Until now, climate modellers and researchers debating the possible future effects of global warming have looked fairly confidently to the past for clues. The pattern of the Earth’s climate in the past seemed fairly well mapped out. But the first analyses of a new climate record, in the shape of a 3029-metre long ice core drilled from the summit of the Greenland icecap could change everyone’s ideas. For it seems that, far from being exceptional as they thought, sudden swings in the world’s climate were the rule over many thousands of years. Climate researchers now find themselves pondering an uncomfortable question: could greenhouse gases force our climate, stable for the past 10 000 years, into this sort of instability?
Ice cores are unique archives of past climate and atmospheric chemistry. The gases, acids and dust present everywhere in the atmosphere are also contained in the snow falling on the ice sheets of Greenland and Antarctica. There, the chemical fingerprint of each year is preserved unchanged as the annual layers move gradually downward in the icecap. Even more importantly, the stable isotopic composition of the ice depends on the air temperature at the time when the snow was formed and deposited, so glaciologists can follow temperature changes back in time.
Data from the latest core, drilled by the Greenland Ice Core Project, appeared recently in the journal Nature. Before this GRIP core, two ice cores had been drilled to the bottom of the Greenland ice sheet. The first was drilled by American glaciologists between 1963 and 1966 near Thule in northwestern Greenland, and reached some 120 000 years back in time; the next deep core was drilled at a radar station in southeastern Greenland from 1979 to 1981 and went back about 100 000 years. Two deep ice cores have also been recovered in Antarctica: the American Byrd core from West Antarctica dating from 1968 spanned the last 70 000 years, while the Vostok core from East Antarctica, drilled in the early 1980s by Soviet engineers and analysed jointly with French scientists held the previous record, covering 160 000 years.
Advertisement
The new ice core was drilled from 1990 to 1992 by GRIP researchers from Denmark, Switzerland, France, Germany, Iceland, Britain, Belgium and Italy, working under the auspices of the European Science Foundation. It is the most detailed climate record of the past 250 000 years ever obtained, and was the first to span two full glacial cycles – it contains ice from two ice ages and three warm interglacial periods.
It was also the first ice core to be drilled at an ideal location, right at the highest point of the Greenland ice sheet, where the snow deposition is reasonably high and where the horizontal movement of the ice is minimal. This gives a much better resolution of the old layers, makes the dating and analysis of the ice core more reliable and the climatic data easier to interpret. And, most importantly, many more old layers are preserved at the bottom of the ice sheet, which enabled the GRIP team to push the ice core record back another 90 000 years.
Before the latest core was analysed, climatologists had already formed a detailed picture of one of the most important periods for investigating climate change. This period, called the Eemian, is the last interglacial, which ended 114 000 years ago. The Eemian is a prime target for research into past climate, because this period was warmer than our present climate and many researchers think it may provide a good model for the way our climate will evolve in the future.
Based on evidence from sea sediment and pollen records, climatologists had worked out that in the Eemian the climate was more humid and warmer than now, perhaps warmer than at any other time in the past 2 to 3 million years. Hippopotamuses were living along the Thames and the Rhine, and lions and elephants roamed Cornwall. Most of the ice sheet in West Antarctica had melted away, and huge ice sheets covered only Greenland and eastern Antarctica. In the northern hemisphere sea ice was present only in the Arctic Ocean and the water level in the oceans was 6 to 8 metres higher than it is now. The chemical composition of the atmosphere was very much like in our own interglacial, which began 11 500 years ago, and the levels of greenhouse gases such as carbon dioxide and methane were similar to pre-industrial levels. Climatologists could pin down the Eemian interglacial as a period of stable climate, lasting between 10 000 and 12 000 years, when the temperature was about 2 degrees warmer in Europe than now.
But the more detailed results from the GRIP core cast severe doubts on this picture. Climatologists’ picture of the Eemian now looks rather different. Even the length of the Eemian is in doubt. The overall time scale of the GRIP core correlates well with the Vostok ice core from Antarctica, the only other ice core which covered the whole Eemian period. But the GRIP team define the Eemian as the period between the first and the last years with higher temperatures than now, which gives a length of some 20 000 years, interrupted by several cool spells lasting from 70 to 5000 years. ‘The disagreement between the timescale derived from sea sediment and pollen records and the GRIP core may be explained, at least in part, by climatic instability in the early stages of the Eemian,’ say Willi Dansgaard and Claus Hammer from the Niels Bohr Institute at the University of Copenhagen, members of the GRIP team which analysed the core. A longer Eemian period also fits in with a new climate model just developed by Andre Berger and his colleagues at the Institute of Astronomy and Geophysics at Louvain-la-Neuve, Belgium, which gives a length for the interglacial of some 15 000 years.
But the most surprising finding was the climatic oscillations. According to the GRIP team the early part of the Eemian was dominated by several oscillations between warm and cool stages. The temperature dropped by as much as 10 degrees, sometimes within as short a time as ten to thirty years. Some cold spells lasted a few decades, while others lasted several hundred years. After 8000 years of fluctuating conditions, the climate settled into a period of stable warmth lasting some 2000 years. This warm period ended abruptly, and for the next 5000 years the temperature in Greenland fell, to 5 degrees colder than it is now. This temperature is similar to one of the milder periods in the middle of an ice age. Then the climate turned warm again for 1000 years, then cold again, and finally warm, until new climatic oscillations preceded the sudden end of the Eemian period 114 000 years ago.
This is the first time such rapid climatic changes have been observed during an interglacial. According to geochemist and climate modeller Wally Broecker of Lamont-Doherty Geological Observatory at Columbia University, New York, the GRIP results came as a total surprise and could not have been predicted by existing theories or climate models. The new picture is that of an interglacial climate changing from warm to cold within decades. It shows that a warm climate such as the Eemian may be very unstable. This is a worrying observation, even though the enhanced greenhouse effect is not expected to last more than a few centuries at most.
Too simplistic
According to Dansgaard and Hammer, the present models used for predicting a future greenhouse climate may be too simplistic. In general, such models predict that the Earth’s temperature will rise slowly and steadily in the coming decades as a result of the increasing accumulation of greenhouse gases in the atmosphere. ‘The new analyses of the GRIP core tell us that a climate just a few degrees warmer than now may change very suddenly either turning colder or warmer. So by polluting the atmosphere with greenhouse gases, we might force the climate system into an unsteady state, where large natural climatic changes could strike almost like lightning,’ they warn. Not surprisingly, researchers are now keen to find out what triggers such events.
According to the GRIP team one possible cause of the instabilities could have been major changes in atmospheric circulation patterns, such as movements of the Polar Front, the boundary between the cold, dry air masses originating from the polar areas and the warm, humid air masses coming from lower latitudes. Shifts in the ocean circulation could have acted as a trigger. The 5000-year-long cool period in the middle of the Eemian can also be traced in the Vostok ice core. But the shorter climatic shifts can not, suggesting that the fast climate changes are more likely to occur in the North Atlantic Region than around Antarctica. This has led the GRIP team and other researchers to suggest that the instabilities could have been triggered by the North Atlantic current switching on and off. A sudden weakening of this warm current could have allowed the cold Greenland current to carry icebergs much farther south. They say that the length of many of the cold periods is probably connected with a re-ordering of the oceanic circulation.
Catastrophic cooling
Major movements of icebergs, possibly from the West Antarctic ice sheet, could have triggered the catastrophic cooling at the end of the Eemian, when the temperature in Greenland dropped about 14 °C within ten years, the GRIP team suggests. However, the speed of cooling in Greenland would then be rather surprising. ‘The new detailed knowledge from the GRIP core, in combination with the awaited results from the GISP-2 core (drilled in Greenland by an American team) may enable us to narrow down the range of possible changes in the ocean and in the atmosphere that favoured this climate instability’, say Dansgaard and Hammer.
So what would it take to introduce such instability into our present climate, and what does this mean for climate modellers? In its paper, the GRIP team gives this answer: ‘Man is already perturbing one of the factors that may have been involved, the greenhouse gases, and our first tentative measurements indicate that the greenhouse gases may be linked to the climatic changes within the Eemian.’ Broecker agrees that pollution of the atmosphere with greenhouse gases could be more dangerous than we thought. However, Berger does not think that the Eemian is a perfect model for what our climate will be like over the next few millennia. This is because the ‘astronomical forcing’ of the Earth’s climate then was very different from today: that is, changes in the Earth’s orbit around the Sun and in the axis of rotation of the Earth have altered the amount of solar radiation reaching the Earth’s surface (insolation). ‘During most of the Eemian the north and mid latitudes of the northern hemisphere received more energy from the Sun in summers than today, but there were also periods when the insolation dropped far below present levels. In comparison, the insolation has remained fairly stable during our present interglacial,’ says Berger.
However, Hammer points out that these differences in insolation are not sufficient to explain the abrupt climatic changes during the Eemian, nor the exceptional climatic stability of the past 10 000 years. ‘As long as we don’t know why the climate of our present interglacial is so remarkably stable, we should not take the risk of pushing the climate system into an unsteady state with greenhouse gases. Especially not, when we are now aware that climatic instability not only has been the rule in the North Atlantic for the last 250 000 years, but also was a characteristic feature of the warm Eemian interglacial,’ he says. But Berger too considers the GRIP results very important for climate modelling. ‘Any model which is able to generate fast climate changes at a timescale of few centuries to some thousand years will have to be tested against the GRIP data,’ he says.
A new core drilled at the Russian Vostok Station in Antarctica and analysed by Jean Jouzel from the Laboratory for Modelling Climate and Environment in Saclay supports the GRIP finding that the Eemian lasted longer than people previously thought. It takes the Antarctic ice core record back to 220 000 years. John Imbrie, a palaeoclimatologist from Brown University, Rhode Island, whose research interests centre on deep-sea sediments, says that this result and the GRIP analyses give a new perspective on the dynamics of climate change. ‘The conventional picture of the Eemian in this part of the world as a relatively stable, warm period is gone with the wind – which apparently is not the case for the Antarctic region.’ He is not surprised that such abrupt climate changes took place in Greenland, but not in Antarctica: ‘After all,’ he says, ‘the North Atlantic is the dynamic part of the Earth’s climate system. But it is a surprise that the temperature changes were so large.’ Imbrie agrees with the GRIP team’s conclusion that rapid changes of this kind must reflect fast shifts of the atmosphere and oceanic fronts in Greenland, the Nordic Seas and the North Atlantic: ‘These climatic changes can’t reflect major changes in the entire mass of the great northern hemisphere ice sheets.’
A major barrier to understanding the global climate system is that different parts operate at different rates. Given some stimulus, the atmosphere will respond very quickly, in a matter of weeks, whereas parts of the ocean take 1000 years to respond fully, and the great ice sheets take more than 10 000 years. These differences are reflected in the records of ice cores (air temperature) and deep-sea sediment cores (sea temperatures). ‘The big challenge now,’ says Imbrie, ‘is to integrate these different sources of information. It will be a fascinating and challenging task, and it may certainly take years to sort this out’.
Evidence is already accumulating. Imbrie points out that the V19-30 core drilled in the Eastern Equatorial Pacific shows that the ice sheets melted and the temperature of the deep water rose some 130 000 years ago, and this – much more modest – change in the Pacific may correspond to the first warm period of the Eemian recorded in the GRIP core. Another core from the Norwegian Sea records that some ice sheets melted in the Barents Sea on two occasions, early in the Eemian and at its final peak.
Greenhouse gases
The GRIP team now eagerly awaits the results of analyses of gases such as carbon dioxide and methane being carried out on the new ice core by Bernhard Stauffers’ team at the Physical Institute of the University of Bern and by Jerome Chappellaz and his colleagues from the Laboratory of Glaciology and Geophysics of the Environment in Grenoble. The two teams analyse the ice core independently, then compare their results. They hope to pinpoint the role of these greenhouse gases in past climatic changes. The first measurements of carbon dioxide carried out on the GRIP core confirm the established increase from pre-industrial to present levels and the ‘natural’ increase which occurred at the end of the last glaciation. ‘By then the temperature and the atmospheric concentration of carbon dioxide rose simultaneously, but we hope that further detailed analysis of the GRIP core will make it possible to find out whether the temperature or the carbon dioxide levels started to rise first,’ says Stauffer.
They are also awaiting the results of the American GISP-2 core, which has been drilled 30 kilometres from the GRIP site at the top of the Greenland ice sheet. The American glaciologists reached the bedrock this summer. ‘It is very important whether the American ice core confirms these results or not,’ say Dansgaard and Hammer.
Meanwhile, the Danish glaciologists are planning to drill an ice core in the Hans Tavsen icecap in Peary Land in the northernmost part of Greenland. Existing climate models combining atmospheric and ocean circulation predict that this area should be particularly sensitive to climatic changes. Also, the evidence from pollen and insects in land-based sediments suggests more drastic climate changes here in the past than in central Greenland. They hope that this core will yield important information about changes in the nearby Arctic Ocean, for example whether the sea ice disappeared at any time in the past. Signs of an opening of the Arctic Ocean could show up in the core as an excess of sea salt from evaporated seawater.
The next big deep drilling project, however, is likely to be car-ried out in Antarctica. Hammer says the Americans, Europeans and Japanese have all expressed interest in a project to recover an ‘ideal’ ice core from the southern hemisphere. Meanwhile, it seems clear that if we continue to increase emissions of carbon dioxide, we could be playing with fire – or with ice.
Rolf Haugaard Nielsen is a freelance science journalist based in Copenhagen, Denmark.
* * *
Getting a GRIP on past climates
The GRIP core was dated by counting annual layers in the core, as well as by calculations. Going back 14 500 years the annual layers were counted in much the same way as people count growth rings in a tree. This is possible because the acid content of the snow that falls on the icecap varies with the seasons. For instance, there is most acid in summer snow, so there is a peak in acidity each summer. Another dating method is based on the dust content of the ice, which peaks during winter and spring.
The rest of the core has been dated by calculations based on ice flow models describing the stretching and thinning of the annual layers as they move downward through the ice sheet, as well as knowledge of how rainfall decreases with cold climate and increases in warm periods.
The calculations are calibrated by two well-known fixed points in the climate record. The first is the cold period known as the Younger Dryas, which took place just before the end of the last ice age, 11 500 years before present. The other fixed point is the extremely cold glacial period which followed hard on the heels of the Eemian interglacial 113 000 years ago.
Back 160 000 years, corresponding to a core depth of 2900 metres, the annual layers of the GRIP core are undisturbed and lie almost perpendicular to the core axis. Cloudy bands are visible, probably indicative of former ice surfaces. These cloudy bands are made of tiny air bubbles surrounding grains of dust. The deeper layers down to 2954 metres of the core, corresponding to 210 000 years ago, shows a disturbed stratigraphy due to local tensions in the ice probably caused by small elevations in the otherwise flat and gently sloping bedrock.
Finally, the regular layer sequence is re-established at a depth of 2954 metres and continues down to a depth of 2981 metres corresponding to 250 000 years ago. The deepest 48 metres of the ice core has not been dated yet. ‘Special caution must of course be applied to the timescale prior to 160 000 years before present, but the American GISP-2 core will soon give a check on this,’ says Henrik Clausen from the University of Copenhagen and Sigfus Johnsen from the University of Iceland in Reykjavik.
The measurements of variations in the air temperature of the past 250 000 years are based on the oxygen-18 method developed by Willi Dansgaard of the Niels Bohr Institute at the University of Copenhagen. It works like this. The snow which falls on the icecap contains two isotopes of oxygen, the normal oxygen-16 and the rare heavy isotope oxygen-18 which has two extra neutrons in its nucleus. When the climate gets colder, the concentration of the heavy isotope in the snow decreases, and when the climate gets warmer, it increases. The relative concentrations of the two isotopes are measured by mass spectro-metry and in this way climate variations can be traced back in time.
The dating and temperature profile of the GRIP core show that, apart from the past 10 000 years, instability with shifting warm and cold spells has dominated the North Atlantic climate over the past 250 000 years. This applies to the last ice age, the Weichsel glaciation (113 000 to 11 500 years ago), to the Eemian interglacial (133 000 to 114 000 years ago), to the previous ice age, the Saale glaciation, and to the last part of the previous interglacial, the Holstein (see also figure on previous page).
Another finding was that both glaciations and interglacials ended and began very abruptly. The end of the two ice ages were preceded by oscillations between cold and mild spells; if for example, during the last glaciation, when the climate shifted from cold to mild and back again on 24 separate occasions.
The last cold spell was the Younger Dryas which lasted some 1000 years. It ended by very rapid warming – the temperature in Greenland rose about 7 degrees over only about 50 years – and our present interglacial began.
