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

by Jean Jouzel

  We can note, however, the limits of the historical approach, even if only from a geographical point of view. Other observations—such as the spread of mountain glaciers, which is affected by temperatures and precipitation, or the temperature of the ground, which is accessible in many drilling spots and at great depth retains a certain memory of the climatic past of the region—are sources of climatic information on these timescales. Rather difficult to interpret, these data nonetheless confirm the warming of the twentieth century as compared to those that preceded it, but they do not enable us to follow its evolution in detail. Fortunately, Nature, whose development is marked by the rhythm of the seasons, offers us ancient archives, trees on the continents and coral in the oceans, which are easy to date over the years and whose growth bears witness to climatic conditions in the past.

  Dendroclimatology is a discipline that examines the thickness of annual growth rings of trees as climatic indicators. It has been practiced since the beginning of the eighteenth century, with the observation, in Europe, of very small growth associated with the exceptionally cold winter of 1708–9. It is essential that the age of each ring be known precisely, and much care is given to the establishment of a firm chronology on a given site on living trees. Then, step-by-step, we identify periods showing overlaps with the help, for example, of construction wood from old buildings. Chronologies of this type enable us to go back more than a thousand years, and much longer continuous series, up to around 12,000 years, have been established from fossilized trees well preserved in riverbeds or in peat bogs. The thickness of the rings depends on the type of tree, its age, its immediate surroundings, and the availability of nutritive elements, as well as on a certain number of climatic parameters and their variation throughout the year: sun, temperature, precipitation, humidity, wind, and so forth. Climatic information can also be extracted from the density of the wood. Appropriate mathematical methods enable us to identify the climatic variables that play the most important role—for example, the temperature and precipitation during the summer—and to reconstruct climatic variations.

  Dendroclimatology is nevertheless subject to calibration problems in the association of different series, and it proves above all adapted to mid and high latitudes because the trees in tropical and equatorial regions, which do not have annual growth rings, are of more marginal use.

  Finally, there is an increasing number of studies on the isotopic composition of the cellulose of trees; the mechanisms of fractionation are in this case very complex, but the combined analysis of carbon 13 and oxygen 18 gives access to parameters such as temperature, relative humidity of the air, and the availability of water. Thus a study carried out on the oak trees and wooden beams in Brittany has recently revealed periods of drought over the last four hundred years.7

  In the oceans, the coral that form many reefs between 30°N and 30°S—outside those latitudes the temperatures are too cold—also display annual growth bands. Such bands can cover several centuries—as many as eight hundred years in Bermuda. Bands result from differences in density; the densest layers are generally formed when temperatures are highest. But the composition of these corals, formed of calcium carbonate, provides additional information. Certain elements such as strontium and magnesium, which have a chemical structure similar to that of calcium, are incorporated into the calcite in the form of traces and in proportions that depend on the temperature of the ambient water. Even if other factors, such as the speed of the growth of the coral, also intervene in these concentrations, analyzing these elements is one way we can estimate the temperature at which it was formed. But the isotopic approach is probably the most fruitful. The composition of precipitate carbonate depends on the composition of oxygen 18 of the seawater and on the temperature of the water. Under certain conditions an analysis of coral can enable scientists to reach that parameter. In other cases, we attempt to reconstitute the isotopic content of the seawater, which is affected by evaporation, by strong precipitation, or even, in coastal regions, by the arrival of freshwater from rivers and streams.

  The Distant Past

  The advantage of dendroclimatology is that the climatic information that comes from various trees, both living and fossil, can be combined to obtain records that go back fairly far. But the approach has its limits, and it will probably be difficult to go much further than the 12,000 continuous years covered in Germany. To go further back in the past, scientists can use sedimentary layers that are deposited very slowly on the ocean floor. On the continents, one also uses sediments, deposits of loess, lake sediments, or those preserved in peat bogs, but paleoclimatology also relies on geological observations and on the analysis of matter such as stalagmites. Finally, in all cases, we must call upon other dating methods in addition to the simple counting of annual rings or bands that are so useful for trees and coral.

  Paleoceanography

  The oldest climatic archives are provided by the sediments that accumulate at the bottom of the oceans. In regions with little accumulation they can cover tens of millions of years, but paleoceanographers are just as interested in obtaining shorter series that, in regions where sediment accumulates quicker, enable them to describe with a great deal of detail the succession of glacial and interglacial periods that punctuated the last million years. There are many indicators that enable us to reconstruct the past functioning of oceans, their circulation, and their climatic characteristics.

  Marine sediments are mainly made up of algae and the remains of small organisms, their skeletons, which are composed of either calcium (foraminifera, coccolites) or silica (radiolars, diatoms). Some foraminifera live in warm water, others prefer cold water; there are many species, and their distribution over a site provides an indication of the temperature that existed in the waters where the foraminifers were formed—on the surface for planktonic species, on the bottom for the benthics. This is true for the coccolites, the radiolars, and the diatoms; we’ve learned this via increasingly elaborate mathematical treatments that enable us to establish the transfer functions and to reconstitute the variations of temperature of the surface water and the deep water.

  Paleoceanographers have been at the forefront of isotopic analysis. Knowing that the formation of calcite is accompanied by an isotopic fractionation that depends on the water temperature, in 1948 the Nobel Prize laureate Harold Urey proposed reconstructing this temperature from the concentration of oxygen 18 in the shells of foraminifera. Of course, this concentration depends on that of the ocean, which is variable following the period under study. Thus in a glacial period there was much more ice on the continents, 50 million additional km3, found in large part in North America, and which contributed to the lowering of the sea level by more than 100 meters. Due to the fractionation that accompanies the formation of precipitation, these glacial ice sheets are depleted in oxygen 18 compared to the ocean. The ocean becomes more enriched with oxygen 18 when the quantity of ice on the continents is greater and when the sea level is lower. The interpretation of concentrations of oxygen 18 in foraminifera is a delicate exercise because the isotopic fractionation also depends on which species is being considered, even on its size. But by carefully choosing the sites for sampling—in places where the temperature varies little or its variation is known by another indicator—the concentrations provide information about the level of the sea in the past and thus about the rhythm of the large glaciations. These records concur very well with the variations in sea level deduced from observing the level of coral reefs that have formed “terraces,” which are found in places where the coasts rise regularly due to the effects of subduction.

  There is also a clear connection between the concentration of oxygen 18 of the surface ocean waters and their salinity because each of these parameters is influenced by the intensity of evaporation on the one hand and by the amount of freshwater introduced (precipitation, river water and ice sheets, melting of the sea ice) on the other. Jean-Claude Duplessy and his team at Gif-sur-Yvette developed a method from this obs
ervation that enables salinity to be measured, a parameter that proves crucial when dealing with variations in ocean circulation. We have already indicated how such variations can be documented by analyzing the carbon 13 of foraminifera; the variations in temperature can also be seen from an analysis of concentrations of certain chemical elements in the same foraminifera.

  As is the case with coral, the concentration of certain chemical elements is influenced by the temperature; this is true of the proportion of magnesium to calcium, which is commonly used as an indicator of the temperature. Organic molecules are also sensitive to this parameter: some organisms respond to a change in temperature by changing the composition of their membrane at the level of molecules called alkenones, favoring nonsaturated varieties when the temperature becomes colder. The alkenones are well preserved in marine sediment, and measuring their level of saturation is another means of determining the temperature of the ocean.

  Continental Archives

  As we can see, the paleoceanographer is a detective on the lookout for the slightest climatic sign. The same is true of specialists of continental archives, who use almost the same range of indicators but in very different environments.

  The analysis of species has an equivalent in palynology, the discipline that looks at pollens and spores produced by plants during blooming and at their dispersal in the environment. In a given site, the distribution of pollens is an indication of the ambient vegetation and thus of the climate, primarily of precipitation and temperature. These pollens are found in many sediments at the bottom of lakes, in peat bogs, and even in marine sediments of coastal regions where they have been transported by the wind. Several series cover all of the Quaternary, but most stop at the end of the last glacial period. In France, the variations in temperature and precipitation have been reconstructed from pollen series that cover the last 150,000 years in the Vosges8 and extend over 400,000 years in the Bresse9 and in the Velay.10 Indicators other than pollens, such as insects and snails, can contribute to paleoclimatic reconstructions.11

  Like marine sediments, lake sediments lend themselves to the isotopic analysis of species such as the ostracods; an analysis of their concentrations of oxygen 18 allows us to go back to that of the lake water and indirectly to that of the rains watering the catchment areas.12 The relationship we have already mentioned between the isotopic content of the rains and the climatic conditions, more precisely the temperature in the regions under study, enables us to estimate this parameter.

  Isotopic analysis is also essential to the study of limestone concretions formed in caves. The concentration of oxygen 18 in the dripping water that very slowly forms a stalagmite—here, too, through calcite precipitation—more or less reflects that of the precipitations. Thus stalagmites studied in China show the rhythm of monsoons13 because in these regions the quantity of precipitation influences the stalagmites’ isotopic content. A study done on a stalagmite from the cave of Villars in the Dordogne shows that the concentration of carbon 13 can indicate the presence or absence of vegetation above the cave, which is itself influenced by climatic conditions.14

  Finally, climatic indications are also present in the deposits of loess, which, brought by the wind, are found in many regions. But it is more difficult to extract temperature or precipitation data from them. The same is true for many other observations: the alignment of moraines and the extent of the frozen ground in periglacial regions, retreats and advances of glaciers, the altitude at which snow remains permanently, the maximum extension of vegetation in mountain massifs, and the evolution of the lake water level in many regions of the planet.

  Dating Oceanic and Continental Archives

  Lake sediments can show annual layers, or varvae, which can be used for dating purposes. This can also be the case for some stalagmites, but it is exceptional in the case of marine sediments. In any event, this type of dating does not allow us to go back in time over tens or hundreds of thousands of years without interruption. Paleoclimatologists must therefore use their imagination, because what interest is there in determining the climate of a region if we don’t know to what period of the past it corresponds?

  Some isotopes, radioactive in this case, have proven to be of enormous help. When they disintegrate these atoms form new elements: for example, carbon 14 is transformed into nitrogen, potassium 40 into argon; this occurs at a rhythm that varies from one element to another but is well established. The initial radioactive isotope (carbon 14 or potassium 40) is characterized by the duration of its half-life, the time at the end of which half of the element will have disappeared, 5,730 years in the case of carbon 14. The concentration of carbon 14 in a piece of wood, which is measured very precisely with the help of an accelerator, will thus tell us its precise age, provided that the concentration of carbon 14 of the atmosphere at the moment when it was formed is known. Carbon 14, which is formed by the interaction between cosmic rays and nitrogen present in the upper atmosphere, has a concentration that varies little over long periods of time, but the variations linked to solar activity and to the magnetic field must be taken into consideration if we want to achieve precise dating. In practice, the inverse approach is followed: an analysis of carbon 14 in wood coming from trees, whose age we know over the last 12,000 years, enables us to calculate the initial content of carbon 14. We then can use the “calibrating” curve, thus established, to date a whole group of other materials (construction wood, organic matter, bone, shells, stalagmites, coral, and so forth), which, as we know, can be of use in archaeology and in many other domains.

  Another radioactive method extends this “calibrating” curve; it is based on isotopes of uranium and thorium, the analysis of which enables us to date coral.15 It is sufficient then to measure the carbon 14 in a well-dated coral to calculate its initial level. But beyond around 50,000 years carbon 14 dating is no longer viable because concentrations become too weak to be measured precisely. For the paleoclimatologist, the uranium-thorium method then takes over; its principal applications are the dating of ocean corals and that of stalagmites on the continental environment. This method also becomes less and less precise as one goes back further in time and is no longer useful beyond a few hundred thousand years. Methods based on the analysis of luminescence or traces of fission, both connected to the phenomenon of radioactivity, have not been found to be useful in paleoclimatology. In fact, it is astronomy that unblocked the research in this field and established a viable dating of sedimentary series covering hundreds, even millions, of years.

  Beyond the fact that astronomy confirmed Milankovitch’s theory, the proof of the periodicities of astronomic parameters (Figure 3.2) in marine records paved the way for a particularly efficacious dating method. Jim Hays and his colleagues obtained different records from cores drilled from the Indian Ocean. Initially they dated them while assuming a regular accumulation of sediment, which meant the oldest sample would be 450,000 years old. They then observed that the orbital periodicities characteristic of variations of eccentricity, obliquity, and precession were also present in these records and deduced from this that changes in insolation were the cause of the succession of glacial and interglacial periods in the Quaternary. This demonstration, which has been amply confirmed since then, offers, in principle, a means of absolute dating, called orbital tuning. The notion, which is applied to many types of records, consists of establishing a rough dating that is then refined by using the hypothesis that the climatic variable concerned responds with a certain delay to the variations in insolation. A study published in 1990 by Nick Shackleton, André Berger, and Dick Peltier16 illustrates the power of this method. In a marine core, these scientists found the level corresponding to the inversion of the Earth’s magnetic field, called Brunhes-Matuyama, and through orbital tuning concluded it had occurred 780,000 years ago, whereas radioactive dating based on the analysis of isotopes of argon suggested it happened 730,000 years ago. This latter estimate has since been proven wrong, and it appears that the orbital tuning estimate is c
orrect. Furthermore, this method has been applied to all of the Quaternary and well beyond, as the variation in insolation, regardless of the period under consideration, seems to be one of the driving forces in past climatic changes, although it is not always the principal one.

  A Cornucopia of Results

  Ever since the first scientific observations of the climate and then the demonstration by geologists of the existence of glacial periods, there has been a considerable volume of data that have been acquired from oceanic and continental archives. This movement toward accumulating data has accelerated in the last few decades, taking advantage of increasingly sophisticated methods of analysis and dating. Our knowledge of the functioning of the climatic system has become revolutionized because of these innovations.

  Even if we still have many questions regarding the implicated mechanisms, the demonstration of the link between the rhythm of the great glaciations and the position of the Earth on its orbit, which we owe to marine sediments, can be considered one of the most remarkable discoveries. An analysis of marine sediments enables us to describe not only the evolution of the sea level but also that of the circulation of currents both on the surface and deep in the oceans, providing, for example, the evidence of massive discharges of icebergs and their influence on deep water circulation. Furthermore, they offer very long-term records going back as far as 65 million years, whose analysis, as we have seen, indicates a progressive cooling of the planet, the formation of the Antarctic ice sheet, and then the more recent formation of that of Greenland.