This site uses cookies. By continuing to use this site you agree to our use of cookies. To find out more, see our Privacy and Cookies policy.
Skip to the content

IOP A community website from IOP Publishing

Powered by Movable Type 4.34-en

In from the cold: September 2009 Archives

For a long time, the conventional wisdom was that the maximum extent of glaciers during the last ice age was reached in about 16,000 BC. Then it was realized that better calibration of the rate of production of carbon-14, the main radioactive clock we have for that period, required that the date of the maximum be pushed back to 19,000 BC. Now Peter Clark and colleagues, writing in a recent issue of Science, provide further clarity. They show that the maximum extent of the ice lasted from about 24,500 to about 17,000 BC.

Plants and animals don't grow underneath ice sheets, so datable evidence for them means no ice at the site of observation. Assemble several thousand of these dates and you get a picture of the changing extent of the ice sheets. The picture is sharper for some ice sheets than for others, but cumulatively the evidence for several thousand years of near-stability is fairly impressive.

It contrasts with some other guides to the palaeoclimate. Oxygen isotopes in ocean-floor microfossils are a guide (unfortunately) to two variables: sea level (or the volume of the ice sheets, which amounts to the same thing) and water temperature. The heavier of the two common isotopes of oxygen tends to stay behind when molecules of liquid H2O evaporate from the ocean surface, so a) if the light molecules accumulate in the ice sheets instead, the oceanic oxygen gets relatively heavier, but b) this effect is less marked when it is warmer because now the heavier molecules evaporate more readily.

Disentangling these two controls on oxygen isotopes is a major challenge, but accepted pictures of deep-ocean temperature show a quite sharp minimum at about 17,000 BC. What Clark and colleagues have shown, in contrast, is that the ice volume was near to its maximum, equivalent to a sea level about 128 m lower than today's, for several thousand years before that. This agrees with calculated variations of the solar radiation regime due to changes in the Earth's orbit. The calculations are reliable, and show a broad minimum of Northern Hemisphere radiation centred near 20,000 BC. The concentration of atmospheric CO2 also varies only moderately, between 185 and 200 parts per million, during their suggested glacial-maximum span.

This is evidence that the climate can be stable in more than one state. On graphs which span tens or hundreds of thousands of years, the last ten thousand years, the Holocene, stand out as a time of little change. I am not saying that the climate has not changed during this time. We know, for example, that it was a bit warmer in its earlier than its later part, and the cool spell known as the Little Ice Age, which bottomed out about 1850 AD, is well documented. But the Holocene climate was much less variable than the irregularly cooling climate during the preceding hundred thousand years of the ice age. Now Clark and colleagues have shown that peak ice can be stable as well, in the sense of looking rather flat on a graph (but in a much cooler way). We would naturally like to know why.

Another interesting point is that different ice sheets reached their maxima at different dates. The big standout in this respect is West Antarctica, where the maximum was as late as 12-13,000 BC, but the big northern ice sheets, now long gone except for Greenland, had maxima up to a few thousand years apart, reminding me of the smaller glaciers during the Little Ice Age, which peaked as early as the late 17th century in some places and as late as the early 20th century in others. Evidently, as they paced through their stately dance, neither the Little Ice Age glaciers nor the ice-age ice sheets were much good at keeping in step.

Regional variations are therefore a fact of the global climate, about which it is nevertheless legitimate to make on-average statements. One implication is that we should not pay much attention to people who point out correctly that some parts of the world are not warming very much at the moment, but argue wrongly from this that global warming is a myth.

Death and glaciers

| | Comments (2) | TrackBacks (0)

When I was starting my PhD, I met Fritz Müller. He had no stake in my studies, but he insisted on being taken to my field site and delivering a string of inspirational remarks that helped me through the following few years. He was a man of boundless energy and foresight, gifted with the rare ability to make others do what he wanted them to do without them minding (at least, most of them). I keep finding that ideas that come into my head were actually in Fritz Müller's head several decades ago, and it is a tragedy for glaciology that one day in 1980, when he was conducting an excursion for journalists on Rhone Glacier in Switzerland, he felt unwell, sat down and died of a heart attack.

Fritz's wife Barbara suffered the compound tragedy of being widowed by the ice twice. Her first husband, the glacial geomorphologist Ben Battle, was drowned in a meltwater stream on Baffin Island in 1953.

Alfred Wegener may be the best known of all victims of the ice because of his arguments for continental drift, first advanced in 1912. Like Fritz Müller, he was decades ahead of his time, but what many do not realize is that Wegener was actually a meteorologist and glaciologist. Some of his measurements of snow accumulation on the Greenland Ice Sheet are still part of standard compilations. In October 1930, having resupplied the forward camp of his meteorological expedition at Eismitte, near the centre of the ice sheet, he and Rasmus Villumsen began the return journey by dog sled to the western margin of the ice. Villumsen was never seen again, but Wegener's body was found the following May. He seems to have died of overexertion and heart failure.

Perhaps Robert Scott is even better known than Wegener. The staggering story of his 1911–1912 trek to the South Pole, which he reached five weeks later than Roald Amundsen, has been retold times without number. The most recent retelling, Susan Solomon's The Coldest March (Yale University Press, 2001), may well be the best, and not just for the way it sets out the heroism, fading into fatalism, of Scott and his companions. It also offers insight into the role in Scott's tragedy of unusually cold weather, and of some of the physiological implications of low temperature that are not fully understood even today.

Amundsen's journey to and from the South Pole was rather uneventful, but he too died on the ice, in this case the sea ice of the north Atlantic, into which his seaplane is believed to have plunged in 1928 while he was searching for Umberto Nobile. Nobile had flown an airship to the North Pole, but it crashed north of Svalbard on the return flight. Amundsen and his crew joined several of Nobile's crew on the list of fatalities. Nobile himself, and most of his crew, were fortunately found and rescued, not without further loss among the search parties, and Nobile died at an advanced age in 1978.

One of the lessons we learn from these famous fatalities is that your physical condition can have a lot to do with whether you come back alive. Setting out in good health, and in company, and equipped with ways of keeping warm, dry and well-fed, are necessities of safe travel on the ice. But the glacier itself can strike at you without warning. The annual toll taken worldwide by crevasses, avalanches and meltwater is difficult to determine because nobody keeps a centralized record, but deaths on glaciers are reported regularly in the media. You don't have to be famous to fall a victim to the ice.

All of the deaths are tragedies, but many were avoidable. We have learned a lot from the sacrifices of the explorers, and it is a shame that we continue to repeat some of their mistakes. Andy Selters' Glacier Travel and Crevasse Rescue (The Mountaineers Books, 1999), and the freely-available manual compiled by Georg Kaser, Andrew Fountain and Peter Jansson (UNESCO, 2003; 3 megabytes), are two among many sources for an understanding of how to come back from the glacier alive.

When Slartibartfast was given the job of shaping the Earth's surface, the part he enjoyed most was the fiords. I am sure he could explain how to create a fiord much better than I can, but I will have a go. The explanation is informative but disturbing.

To make a fiord, you need fast-flowing ice, which implies that the glacier bed must be a site of intense dissipation of energy. Most of the motion is by basal sliding, which implies that the bed cannot be frozen and also suggests that much of the energy will be used up in entraining and removing bed material. This agrees with the visible fact that fiords are much deeper than the mountain and plateau terrains through which they are threaded. It also agrees with an explanation of what is going on that appears in a recent study.

Michèle Koppes and colleagues measured the volume of sediment beneath the waters of Marinelli Fiord in Tierra del Fuego. These waters have taken the place of the tongue of Marinelli Glacier, which has retreated 13 km since about 1945. Before then, its terminus was stable at the mouth of the fiord. The sediment must have been delivered by the glacier since its retreat began. The implied rate of erosion is an astonishing 39 millimetres/yr, give or take 40%. That is, the sediment implies that the glacier has stripped 39 mm off the land surface during each of the past 50–60 years. It seems certain that most of the erosion must be happening beneath the fast-flowing trunk of the glacier. If so, its bed is being overdeepened very rapidly indeed.

Why was Marinelli Glacier stable before 1945 and unstable thereafter? The answer has something to do with climatic change, but more immediately with the glacier's own behaviour. When it was stable, its terminus rested on the pile of sediment it had itself deposited at the mouth of its fiord. Beyond that there was deeper water, into which the glacier did not advance because the ice arriving at the terminus simply broke off as icebergs. The deeper the water, the faster the iceberg discharge rate. When the climate changed, the glacier was no longer able to deliver the amount of ice needed to keep the terminus where it was. So it retreated.

The trouble is that the retreat moved the terminus down a slope. Not towards the deeper water beyond the glacier's terminal moraine, but towards a deeper part of its own bed, carved by its own erosive handiwork. Just as it was unable to advance into deeper water, so Marinelli Glacier has been unable to stop retreating into deeper water. Its overdeepened bed has aggravated its inability to deliver the ice to keep up the iceberg discharge rate. It has done its best, by thinning and accelerating, with the side effect of increasing its erosive performance dramatically and thus setting itself up for renewed unstable behaviour after the next period of cool climate. But it is not going to settle down this time until it finds a part of its bed that slopes upwards inland, bringing the delivery of ice and the discharge of icebergs back into balance.

Many of the fiords we know about have shallow lips at their mouths, suggesting that they are deglaciated analogues of Marinelli. Most of the spectacular glacier retreats that have been documented in recent years, such as those of Jakobshavn Glacier in Greenland, and Columbia Glacier and the glaciers of Icy Bay in Alaska, have, like that of Marinelli, been from shoals produced at the heads of fiords by the glaciers themselves during less benign climatic times. But the scariest fiords of all are the ones we can't see because they are still full of ice, and the scariest retreats are the ones that haven't begun yet.

Marinelli Glacier is a shrimp (area about 160 km2) beside the leviathans of West Antarctica. Pine Island Glacier, draining about 185,000 km2 of the Antarctic Ice Sheet, resembles Marinelli in having a deep channel and a bed that slopes inland. As yet its terminus position hasn't changed by much, but in a study just published Duncan Wingham and colleagues report that the ice just upstream from the terminus of Pine Island Glacier is now thinning at about 80 metres/yr, more than ten times the rate of just ten years ago.

The technology for finding out more about glaciers keeps getting more diverse. In addition to bouncing laser beams and radar waves off them to find out about changes in their shape, and to finding out about changes in their mass by weighing them, an intriguing recent innovation is to listen to them.

At glaciers that terminate in tidewater, the noise of great chunks of ice cracking off and collapsing into the water inspires awe if you happen to be within earshot, but I am thinking here of noises that are detectable not by the human ear but rather by seismographs. Most earthquake waves are sound waves, that is, more or less rapid fluctuations of pressure. Except that the seismographs are picking up waves that have travelled through the earth rather than the air, seismic waves are not different fundamentally from those to which our ears are sensitive. But the study of icequakes is still in its infancy.

Icequakes are fluctuations of pressure that originate in sudden motions of glacier ice, not of the rocky earth. They have been known for quite some time, but their interest and potential were first highlighted by Göran Ekström and colleagues (here, but you can't get back to this page from there for some reason), who filtered the records from seismic observatories and identified several hundred long-period (that is, rumbly) events that did not look like ordinary earthquakes. Of the 71 found in areas usually regarded as seismically quiet, 46 were from glaciers, and of these 42 were from fast-flowing outlet glaciers in southern Greenland.

People sat up and took notice. Since the Ekström report in 2003, at least three different ways in which glaciers can make a noise have been documented. Their original suggestion can be interpreted as stick-slip motion at a patchily-frozen glacier bed. The ice lurches jerkily downstream, and some of the energy thus released finds its way to the seismic observatories, or to specially-deployed arrays of seismometers in the neighbourhood of the icequake.

Abrupt failure at the propagating tip of a water-filled crevasse can do essentially the same thing. This is an explanation favoured by Shad O'Neel and Tad Pfeffer, and is interesting because such failures seem to be possible precursors of even bigger events due to the calving of icebergs.

That brings us to the third mechanism, the calving itself, and to a recent analysis by J.A. Rial and colleagues that may be the most interesting of all. These authors studied Jakobshavn Glacier in west Greenland, a major ice-sheet outlet of which the terminus is falling to pieces dramatically. The rumblings they observed are consistent in many ways with earlier explanations, and in particular with the idea that things start with an iceberg breaking off the end of the glacier. But the rumblings go on for tens of minutes, and end with a large "culminating event" that can often be pinpointed to a part of the glacier margin 10-12 km upstream from the calving front. Evidently the loss of back-pressure due to loss of the iceberg leads to abrupt release of stress about half an hour later, at the frozen contact between the glacier and its valley wall.

This concern with the physics of how glaciers lose icebergs may strike you as finicky. What makes it worthwhile for the authors, and for their readers, is given away when they write that this sort of pattern suggests "a highly repeatable process of local glacier dynamics currently unknown to us". In other words, something genuinely new is awaiting efforts to make sense of it.

What is more, it offers the prospect of finally being able to measure the calving rate. Seismic measurement of calving rates is not around the corner, but it would clear away one of the major obstacles to quantifying the mass balance of large tidewater glaciers. The rate at which the terminus advances or retreats is not the main part of the problem. Rather, we need to know the rate at which ice is arriving at and discharging from the terminus. At present the best method, at least for regular monitoring, is to watch the icebergs falling off and guess at their sizes. Listening to them might turn out to be a better idea.