The White Planet: The Evolution and Future of Our Frozen World

by Jean Jouzel

  The verdict, pronounced in 2004 upon seeing the profile of deuterium in the ice analyzed at Saclay, was unambiguous: the rhythm of the climate of Antarctica also changed around 400,000 years ago (Figure 8.1) in a more pronounced way than did the sea level, or continental ice volume, recorded in marine sediments. Earlier the warm periods were less warm, but they lasted longer and, except for the one that happened 650,000 years ago, the glacial periods were less cold. The following year, researchers from Bern and Grenoble provided another important confirmation: the content of carbon dioxide and, to a lesser degree, of methane also varied differently more than 400,000 years ago. The maximal values were also less elevated than those reached since, and the variations in the carbon dioxide varied just as consistently with the temperature as in the 420,000 years at Vostok. The correlation between these two parameters—and this was a very anticipated test, including by skeptics of the greenhouse effect—remained just as strict before that period as after it. The results henceforth covered 800,000 years, and the quality of this relationship between variations in greenhouse effect, dominated by that of carbon dioxide, and climate is truly impressive. One might imagine that the variations in greenhouse effect were at the origin of the change of the climatic rhythm observed at around 400,000 years ago. We can also argue that, just as for a deglaciation, the greenhouse effect participated in that change of climatic rhythm through a modification of the radiative forcing, but the primary cause was likely linked to the modifications in insolation. The insolation received each year at Dôme C presented maxima that, like those of the temperature reconstructed at that site, have been greater since 400,000 years ago than they were before 400,000 years ago (Figure 8.1); proceeding from there to seeing the origin of the change in climatic rhythm observed 400,000 years ago requires only a step.

  CHAPTER 9

  Rapid Climatic Variations

  If we were to rank in importance the climatic phenomena we’ve been discussing, the existence of rapid climatic variations would be at the top of the list. Who hasn’t heard of the halting of the Gulf Stream or, wrongly, of a return to a glacial-like climate that would affect the regions that border the North Atlantic? The media have seized this image and it has even been portrayed in theaters; it is the theme of The Day after Tomorrow, a sci-fi thriller from 2005. Let’s be clear: this is science-fiction; the icing over of a large part of the United States in a few days is, fortunately, not at all realistic. But the first scenes in which a glaciologist is at work are seductive.

  As we have already mentioned, the analyses carried out over the ice in Greenland are at the heart of this notion of rapid climate variation. This idea already had some followers before the first deep ice core drilling took place at Camp Century in northwestern Greenland. But the opinion that was very widely accepted in the first half of the twentieth century was that the climate could only change gradually. That our planet has lived a succession of glacial and interglacial periods was, for many scientists, a concept that was difficult to accept, but to suggest rapid changes was absolutely out of the question, if only because it takes time for a glacial ice sheet to form or to melt.

  The First Indications

  In the 1930s the study of pollen preserved in peat bogs and lakes in Scandinavia challenged what was believed up to that point: in those regions the end of the last glacial period was clearly marked by oscillations: an initial warming, which ended in a period named Allerod, was followed by a cooling around the period of the Younger Dryas (a name derived from the name of a flower from the Arctic tundra, Dryas octapetala). In 1955, a few years after its invention, which earned Willard Frank Libby the Nobel Prize in chemistry, the method of carbon 14 dating was applied to those sediments; it enabled the dating of that oscillation at 12,000 years. Hans Suess, another carbon 14 specialist, confirmed its existence from marine sediments. However, the idea we had at that time of a rapid climate change was quite far from what we understand today since the periods evoked were on the order of a millennium rather than a dozen years. One of the reasons for this is that the carbon 14 dating method was not very precise at that time, indeed on the order of a thousand years or so. And the sediments analyzed, whether continental or oceanic, did not enable us to follow the variations in climate on the scale of decades.

  In 1960 Wally Broecker and his colleagues at Lamont near New York were the first to point out the importance of these oscillations, which went from 5° to 10°C. And in the 1970s people began to talk of centennial changes with multiple indications from the point of view of both the archives that provided them—pollen, marine foraminifera, the remains of insects or snail shells—and their geographic distribution. The Younger Dryas was identified in North America as well as in central Europe and in many ocean core samples.

  The core drilling at Camp Century provided a great deal of data that support the idea of rapid climate change. In 1969 Willi Dansgaard, Chet Langway, and their colleagues published an ice core record that covered around 100,000 years. At the bottom of the drilling site, they identified ice that could have been formed during the preceding warm period 130,000 years ago. The detailed profile of oxygen 18 content became available in 1972; the recent Allerod/Dryas sequence is very clearly marked in it. More unexpected was the discovery of a series of events, which Dansgaard described as violent, that occurred throughout the glacial period. The comparison with the Byrd records that was done at the time showed a different occurrence in Greenland, visibly very perturbed during the glacial period, and in West Antarctica, which had a much smoother isotopic profile and was thus probably exempt from oscillations of that type.

  This increasing number of indications at that time did not draw much attention from the scientific community, which was more interested in the description of conditions that existed in the Last Glacial Maximum and in the establishment of a link between the long-term climate changes and astronomical parameters. During the 1970s there was also speculation about the arrival of the next glacial period, which was thought to be rather imminent—a question of a few centuries, according to some. The fact that the climate could change over such a short period stoked what, in retrospect, appeared to be an unfounded fear that relied on the idea that all interglacial periods are relatively short and on the observation of a cooling that was then taking place in the Arctic. That idea was most likely erroneous, as we have seen, as our interglacial period has every reason to continue over another 20,000 years.

  Increasingly Clear Indications

  The ice core drilling at Dye3, the first results of which were published in 1982, was an important stage in the history of rapid climate variations. Our Danish colleagues analyzed the results of the drilling in great detail: the current Holocene period covers more than 1,500 meters, and measurements of the oxygen 18 concentration were taken over more than 70,000 ice samples, which enabled the dating of the core year by year over close to 8,000 years. The last climatic transition, represented by close to 50 meters of ice, can be described precisely. Hans Oeschger pointed out that it resembled almost identically a record obtained in the sediments of a Swiss lake, reinforcing the idea that the rapid warming associated with the end of the Younger Dryas resulted from a retreat of waters of polar origin in the North Atlantic with climatic effects somewhat parallel at the center of Greenland and Western Europe. Willi Dansgaard, Jim White, and Sigfus Johnsen1 examined that transition under a magnifying glass; isotopic analyses showed that this warming, which they determined to be of 7°C, happened in less than fifty years and that the displacement of the polar front was even more rapid, on the order of a dozen years.

  Just as important from the point of view of rapid climate variations is the similarity in the climatic curves of Camp Century and Dye3 throughout the last glacial period. Almost all the violent events identified in the former are the same in the latter, in spite of the 1,400 kilometers that separate the two sites. These oscillations were thus not linked to purely local instabilities in the glacial ice sheet; they have at least a regional character. Dan
sgaard and his team attributed them to changes in the atmospheric circulation in the North Atlantic between two stable states. This explanation was not accepted by Oeschger and Broecker.2 Both designated the ocean as the first guilty party, all the more so since Oeschger’s team,3 which participated in the Dye3 project under the leadership of Bernhard Stauffer, discovered the existence of rapid variations in the concentration of carbon dioxide in the air trapped in the ice. Those variations appeared to systematically accompany the abrupt variations in temperature recorded in the ice. In fact, as we will see later in this chapter, we know now that they have been produced by an artifact and hence are not representative of the atmospheric composition in CO2.

  A Connection with Ocean Circulation?

  Oeschger, Broecker, and their teams formed the hypothesis that ocean circulation was less active in glacial periods than it is currently. The feeding of the deep ocean occurs from surface waters, but those sink only in well-defined regions of the ocean, where they are sufficiently dense. For that, they must be cold and salty. Today these conditions are encountered only in the North Atlantic, in the Norwegian Sea in particular, and in the south in the Weddell Sea around Antarctica; there is no formation of deep waters in the Pacific Ocean, and the same was probably true in the North Atlantic during a glacial period. Schematically, the ocean would have had two stable states, one corresponding to current conditions, with the production of deep waters in the North Atlantic, the other without. Broecker suggested the possibility of an intermediary mode whose characteristics would be specified a few years later: the surface waters sink more in the south and less deeply.

  The events discovered by the Danish scientists corresponded to the passage of one mode of functioning in the ocean to another, which, at the same time, would have explained the variations in the carbon dioxide content, of which the ocean is a huge reservoir, and those of the climate. In fact, the waters that currently sink in the Norwegian Sea are brought there on the surface from tropical regions, and they warm our lands—this is the famous Gulf Stream. When that circulation stops, the arrival of heat is also interrupted, which accentuates the glacial nature of the climate. Broecker quickly became the most ardent defender of the key role of ocean circulation in the oscillations recorded in the ice of Greenland. He also pointed out the role of the atmosphere because the density of the surface waters depends on the amount of freshwater, including that which evaporates and that which is provided by precipitation, the flow of rivers, and the melting of ice. The article he published in 1985 addresses the functioning of this ocean-atmosphere system. A talented scientist, Broecker also knew how to express his ideas to the general public. It is largely to him that we owe the popular image of the conveyor belt with which we associate the Gulf Stream: it circulates on the surface, then the waters that sink into the North Atlantic travel through the deep Atlantic waters from north to south, turn around at Antarctica, bathe the entire Pacific, and then return to the surface.

  However, something went wrong. Not so much the idea of the Gulf Stream that stops and starts again—even if the idea was criticized at the time by purists, it has stood the test of time. But the rapid variations in carbon dioxide were troubling, inasmuch as they occurred in the same samples as the climatic variations seen in ice. This should not have been the case: changes in climate and in the composition of the atmosphere that occur at the same time are not recorded at the same depth because the air bubbles are trapped at a hundred meters from the surface. Our Bern colleagues immediately raised the following question: Wouldn’t the analyses of carbon dioxide be perturbed by the partial fusion of the ice in the summer or by the greater or lesser amount of carbonates, which when they decomposed in fact produced carbon dioxide? The questions they raised were justified: the ice of Greenland contained enough impurities for the carbon dioxide analyses to be erroneous. Fortunately, those of Antarctica are much purer and are very viable archives of the variations in carbon dioxide in the atmosphere of the past. Forgotten were the variations in the amount of carbon dioxide during the glacial period; there were indeed fluctuations, but they were much smaller and less rapid than those seen at Dye3. However, the isotopic variations recorded in the ice were indeed real, and the core samples of GRIP and GISP2 unquestionably confirmed the rhythm, the rapidity, and the great spikes in temperature that were associated with them.

  Confirmation

  August 1991: ice core drilling operations were in full force in the center of Greenland at its highest point, the Summit region. At GRIP the Europeans had reached 2,321 meters. They knew they had ice from the last glacial period. A simple calculation showed that it was more than 20,000 years old—and much older if it was proved that in a cold period the accumulation was smaller than that which currently prevails. Isotopic analyses were moving along. Priority was given to the bottom part of the core. The stated objective: to confirm the existence of rapid climate variations at the end of the last glacial period and the abrupt transition toward current climate conditions, both of which were demonstrated in the ice core of Dye3.

  The dating of the GRIP core raises some issues, however. A direct consequence of the relationship with the temperature, the isotopic content of the snow is higher in the summer than in the winter. This seasonal cycle offers a means to date the successive layers year by year. The method could be applied to the ice core of Dye3 over nearly 10,000 years. Further back the seasonal indications faded then disappeared under the effect of diffusion. To our disappointment, this disappearance was faster on the GRIP ice core. Beyond 3,000 years the seasonal isotopic variations were of no help in dating the ice. The great volcanic eruptions, which leave an easily detectable chemical imprint in the ice, were not very useful there because their chronology was increasingly uncertain the further one went back in time. Fortunately chemists took over from there. The content of dust and chemical elements such as calcium, nitrate, and ammonium have a well-marked seasonal cycle. It was work that demanded extreme patience, but the results were there, even if precision decreased with depth. The age of the core was determined within 60 years at 10,000 years and within 800 years at 20,000 years. This ice was located at a bit less than two kilometers in depth, and an age of 40,000 years was thus attributed to the deepest ice, at 2,321 meters.

  To reach that goal, the chemists almost continuously perfected methods that enabled them to analyze certain properties of the ice in the field. They worked in a true underground laboratory, which we called the “scientific trench.” It was installed under a few meters of snow so that the temperature remained very cold on summer days when the thermometer could climb above zero in the sun. In the trench it was around −15°C, but it was a hive of activity in which about forty people happily worked together. Once the ice core samples, which in general measured two to three meters in length, were extracted, they were carefully inventoried and then stored for a few days to stabilize them from thermal and mechanical disturbances. They were then ready to be examined—one could note with the naked eye layers of ash and even variations in structure that showed annual variations—and to be cut into 55-centimeter pieces that were easily handled and calibrated for the containers that would take them to Copenhagen. The electrical conductivity of the ice was analyzed; it varied depending on the impurities that were present in it. This measurement enabled the scientists to note all of the volcanic eruptions: they are not all visibly seen in a layer of ash. They could also identify precisely the transition between the last glacial period and the current climate. Chemists brought their chromatographers and developed a very slow ice-melting method that enabled them to analyze it on the scale of a centimeter. Others, who had to set up a cave for themselves where the temperature was close to −30°C, studied the physical properties of that ice; for them, observation in polarized light, which revealed the distribution and the size of magnificent crystals, was an essential source of information.

  However, some scientists, working for instance on ice isotopes or gases enclosed in ice, were not fortunate enough
to have results as soon as the ice was extracted. They had to be patient because the samples had to be brought back to their laboratories. Although it might have been possible to install a mass spectrometer or a line of extraction and gas analysis on the site, it would not have been cost-effective to do so and there would have been the risk of insufficient analytical precision. For the specialists who analyzed the gases and isotopes in the ice, work in the field was limited to cutting and packing the precious samples following specific handling procedures. But we were all impatient to have access to the results. Our Danish colleagues who were isotope specialists did their utmost to make results quickly available; the mass spectrometer in Copenhagen worked at full speed, and a few weeks after the end of the summer campaign we had the detailed profile of the variations in oxygen 18 content for the period from 10,000 to 40,000 years in the past.4 The isotope specialists told us that we had here a record of variations in temperature at the center of Greenland. We needed to keep in mind that a variation of 1/00 in oxygen 18 corresponded to a change in temperature of 1.5°C—at least that was what we asserted at the time. Two intervals were particularly interesting because they were very perturbed: the deglaciation between 15,000 and 10,000 years ago and the period from 40,000 to 25,000 years ago.

  There was a lot of discussion among scientists concerning the age of the events that marked the last deglaciation. Up until 1990, the carbon 14 dating placed the end of the Younger Dryas at 11,000 to 10,000 years ago. The methods that on the GRIP ice core gave access reliable dating, as discussed above, seemed to indicate a greater age. This discrepancy arose from the fact that the rate of carbon 14 production did not remain constant throughout the ages, a difficulty that was avoided thanks to the work of a Franco-American team which, at the same time, was implementing a second method of dating based on the analysis of isotopes of uranium and thorium in their study of a series of coral from Barbados; in 1990 Édouard Bard and his colleagues showed that the carbon 14 ages had to be seriously corrected and that the Younger Dryas ended 11,500 years ago.5 Confirmation of the end of the Younger Dryas was provided by both European and American glaciologists: the age estimated from the counting of annual layers was 11,550 years at GRIP and 11,640 years at GISP2.6 The rapidity of the transition was confirmed, whereas the cooling that led to the Younger Dryas was much more gradual. Beyond these aspects, the similarity between the GRIP and the Dye3 records is quite striking.