Frozen Earth: The Once and Future Story of Ice Ages

by Doug Macdougall

  Still, notwithstanding the seashell argument, even some of the opponents of the glacial theory had to admit that it would be difficult, if not impossible, to transport large erratic boulders in water over long distances, no matter how violent the storm or flood. It could be shown by simple physics that it couldn’t be done. So they came up with the ingenious solution mentioned in chapter 2: the erratics might indeed have been carried by ice, but ice that was floating on formerly more extensive seas, transporting boulders from a northerly source. If parts of the continents had been submerged in the past, they reasoned, the icebergs could have floated over the sunken land, dropping their rocky burden as they melted. That would explain the presence of ocean shells in the drift. It was the idea of drifting icebergs that first led to use of the term “drift” for the characteristically chaotic sediments left behind by glaciers—sediments that have neither the well-defined layers nor the uniformity of grain sizes that characterize those deposited in water. The term is still in use today. Geologists also refer to such material as being unsorted, because it encompasses materials ranging in size from grains of sand and occasional shells to the erratic boulders themselves.

  Drift and erratic boulders were not the only glacial features studied by early ice age researchers, but because essentially identical materials could be observed directly associated with glaciers in the Alps, these deposits were among the most persuasive evidence of past glaciation. Every existing mountain glacier carries a large amount of rock debris that will eventually become glacial drift. Some of it falls onto the glacier surface from the surrounding valley walls, and some is actually plucked from the bedrock below by the ice itself. Beginning with Agassiz’s systematic studies at his glacier observatory, a series of investigations also showed that glaciers flow, and do so at significant rates. The rock debris is carried along with the flowing ice, and, at the snout of the glacier, dumped in a chaotic pile of large and small boulders, gravel, sand, and silt—a feature known as a moraine. Actually, glaciologists distinguish many types of moraines, but in its most general sense, the term—like the term “glacial drift”—just refers to the debris carried by a glacier. Terminal moraines mark the farthest extent of a glacier, lateral moraines form along the sides of mountain glaciers, and medial moraines in their middles, the result of tributary glaciers entering the main ice flow. Figures 5 and 6 illustrate a few varieties of moraines. Some types—for example, medial moraines—may exist on the ice of an active glacier, but can also be distinguished long after the glacier has melted away, because they form a longitudinal ridge in the middle of a glacial valley.

  Figure 5.That glaciers flow is particularly apparent from the air. This glacier in Greenland flows toward the observer, carrying on its surface ribbons of rocky material that have fallen onto its surface from the valley walls, forming moraines. Because many tributaries join the main glacier, it becomes more and more banded with moraines downstream as each tributary adds its contribution of debris. Photograph courtesy Professor Michael Hambrey, Liverpool John Moores University.

  Ice in a glacier, like water in a stream, flows under the influence of gravity. A mountain glacier accumulates snow at its upper end, where the average temperature is low, and loses ice by melting at its snout. In places like Greenland and Alaska, some glaciers flow directly into the sea, where gigantic chunks break off and form icebergs. If snow accumulation and melting are more or less in balance over a significant length of time, the size of a glacier and the location of its lower end will remain approximately constant, in spite of the fact that the ice is flowing and transporting rock debris all the while. When this occurs, the glacier is in a steady state, and very large terminal moraines can be built up. If the climate warms and the glacier melts away, the moraine remains as a distinctive landform—a great ridge composed of pebbles and boulders, marking the previous terminus of the glacier. Sometimes there is a whole series of these features, tracing out positions where the glacier front remained stationary for varying lengths of time before melting back further.

  Figure 6.A cross section through a lateral moraine in Switzerland illustrates the great range of sizes of material it contains. Someone has sorted out piles of sand and boulders of various sizes near the bottom of the picture. Notice that a forest and a layer of soil has developed on the moraine. Glacial deposits from each cycle of the Pleistocene Ice Age have such soils developed on them, indicating that the ice advances were separated by long and relatively warm interglacial periods. This particular moraine can be traced for many kilometers. Photograph copyright Dr. John Shelton.

  Once the nature of moraines formed by contemporary glaciers was understood, it became clear that much of the enigmatic “drift” so common in northern Europe and North America must have an analogous origin in the now-vanished glaciers of the ice age. The moraines left behind by continental-scale ice sheets are really not much different from those of alpine glaciers, except that they exist on a much grander scale. In places they can be traced for hundreds of kilometers, winding through the countryside and marking an ancient glacial boundary. But the ice sheets of the Pleistocene Ice Age didn’t deposit their rubbly burden only as terminal moraines, easily recognized by their ridgelike shape. Some of that material was simply scattered across the landscape as a layer of gravel and boulders without any particular form. Sometimes the drift was shaped by the moving ice into features such as strange teardrop-shaped hills called drumlins, which usually occur in swarms, lined up parallel with one another. Exactly how drumlins form is unclear, but they apparently take shape beneath the flowing ice, their orientation reflecting the direction of ice movement. In other places, drift occurs as long, sinuous ridges of sand and gravel called eskers, which have occasionally been put to use as beds for railway lines in low-lying marshy areas. Eskers are thought to be essentially “negative streams”—rocky material built up in a confined stream that flowed beneath a glacier. When the glacier finally melted away, they were left standing above the surrounding countryside.

  Starting soon after Agassiz published his Études sur les glaciers, geologists began to map out these features wherever they existed. A primary goal of this mapping was to determine the extent of the ice age glaciers, another to discover how they had flowed. Even now, details are being added to the general picture, which emerged quite quickly. It has become clear that the ice age glaciers did not form a single, gigantic ice sheet that extended southward from the North Pole, as Agassiz and his supporters had initially assumed. Instead, there were centers of ice accumulation, located where temperatures were low and the snow supply was ample. In North America alone, there were several centers of thick ice accumulation, with ice flowing out in all directions and in places coalescing with the glaciers of other centers. But some parts of the far north—for example, parts of Alaska—had no glaciers at all, even during the coldest part of the ice age, because of low snowfall.

  It was the mapping that revealed the multiple glacial episodes of the Pleistocene Ice Age. There is an exceedingly simple but very powerful concept in geology, first formalized in the 1600s and still taught to beginning students in the earth sciences: any geological feature that cuts into or across another is younger than the one it cuts across, and any material deposited on top of something else is younger than the underlying material. To beginning geologists, it often seems silly to formalize such a commonsense principle, yet even quite complex sequences of geological events can often be unraveled by applying this concept. It has been used for everything from exploration for oil to working out the cratering history of the moon. When it was applied to the moraines and other deposits left by glaciers of the Pleistocene Ice Age, it showed that there had been several distinct glacial episodes, separated from one another by significant amounts of time.

  The principle of superposition, as the concept just described is sometimes called, provides information about relative time—one deposit is older than another, or some process occurred before another—but not absolute time in years. That only became
possible more than a century after the ice age theory was proposed, after the discovery of radioactivity and the development of techniques that used radioactivity for dating. But even in the nineteenth century, geologists were able to determine that there had been at least three and perhaps as many as five separate expansions of ice far south into Europe and North America during the Pleistocene Ice Age, and that these had been separated by long periods of time with much warmer climates. European and North American scientists gave these episodes different names, and it was not possible to correlate them precisely between continents; however, it was generally agreed that on each continent, the glacial deposits recorded the same series of cold and warm episodes. The changes in ice age climate had been global, or at least they had affected widely separated parts of the Northern Hemisphere similarly. We now know that the glacial periods identified by mapping their deposits were only the last few of a long string of cold and warm cycles stretching back several million years. This knowledge comes not from studies on land, but rather from evidence of a quite different type contained in deep-sea sediment cores. On land, the evidence for the earlier glacial cycles has been almost completely obliterated by the more recent ones, but in the oceans each layer of sediment buries and preserves the ones that preceded it.

  How did the early investigators, without the help of radioactive dating methods, conclude that long time intervals separated the glacial periods? It was a task that required a certain amount of ingenuity. In many localities, it was fairly straightforward to use the principle of superposition to determine that there had been several different glacial advances. In places, younger drift could be observed deposited on top of earlier glacier debris, and in other localities, older moraines had been broken through and partly scoured away by more recent glaciers, which had deposited their own debris far beyond. However, determining just how much time had elapsed between these various events was a difficult problem. An important clue was that between successive glacial advances, soil had developed on the moraines and drift deposits. Soil forms anywhere rocks are exposed to rain—water is an effective solvent, and it also promotes chemical reactions with the minerals contained in rocks. The result is that solid rock dissolves and crumbles and is transformed into the soft clay of soil. Plants, insects, and microbes appear, churning the soil, facilitating even more chemical reactions and adding organic matter. In tropical climates with heavy rainfall, soils have been observed to form on fresh volcanic lava flows within a few hundred years or even less. But in the colder regions from which the ice age glaciers retreated, soils formed much more slowly. Soil layers that developed on moraines and drift between glacial advances indicate that the cycles were separated by relatively long periods of moderate climate. Fossils of plants and animals in the soil paint a similar picture. The interglacial periods were long enough for there to be a complete change of fauna and flora, and the new species were characteristic of temperate rather than arctic regions. When radioactive dating methods became available, it was discovered that through the last six or seven glacial cycles, the times of maximum ice advance were separated by roughly one hundred thousand years, and the warmer periods, with temperatures similar to today’s, lasted ten to twenty thousand years.

  When temperatures rise above freezing, large amounts of meltwater flow across the surfaces of glaciers, along their edges, at their bases, and even through the ice itself. During one of his field sessions on an Alpine glacier, the ever-curious Louis Agassiz had himself lowered down an almost vertical tunnel that had been cut by summer meltwater. It was one of his more foolhardy experiments; the tunnel got narrower and narrower, and eventually bifurcated, and Agassiz lost voice contact with his colleagues on the surface. They kept lowering him, right into an icy torrent deep in the glacier. Fortunately for Agassiz, the glacier-bound stream wasn’t very deep, and eventually his friends became concerned and hauled him up. But it could have been much worse—the amounts of water coursing through melting glaciers can sometimes be huge.

  The meltwater flowing away from a glacier carries with it grains and fragments of rock that were originally embedded in the ice. Ice is indiscriminate about what it carries, but the running water is quite efficient at sorting out the chaotic jumble of particles according to size and weight. It winnows the unsorted glacial drift, dumping the largest pieces at the base of the glacier or close to its boundary, and carrying the smallest grains in suspension over long distances. The meltwater streams build up sand bars in some places, gravel bars in others, and when they are flowing at full force, they sometimes carry quite large boulders along with them. In the northern United States and in Canada, in Scandinavia and northern Europe, man has taken advantage of this combined production and sorting process that is a relic of past glacial action. The sand and gravel deposits of the meltwater streams are scooped up by the truckload and used as construction materials. In overall economic importance, these deposits overshadow all other kinds of mining activity. In these same regions, the action of ice and meltwater has left an aesthetic legacy in addition to a practical one: the undulating topography (and the sand traps) of many a well-groomed golf course.

  The very finest particles of rock carried by the meltwater streams, much smaller than sand grains, are sometimes called rock flour. They are produced by the scouring action of the ice on underlying bedrock, and they are so fine that they remain suspended for very long periods of time and give glacial lakes their characteristic brilliant blue-green color. The scraping and scratching and polishing that produces rock flour leaves very distinctive telltale marks on the underlying bedrock. But it is not the ice itself that does the grinding; even hard, brittle ice at temperatures well below freezing cannot gouge out scratches and grooves in solid rock—it is simply not hard enough. Yet Agassiz and other careful observers of the Swiss glaciers found such features on rocks near the edges of glaciers. They noticed that the scratches were most prominent in areas where the ice had recently retreated, but, like the erratics and drift, could also be found far from any contemporary glaciers. They soon realized that it was actually the rock debris carried by the ice that was producing the scratches. Rocks and pebbles embedded in the ice were being dragged across the underlying surface; the glaciers were like gigantic sheets of sandpaper smoothing out the rocks beneath. In the process they produced the rock equivalent of sawdust: glacial rock flour. When scientists were eventually able to map out the movement of ice within glaciers, they discovered that the base of a glacier is continually being renewed with ice from above, complete with its embedded grit. The natural sandpaper is constantly being refreshed.

  In most places that were glaciated during the Pleistocene Ice Age, scratches and grooves and polished rock surfaces are very abundant. Once you know what to look for, they seem to pop up everywhere. Recently, I walked along the wonderful meandering stone wall built by the artist Andy Goldsworthy at Storm King, an art park not far from New York City. The wall was made from local stones, with those in the top layer chosen for their flat surfaces. It was the Pleistocene glaciers that left them with these surfaces—most show the characteristic scores and scratches of glacial scouring. They are a reminder that just twenty thousand years ago the region was thick with ice.

  Figure 7.A cartoon sketch of Professor William Buckland by the mining engineer Thomas Sopwith titled “Costume of the Glaciers,” showing Buckland dressed for fieldwork. The numerous captions are difficult to read, even in the original, but the lines at Buckland’s feet are noted to be “Prodigious Glacial Scratches” produced by “the motions of an IMMENSE BODY.” Other captions are summarized in the text. From Mrs. E.O. Gordon, The Life and Correspondence of William Buckland (London: John Murray, 1894).

  In Britain, when the debate over the ice age theory was raging in the middle of the nineteenth century, a well-known mining engineer named Thomas Sopwith gently poked fun at the subject of glacial scratches. He sketched the Reverend William Buckland, professor of geology at Oxford, equipped for a field expedition, nattily attired in
long coat and high boots, with scratched and grooved rocks at his feet (see figure 7). A caption indicates that one set of scratches had been produced by glaciers thousands of years ago. Another set, the label says, had been made by the wheel of a passing cart “the day before yesterday.” Sopwith signed his cartoon with the words “Scratched by T. Sopwith.” Buckland was something of an eccentric, a larger-than-life figure, well known to the public and also a highly respected scientist. He was impressed by his friend Agassiz’s work on fossil fish, but initially unconvinced about ice ages. However, after Agassiz personally showed him glacial features in the Alps, and a few years later accompanied him to study moraines, erratics, and glacial scratches in Scotland and England, Buckland was won over and became one of Agassiz’s allies in the ongoing controversy. The public, too, was following the debate—and perhaps chuckling over the seriousness with which learned men treated these little scratches on the rocks.

  The scratching and grinding that produces these marks is also the process ultimately responsible for a peculiar type of sedimentary deposit that was not known in Agassiz’s day to be related to glaciers, but which is now recognized as a key indicator of past ice ages. Characteristically, the sediments of glacial lakes contain what are called varves, a term derived from a Swedish word for periodic repetition. The distinguishing feature of these sediments is a repeated pattern of layers. In detail, in a typical case, the layers are fairly uniform in thickness and sharply separated from one another. They consist of a lower layer of relatively coarse, silty material, overlain by a layer of much finer-grained particles, and then the pattern repeats itself. Each coarse-fine pair constitutes a single varve, which may be a few millimeters to a few centimeters thick. It is now known that each varve represents one year of sediment accumulation. There are some rare circumstances in which similar deposits can be formed in nonglaciated areas, but the vast majority of varves are the product of deposition in a glacial lake. In summer, when meltwater is abundant, streams carry the products of glacial scratching and grinding into the lake, where the coarsest particles sink to the bottom fairly quickly. The finest material—the rock flour—mostly remains suspended, in part because winds keep the water stirred up in summer, and in part because the particles are so small that they settle only very slowly. As winter approaches and the temperature drops, the streams of meltwater coming from the glacier gradually dwindle and eventually stop. There is no longer a supply of new sediment, but over the winter, with a frozen surface to keep the water still, the fine suspended particles slowly sink to the bottom to form the fine-grained part of a varve couplet. A giveaway that the varves are indeed of glacial origin is the presence of dropstones, occasional large pebbles or rocks that are embedded in the fine-grained varves. A bit like erratic boulders on land, they seem, because of their size, to have nothing to do with the varved sediments in which they are found. They are, in fact, quite literally dropped in. Carried out into the glacial lake on pieces of ice, they fall into the fine-grained sediments when their transport melts. If the sediments of a glacial lake harden into solid rock, they can survive long after moraines, drift, and other surface deposits have vanished. Often, preserved varves are among the few remaining indicators of past ice ages, as is the case for one of the Earth’s earliest, dating from about 2.2 billion years ago. Varved sediments containing dropstones still survive from this episode at several places around the globe.