water; while above the ice, volcanoes that poked their hot heads through the frozen veneer spewed their gases into the atmosphere.
And there they stayed. No rain washed these gases out of the air and into the seas, and carbon dioxide began building up. It is a fairly safe uniformitarian assumption that Neoproterozoic volcanoes gave off at least as much gas as volcanoes today. From this it can be calculated that after a snowball event lasting ten million years, levels of carbon dioxide in the atmosphere would have risen possibly as much as a thousandfold. Earth’s inner fire was about to save the world from the reign of the ice.
The end would have come suddenly. As the greenhouse effect kicked in, temperatures swung wildly upwards, to perhaps as much as 50°C at the ocean surface. Evaporation began again, further enhancing the greenhouse, since water vapour is one of the most powerful greenhouse gases of all. The water cycle now went into overdrive; torrential rainfall washed carbon dioxide out of the atmosphere, creating acid rain that landed on the newly exposed land surface (strewn with glacial rock flour) and so dissolved its minerals even more quickly. The reborn rivers returned huge quantities of bicarbonates to a sea already saturated by ten million years’ worth of volcanic carbon dioxide pumped into it by submarine volcanoes.
Massive limestones were then deposited in shelf seas worldwide: seas that were also progressively deepening as the liberated water of the ice sheets filled up the ocean basins. Carbonates precipitated out on the seabed, depositing thick limestones directly above glacial deposits, with no sign of a time-break. Some of these limestones contain crystals of the calcium carbonate mineral aragonite, which normally only precipitates today from supersaturated pore-fluids as microscopic, needle-like crystals. Aragonite crystals that formed on the Neoproterozoic seabed took on gigantic dimensions as giant fans, some as tall as a man.
The ice-break of a snowball would have been Earth’s most dramatic spring. The gradual release of the last Ice Age’s grip 10,000 years ago must have been as nothing to the chaos that prevailed as the Neoproterozoic icecaps retreated. Persistent winds of over 70 kilometres per hour blew over much of the Earth’s surface as a result of the vast air-pressure differences between the thawing tropics and the polar caps. These winds were not passing storms but lasted for years, causing huge oscillation ripples to form in sediments accumulating as deep as 400 metres: nowadays far beneath the reach of even the biggest storm waves. The ocean-surface waves generated by these global hurricanes probably exceeded seventeen metres in height, over huge tracts of sea. Below, meanwhile, the dissolved volcanic iron built up during the snowball finally oxidized and precipitated, and Banded Iron Formations suddenly made a comeback after a hiatus of a billion years.
Can all this really have happened? The geological evidence – glacial deposits at low latitudes, directly overlain by limestones, and the reappearance of BIFs – all fit the hypothesis. Moreover, they make sense of phenomena that have long been seen as ‘anomalies’, giving the Snowball Earth model immensely persuasive explanatory power.
As early proponents of continental drift had found, ‘old-fashioned’ geological evidence is qualitative, and often open to several possible interpretations. Powerful hypotheses, especially when they are ‘non-uniformitarian’ in the sense that they have no precise modern analogue on the planet today, are at once exhilarating and disturbing. Doubters mutter that a powerful ruling theory is driving interpretation. Could the whole snowball story be no more than a ‘selective search after facts’? What of ‘multiple working hypotheses’?
We have already seen some of the many different isotopes of carbon, the element of life, and how their different atomic weights lead them to behave differently in the natural environment. Carbon 12 is the common form, but there is an unusual heavy isotope, carbon 13, which gains its extra unit of mass by having an extra neutron in its nucleus. The carbon that comes out of volcanoes (mostly as carbon dioxide) contains both isotopes in a well-known ratio. But any carbon that has been involved in life processes has a different signature, because photosynthesis, where everything begins, prefers carbon 12. Living tissues therefore contain lower than average amounts of 13C; but conversely, limestones (principally calcium carbonate) that form at times when life is thriving have above average values of 13C because they are made out of the carbon left over in the environment after life processes have taken their share.
If you test the carbon-isotope ratios in limestones immediately underlying and overlying the snowball’s glacial deposits, remarkable changes are revealed, bringing impressive support to the snowball model. When pre-snowball limestones were laid down, life was thriving; so 13C values begin high but drop steeply as the contact with the glacial deposits approaches. This says that life was shrinking back with the onset of snowball conditions, leaving more 13C around in the seawater to be incorporated in limestones. For ten million years or so that the snowball lasted, no limestones were laid down. However, the first limestones deposited after the snowball – the ‘cap carbonates’ – remain low in 13C because life had yet to recover. Then, gradually, 13C values rebound as resurgent life in the recovering shallow seas fractionated ever more 12C into living tissues.
Between 710 million and 580 million years ago, the snowball cycle happened twice, possibly three and (some say) even four times during this never-to-be-repeated interval in our planet’s histo
ry, as the supercontinent Rodinia split apart. So why did they stop happening? Why did snowballs not happen, for example, during our most recent glaciation, commonly known as ‘the’ Ice Age, which ended between ten and twelve thousand years ago?
The reason lies in the fact that, unique among supercontinents, Rodinia seems to have straddled the Equator, meaning that the world had no land at either pole. Never has that coincidence of low greenhouse gases, weak Sun and tropical concentration of landmasses come about again. Since the end of the Neoproterozoic, despite ice ages aplenty, none has ever gone to anything approaching a snowball. Since the break-up of Rodinia, the safety catch has been back on. Iceworld was finished.
Snowball or slushball?
In structuring the story as I have, however, I have taken one particular route through a mass of scientific evidence treading a line of stepping stones across a torrent of argument. The Snowball Earth hypothesis remains controversial and contested.
Take, for example, the one big question mark over the whole idea of the snowball model as advocated by Paul Hoffman and his supporters. How could a total white-out, involving a global ice-cover perhaps many kilometres thick, ever have allowed photosynthetic organisms in the oceans to survive? Clearly, life did survive successive snowballs and therefore, argue the theory’s critics, each so-called snowball must really have been a ‘slushball’, preserving ice-free refuges at the Equator.
The slushball model, on the other hand, produces a slower deglaciation, with no sudden ‘flip’ from icehouse to greenhouse, and has atmosphere and hydrosphere remaining in balance throughout. A Slushball Earth might not even have had uniformly anoxic oceans. Can this explain the recurrence of BIFs? Some geologists have found evidence for erosion and deposition, with glacier advance and recession at this time. Those who believe in the Hard Snowball have to say this all took place during the short deglaciation phase, because their model predicts a total shutdown of such hydrologically dependent activity for ten million years. This may be true; but is it?
Could computer models help resolve these issues? They tend to produce a runaway snowball effect when levels of atmospheric carbon dioxide are about the same (or only slightly lower) than today. However, not all climate computer models display this kind of flip-flop instability. For example, models in which no heat is transported across the ice-line are less likely to go to total snowballs because the tropics stay warm enough to remain melted. Programs that couple atmosphere and ocean circulations are also resistant to global glaciations, because they allow ocean convection at the ice-free tropics. So climate models, on their own, do not provide conclusive evidence because they can be tweaked, within quite reasonable bounds, to support a number of plausible outcomes.
So although models that result in ‘slushball’ solutions keep biologists happy, they do not appear to explain the geological evidence, especially the ‘cap carbonates’, and the temporary return of BIFs, quite as well. The full snowball model also demands that deglaciation be very rapid, which is consistent with the way cap carbonates seem to have been deposited, without any time gap, on top of glacial deposits.
Resolution of this impasse hinges on one crucial question: exactly how thick did the ice of Snowball Earth get? We know how sea-ice thickness and surface temperature are related in modern oceans, and we know from computer models roughly how cold it would have been during a snowball episode. Applying these simple formulae in a uniformitarian way suggests that, during a full snowball, ice at the tropics should have exceeded a kilometre in thickness, far too thick for any light to get through. How, under a full snowball, could slimeworld have survived even one ten-million-year-long night, let alone two (or more)? Clearly, somehow it did.
In 2000 a new suggestion came along to break the impasse, involving a more biology-friendly ‘thin ice’ model. In this model the ice-cover is total, but just a few metres thick at the tropics: thin enough to allow light through while providing enough of a seal to restrict the hydrological cycle to a minimum and prevent the oceans breathing. The idea came from David Pollard and James Kasting, of the Earth and Environmental Systems Institute at Pennsylvania State University.
With ice, thinness and transparency go hand in hand. Opaque ice is compelled to be thick. You could call this another ‘greenhouse effect’, because transparent ice, on the other hand, traps heat like a greenhouse’s glass panes, melts itself from below, and stays thin.
Today’s sea ice is full of inclusions that scatter the light and make it highly opaque. Critics of the ‘thin ice’ idea were quick to point this out; for if ice at the Neoproterozoic tropics was like modern sea ice, it would have been opaque, hence also thick. However, Pollard and Kasting are not so sure about this uniformitarian approach. According to them, the ice at the tropics would have formed by a combination of two processes: the Equatorward flow of ice from high latitudes, forming ‘sea glaciers’, and water that simply freezes on to the bottom of the sheet.
When sea ice grows in the modern-day Baltic, for example, water freezes to the underside of the ice sheet, trapping pockets of brine that make the ice opaque. But Pollard and Kasting’s model suggests that Neopoterozoic Snowball Earth ice would have formed at a mere seven millimetres per year – much more slowly than modern sea ice. Such slow freezing would have produced much clearer ice.
As for ‘sea-glacier’ ice, its nearest analogue today is the ice seen on land glaciers, like those in Antarctica, which form by the accumulation of snow. In land-glacier ice the main light-scattering inclusions are bubbles of air, which originally lay between the snowflakes before they were annealed together by pressure. According to Pollard and Kasting, for Neoproterozoic sea-glacier ice to have been clear enough for it to stay thin, it must have had a bubble density of no more than about 0.32 per square millimetre, and that lies well within the range of bubble densities seen in the upper parts of ice cores taken from Antarctic ice sheets today.
So it appears that life could indeed have survived ten million years in the chiller, because when the freezer door closed the light didn’t, for once, go out. Slushballers, on the other hand, regard the thin ice idea as an unnecessary sophistication. They do not believe that total ice cover is required to ensure that the oceans stagnate and so accumulate ferrous iron in solution. Evidence emerged in 2005, from organic-rich rocks dated to 700 million years ago, that suggested to Alison Olcott of the University of Southern California in Los Angeles and her colleagues that not only was photosynthesis operating during the snowball, but it was widespread, tropical and happening in stagnant water. If their interpretation of the biomarkers in these rocks from south central Brazil is correct, even the photic zone of the Neoproterozoic ocean was oxygen-free. Perhaps there was a thin ice cover, as Pollard and Kasting suggest; but perhaps there was simply no ice at all. Perhaps a tropical, ice-free waistband around the Earth was just not enough to break up the stratification of the global ocean.
As I finish writing this book at the beginning of 2006, another international conference, this time in Switzerland, is planned for the summer. More evidence, from all over the world, will be presented in support of new interpretations of these pages from the greatest palimpsest that may settle the controversy between snowball and slushball. But just as Lemuria finally sank beneath the waves of new knowledge, today’s closest approximation to truth slides into myth as the latest ideas are subjected to the evolutionary pressure exerted by the realities of new evidence.
Hope and glory
Can it be entirely coincidental that, after three billion years during which the pinnacle of evolution was green slime, complex life burst into existence just as the last snowball melted away never to return? Could it be that, if a low-latitude supercontinent called Rodinia allowed the snowballs to happen, and if the snowballs somehow gave life the kick in the genes it needed to develop complexity, Rodinia really was our motherland? Could the vagaries of the Supercontinent Cycle be the main reason why, in place of universal lasagne, we have in Jacques Prévert’s words, ‘New York
passions, Parisian mysteries, the little canal at Ourq, the Great Wall of China, the river at Morlaix, legionnaires, torturers, rulers, bosses, priests, traitors, pretty girls and dirty old men’? Could Rodinia be the reason we have such a diverse world? Could Rodinia be the ultimate reason we are all here today, doing what only humans do: wondering how we got here?
To try to answer this question, we need to know something rather accurately. When exactly did complex life first develop? Only when we know that can we hope to judge whether there is a case here to answer.
On 12 November 1931, three years before his death at the age of seventy-six, Sir Edward Elgar and the London Symphony Orchestra performed the trio from the Pomp and Circumstance March No. 1 (‘Land of Hope and Glory’) for the opening of EMI’s new recording studios at 3 Abbey Road, not far from a certain pedestrian crossing later made world famous by the Beatles.
The Abbey Road recording was not Elgar’s first foray into this newfangled technology; he made his first recording in 1914, weeks before the First World War broke out. But despite the poor quality of contemporary reproduction (imagine a wind-up gramophone playing 78rpm shellac discs using a needle and a big horn as acoustic amplifier), these ancient recordings actually contain a wealth of sound detail that was invisible – or rather, inaudible – to the old technology. With digital remastering all kinds of unimagined detail can now be heard. All that information was always there, but only the new tools allow it to be revealed. The same is true of the geological record.
When geologists began looking systematically for fossils for the first time in the nineteenth century, using William Smith’s discovery that you could identify and correlate strata of any given age by the fossils they contained, they noticed that rocks below the Cambrian system were barren. The term they gave to the whole ‘Cambrian and younger’ geological record was ‘Phanerozoic’, which means ‘evident life’. The apparent suddenness of life’s appearance posed a great problem for evolutionary theory because at 542 (plus or minus one) million years ago, and seemingly from nowhere, nearly all the main animal body plans (arthropods, molluscs, echinoderms and so on) seemed to burst on to the scene and hit the ground running, swimming and burrowing like there had been no yesterday. It all gave Darwin sleepless nights.