The last year of the last millennium was its hottest. There was no global drought; no summer heatwave swept across the planet. Yet enough months were, quietly, that little bit above average to add up to a record breaker, for Britain at least. What conclusions should we draw? The most popular one is that this is the result of a human-induced greenhouse effect. That is almost certainly true, but the connection is not as obvious as it seems. Why scientists have been hedging their bets on the matter, when there is evidence of global warming by about half a degree centigrade over the past century, seems puzzling-until you appreciate the full complexity of the earth's climate.
Climate has no big idea. Evolution has Darwinian natural selection; cosmology has the big bang; genetics has DNA. Climate is just "one damned thing after another"; worse, it is lots of damned things at the same time. This makes for messy history. For example, it is not true that climate was steady and comfortable until we started pouring greenhouse gases into the atmosphere. Consider how global temperatures have fluctuated over the past 1m years: the graph looks like the jagged profile of the Dolomites. Focus in on the past millennium or even the past century and you see the same pattern: a series of peaks too fine to have been discernible in the million-year record.
We can make only one sweeping statement about the climate system: the average temperature of the earth's surface depends on a balance between how much heat it receives from the sun and how much it radiates back into space. Changes in global mean temperature are ultimately caused by changes in the amount of heat into or out of the planet.
The complexity arises from the fact that so many phenomena induce such changes. One is the greenhouse effect. The earth's atmosphere is roughly four fifths nitrogen and one fifth oxygen; but 1 per cent or so is made up of a m?nge of other gaseous compounds, many in such small quantities that they were detected only over the past few decades. Most of this is argon, and the rest is primarily water vapour and carbon dioxide. Yet these tiny proportions exert a huge effect. Water vapour, carbon dioxide and other greenhouse gases such as methane absorb some of the heat radiated by the warm surface of the earth. So "greenhouse gases" retain some of the sun's energy which would otherwise be absorbed and then re-radiated to space. This warms up the lower atmosphere, and so surface temperatures rise.
Contrary to common belief, the greenhouse effect is not primarily human-made. There are natural sources of greenhouse gases. For example, carbon dioxide is generated by decomposition of plant matter, respiration of plants and animals, and volcanic emissions, while methane is produced by decomposition in air-free environments. These effects raise the temperature by about 35 degrees: most of the world would be below freezing if there was no natural greenhouse effect. But industry and agriculture add to the effect by producing significant amounts of greenhouse gases.
Another factor which influences the earth's heat is brightness. Clouds and ice sheets reflect sunlight and so act to cool the planet. Vegetation generally darkens dry ground and so reduces its reflectivity. The overall reflectivity of the earth is called its albedo. Both the greenhouse and the albedo effects are subject to feedbacks. For example, if the world cools and ice sheets grow larger, the albedo increases-more sunlight is reflected back into space-which induces further cooling. (The converse is also true.)
James Lovelock's Gaia hypothesis, advanced in the 1970s, suggested that the strongest influence on natural climate change comes from living systems, which display feedback loops which counteract any small changes, either cooling or warming. Lovelock and others proposed that albedo changes caused by clouds are regulated by marine algae which exude a sulphur-containing gas called DMS. This gas is converted in the atmosphere to sulphate, which is a component of the tiny salt-like particles on which water droplets condense to form clouds. If the earth warms up, they said, the algae are stimulated into producing more DMS, which results in the formation of more clouds, a decrease in the earth's albedo-and a compensatory cooling effect. By such means, say Gaia supporters, life acts as a kind of thermostat for the planet. There is no strong evidence that this hypothesis is true, although Lovelock's idea has had the effect of focusing more attention on the (undoubted) role that life does play in climate regulation.
The Gaia hypothesis looked as if it might be climate's big idea, but there is just too much else going on. The greenhouse effect might explain why this century was warmer than the last, but not why this summer was warmer than the previous one. On time scales of billions of years, the amount of heat received from the sun changes because its structure and composition as a star alters. (In about six billion years it will grow to fry and then to swallow the earth.) And the sun's heat output rises and falls with an 11-year pulse, accompanied by the proliferation of sunspots at each maximum. But the change is small, and it is not obvious how that could manifest itself as a big shift in climate.
This hasn't prevented some scientists from suggesting that all recent climate change can be explained by changes in solar activity. Certainly, a period of cold climate, from the 14th to the 18th century, coincided with episodes of lower-than-average sunspot counts. Such links have been trumpeted in the US by the Global Climate Coalition, an industry-funded body which opposes regulations on emission of greenhouse gases and seeks to find natural explanations for recent warming.
the big freeze
Looking back over longer time scales, we run into the pseudo-rhythmic cycle of the ice ages. The last ice age ended 11,000 years ago, having persisted for some 100,000 years. Before that occurred a warm period like today's, called an interglacial, which began at the end of the previous ice age about 140,000 years ago. And so it goes on-a succession of ice ages and interglacials repeating roughly every 100,000 years for the past half a million years.
During an ice age, the polar ice caps grow towards the equator. Water evaporated from the oceans falls on the ice as snow and is gradually compacted, under successive layers, until it forms new ice. This means that the oceans don't get the water back again as rain, and so they become shallower. Across the globe, sea levels plummet. At the height of the last ice age, some 18,000 years ago, the sea level was so low-on average about 120 metres below its present level-that Australia was connected to New Guinea by dry land, and Asia to Indonesia. The ice sheets in the northern hemisphere covered northern Germany and half of Britain, as well as most of southeast Asia and South America. The global mean temperature then was some six degrees lower than today. (Relatively small differences in the global mean can be accompanied by huge differences in regional environmental conditions.) And significantly, quantities of the greenhouse gases carbon dioxide and methane in the atmosphere were much reduced. This is consistent with a colder world; but what was cause and what was effect? Did a cold climate suppress the natural sources of these gases, or did some change in those sources trigger the cooling? This is a key question for scientists trying to understand what future changes in greenhouse gases have in store for us.
Dramatic climate shifts like the ice ages leave telltale signs. These were the catalyst for the revolution in understanding of climate which took place in the late 19th century. It all began with "erratic" rocks-boulders in mountain landscapes where they were not supposed to be. Often the only known bedrock of the same type was miles away. What could have carried such boulders so far? The answer was clear to most geologists of the early 19th century: the Biblical flood, which was believed to have rearranged the global landscape. But Swiss geologist Louis Agassiz revived the old idea of Scottish geologist James Hutton: that glaciers had borne the erratic rocks far afield before the ice melted.
Whereas it had long been assumed that the earth's climate had always been more or less as it is in modern times, geologists were now forced to conclude that in days past it had sometimes been much colder. What had caused these cold spells? The answer, it seemed, lay not in the workings of the earth's own climate, but in astronomy. In 1842 Joseph Alphonse Adh?r, a French mathematician, realised that a repeating cycle of ice ages might result from variations in the earth's axis of rotation. At present this axis stands at an angle of about 23.5 degrees relative to the plane of the earth's orbit around the sun. This tilt is responsible for the seasons: summer occurs in the hemisphere tilted towards the sun at that time. Adh?r knew that the earth actually wobbles around its axis, rather like a spinning gyroscope. This wobble is responsible for the precession of the equinoxes, and makes the apparent positions of the stars change slowly over many years. It takes 23,000 years to complete one wobble.
Another effect of the wobble is to make the length of the winters in the two hemispheres different. Adh?r proposed that ice ages might be triggered when the winters were longest-once every 11,000 years or so in each hemisphere. Adh?r succeeded in establishing the idea that periodic changes in the earth's orbit could trigger periodic changes in climate. But it was not until 1920 that a Serbian mathematician, Milutin Milankovitch, finally calculated the sums correctly. There are two other changes in the orbit which also affect the heat balance of the earth. First, the tilt angle itself changes-the earth tips back and forth-over a period of 40,000 years. Second, the shape of the elliptical orbit around the sun (called the eccentricity) alters on a 100,000-year time scale, becoming alternately longer or shorter. Milankovitch showed that a combination of these three oscillations will alter the amount of heat the earth receives over these time scales, and thus could trigger ice ages. Because the three rhythms are out of step, their combined effect is only semi-regular, and so ice ages of varying coldness are anticipated.
The predominant pulse of the glacials over the past 700,000 years has been on a 100,000-year scale, implying that changes in eccentricity have been the main driving force. Looking still further back in time, the rhythm of the ice ages changes to a beat of 40,000 years, suggesting that changes in the tilt angle were then more important.
dead plankton do tell tales
How can we know about climate change over these huge stretches of time? We must dig. In the 19th century, James Croll, a Scottish scientist who had improved on Adh?r's astronomical theory of the ice ages, suggested that a record of past climate might be compiled by examining the remains of marine organisms in "the deep recesses of the ocean." The seas are full of microscopic creatures such as the plankton called forams, which construct elaborate shell-like skeletons for themselves from dissolved calcium carbonate, the stuff of chalk. When the forams die, their bodies settle to the sea bed, and the robust shells accumulate in the muddy sediment. Different species of foram live in waters of different temperatures, so if we can identify changes in the types of foram we may infer something about the changes in ocean temperature at that location when the sediment was deposited. In the 1920s, the German scientist Wolfgang Schott showed a progression in Atlantic sediments from warm-water species of forams in the upper layer, to cold-water species and then back to warm-water species lower down.
The development of radiocarbon dating meant that precise ages could be assigned to the different layers of sediment. This showed that the uppermost warm-water species disappeared about 11,000 years ago-just when, we now know, the switch between a glacial and an interglacial episode occurred. But these dead sea creatures have still more tales to tell. In the late 1940s, scientists discovered that the oxygen in the carbonate shells of cold-water forams is enriched in oxygen's heavier isotope, oxygen-18, relative to warm-water forams (which contain a higher proportion of the lighter oxygen-16). By measuring the oxygen-isotope ratios in sediments, we have a thermometer of the water temperature when the sediments were laid down.
In the 1950s, using this isotope method, the geologist Cesare Emiliani deduced that there had been seven ice ages over the past 300,000 years. In the 1970s, climate scientists initiated an international programme, called Climap, to map the climate record of the past 700 millennia from columns of sediment drilled from sea beds throughout the oceans. This record confirmed the broad features of Emiliani's work and showed that ice ages had recurred at intervals of roughly 100,000 years. Overall, the Climap record matched the pattern predicted by Milankovitch, and some scientists thought that the riddle of the ice ages was solved.
But it is not quite that simple. There seem to have been periods in the distant past when there was no cycle of ice ages at all. The Cretaceous period, 144m to 65m years ago, is thought to have been warmer than today, perhaps by as much as 60 degrees at the south pole. This warm spell may have persisted until about 30m years ago. Some scientists believe that, over time scales of millions of years, the slow movements of the continental plates may control global climate. For one thing, ice caps at the poles can form most readily if there are continents at the poles. Because of continental drift this has not always been the case-Antarctica may have been at the equator about 900m years ago, for instance. The cooling 30m years ago, meanwhile, might have been triggered by the collision of the Asian and Indian plates which rumpled up their edges to create the Himalayas. The formation of snow on this high ground increased the earth's albedo, and the increased "weathering" of the high rocks by rain would have had the effect of removing the greenhouse gas carbon dioxide from the atmosphere.
So Milankovitch cycles are not the only story in historic climate change. Further, the changes in earth's orbital characteristics provoke only very small changes in the amount of heat from the sun, and it is not clear that this is sufficient to trigger a switch in global climate-particularly on time scales as rapid as sometimes observed. Milankovitch's theory predicts gradual change but we often see abrupt shifts between glacials and interglacials.
Much of what we know about climate change now comes not from marine sediments but from the polar ice caps. The reason foram shells are enriched in oxygen-18 in times of cold climate is that this heavier isotope evaporates less readily from the seas. So when the world's water gets locked away in ice during glacial times, and the sea level sinks, there is proportionately more oxygen-18 in the seas-and correspondingly more oxygen-16 in the water of the polar ice sheets. By drilling cores out of these sheets-some almost five kilometres thick and 250,000 years old at the base-we can reconstruct a climate record over this period from isotope measurements. Better still, the ice is peppered with tiny bubbles which contain samples of the atmosphere from the time the ice was deposited as snow. This means that we can study how changes in the concentrations of greenhouse gases have varied with changes in climate, and deduce something about cause and effect.
throwing the ocean's switches
In the early 1990s another dramatic discovery from ice-core climate records was that shifts in climate can be disconcertingly rapid. During some periods the temperature seems to have risen and fallen significantly over only a few decades. One fast switch happened as the world was emerging from the last ice age 11,000 years ago. It was warming gradually until, about 10,500 years ago, the world plunged back into ice-age conditions over a period perhaps as short as 50 years. The change was especially pronounced in the north Atlantic region. One explanation for this rapid cooling, called the Younger-Dryas event, involves ocean circulation.
The water in the oceans is constantly on the move. It doesn't simply slosh back and forth with the tides, but circulates steadily around the globe in huge currents. The surface currents, down to a depth of about 100 metres, are driven by the winds. But below this is a huge conveyor-like circulation of deep water which rises in the north Pacific, passes eastwards across the equator and around Africa, and travels north, to sink in the north Atlantic and make the return journey via the Southern Ocean. The Atlantic sinking happens because the water becomes colder; cold sea water is denser than warm sea water. The effect is accentuated by water freezing into ice at the poles: ice rejects salt, and the salt left behind in the water makes it denser. So the circulation of the oceans is driven by changes in heat and salt-the thermohaline circulation.
This conveyor belt carries warm water from the tropics towards the poles, and thus helps to redistribute heat across the planet. If it were to cease turning (or turn less vigorously) the high latitude regions such as northern Europe would be much colder. This may be what happened during the Younger-Dryas event. As the northern ice sheets melted at the end of the ice age, huge quantities of fresh water were added to the north Atlantic. This freshening reduced the density of the sea water and may have stopped it from sinking as it flowed northwards. That would have put a brake on the entire thermohaline circulation, depriving the north Atlantic region of heat and plunging it back into a short-lived ice age.
On time scales of decades the circulation of the oceans and atmosphere holds the key to climate. The tropics are warmer than the poles because they are struck squarely by the sun's rays; at the poles the rays arrive at an oblique angle and so the same amount of heat is distributed over a greater area. But, between them, atmospheric and ocean circulation constitute a worldwide heat distribution service which redresses some of this inequality.
What about the atmosphere? Air rises where it is warmed by heat radiated from the earth's surface; it sinks when it cools again and becomes denser. This is called convection. Crudely speaking, the earth is encircled by two bands of convective circulation in the tropics, in which warm air rises around the equator, moves outwards towards the respective poles, and then cools and sinks again at latitudes of 30oN and 30oS. The sinking air is then carried back towards the equator, where the circulations from the north and south converge in a region called the intertropical convergence zone. As the air rises and cools in this zone, the water vapour it contains, from evaporation of the equatorial oceans, condenses into droplets, creating towering cloud stacks which produce rain.
Interactions between the ocean and the atmosphere give rise to climatic phenomena such as El Ni????nd tropical hurricanes. El Ni????vents occur every two to ten years. It is still not known exactly how they arise; but their signature is a warming of surface water in the central and eastern equatorial Pacific Ocean. The effects of an El Ni????vent vary. Rainfall in central America, Brazil, Australia and Indonesia diminishes, causing drought, whereas rainfall is anomalously high in southeast Africa, Peru and Ecuador, bringing floods and landslides.
Predicting weather patterns-including freak events such as hurricanes-depends on our ability to simulate in computer models the interactions between the atmosphere and oceans. Predicting rainfall, for example, is a matter of estimating patterns of evaporation, convection, and transport of moist air. But even the best models cannot predict weather patterns for longer than about ten days ahead, because weather systems are intrinsically chaotic. Fluctuations too small to include (no matter how detailed you make your model) can turn out to have disproportionate consequences.
predicting climate
Nevertheless, climate trends on longer time scales can be predicted-because we are then asking broader questions. No one cares whether it will rain in Chelmsford 30 years from next Friday; but we do want to know whether northern Europe will be on average warmer, wetter, stormier, than at present. Different considerations come into play over different time scales: for example, predicting climate over the next century means that changes in the extent of the ice caps and mountain glaciers must be included in computer models. Ideally, these would also consider how ecosystems might respond to changes in temperature, rainfall and so on, and how this in turn affects the quantity of greenhouse gases released into or absorbed from the atmosphere by biological processes.
But this kind of climate modelling requires an understanding of feedback mechanisms which is at present only partial. We would like to believe that we now know all the main influences on climate over at least century-scale periods-but even that is by no means sure. This is why the predictions of future climate change over the 21st century made by the Intergovernmental Panel on Climate Change are uncertain. The world might warm by as little as 1.5 degrees, or by as much as 4 degrees. The mean sea level might rise by a foot or a yard. Legislating on climate change in the face of such unknowns is a tricky business-especially with vocal lobby groups eager to conflate uncertainty with ignorance. The current trend in formulating strategies on emissions of greenhouse gases seeks to include socio-economic factors and possible developments in future technologies as part of the climate models. But the fact is that the time scales of climate change are not those in which governments are accustomed to think, nor in which economists and technologists feel easy about making predictions. This is why it is vital to study the climates of times past, in order to see ahead more clearly.
