Underworld: The Mysterious Origins of Civilization

  It also happens to coincide, almost exactly, with Plato’s date of around 11,600 years ago for the sinking of Atlantis, when, the reader will recall, ‘There were earthquakes and floods of extraordinary violence, and in a single dreadful day and night … the island of Atlantis was … swallowed up by the sea and vanished.’63

  ‘You remember only one deluge …’

  I’m not trying to ‘find’ Atlantis, or even to guess where it might have been located – if it ever existed at all – since it is well known that such inquiries lead to madness. I prefer to treat it like any other archaic flood account, whether in the form of myth or purporting to be history, and to consider it solely in terms of its general level of plausibility – a task made easier by its unusual detail and precision. What it tells me at that level is at least the following:

  A devastating global flood occurred around 11,600 years ago. This is interesting, the date coincides with the second of John Shaw’s super-floods and with Cesare Emiliani’s data from the Gulf of Mexico.

  The flood was accompanied by enormous earthquakes. This is plausible because of the close correlation between huge earthquakes, enhanced volcanism, rapid ice melting, and fast post-glacial flooding.

  The island of Atlantis was swallowed up by the sea and vanished in a day and a night. We have seen how isostatic rebalancing sometimes occurred very rapidly and cataclysmically at the end of the last Ice Age and how it is theoretically possible that intense isostatic subsidence in a suitably weakened area of the earth’s crust could have brought about just such a sudden collapse as Plato describes.

  There is one further element of the story that also resonates with scientific evidence, and this is that the flood that destroyed Atlantis 11,600 years ago was but one of many floods …

  Remember that the source of the Atlantis tradition is supposed to have been an ancient Egyptian priest, in conversation with Plato’s ancestor Solon. Here’s how Plato reports the exchange in the Timaeus:

  Egyptian priest: Oh Solon, Solon, you Greeks are all children, and there’s no such thing as an old Greek.

  Solon: What do you mean by that?

  Egyptian priest: You are all young in mind, you have no belief rooted in old tradition, and no knowledge hoary with age. And the reason is this … With you, and others, writing and the other necessities of civilization have only just been developed when the periodic scourge of the deluge descends and spares none but the unlettered and the uncultured – so that you have to begin again like children, in complete ignorance of what happened in early times … You remember only one deluge, though there have been many …64

  As a general synopsis, I have to say that the priest’s comments fit reasonably well with the three global superfloods and countless lesser deluges that we now know did occur at approximately 15,000, 11,000 and 8,000 years ago. Moreover, his placing of the Atlantis flood anywhere in this period (the only period in the last 125,000 years when there actually were floods of the kind described) is – if you stop to think about it – quite an achievement in itself.

  An aggressive little bugger from Yorkshire …

  We’ve seen that it was Cesare Emiliani who first drew serious attention to the possibility of post-glacial superfloods. In a paper published in Science magazine in 1975, he and a group of colleagues presented startling evidence from deep-sea cores from the north-eastern part of the Gulf of Mexico. The evidence revealed ‘a 2.4 per cent isotopic anomaly between 12,000 and 11,000 years ago’, which the authors correctly interpreted as having been caused by ‘the occurrence of major flooding of ice meltwater into the Gulf of Mexico … centring at about 11,600 years before the present’.65

  At the time Emiliani’s ideas were not well received. As Isaac Asimov was later to comment: ‘The suggestion was largely ignored because it was difficult to imagine the ice melting that fast, but in 1989, John Shaw … made a suggestion as to just how such floods might come about …’66 I thought that I had already fully understood Professor Shaw’s catastrophic scenario of how the three great deluges were caused by sudden releases into the world ocean of pent-up meltwater from behind ice dams. But as I looked more closely at his research, and at the transcript of the lengthy interview he had given us in February 1999, I began to realize that his story had hidden complexities and that the cataclysms he described could have been far more severe than I had initially supposed. For it was not just a matter of very rapidly rising seas submerging and washing away low-lying coastal areas – although there was an immense amount of that! – but also of the true character and extent of the run-off floods on land as the ice-caps melted down and the glacial lakes burst their ice barriers.

  Shaw’s interest in this problem does not begin with floods but with drumlins:

  Drumlin: elliptical, streamlined hill composed of till [unstratified glacial deposit consisting of boulder clay and rock fragments of various kinds] deposited beneath moving glacial ice. Drumlins commonly are found in clusters with their long axes roughly parallel to the direction of the ice movement. They slope steeply in the direction from which the glacier came and gently in the direction in which it moved. They vary in height from 6 to 60 metres and in length up to several miles … Drumlin fields may contain as many as 10,000 drumlins; one of the largest fields is in the north-western plains of Canada.67

  Based at the University of Alberta, Professor Shaw has Canada’s drumlins at his doorstep, at least in a manner of speaking, so it’s not surprising – as a geologist – that he should have views about them. But the reactions that his views have elicited amongst other geologists are harder to understand:

  When I go to conferences, people yell at me, people get angry and they yell and scream, and are constantly bringing in diversions because they don’t want the story to be told. And being an aggressive little bugger from Yorkshire anyway, I tend to fight back.68

  At a recent conference in Sweden a senior Quaternary geologist instructed Shaw: ‘Don’t bring your ideas here’:

  So I looked at him and grinned, and next day I gave the paper. And then it was rejected and not published in the conference proceedings so I put it on the Net, and that’s where it is now … If I were a young assistant professor I wouldn’t be kept and I wouldn’t have published either and people would say my ideas were barmy.69

  What, one might ask, is all the fuss about? It seems hard to believe that geologists could come close to excommunicating such a senior and widely respected colleague as Professor Shaw simply for expressing an original scientific opinion on the matter of elliptical, streamlined hills. I mean, who cares?

  In fact, we should care, says Shaw, because the drumlins and other ‘hummocky’ landforms strewn across Canada are evidence of continental floods of biblical proportions – floods of water in some cases hundreds of metres high-that roared out from beneath the ice-caps during the last deglaciation, destroying or mangling everything in their path. Shaw explicitly suggests that many elements of the universal myth of the deluge may be explained by such floods pouring down off the land – intimately linked, as they were, to the episodes of sudden and ferocious sea-level rise that took place between 15,000 and 8000 years ago.70

  Slow and gentle or fast and furious?

  Although there is no single explanation for the formation of drumlins to which all geologists subscribe, most see them as the result of a relatively slow subglacial process involving first the lodgement of a huge mass of ‘till’ on the bedrock beneath the glacier and subsequently its moulding into the classic ‘streamlined-hill’ shape by the flow of the ice itself.71 Such gradualistic theories have dominated the earth sciences and archaeology since the end of the nineteenth century, creating an exceptionally difficult environment in both disciplines for the exploration of alternative hypotheses requiring any kind of sudden change or catastrophic agency. Because John Shaw’s theory requires both, it was inevitable that it would face stern opposition. Nevertheless, he has stuck to his guns since first putting his ideas forward in the 1980s and has gradually seen
a convergence of evidence building up in his favour, including ‘subglacial landforms, surface water isotopic composition of the Gulf of Mexico, and the sedimentology of cores from the Gulf’.72

  At risk of reducing a massively documented and complex argument to statements of ludicrous simplicity, I think it is fair to say that Shaw himself does not claim to have found any definitive, all-inclusive explanation for the formation of drumlins but believes them to be features that are caused in different ways by different kinds of cataclysmic floods and not, as has traditionally been thought, by ice moulding. For example, ‘on the evidence of form and structure’, his interpretation of the Livingston Lake drumlins in northern Saskatchewan is that they are ‘infills of inverted erosional marks scoured in the ice-bed by subglacial meltwater’.73 In other words, forget about the old notions of ‘lodgement’ and ‘moulding’ that generations of geologists have had hard-wired into their logic-circuits. Consider the possibility, instead, that the end of the Ice Age was much less genteel – as, indeed, we already know that it was in almost every other measurable characteristic that we have encountered – and that the vast drumlin-fields at Livingston Lake were created by apocalyptic meltwater floods.

  This is precisely Shaw’s scenario and he believes that the ‘subglacial land-forms’ – i.e. the drumlins themselves – are his most powerful evidence:

  When I first looked at drumlins – this is how it all started for me really – I thought, My, they look just like erosional forms on the sea-bed – which are negative forms of course – but these ones are positive. How can that be? Then the idea came to me, OK, if you erode upwards into the ice and then fill in the cavities with sediment that’s what you would get. And so we went and dug holes and found out that the sediment corresponded to filling in from below and very catastrophically.74

  In brief, Shaw’s argument is that at certain stages during the collapse of the Laurentide ice-sheet between 15,000 and 8000 years ago, parts of the slowly moving ice-mass – more than 3 kilometres thick and weighing as much as a giant mountain range – must have rested not on bedrock but on a deep layer of meltwater moving at high speed and under enormous pressure. These ‘turbulent-flows’ would have carried with them tremendous volumes of sediment ranging from finely grained clays to huge stones and boulders, and it is easy to see how a cavity eroded into the base of the ice-mass – where it rested on the running water – would quickly have become filled up and densely packed with sediment forced in from below. The result, like any object created in a mould, would have taken on the characteristic shape of the mould – which in the case of this kind of erosion is streamlined, elliptical and hill-shaped – and might then have been sealed within the ice, and carried further by it, until it was ultimately released by generalized melting.75

  Different kinds of landforms created by subglacial meltwater floods of varying depths. Based on Shaw (1998).

  Take a few thousand such objects of varying sizes, dump them in northern Saskatchewan, and you have the Livingston Lake drumlin-field.

  Shaw believes that other drumlin-fields in Canada have been created in a different way – again involving glacial meltwater rather than ice, but this time as a direct erosional agent on bedrock or depositional landforms:

  Drumlins around Peterborough and Trenton, Ontario, are mainly erosional; their internal stratigraphy is relatively undisturbed … Drumlins in Ireland contain complex glacigenic sequences … The form of these Irish drumlins … is almost entirely erosional.76

  Returning again to his notion of powerful floods running under immense pressure at the base of the ice-sheets, Shaw draws attention to the drumlins of Beverley Lake field in Canada’s Northwest Territory, which he suggests were sculpted by these floods, and to erosional marks – also caused by floodwater-in the bedrock near Kingston, Ontario:

  Concerning the depth of the flow, it is clear that the [Beverley Lake] drumlins … must have been submerged in the formative flow … minimum depths of about 20 metres were required … Erosion marks in the bedrock in the Kingston area, Ontario, indicate subglacial meltwater flows that have widths of more than 60 kilometres. Spectacular erosional marks along the north shore of Georgian Bay, Ontario, also indicate broad subglacial meltwater flows. On a helicopter traverse along the north shore of Georgian Bay, a single field of bedrock erosional marks was noted that had a width of at least 50 kilometres … [These] drumlins and erosional marks indicate meltwater floods that were competent to remove the largest boulders … Flow widths, equal to the widths of drumlin and erosional-mark fields, were in the range of 60 to 150 kilometres.77

  I think it is worth re-emphasizing Shaw’s figures, and their implications. He is talking about turbulent, energetic floods 20 metres deep flowing in vortices at high speed and pressure, under the main ice-sheets, across fronts up to 160 kilometres wide. Only floods on such a scale and of such violence could have sculpted the drumlin-fields and hummocky terrain and tortured pitted scablands of Canada and the United States and carved out other remarkable features such as the extremely large through valleys – including those containing the Finger Lakes – that lie to the south of drumlin-fields in northern New York State.78 ‘Volumes of water required to sustain such floods’, observes Shaw, ‘would have been of the order of one million cubic kilometres equivalent to a rise of several metres in sea-level over a matter of weeks.’79

  Drowned coral and floating ice

  Of course, when water flows under ice, severing its connection to bedrock, the ice begins to move – ‘surge’ is the technical term:

  Subglacial meltwater sheets with thicknesses of several tens of metres occurred over vast areas of the Laurentide Ice Sheet. The decoupling of glaciers from their beds as a consequence of increased water pressure is used increasingly to explain their rates of sliding. The scale of this process implied here is much larger than that considered for modern glaciers. Nevertheless, the effects should be similar … In short, the glacier is expected to surge.80

  There is indeed compelling evidence of a series of massive glacial surges at the end of the last Ice Age. These correlate with meltwater pulses and peaks of sea-level rise, recorded, for example, in ‘drowned’ reefs of Acropora palmata from the Caribbean-Atlantic region near the island of Barbados. Acropora is an efficient tracker of rising sea-level because it is a light-loving coral that dies at depths greater than about 10 metres. The Barbados reefs were drowned three times at the end of the last Ice Age – at approximately 14,000, 11,000 and 8000 years ago81 – and so suddenly and deeply on each occasion that they now form three distinct steps, one for each flooding peak (rather than having crept towards shallower water as would have been the case with more gradual sea-level rises). Shaw and his colleague Paul Blanchon at the University of Alberta conclude in a 1995 paper in Geology that the reef data confirm:

  three catastrophic, metre-scale sea-level rises during deglaciation. By converting radiocarbon-dated marine and ice-sheet events to a sidereal chronology we show that the timing of these catastrophic rises is coincident with ice-sheet collapse, ocean-atmosphere reorganization and large-scale releases of meltwater.82

  There is also evidence that a cataclysmic feedback mechanism may have been at work between even relatively small eustatic sea-level rises due to meltwater alone and much larger and more sudden events caused by the destablization of entire ice-sheets extending over continental shelves.83 Indeed, in an article in Nature, geologists D. R. Lindstrom and D. R. Macayeal go so far as to identify ice-sheet mechanics ‘as a controlling factor in meltwater production’.84 They then make the very radical and original suggestion that:

  sudden and significant changes in sea-level due to the floating of formerly grounded ice-sheets and attendant ice-dome drawdown might have accompanied the meltwater pulses and these ‘jumps’ in sea-level might not have been recorded in the reef accretion data. Thus a logical mechanism exists by which sea-level may have risen faster and to higher levels than represented by the reef-accretion histories at Barbados.85

/>   In other words global floods that already appear to have been extremely sudden and severe on the basis of the coral-reef data alone – and each ‘drowning’ event required a minimum instantaneous sea-level rise of 5 metres before it would take effect86 – may temporarily have been several magnitudes more severe than the coral-reef record shows. Shaw and Blanchon suggest that a global eustatic hike in sea-level of between just two-tenths and four-tenths of a metre in a period of a few weeks would have been ‘sufficient to free grounded ice and stimulate further ice-sheet wasting, additionally elevating sea-level on the order to 5 to 10 metres or more’.87

  Armadas of icebergs

  Induced by sudden sea-level rises, such sudden wasting at the sea-margins of the ice-sheets would have manifested in equally sudden launchings of fleets of gigantic icebergs. In 1988 the German oceanographer Hartmut Heinrich was the first to come up with the firm geological evidence for such a cataclysmic ‘iceberg-calving’ process during the last Ice Age. By examining deep-sea drill cores sampled at various points across the North Atlantic he demonstrated the existence of widely dispersed layers of ‘ice-rafted detritus’ – millions of tonnes of rocks and rocky debris that had once stood on land, that had been clawed up by the ice-sheets and that had ultimately been carried out to sea frozen into huge icebergs:

  As they melted they released rock debris that was dropped into the fine-grained sediments of the ocean floor. Much of this ice-rafted debris consists of limestones similar to those exposed over large areas of eastern Canada today. The Heinrich layers as they have become known, extend 3000 kilometres across the North Atlantic, almost reaching Ireland.88