杏吧原创

WHEN global warming takes hold, who will suffer most? Conventional wisdom has it that the high latitudes and polar regions are the most vulnerable. The tropics are supposed to remain more or less unscathed. But this reassuring picture is fading in the face of accumulating evidence that global warming could, after all, wreak havoc in the tropics. Within a century, temperature increases may disrupt climate in a band that circles the globe and stretches from southern Europe in the north to South Africa in the south putting 350 million people at risk of famine (see Diagram).

Double trouble

Despite international efforts to limit emissions of carbon dioxide (This Week, 18 March), the amount of CO₂ in the atmosphere is expected to effectively double by the middle of next century. The full effects on the global climate will come later, and even if the amount of CO₂ in the atmosphere stabilises at double today鈥檚 levels the International Panel on Climatic Change (IPCC) estimates that by end of the 21st century the global temperature will have increased by between 1.5 掳C and 4.5 掳C. And if no replacements for fossil fuel are found, the temperatures could continue to escalate.

Until recently, most climate researchers were predicting that the tropics would escape virtually all the effects of this warming. This comforting view was based on a mass of evidence from prehistoric periods, which suggested that in the past temperatures in the tropics have remained stable while the rest of the world became warmer or cooler. Some of the most persuasive of this evidence came from the last ice age, which peaked some 20 000 years ago, when the world as a whole cooled by 4 or 5 掳C. As tiny changes in the Earth鈥檚 orbit around the Sun reduced the amount of solar radiation that arrived, vast ice sheets crept across much of North America and northern Europe. But according to an assessment published in 1981 by the Climate Mapping Project (CLIMAP), temperatures in the tropical ocean hardly changed during this time. The CLIMAP researchers studied the remains of microscopic shelled organisms known as foraminifera and diatoms which lay buried at the bottom of tropical oceans. They found that roughly the same species were present during the ice age as thrive there today, and from this they deduced that the ocean temperatures, too, were the same.

The more limited evidence available from around 3 million years ago 鈥� the last time the Earth was a few degrees warmer than today 鈥� led to similar conclusions. For several decades researchers have sifted through the few foram and diatom shells that have survived from this period, and like the CLIMAP scientists they deduced ocean temperatures from the distributions of the different foram species. They also examined the ratios of different isotopes of oxygen in the creatures鈥� shells, as this is a sensitive measure of the temperature of the water they lived in. Although the data are sparse, and the conclusions tentative, it seemed that temperatures in the equatorial regions were much the same as they are today, while the high latitudes were up to 10 掳C hotter then than now.

Another question is what caused global temperatures to be so warm then. Could it be that CO₂ was responsible? In the 1980s, Bob Berner from Yale University developed a method for modelling past levels of CO₂ in the atmosphere. His method used the knowledge that the process of creating new ocean floor releases CO₂ to the atmosphere, and that weathering of the land reduces atmospheric CO₂ levels. Three million years ago, the ocean floor was being created faster than it is today, and less of the Earth鈥檚 surface was covered by land, so Berner deduced that there must have been more CO₂ in the atmosphere then than now. Most scientists put two and two together and concluded that the CO₂ was the cause of the warmer temperatures 鈥� just as it is expected to be in the future.

There was even an explanation for why tropical temperatures should have remained stable. V. Ramanathan and colleagues at the National Center for Atmospheric Research in Boulder Colorado, reasoned that the extra water vapour that would be released by any warming of the tropical seas could have two competing effects on climate. Clouds cool the Earth by reflecting sunlight back out to space; the thicker they are, the more sunlight they reflect. But water in clouds, or water vapour free in the atmosphere, can cause warming by absorbing radiation from the Earth and trapping it, just like other greenhouse gases such as CO₂. For thin clouds the greenhouse effect tends to dominate, but at a certain cloud thickness the reflecting, cooling effect takes over.

Cirrus clouds up to 16 kilometres above the tropical ocean tend to be thin, so they mainly act to warm the Earth. Ramanathan reasoned that if the tropical seas warmed, the extra heat energy would allow warm, moist air to rise higher into the atmosphere. This extra convection would cause the high-altitude cirrus clouds to thicken to the point where the reflection mode would begin to overtake the greenhouse mode, producing a net cooling rather than warming. In effect, this mechanism would act as a thermostat, preventing any large temperature swings. Because convection is much stronger in the tropics than anywhere else, this thermostat effect should be restricted mainly to tropical latitudes.

Tropical turmoil

Ramanathan鈥檚 ideas have recently started to fall from favour as evidence has come in showing that warm sea surface temperatures do not tend to coincide with thick high-level clouds. But at the time, everything seemed to point to stable tropical climates. This picture was perhaps all the more persuasive because it fitted well with our day-to-day perspective. Temperatures in the tropics vary little from season to season and from one day to the next. At higher latitudes, wild temperature swings are common.

But while observational studies were coming up with reassurance for the tropics, computer models were telling a different story. In particular, the computer number-crunchers suggested that as the climate started to warm, the oceans would release more water vapour into all levels of the atmosphere. Rather than acting to thicken clouds and so reflect sunlight, this additional vapour would spread itself widely and act predominantly as a greenhouse gas. This would further accentuate the warming at all latitudes, including the tropics. When the level of carbon dioxide in the model atmosphere was doubled, numerical models of the climate showed a significant tropical warming 鈥� anything from 1掳C to 4掳C. Because these models were built on rather shaky foundations 鈥� no one could be sure of the precise mechanisms associated with water vapour transport and cloud generation, for example 鈥� many researchers assumed that they must be wrong. Those scientists already convinced by the observational data that the tropics would not warm, suspected that the models were flawed and were coming up with the wrong answer.

Now the balance is swinging the other way as observations of the surface air temperature come in. Data from NASA鈥檚 Goddard Institute for Space Studies (GISS) in New York and the British Meteorological Office show that tropical temperatures are on the increase. The 1980s were the warmest decade on record, and this was primarily because temperatures rose in the tropics. Globally, the years 1981 to 1990 were close to 0.5 掳C warmer than a century earlier, and 0.3 掳C warmer than the 1951 to 1980 average. The tropical ocean temperatures were between 0.25 and 0.75 掳C warmer compared with 1951 to 1980. Since 1976, the eastern tropical Pacific has been more than 0.5 掳C warmer than in the previous decades. No one knows whether this change is due to global greenhouse warming, but whatever the cause, it is certain that tropical climates are not quite as unchanging as we had thought.

Ice in Hawaii

Meanwhile, the CLIMAP results are being challenged by more recent attempts to deduce conditions during the ice age. Observations at a wide range of tropical locations have revealed that glaciers crept down the mountains by some 900 metres. There is evidence of glaciation at Mauna Kea in Hawaii, for instance, where today the freezing level is hundreds of metres above the top of the mountain. Vegetation zones also crept lower as species accustomed to warmer climates migrated to lower altitudes. Evidence of past glaciation and changes in the vegetation imply that the land in the tropics chilled by 5 掳C at an altitude of 3 kilometres during the last ice age.

Last year Tom Guilderson and Rick Fairbanks of Lamont-Doherty Earth Observatory in New York State returned to the question of sea surface temperatures during the ice age. In corals, the ratios of strontium to calcium and ratios of the isotopes of oxygen are both sensitive indicators of the temperature of the sea in which the organisms lived. So Guilderson and Fairbanks sifted through coral remains from the tropical Atlantic until they found samples from the right period, then measured these ratios. Their results suggest much greater cooling than CLIMAP more in line with the land figures.

Around the same time, Martin Stute and his colleagues, also at Lamont-Doherty Earth Observatory, looked at the amounts of the noble gases neon, argon, krypton and xenon dissolved in groundwater in the southwestern US and eastern Brazil 鈥� water that started to filter down from the surface during the last ice age. Because these gases become less soluble as the temperature rises, the quantity dissolved in the water reflects the temperature of the water as it disappeared underground. Stute and colleagues found large concentrations of noble gases, corresponding to substantial cooling of around 5 掳C and providing further evidence that the ice age affected the tropics as well as higher latitudes.

On top of these new doubts about the fate of the tropics during the ice age, questions are being raised about the stability of tropical temperature during the warmer spells too. Leaving aside suspicions about reliability of data from millions of years back, one key question is whether the climate鈥檚 behaviour during past periods of warming is relevant to the future CO₂-warmed world. For one thing, it seems that CO₂ levels some 3 million years ago may not have been as high as was thought.

Greg Rau of the University of California at Santa Cruz and Maureen Raymo of the Massachusetts Institute of Technology have used measurements of carbon isotope ratios to deduce how much CO₂ there was in the ancient atmosphere. Rau had already shown that the ratio of heavy to light isotopes of carbon in organic matter floating in the ocean seems to depend on the level of CO₂ dissolved in the water, which in turn reflects the atmospheric CO₂ level. From the carbon isotope ratios in organic matter buried in ocean sediments, the researchers concluded that the CO₂ concentration in the atmosphere 3 million years ago was probably not much higher than it is now.

But what led to higher global temperatures in the past if not greenhouse warming caused by CO₂? A more vigorous ocean circulation than today may have been the culprit. In 1991 Mark Chandler and I published the results of modelling studies that showed how vigorous poleward ocean currents could lead to a climate that is much warmer than today鈥檚 at high latitudes but unchanged at the tropics. As the oceans transported more heat to high latitudes the polar ice caps and regions of floating sea ice would melt slightly. This would reduce the amount of sunlight that the white ice reflects back to space, allowing more energy in to warm the Earth. In other words, these stronger currents would not only redistribute the warmth, they would also make the planet as a whole warmer. In 1992, Raymo reported chemical signs from sediments on the North Atlantic ocean floor which support the idea that poleward currents were unusually active 3 million years ago.

The year before, I took part in several studies analysing data on water vapour from the atmosphere gathered by the SAGE II (Stratospheric Aerosol and Gas Experiment) satellite. These found that there is more water vapour in the atmosphere over the tropical western Pacific Ocean than over the cooler tropical eastern Pacific, and that there is generally more in summer than in winter. In other words, in warmer conditions, water vapour increases in the atmosphere at all levels, exactly as computer models predicted.

But this extra water vapour does not seem to make clouds thicker. In fact, there is even evidence that it might have the opposite effect. A study using satellite observations from the International Satellite Cloud Climatology Project, published in 1993 by George Tselioudis and colleagues from the GISS, shows that everywhere except the polar regions, low-level clouds actually become thinner as the temperature increases. One possible reason is that the extra water vapour might increase the size of the water droplets in the clouds, and thus make them more likely to precipitate as rain. Low-level clouds are usually very thick, so they act to cool the climate by reflecting solar radiation back to space. If they become thinner and less reflective as the climate warms, more sunlight will get in, and low latitudes will warm even more.

It is results such as these that are leading many researchers to abandon the idea that the tropics are not affected by global climate change. Instead they are coming to the conclusion that the computer models may have been right all along about tropical warming. In fact, the models may have underestimated it. The consequences of a rise in temperature in the tropics could be devastating. As land and air temperatures increase, the atmosphere can hold more moisture. In tropical regions, an increase of 4 掳C in air temperature means that around 30 per cent more moisture can be evaporated from the ground. The oceans warm more slowly, because their heat capacity is much greater, so the increase in evaporation from the warming ocean is much less. Computer models show that a 4 掳C rise in global air temperature would lead to a 12 per cent increase in evaporation. The oceans provide the water for most of the planet鈥檚 rain, so this leads to a similar increase in global precipitation. However, a 12 per cent increase in rainfall would not be enough to make good the attempt by the land to lose 30 per cent more of its water by evaporation, so the land would dry out, especially at low and subtropical latitudes. In 1990, I took part in several studies that modelled climate change during the next century. These concluded that a 2 掳C increase in global temperature would bring frequent severe droughts to tropical and subtropical locations where such droughts now occur only 5 per cent of the time. A 4 掳C warming would bring frequent droughts to middle latitudes as well, and arid climates would extend about 35掳 north and south of the equator.

Food shortages

Even though increased CO₂ levels can fertilise crops, the net effect of increased CO₂ and increased temperature would cause a 10 per cent decline in the production of wheat, maize, soya beans and rice in developing countries, according to a study published last year in Nature by Cynthia Rosenzweig from the GISS and Martin Parry from Oxford University. Estimates for the numbers of people who in 2060 will be affected by famine due to climate change range from 50 million to as many as 350 million. This is in addition to a baseline population at risk of hunger, which will already have been swelled by population growth to some 640 million.

Significant tropical warming would also bring increased hazards from severe storms and hurricanes, which feed on the energy unleashed as water vapour from warm oceans condenses into rain. A multitude of factors influence hurricane development, including temperature, wind and moisture, which make it impossible to predict how climate change will affect the pattern of hurricanes. But since tropical storms appear to form only at temperatures above 26 掳C, it seems likely that warmer seas will fuel more of them and that they will be more intense.

Another unknown is the effect of global warming on El Ni帽o, the erratic reversal of the warm currents in the Pacific Ocean. El Ni帽o events can cause climate chaos round the world, and are unpredictable at the best of times (see 鈥淓l Ni帽o goes critical鈥�, New 杏吧原创, 4 February). This year El Ni帽o appeared for a record fifth year running, and the suspicion is creeping in that climate is already changing. Neither day-to-day experience, nor the prehistoric climate record, nor even our best climate models can tell us what the outcome will be. So it is vital that we keep track of tropical temperatures and watch how they change. Surface temperatures are being monitored around the world, and satellites can now provide a global picture of temperature change at different levels in the atmosphere. The planet is likely to be slow to warm, but once warmed is likely to be difficult to cool. It would be wise to bear this in mind when deciding how we should curb our emissions of greenhouse gases. By the time the alarm bells are clearly heard, it may well be too late for rescue.