by Jean Jouzel
The eccentricity is quite simply the degree of flattening of the Earth’s orbit. It varies from zero, when the orbit is circular, to around 6%, with periodicities close to 100,000 to 400,000 years. It is currently 1.7%, and the Earth is closer to the Sun in December than in July. The Earth’s axis is more or less inclined in relation to the plane of the orbit, and that incline is called obliquity; its current value of 23° 27´ varies at a range of 2° 30´ following a cycle of 41,000 years. Precession is a more subtle parameter. It describes the fact that the place where our planet is located at a given moment of the year moves over time: 11,000 years ago the Earth on June 22 passed by the point nearest the Sun, the perihelia; now this passage takes place on December 22. It will be on June 22 again in 11,000 years, which corresponds to a periodicity a bit greater than 20,000 years.
Figure 3.2.Variations in the Earth’s orbit and astronomic parameters, eccentricity, obliquity, and precession. The lowest curve represents the insolation of the summer solstice at a latitude of 65°N.
At any given moment, the insolation varies only as a function of latitude. On average, over long periods of time it depends only on eccentricity and varies very little—by less than 0.2% during the last million years. Its yearly variations are modulated by obliquity with opposite signs between the low and high latitudes, but these yearly variations remain relatively weak, not going beyond 6 Wm-2 (watts per square meter) at any latitude. In contrast, where there are large variations, they are on a seasonal scale: linked to precession, they can reach 20%. Let’s take the example of Paris, which at the beginning of summer, 125,000 years ago, benefited from 458 Wm-2 but, 11,000 years later, had to be content with only 385 Wm-2; today we are between the two with a value of 411 Wm-2. Let us add that this seasonal variability is itself influenced by the shape of the orbit: the variability increases the more elliptical the orbit.
Adhémar and Croll were correct in suggesting that the way in which the Earth turns around the Sun helps explain the onset of the glacial periods. But neither of them succeeded in establishing a theory that stood the test of time. Adhémar postulated that the glacial periods were simply ruled by precession and its close to 20,000-year periodicity and that they occurred in the hemisphere in which winter was the longest, returning alternately in the north and the south. Croll believed only the distance between the Earth and the Sun, and that value on December 21, the winter solstice, played a role; his idea was that when that distance exceeded a certain value, the winters in the Northern Hemisphere were sufficiently cold for a glacial period to begin. Both Adhémar and Croll correlated glacial periods with very cold winters in high latitudes. This was reasonable, it seems, but it does not stand up to analysis: temperatures are currently cold enough for Arctic regions to be snowy in the winter without our being in a glacial period.
A decisive step toward determining the origin of glacial periods was taken a few decades later by the Serbian mathematician Milutin Milankovitch, who spent more than thirty years developing the theory that bears his name. One of his objectives was to calculate variations in the insolation over long periods of time for different latitudes, which is not a small task. Even though he had the values of the three astronomical parameters over the last million years, whereas in Croll’s time they had only been calculated over 100,000 years, he had to devote almost ten years to his calculations, since at the time he had just paper and pencils. In 1920, he achieved his goal and the results were published; they attracted the attention of the German climatologist Wladimir Köppen, famous for the classification of climatic zones from the types of vegetation found in them. With his son-in-law, Alfred Wegener, the father of the theory of continental drift, he was interested in the history of climates and invited Milankovitch to Germany. The collaboration that resulted proved fruitful since Köppen provided Milankovitch with the key to his theory: the triggering factor was a succession of cold summers in high latitudes of the Northern Hemisphere, cold enough so that the snow of the preceding winter did not melt. The snow, with its strong reflective power, amplified the cooling to the point that a glaciation could commence. Milankovitch began to calculate the insolation in the summer at 55°, 60°, and 65° N, and Köppen showed that the data had enough similarities with those of Penck and Bruckner (who described the succession of glacial periods in the Alps) such that Milankovitch continued to put the finishing touches on his theory and in 1941 published his defining book: Canon of Insolation and the Ice Age Problem.5
However seductive it may have been, the validity of this theory was not proven. There were no reliable chronologies, and there was no continuous record of past climatic conditions. These obstacles would be overcome thanks to the analysis of marine sediments that provided such a record over hundreds of thousands of years and could be correctly dated. In 1976, Jim Hayes, John Imbrie, and Nick Shackleton6 provided this proof of a connection between the insolation and the rhythm of glaciations by demonstrating that the periodicities corresponding to precession and obliquity were present in the climatic series deduced from the analysis of marine sediments. Later the calculations of the amount of insolation were refined in relation to those of Milankovitch by André Berger in Belgium and Jacques Laskar in France. Berger showed that precession is in fact characterized by two periodicities, around 19,000 and 23,000 years, and that double periodicity was also present in the data.
Milankovitch earned well-deserved recognition, but his work was far from establishing the theory of paleoclimates. Many questions remained. And then, a few years later, the deep ice core drillings in Antarctica provided an additional dimension by revealing a very close connection between the composition of the atmosphere and the earthly climate over that time period. The drilling carried out in Greenland revealed extremely rapid variations in the climate that could occur on the scale of a few decades, or less, too rapid to be explained by the much slower variations in the Earth’s orbit.
We will now turn to this fascinating scientific adventure: deciphering glacial archives.
PART TWO
POLAR ICE: AMAZING ARCHIVES
CHAPTER 4
Reconstructing the Climates of the Past
Human nature, it seems, has caused people to explore the story of our past ever since the dawn of time, whether it has been the history of nations, that of civilizations, or that of our planet and solar system. The history of past climates is no exception. This need of memory has become even more crucial over the last few decades, as an understanding of the climate and its evolution has become of major concern for the near future of our civilization.
To aid us in this task, we have meteorological measurements, which we’ve had for around a century. More recently we’ve been able to use observations from space. On the other hand, glacial archives, marine and lake sediments, fossil fauna and flora, soil, trees, limestone formations, and coral, have preserved a record of the past. When dated, these sources enable us to explore the past and to reconstruct continuously certain phenomena that, depending on the situation, describe the seasons, centuries, and millennia of the climate, of our atmosphere, of the oceans, and of the continents. The information thus archived requires a great deal of deciphering, in a certain sense analogous to the task of Jean-François Champollion in the nineteenth century when he unraveled Egyptian hieroglyphs. The climates of the past are the subject of a distinct scientific discipline: paleoclimatology.
The guiding principle of paleoclimatology rests on a simple idea: many characteristics of what exists on Earth, whether in the living world, in the mineral world, or in the water cycle, depend on the climatic conditions that existed when the different elements were formed. If these diverse materials accumulated layer by layer in such a way that the sequence of time of the deposits has been preserved, at least for the most part, we have then a climatic archive that can be used, provided that it is physically accessible, that it can be dated, and that some of its properties can grant us access to information that enriches our understanding of the climates of the past. The method
s that are used to analyze these archives, adapted to the environment and the period under study, are extremely varied.
The simplest method, at least in principle, is studying the distribution of species, whether animal or vegetal, or for a given species, its annual growth pattern. We have here an obvious potential climatic indicator. Another method we can use, perhaps less concrete for laypeople, is the chemical composition of a matter. Whether it comes from the mineral world or that of the living, it is likely to be influenced by climatic parameters and is a very interesting source of information. We imagine that, for many readers, the term isotope is shrouded in mystery. And yet isotopic analysis is, by far, one of the most efficacious tools of the paleoclimatologist, whether he or she is studying continental, oceanic, or glacial archives.
The Round of Isotopes
We are all familiar with the terms atom, hydrogen, oxygen, carbon, and nitrogen, as well as the elements that make up Mendeleev’s table. Each one is characterized by an atomic number that represents its number of protons and electrons: 1 for hydrogen, 6 for carbon, and so forth. A certain number of neutrons can be added to the number of protons (the atomic mass corresponds to the sum of protons and neutrons). A given atom can have various values: these are called isotopes. Let’s take the case of hydrogen. It corresponds by more than 99.985% to an atom formed by a proton and an electron, but in around 0.015% of the cases a neutron is added to form an isotope of hydrogen that is commonly called deuterium, the one that creates so-called heavy water. Add another neutron and we have tritium, a radioactive isotope present in infinitesimal quantities in nature, whereas deuterium is a stable isotope. Carbon also has three natural isotopes: the main one is carbon 12 (with 6 neutrons), and the two others are carbon 13 and carbon 14; the latter is radioactive. Oxygen’s three natural isotopes are oxygen 16 (8 neutrons), which is the principal atom, oxygen 18, and oxygen 17; the latter two are stable.
The chemical properties of the isotopes of the same element are identical, but their physical properties are slightly different. These minor differences put their stamp on all the mechanisms that have taken or take place in our universe whether during the formation of our planet or, nearer to us, when a cloud creates rain, hail, or snow. The radioactive isotopes are able to tell us the age of an event, to date it, and in some cases to describe its duration. In fact, a true discipline has been built around isotopes—isotopic geochemistry—whose applications are particularly prolific in the realm of paleoclimatology, thanks primarily to the isotopes of hydrogen, carbon, and oxygen that we have just mentioned.
Let’s look at water. Most water is made up of H216O molecules, formed by two hydrogen atoms (of mass 1) and one oxygen atom of mass 16. The existence of deuterium and oxygen 18, the most abundant isotopes of hydrogen and oxygen, leads to “isotopic” forms of the water molecule, HDO and H218O, which accompany H216O in all reservoirs where water is present in average concentrations close to 0.03% for HDO and of 0.2% for H218O. When we analyze a sample of rain or snow using a mass spectrometer, it is never that average value that is measured. Each rainfall or snowfall has its own unique composition of deuterium and oxygen 18, which result from its history. We owe this characteristic to the fact that the physical properties of these three molecules are slightly different: the heavy molecules HDO and H218O have, predictably, more difficulty evaporating, but they also have a tendency to condense more easily during the formation of rain or snow.
The result of this isotopic “fractionation” is that each time little drops or snow crystals form in a cloud, there is a lessening of the amount of water vapor that created them, as well as of deuterium and oxygen 18. Gradually, as an air mass approaches polar regions and enters them, it cools, creates more precipitation, and is again weakened in terms of the heaviest isotopic molecules. So it is in the coldest region of the globe, on the Antarctic plateau, that one finds samples of snow that contain the least deuterium and oxygen 18. Thus when one goes from the coast to the center of Antarctica, the surface temperature regularly decreases. For each cooling of 1°C, the concentration of deuterium in the snow diminishes by 0.6%.1 This characteristic is well explained by an isotopic model in which this simplified description of the life of a mass of air is translated and by isotopic models that account for the complexity of the atmosphere.2 This is the foundation of what we call an “isotopic thermometer”: the colder the temperature, the weaker the isotopic content; the same is true for the inverse (Figure 4.1).
Another such example is found in carbon 13. This time, the fractionation is linked to biological processes: the amount of carbon 13 compared to carbon 12 in the biomass is smaller than that found in the carbon dioxide of the air, CO2, or in the ocean because during reactions of photosynthesis, plants less quickly assimilate molecules of “heavy” carbon dioxide such that carbon 12 is replaced by carbon 13. This property has many applications, including that of reconstituting ocean circulation on a global scale: as the water masses sink into the global ocean, they receive organic particles that are decomposed while forming carbon dioxide. The older a body of water, the longer it has circulated in the deep ocean, the richer it becomes in carbon dioxide and thus more depleted in carbon 13. This loss is reflected in the carbon 13 composition of the benthic foraminifers, small organisms that grow on the ocean floors using that carbon dioxide to form their skeleton.
Figure 4.1. The isotopic thermometer is based on the relationship that exists, for the current period, between the average amount of deuterium and the temperature of the site of sampling. SMOW: Standard Mean Ocean Water. Source: Adapted from Claude Lorius and Liliane Merlivat, “Isotopes and Impurities in Snow and Ice,” in Proceedings of the Grenoble Symposium Aug.–Sept. 1975 (Vienna: IAHS, 1977), 125–37.
Finally, the oxygen cycle is influenced by both biological and physical fractionations. During the process of breathing, there is preferential consumption of lighter oxygen 16; thus the oxygen remaining in the air or the water becomes enriched with oxygen 18. However, photosynthesis is not accompanied by fractionation; the isotopic composition of the oxygen produced is identical to that of oxygen present in the water used. But as we have seen, the isotopic content of that water itself depends on fractionation processes connected to physics. The signal that results, accessible from the analysis of the oxygen in the air bubbles trapped in the polar ice, is complex but very interesting from a paleoclimatic point of view.
Later we will discuss other examples of the way in which the “stable isotopes” bear witness to past climates, but these examples are the most illustrative.
Going Back in Time
As glaciologists, we pay particular attention to the glacial archives that are at the heart of this work and tell us a lot more than simply the history of the climate in polar regions. The chapters in part 2 are devoted mainly to those archives. We will first look at all the approaches that enable paleoclimatologists to describe the conditions that prevailed throughout time on the surface and in the deeper layers of the oceans, that motor of the climatic machine, and on the continents where our civilizations developed. We invite the reader to explore them by going back in time.
The Recent Period
Although some temperature records cover several centuries—the longest series is that from the center of England, which began in 1660 3—it was only in the second half of the nineteenth century that more systematic measuring was put into place. Gradually there were enough of these measuring networks, which were also interested in atmospheric pressure and precipitations, that mean values could be established for a given region and over the entire planet. To that end, it was important that the data be homogenized because the instruments and measuring methods were modified over time; also many stations moved or witnessed their immediate environment change with the rhythm of urbanization. This is how Jean-Marc Moisselin4 and his colleagues at Météo France produced a very reliable map of the evolution of the temperature of France during the twentieth century. It indicates an average warming close to one degree with s
lightly higher values in the southwestern quarter.
What were the conditions in the preceding centuries? Historical chronicles, popularized in France by Emmanuel Le Roy Ladurie,5 very well documented in many European countries, as well as in China and Egypt, with series of observations that cover millennia, carry on the work. They are of a miscellaneous nature: the freezing of the canals in the Netherlands, the number of days of frost or snow, flooding and droughts, the strength of the wind and storms on the continents and oceans, the date of the migration of birds, of the blooming of trees, of grape harvests, and of the volume of harvests; these observations provide an idea of what the weather was like in those times. Even if these historical data are often of a local, largely discontinuous nature, and although they have a tendency to favor the observation of climatic extremes, they are precious sources of information. Our knowledge of the climate during the last millennia, for instance, in France, largely depends on them. The same is true for the identification of a warm period in the eleventh and twelfth centuries and of the Little Ice Age, which prevailed in Western Europe, though it was nevertheless interrupted by milder periods, from the sixteenth to the nineteenth centuries.
Data such as the dates of grape harvests, of which many chronicles give a continuous, chronological account, lend themselves to a precise estimate of summer temperatures. We then use a so-called phenomenological model that describes the evolution of the vine as a function of climatic conditions. This method has been recently applied in Burgundy, France, where dates of grape harvests have been recorded since 1370.6 Even though years as hot as those we experienced in the 1990s had already occurred since the fourteenth century, this study confirms the exceptional nature of the summer of 2003.