by Jean Jouzel
It has now been nearly fifty years since the first glacial deep core drilling began at Camp Century. Fifty years during which the community of glaciologists has not been spared failure and disappointment but which, above all, have been full of success, to which collaborating drillers, logisticians, and scientists have all contributed. This has been a period marked, perhaps more than for other scientific research fields, by the increasingly international nature of the programs. Thus, on the French side, in addition to the collaborations with the Russians, then the United States in Vostok, with a European consortium for GRIP and EPICA, within the framework of a multinational program for North GRIP and NEEM, there have been other partnerships such as those we have established with Australia in the drilling at the Law Dome and with Japan in that of Dôme Fuji. We should also mention the first drilling achieved on James Ross Island, off the Antarctic Peninsula, in collaboration with Argentina, as well as the projects with the Chinese. This diversity has enabled French teams to be involved in almost all the great advances that have been made in the study of ice core samples, and in particular in the two scientific discoveries that we discuss later in the book: the existence of a relationship between greenhouse gases and the climate in the past (chapter 11) and that of rapid climate changes (chapter 9).
CHAPTER 7
Vostok
THE CORNUCOPIA
The term cornucopia, attributed by the journal Nature to the Antarctic Vostok ice core drilling, deserves explanation. In the mid-1980s, our knowledge of the great glacial/interglacial cycles that marked the Quaternary essentially rested on the study of marine sediments. Thanks to them the astronomical theory, which stipulates the existence of a connection between the variations in insolation linked to the slow evolution of the Earth’s orbit and those great climatic cycles, had been very widely accepted. In the wake of the article published in Science in 1976 by Jim Hays, John Imbrie, and Nick Shackleton, who established the connection, paleoceanographers put an ambitious program into motion: the Climate/Long Range Investigation Mappings and Predictions Project (CLIMAP) aimed to map the climatic conditions that ruled over all the oceans in the Last Glacial Maximum.1 The data, published during the 1970s and the beginning of the 1980s, fully reveal how marine sediments have aided in our knowledge of past climates.
Ice core drilling and science in Greenland and Antarctica were still in their infancy at this time. Thanks to high accumulation, ice core drilling at Byrd offered a very great thickness of ice for a detailed study of the last deglaciation, which began around 20,000 years ago, and of the warm period in which we have been living for more than 11,000 years. On the other hand, the ice of the last glacial period proved difficult to date beyond 60,000 years. The situation was rather similar for the two core samplings from Greenland that were available then, those of Camp Century and Dye3, for a simple reason: the oldest ice corresponded to the deep part of the core, the final 150 to 200 meters near the bedrock. The flow of the ice becomes increasingly complex the closer one gets to it, to the point that it is impossible to develop a flow model sufficiently reliable to establish a chronology.
The first ice core drilling at Dôme C escaped that problem: the maximum depth reached, 903 meters, was located more than two kilometers above the base. Calculating the flow was simple, especially since the site is located on a dome. The flipside was that this core sample from Dôme C did not go back very far in time. Even if the age of 30,000 years that we had initially attributed to it later proved to be too young (it has henceforth been established that the core retrieved in 1978 covered the last 45,000 years), we were still far from the beginning of the last glacial period, around 110,000 years ago. Thus neither the drilling at Camp Century or Byrd, done during the 1960s, nor those of Dôme C in 1978 or Dye3 in 1981 could rival the marine sediments available at the time. The oldest ice was too young for the results it provided to revolutionize our knowledge then of the succession of glacial and interglacial periods that extended over hundreds of thousands of years.
However, those ice cores did enable us to document the climate of Greenland and Antarctica in an extremely detailed way. The isotopic records obtained in Greenland put the researchers of Copenhagen and Bern on the path of rapid variations. Other parameters such as those corresponding to the fall of desert dust and of sea salt were also recorded in various ice cores; they bore witness to a very active atmospheric circulation during the Last Glacial Maximum. What is more, the core drilling at Dôme C was the occasion of many firsts. Among them we have already cited the possibility demonstrated by the team from Orsay of analyzing beryllium 10, an isotope of cosmogenic origin produced through the action of cosmic radiation on upper layers of the atmosphere, in the polar ice. There was also the isotopic analysis of the oxygen of air bubbles extracted from ice using a method perfected by our American colleague Michael Bender during a stay at the Centre des faibles radioactivités in Gif-sur-Yvette2 and the development of a joint profile of concentrations of deuterium and oxygen 18 in the ice that was carried out at Saclay.3 The most important result was that obtained by Robert Delmas and his colleagues at the LGGE: in the last glacial period, concentrations of carbon dioxide were 30% weaker than those that existed just before the industrial revolution.4 With some distance, all of these approaches proved important and were at the origin of much later work, but our community of glaciologists was a bit behind because none of the ice core drilling went beyond the last glacial period.
A Complete Glacial-Interglacial Cycle
The situation changed dramatically between 1985 and 1990 for two reasons. Thanks to the ice core drilling at Vostok, French researchers had access to a core sample that had gone beyond a depth of two kilometers. It was enough to obtain the ice formed during the preceding glacial period, 150,000 years ago. Despite everything, that was far enough from the bedrock for a viable chronology to be established. And then those new methods put into place at the ice core drilling at Dôme C could be applied on that core sample as well as new approaches, especially since the LGGE in particular, thanks to Jérôme Chappellaz, quickly mastered the analysis of the composition of methane that was initially developed in Bern.5 The unique possibility that the Antarctic ice provided in the reconstruction of the history of the two principal greenhouse gases whose composition was affected by human activity meant the work carried out on the Vostok ice would have many repercussions well beyond the relatively small circle of scientists who were interested in the history of our climate.
The reconstruction of climatic variations based on the isotopic analysis of the ice drew attention in 1985. At the coldest point of the last glacial period the temperature was close to −65°C (Figure 7.1); it was thus close to 10°C colder than it is currently (in Vostok the mean annual temperature is −55°C). It is interesting to note that the cold periods have a periodicity of close to 40,000 years, a periodicity characteristic of obliquity, a parameter of the Earth’s orbit that determines the annual variations of insolation. Further, these cold periods occur at the moment when local insolation is the weakest. Thus part of the mechanisms put into play are different than those concerned in Milankovitch’s theory, based on the variations of summer insolation in the Northern Hemisphere governed by the change in precession. But from this record of temperature at Vostok we had a clear indication of a connection between insolation and climate. The last interglacial period, around 130,000 years ago, was warmer at Vostok than it is today (by 3–5°C), and it probably lasted longer than in other regions of the planet. Furthermore, the variation in temperature at Vostok presented similarities with that of the sea level, which confers a global character, at the very least from a qualitative point of view: when the sea level is low due to a large amount of ice on the ice sheets of the Northern Hemisphere, it is cold in Antarctica, and vice versa.
Figure 7.1. Vostok ice core. The middle curve indicates the variation in temperature at Vostok throughout the last 420,000 years. The upper and lower curves represent concentrations of CO2 and CH4 over the same period,
indicating the recent spike of these two greenhouse gases.
Climate and Greenhouse Effect Go Hand in Hand
Was there a high concentration of CO2 during that warm period 130,000 years ago? Was the cold period that preceded it, the next-to-last glacial period, characterized by concentrations comparable to those of 20,000 years ago? The response so anticipated by the scientific community was positive and unambiguous (Figure 7.1). But because data bear on an entire climatic cycle, the ice of Vostok provided a much stronger message: the records obtained along the more than two-kilometer ice core indicated that throughout the last 150,000 years the concentrations of carbon dioxide were, in general, faithfully correlated with the variations in temperature deduced from an isotopic analysis of that ice; the colder it was, the weaker the concentrations, and vice versa. On October 1, 1987, that which would be unanimously greeted as a discovery made the cover of Nature, which uncharacteristically devoted three articles to the Vostok ice core in that issue. The first presented and discussed the detailed profile of the variations in temperature at Vostok; the second focused on the record of carbon dioxide concentrations; and the third analyzed the connection between those two parameters, which vary in a parallel way, and orbital forcing. These three articles, already cited in previous chapters’ notes, were accompanied by two commentaries, one by Eric T. Sundquist, a specialist in carbon cycles, the other by the journal’s editor, Philip Campbell, to whom the ice core drilling of Vostok owes the name “cornucopia.” The media, recognizing the importance of the relationship between carbon dioxide and past climate change, seized the results that, three years later, were reinforced with data relative to the concentration of methane; it, too, appeared strongly correlated to the temperature in Antarctica, with values twice as great in the warm period as in a glacial period. These curves of Vostok (such as those depicted in figure 7.1 showing the relationship between greenhouse gases and climate) traveled around the world through political as well as scientific circles.
We have mentioned the confirmation provided by the temperature curve at Vostok for the existence of a link between insolation and climate, but the analysis of the air extracted from the ice also very strongly supports the idea that the variations in greenhouse effect have played an important climatic role in the past. An awareness of variations in concentrations of carbon dioxide and methane appears to fill two of the main gaps in the astronomical theory of paleoclimates, which faces a dual difficulty. How can we explain the 100,000-year cycle that dominates climatic records but is present in the astronomic data only through the eccentricity whose influence is very weak? To what is owed the relative synchronism of the climatic events between the two hemispheres when variations in insolation are not synchronous? It is in fact toward the idea of a climate whose astronomical parameters would be the metronome—and where the greenhouse gases, and especially CO2, would play the role of amplifiers of changes in insolation—that we have been led by Antarctic records. However, the explanation of these natural variations in concentrations of carbon dioxide and methane is only partial at this point. It has still not been settled convincingly, but we know that, in each case, the variations of these two greenhouse gases put into play interactions between physical parameters of the climate and geochemical cycles linked to the living world. For carbon dioxide, these variations certainly imply the circulation of the oceans and their productivity and, to a certain degree, interactions with the continental biosphere. On the contrary, variations in methane essentially come from the impact of the climate on its earthly sources, including swamp zones and perhaps the frozen ground of high northern latitudes.
Whatever the causes might be, the natural variations in concentrations of these two compounds bring about an increase in the greenhouse effect by a bit more than 2 Wm−2, thus the equivalent of that linked to human activity during the last two hundred years. If the climatic system was exempt from feedbacks, that direct radiative effect would produce an increase in temperature of only 0.6°C. But sea ice, water vapor, and clouds in a glacial period—and as is the case today and as it will be in the future—modified the response of this system. Indeed, the sensitivity of the climate to these so-called rapid feedbacks appears to depend weakly on the climatic period under consideration, and we proposed to estimate this parameter from Vostok and other data of the past. This approach does not require that the complexity of the mechanisms driving large climatic changes be completely deciphered or that a definitive response to the “chicken or egg” question—which is the cause, which is the effect—be answered, which is so often raised on the subject of respective variations of the climate and of greenhouse gases during the past. When we are looking only at the sensitivity of the climate, it is enough that the various forcings that operate on the scale of large observed climate changes can be correctly estimated; then we evaluate the role of the various factors that can account for the succession of the glacial and interglacial ages.
Although a simplistic approach, we used a means to statistically compare the temperature profile of Vostok with different forcings able to explain it: the astronomical contribution represented by the variation in volume of ice accumulated in the Northern Hemisphere and by that of local insolation, radiative forcing linked to the variations in CO2, modifications of the radiative budget resulting from changing the content of atmospheric aerosols (dust and sulfates). This statistical analysis confirms the visual impression: a contribution linked to the astronomical forcing (via the variations in volume in the ice in the Northern Hemisphere) and greenhouse effect each explain 40 to 50% of the variability in Vostok temperature; the influence of other possible causes remains marginal. Thus an increase in the content of greenhouse gases would explain up to 2°C of the 4–5°C of average warming of the Earth during deglaciation, indicating that the direct radiative forcing was increased, through feedbacks linked to sea ice, water vapor, and clouds, by a factor close to 3 (~2/.6). In the case of a doubling of the content of carbon dioxide compared to its approximate current value, the forcing would be double that observed since the Last Glacial Maximum (4 Wm−2), and we have deduced from the data of the past that the warming would then be 3–4°C. This approach has limits, both on the level of statistical method put into place and in terms of estimating accurately the variation in mean temperature of the planet in the past. However, it illustrates the key result provided to various degrees by the climatic models whose most recent results concur with the deductions we published in 19906 from data issued from polar ice: these are mechanisms of amplification vis-à-vis the radiative forcing connected to the anthropogenic greenhouse effect, which should operate during the coming decades. Furthermore, the results obtained at Vostok have undeniably played a role in the awareness of the link (discussed in the last part of this work) between the evolution of our climate and the increase in the greenhouse effect related to human activity.
One, two, three, and then four climatic cycles were finally reached in 1999. The records we have obtained, along with our Russian and American colleagues, thus cover 420,000 years. The theory put forth a dozen years earlier was fully confirmed: the climate of Antarctica and greenhouse gases go hand in hand throughout the period characterized by the succession of four very marked glacial/interglacial periods.7 This extension of the records highlighted the role of human activity during the recent past. As observed along the Vostok core, never in 420,000 years have the quantities of carbon dioxide and methane present in the atmosphere been as high as they are today (Figure 7.1); never have they risen so rapidly.
Thanks to the comparison of records obtained in the Vostok ice and those provided by the analysis of the ocean sediments, we began to be able to better decipher the complexity of the mechanisms of large climate changes. Thus we henceforth had a better view of the sequence of events as it was repeated during the transitions between a glacial and interglacial period. Such a transition corresponds to a global warming on the order of 5°C on average, accompanied by the melting of most of the ice accumulated on the c
ontinents of the Northern Hemisphere, a melting that provoked a rise in sea level by about a hundred meters. Somewhat surprisingly, things may have begun to move in the southern regions: in Vostok the first signs of warming were observed—probably linked to subtle changes in insolation—but, although the relative timing between temperature and CO2 at Vostok is difficult to estimate, the increase in the concentration of carbon dioxide possibly began a few centuries or a thousand years later. Whereas the temperature in Antarctica and the amount of carbon dioxide (and methane) increased, it was only after a few thousand years that the ice sheets of the Northern Hemisphere began to melt so extensively. This process ended several thousand years later. Was it the local insolation, that which prevailed in Antarctica, that gave the signal to begin? Or did that result indirectly from changes in insolation in the Northern Hemisphere, as proposed by the classic astronomical theory? This question has not been definitively answered.
Everything leads us to believe that these changes in insolation modified the conditions prevailing in the Southern Ocean, a large reservoir of carbon, which was then translated into the observed increase in the concentration of carbon dioxide. Once in play, it participated in the warming by increasing, through the radiative forcing associated with it, that which was initiated by the change in insolation (Figure 12.5). The sequence of events was such that the increase of the greenhouse effect then participated fully in the melting of ice sheets in the Northern Hemisphere, itself amplified by the feedback linked to the change of albedo (the gradual disappearance of ice sheets diminishes the reflective surfaces and accelerates their melting). To the question “Which came first, the chicken or the egg?” the answer is not so simple. In our opinion, it was indeed radiative forcing that was the initial cause of deglaciation, but once the greenhouse effect began to increase, it in turn became a true agent in deglaciation—to the point of being half responsible for the mean increase in the temperature that accompanied these major climatic transitions.