SINCE HUMAN beings emerged as a species, Homo sapiens, half a million years ago, the biggest natural changes in climate have been the switches into, and out of, ice ages. During that time (and for several million years before) the Earth has been slightly colder than today. There were full ice age conditions with extensive icecaps in both hemispheres. Many pieces of geological evidence show that there has been a regular rhythm of changes in temperature in which an ice age roughly 100 000 years long is succeeded by a slightly warmer interval, called an interglacial, which lasts between 10 000 and 20 000 years. The pattern has repeated 10 times in a little more than a million years and was only slightly different in the 2 million years before that. Today, we are living near the end of a natural warm spell, an interglacial which began a little over 10 000 years ago.
These rhythms are now well understood. Studies carried out in the second half of the 1980s have established that the ice age rhythms are linked with changes in the orientation of the Earth as it orbits around the Sun, and that the amount of carbon dioxide in the atmosphere amplifies these changes. The natural greenhouse effect seems to have an adjustable thermostat, whose temperature setting is partly determined by astronomical influences. The discovery is important because it shows how changes in the concentration of carbon dioxide in the atmosphere affect the temperature of the globe. It is alarming because the build-up of carbon dioxide caused by human activities is bucking against the natural trend, throwing a spanner in the works that may cause unpredictable changes in the working of the weather machine.
Advertisement
The astronomical theory of ice ages goes back to the 19th century, and the work of James Croll, a Scot who was born in 1821 and published his ideas in the 1860s. Croll’s first paper on ice ages was published in the Philosophical Magazine in 1864. By the time the theory became established, in the first quarter of the 20th century, Croll’s name, although not by any means forgotten, had been relegated to the footnotes of history. Today, most accounts of the astronomical theory call it the ‘Milankovic Model’, after a Yugoslav astronomer, Milutin Milankovic, who refined and improved the theory in the decades before the Second World War.
Whatever name it goes by, the three main cycles that are the basis of the astronomical theory are easy to understand. The difficulty lies in explaining how the modest seasonal changes in temperature that they produce can cause the pattern of ice ages and interglacials. This is where carbon dioxide comes in.
The weather changes with the seasons because the Earth is tilted relative to the plane of its orbit about the Sun. Imagine a line drawn from the centre of the Sun to the centre of the Earth. This line lies in a plane that astronomers call the plane of the ecliptic. If our planet were upright, it would be spinning in such a way that the North Pole lay directly above this line, so that a line from the pole to the centre of the Earth would meet the ecliptic plane at right angles. If that were the case, everywhere on Earth there would be 12 hours of daylight and 12 hours of night, every day of the year. There would be no seasonal variations. In fact, the spinning Earth leans over slightly, 23.5 degrees out of the vertical, that is, 23.5 degrees away from the perpendicular to the ecliptic. Because the spinning Earth acts like a gyroscope, it maintains this orientation as it moves around the Sun. In the course of a year, although the Earth’s position in its orbit around the Sun changes, the direction in which it is tilted stays the same, with the North Pole always ‘pointing to’ the Pole Star. This is what causes the seasons.
When the Sun is on the same side of the Earth as the Pole Star is, the northern hemisphere is tilted towards the Sun. There are 24 hours of sunlight at the pole, and more hours of sunlight than dark each day over the whole northern hemisphere, where it is summer. Six months later, the Earth is on the other side of the Sun, but still tilted towards the Pole Star. Now, the northern hemisphere is pointing away from the Sun. The pole is dark, and there are long nights everywhere in the northern hemisphere, where it is winter. The seasons are reversed in the southern hemisphere, and the two intermediate seasons correspond to intermediate positions of the Earth in its orbit around the Sun.
All this happens because the tilt of the Earth is constant over a year. But the spinning Earth is like a gyroscope, and anyone who has watched a child’s spinning top will know that a gyroscope can wobble as it spins. The Earth wobbles in a similar fashion, chiefly as a result of the gravitational influence of the other planets. But because the Earth is such a big ‘spinning top’, the wobbles are rather slow. They are so slow that they do not affect the seasonal patterns over a year, or even 100 years; but they begin to become important when we look at stretches of time measured in thousands of years. These wobbles make up the Milankovic cycles.
The most rapid change that we have to worry about stems from the fact that the direction in which the North Pole points actually traces a circle on the sky. This circle follows a cycle that varies slightly in length but has two main components, near 23 000 and 19 000 years. Ten thousand years ago, the present-day Pole Star would not have provided an accurate indication of the direction of north. The shifting orientation of the pole causes an apparent drift of the stars across the sky, as viewed from Earth, which is called the precession of the equinoxes; in fact, it is the spinning Earth that is precessing, and one result of this is that the pattern of the seasons changes slowly as the millennia pass. This is one component of the Milankovic Model.
At the same time as the axis precesses, the angle at which our planet is tilted varies, nodding up and down between 21.8 degrees (more nearly upright) and 24.4 degrees (more tilted) over a cycle roughly 41 000 years long. This makes the second part of the Milankovic Model. The present day tilt, which is actually closer to 23.4 than 23.5 degrees, is about halfway between these extremes. At present the tilt is decreasing, which means, among other things, that the differences between summer and winter are less today than they were a few thousand years ago. Summers are a little cooler, and winters a little warmer.
The third component of the Milankovic Model is a slightly different effect, linked with changes in the shape of the Earth’s orbit around the Sun. This is also caused by the shifting interplay of gravitational forces in the Solar System; as a result, the Earth’s orbit stretches slightly from circular to elliptical and back again with period of roughly 100 000 years. When the orbit is circular, the Earth receives the same amount of heat from the Sun every day of the year. When the orbit is more elliptical, on some days of the year our planet is closer to the Sun, and receives more heat, than on other days in the same year. But the total amount of heat received by the whole planet over a whole year is always the same (as long as the Sun’s output is constant); the three astronomical cycles only redistribute heat between the seasons, and do that only by a very modest amount. At the time of Croll, or even of Milankovic, it was impossible to test the astronomical theory of ice ages properly, because nobody knew the exact dates of the advances and retreats of the ice over past millennia. Most climatologists regarded the theory as extremely implausible, because the astronomical calculations show that the change in heating from the Sun (insolation), over the seasons is small. In one of his earliest calculations, Milankovic himself had presented the figures in terms of the way that summer sunlight has varied at latitude 65Degree north. Compared with the present-day insolation at different latitudes, the summer insolation at 65Degree north has varied, over the past 600 000 years, between the equivalent of present day summer insolation at 70Degree north and present day summer insolation at 60Degree north. But, of course, in the years when summers were hotter, winters were colder (and vice versa). The effects exactly balance, so that the band around the globe at 65Degree north, like all latitudes, always received the same amount of insolation over the whole year as it does today. The Milankovic Model remained an intriguing but unproven hypothesis, with only a few ardent supporters, until techniques were developed to trace the temperature of the globe from millennium to millennium down through the ages. The breakthrough came in the 1970s, from studies of the remains of the chalky shells of planktonic Foraminifera, tiny protozoans, in deep-sea sediments.
These sediments are constantly being laid down, as plankton die and fall to the seabed. Older remains lie deeper below the surface and younger sediments near the top. Geologists can extract sediments from the seabed in the form of long cores drilled from ships, but they cannot simply read off from the core the age of the sediments at any particular depth. There is no straightforward equivalent to the annual growth rings of trees, which can just be counted to find the age of a particular piece of wood. Instead, you have to date the sediments by comparing the magnetism locked up in the sediments when they formed with the magnetism of continental rocks that have already been dated by other means.
The Earth’s magnetic field varies considerably over geological time, sometimes being weaker, sometimes stronger, and intermittently reversing direction completely. This pattern of changes, especially the reversals, provides a distinctive fingerprint. A stretch of sediments drilled from the seabed shows a particular magnetic fingerprint that matches part of the dated pattern of magnetism found on the continents. This method gives precise dates for sediments on the sea floor. So the geologists knew that a particular layer of sediment had been laid down between, say, 35 000 and 36 000 years ago, that another layer was deposited between 127 000 and 128 000 years ago, and so on. In fact, by the mid-1970s they had a continuous record of sediments covering some half a million years back from the present day, extracted from the Antarctic Ocean. The next stage was to use this sediment core as a fossil thermometer, tracing the temperature at the time each layer was being laid down.
The trick depended on some subtle measurements, and an understanding of how the biochemistry of Foraminifera is affected by changing global temperatures. There are two common types of oxygen atoms (two isotopes) known as oxygen-16 and oxygen-18. Both are present in the air that we breathe, and also in sea water (H2O). Oxygen-18 is heavier than oxygen-16 so some molecules of sea water are heavier than others. Heavier molecules of water freeze more readily than lighter molecules, so during an ice age proportionately more of the heavy molecules are locked away in ice. As sea creatures take up oxygen from the surrounding water to help to build their shells, during an ice age the shells contain proportionately more of the lighter oxygen isotope, which has not been locked away in the ice sheets. By measuring the way the proportions of the two oxygen isotopes vary in different layers of sediments, geologists can infer not just the exact temperature at the time the sediments were laid down, but the extent of global ice cover. The sediments provide a direct record of the pulse of ice ages.
Carbon dioxide, the missing link
Finding the right sediments to analyse, getting the timescale right and extracting the isotope data were all difficult tasks. From the 1940s onwards, researchers tackled various aspects of the problem. The breakthrough came in 1976, when Jim Hays, of the Lamont-Doherty Geological Observatory in New York State, John Imbrie, from Brown University in Rhode Island, and Nick Shackleton, of the University of Cambridge, analysed the changing pattern of ice cover revealed from the isotopes in core sediments spanning half a million years. The analysis showed that the rhythms of advance and retreat of ice cover over the past few hundred thousand years are dominated by three cycles, respectively 100 000 years, 41 000 years and a little more than 21 000 years long. The Milankovic Model had well and truly come in from the cold. But the puzzle remained: how could such small, seasonal effects produce such profound changes in global climate? In the 10 years following the work by Hays, Imbrie and Shackleton, many studies strengthened the link between ice ages and astronomical cycles. The main contribution came from a group of researchers from five universities, headed by Imbrie. They called themselves SPECMAP, from the term ‘spectral mapping’. The ‘spectrum’ of a pattern of variations is what reveals the presence of regular, periodic contributions to the overall picture of confused ups and downs of, in this case, temperature. The biggest surprise to emerge as they extended this kind of study further back into the past, using longer stretches of sediment core from the seabed, was the dominant role of the 100 000 year cycle. This cycle ought to be the weakest of the three astronomical influences, and produces variations in insolation amounting to no more than 0.1 per cent of the annual total. Yet by 1982 the analysis had shown that ice age rhythms had marched precisely in step with this long-term cycle for at least the past 800 000 years. The two shorter cycles are also present in the ‘signal’ from the same sediments, and together the astronomical influences account for about 80 per cent of the variation in climate over timescales between 20 000 and 100 000 years long. Of course, other influences are at work, volcanoes erupt, ocean currents shift position, and so on. But the most important set of influences determining the pattern of ice age and interglacial rhythms is indeed the Milankovic mechanism.
In the 1980s the study of astronomical cycles in the geological record went from strength to strength. Microscopic life forms in the water off the Arabian peninsula, including the ubiquitous Foraminifera, are affected by temperature changes in the water that are themselves linked with variations in the Indian monsoon. Warren Prell, one of Imbrie’s colleagues at Brown University, headed a study which found that the intensity of the monsoon is affected by the astronomical cycles. But identifying the existence of the rhythms is quite a different task from understanding how they work to control climate. Even taking the three basic cycles alone, it is clear that the combined influence on the Earth is never exactly the same from one ice age (or one interglacial) to the next. Although the Earth is warm today, and it was warm about 120 000 years ago, the warmth of the two interglacials was produced by two different combinations of the three cycles. Because neither 23000 nor 41000 divides exactly into 100 000 years, the exact pattern does not repeat for literally billions of years. Fortunately, that need not bother us now. We are interested in what happened during the most recent ice age, and maybe the one before that, and on that sort of timescale there is now clear proof that the astronomical cycles exert their strong influence on the climate of our planet through an intermediary, the carbon dioxide greenhouse effect.
Solid evidence that the greenhouse effect, or the lack of it, might help to explain the pattern of ice ages came in 1982, when researchers from the University of Bern, in Switzerland, reported their analysis of bubbles of carbon dioxide trapped in the icecaps. When snow falls on the ice, it is light and fluffy and there are plenty of gaps between the snowflakes. These gaps are full of air. As more snow falls and the years pass, each layer is squeezed down in its turn and becomes ice, but some of that air stays trapped in the ice as bubbles. On a polar icecap (or, indeed, on a mountain glacier) the youngest ice is at the top, and successively older layers lie deeper below the surface. Because the snow is laid down in regular annual layers, which are still visible when it has been compressed into ice, you can, in some cases, determine the age of a piece of ice from a particular depth by drilling a core of ice from the sheet, and counting the layers down from the surface.
The technique, pioneered by a Dane, Willi Dansgaard, and his colleagues in Copenhagen, is useful for climatologists because the same kind of isotope measurements that reveal temperatures of long ago from the deep sea sediments also reveal how temperature has changed over the period of time covered by the ice core. The Swiss researchers, however decided to investigate not the temperature record in their core sample, but the atmospheric record.
Old air from the ice sheets
Researchers had tried similar things before but with mixed success; it was the Swiss work that opened the eyes of climatologists to the implications of the technique. Careful studies of the composition of the air trapped in the bubbles in the ice showed that about 20 000 years ago, when the most recent ice age was at its coldest, there was much less carbon dioxide in the air. Then carbon dioxide was present only in the range from 180 to 240 parts per million, compared with 280 ppm in the early 19th century, before widespread burning of fossil fuel began, and 350 ppm today. About 16 000 years ago, at the time when, we know from other geological evidence, the ice sheets began to melt, the proportion of carbon dioxide began to increase, and by the end of the ice age it had reached roughly the same concentration as the pre-industrial level. The evidence was persuasive, but many climatologists were worried that carbon dioxide might somehow be leaking out of the old bubbles buried in the ice, or that the tricky techniques needed to extract and measure the tiny samples of old air might be giving a false reading. These doubts were allayed when Shackleton and his colleagues found exactly the same pattern of variations in the level of carbon dioxide from a completely independent technique, using deep sea sediments and another kind of isotope study, this time looking at carbon instead of oxygen.
Microscopic life forms near the surface of the ocean, plankton, act as a pump, extracting carbon dioxide from the air and depositing it on the sea floor. Like oxygen, atoms of carbon come in different varieties. The important ones in this study are carbon-12 and carbon-13. Both are stable, and carbon-12 is the lighter of the two isotopes. Because of the way their biochemistry works, the plankton prefer to take up the lighter isotope to build their shells. So the biological carbon pump not only shifts carbonate to the bottom waters, it also preferentially shifts carbon-12, rather than carbon-13, out of the surface layers. Some organisms, however, live in the cold, deep waters, not near the surface. They build their shells out of the carbonate that is available to them, and this is already enriched with light carbon, so their shells contain even less carbon-13, proportionately, than the shells of creatures that live in the surface layers. The different ratios of carbon isotopes in the two types of shells tell geologists how much carbon dioxide was in the atmosphere when they formed.
By now, it should come as no surprise to find that these studies, too, show that there was less carbon dioxide in the air during the latest ice age than there is today. The ice core studied by the Swiss team extends back over the past 40 000 years, and over that time span the concentrations of carbon dioxide measured in the ice core bubbles and the concentrations inferred from measuring isotopes in sediments on the seabed are exactly the same. But the sedimentary record goes back much further, for hundreds of thousands of years, and provides much more information. The intriguing new discovery is that the changes in carbon dioxide precede the changes in ice cover. Working with Nicklas Pisias, of Oregon State University, Shackleton applied the technique to analyse a sediment core covering a span of 340 000 years, more than three complete glacial-interglacial cycles. They found all three of the Milankovic rhythms present in both the temperature record (determined from oxygen isotopes) and the carbon dioxide record (determined from carbon isotopes). But by comparing their finds with calculations of how the astronomical cycles have varied, they found that the changes in the Earth’s orbit precede the changes in carbon dioxide, and that the changes in carbon dioxide in turn precede changes in climate, or, at least, changes in ice cover. Somehow, the astronomical changes cause the biological pump to increase or decrease its activity. This produces a change in the strength of the greenhouse effect which is then a key element in switching the world in to or out of an ice age, and helps to explain ups and downs of temperature on lesser scales, within an ice age or an interglacial. Carbon dioxide amplifies the changes that the Milankovic process is trying to produce.
The amplification almost certainly operates through changes in the workings of the carbon pump at high latitudes. High latitudes feel the effects of the Milankovic rhythms more strongly than tropical regions do. Some researchers suggest that plankton living at high latitudes form the link between the slight temperature changes of the Milankovic cycles and the start of an ice age. When there is plenty of sunlight, plankton might grow vigorously, taking carbon out of the atmosphere as they do so, leading to cooling. All such proposals are still very tentative. But one clear conclusion is that the world ‘ought’ to be cooling into the next ice age about now. In fact, the build-up of carbon dioxide and other greenhouse gases produced by human activity is making the world unnaturally warm. And there is another, related, problem. Any thinning of the ozone layer at high latitudes may kill plankton in these sensitive areas, thereby reducing the effectiveness of the biological pump; this will lead to yet more carbon dioxide in our atmosphere, and even more global warming.
The only way to get a better understanding of the links between astronomy, biology and climate that control the pulse of ice ages will be by taking a more detailed look at the latest ice age. Studies of air bubbles from the deep Vostok core, drilled in Antarctica (see ‘Frozen assets of the ice cores’, New Scientist, 14 April 1988) are now pointing the way forward. And, because of the carbon dioxide connection, these studies of ancient ice age rhythms will provide invaluable insights into the growing greenhouse effects which threaten to warm the Earth more in the next 50 years than at any time for the past 5 million years. James Croll never knew what he was starting.
