sign in sign up
<<
Home , Little Ice Age

Rapid

New evidence shows that earth’s climate can change dramatically in only a decade. Could greenhouse gases flip that switch?

Kendrick Taylor

ver the course of geologic history, in climate happened slowly and would The dust, chemicals and gases in the Othe earth’s environment has been never affect me. After all, a single cli- reflect the environment along the far from static. Indeed, 600 million mate cycle that includes an and water’s journey from distant sources to years ago the atmosphere lacked suffi- a warm period lasts 150,000 years and the . They record how cold it cient oxygen to support animal life. is controlled by gradually changing or- was, how much snow fell in a year, More recently, as shown by sediments bital parameters of the earth. It did not what the concentration of atmospheric recording conditions over the past seem possible that climate cycles that gases was and what the atmospheric 500,000 years, the planet’s climate var- lasted so long could change perceptibly circulation patterns were. ied between at least two different states. during my lifetime. Even greenhouse- We can identify annual layers in the The record from the past 150,000 induced climate changes are normally ice because the concentration of sea years is particularly well preserved, of- predicted to happen gradually over salts, nitrate and mineral dust and the fering details about these repeated cli- several generations, allowing an oppor- gas content in snow are differ- mate changes. Between about 131,000 tunity for society to adapt. ent than in summer snow. We count and 114,000 years ago there was a My attitude changed profoundly the annual layers to determine the age warm period like today’s climate, re- while I was working on a project fund- of the ice, and by measuring the thick- ferred to in as the or ed by the National Science Foundation ness of the annual layers we can deter- globally as Marine Isotope Stage 5e. to develop a climate record for the past mine how much snow fell each year. This was followed by the Wisconsin ice 110,000 years. By examining ice cores The gas trapped between age, which ended about 12,000 years from , my colleagues and I offers a sample of the ancient atmos- ago when the current relatively warm determined that climate changes large phere, and we can use it to determine period began. enough to have extensive impacts on what the concentrations of greenhouse Although the past half-million years our society have occurred in less than gases such as carbon dioxide and constitutes the current-events period in 10 years. Now I know that our climate methane were long before human be- geologic time, on a human time scale could change significantly in my life- ings measured the atmosphere directly. the events I just described are in the dis- time. We are still a long way from be- General patterns of atmospheric circu- tant past. Because their time scales are ing able to predict such a change, but lation can be reconstructed by using so long, I used to believe that changes we are getting closer to understanding tracers such as soluble chemicals (for how it might take place. A pressing example, nitrate, ammonium, sodium concern is whether anthropogenic and calcium) and rare earth elements Kendrick Taylor is a research professor with the changes to our planet’s atmosphere in insoluble dust particles to determine Water Resources Center of the Desert Research might perturb the climate’s stability. how wind moved air and dust from Institute. He received his B.S. in geophysics from the Colorado School of Mines, his M.S. in geo- the source regions for these com- physics from the University of Wisconsin, Ice, the Museum of Climate pounds to the drilling site. Madison, and his Ph.D. in hydrogeology/hydrolo- One can learn a lot about what controls gy from the University of Nevada, Reno. Taylor climate by studying glacial ice. When Ice as Thermometer currently is the Chief Scientist for the National snow falls, it collects insoluble dust Air temperature is naturally of primary Science Foundation’s WAISCORES program, particles, soluble compounds and the interest when we talk about climate, which is recovering the first of two deep Antarctic water in the snow itself. In some places and fortunately we have three ways to ice cores. Although his major recent research inter- more snow falls in a year than melts or determine what it was in the past. est has been in the use of polar ice cores to develop sublimates away. Annual layers of First, we can measure the isotopic com- paleoclimatic records for use in improving under- snow pile up, with atmospheric gases position of the oxygen and hydrogen standing of the earth’s climate system, he has also led projects assessing groundwater supplies in the filling the open pores between snow in the ice. When water vapor in clouds U.S. and abroad. Address: Water Resources crystals. The weight of accumulating condenses, the ratio of oxygen-18 to Center of the Desert Research Institute, 2215 snow compresses the pores in the oxygen-16 and the hydrogen-2/hydro- Raggio Parkway, Reno, NV 89512-1095. Internet: snow below, turning the snow into ice gen-1 ratio are affected by the ambient [email protected] and trapping the atmospheric gases. temperature; the colder the cloud, the © 1999 Sigma Xi, The Scientific Research Society. Reproduction with permission only. Contact [email protected]. 320 American Scientist, Volume 87 Figure 1. Although earth’s climate has long been known to experience warm and cold phases, earth scientists have assumed that changes in climate take place only very slowly. Surprisingly, however, recent ice-core studies in Greenland and , along with corroborating evi- dence from ocean-sediment cores, have established that as recently as 11,650 years ago, a change in average temperature as great as 10 degrees Celsius may have occurred within 20 years. This realization has led to a new understanding of earth’s climate as existing in three phases, which are relatively stable until perturbed to the point of a precipitous change. The author and other paleoclimate investigators are con- cerned that anthropogenic additions of carbon dioxide to earth’s atmosphere may alter the hydrologic cycle of the North Atlantic sufficiently to trigger a rapid switch from the present warm climate to a cold one. In this snow pit at Greenland’s summit, light from an uncovered, adja- cent snow pit shines through the layers of snow. The thick, dark bands of snow were deposited during the winter; the sequences of thin, lighter bands were deposited during the summer. These visual differences occur because the summer and winter snow have a different den- sity and crystal shape. (Unless otherwise noted, photographs by the author.) lower the ratio. Measuring how the ra- at the time when the ice overburden surface temperature changed during the tios of these isotopes changes along an pressure closes the pores and traps the past several hundred thousand years. ice core gives us a good idea how the nitrogen gas in the ice. Variations in the air temperature changed over time. isotopic composition of the nitrogen The Greenland Report The second way to determine prehis- along a core show when and by how In Greenland, annual ice layers are toric temperatures is to measure the iso- much the surface temperature changed. stacked up like thousands of annual topic composition of the nitrogen gas Finally, because of the large thermal weather reports. In 1982, a European trapped in the ice. At depths between inertia of an ice sheet, the current tem- and American team made the first at- about 5 and 50 meters in an ice sheet, perature distribution in an ice sheet is tempt to read that record, by recover- air can move in interconnected pores strongly influenced by what the surface ing an ice core from southern Green- but is sheltered from mixing by the temperature was in the past. The land. Measurements on the ice core wind. Nitrogen-15 slowly moves to- physics is similar to cooking a large indicated that about 11,700 years ago ward colder locations, and nitrogen-14 frozen turkey. If we move the turkey di- the climate of the North Atlantic region slowly moves toward warmer loca- rectly from the freezer into the oven, the changed from a dry and cold ice age to tions. This process creates a near-sur- outside of the turkey will be done before the current warmer and wetter Holo- face gradient in the nitrogen-15/nitro- the inside even defrosts. By modeling cene. Altogether it took 1,500 years for gen-14 ratio that depends on the the current thermal state of the turkey, or the climate transition to be complete near-surface temperature gradient. The an ice sheet, we can determine the histo- and a few thousand more years to melt resulting isotopic composition of the ni- ry of the turkey’s, or ice sheet’s, surface most of the ice, but the surprise was trogen trapped in the ice depends on temperature. The physics of these three that most of the transition occurred in the difference between the surface tem- approaches is well understood; together only 40 years. This was only one perature and the temperature at depth they allow us to reconstruct how the record, and it came from a single 10- © 1999 Sigma Xi, The Scientific Research Society. Reproduction with permission only. Contact [email protected]. 1999 July–August 321 centimeter-diameter ice core. Still, this finding was impossible to ignore and too puzzling to comprehend. In 1993, Americans and Europeans led by Paul Mayewski of the University of New Hampshire and Bernhard Stauffer of the University of Bern in Switzerland finished recovering two new ice cores from the summit of the Greenland ice sheet. More than 40 uni- versity and national laboratories partic- ipated in the projects. We shared sam- ples, spent time in one another’s labs, replicated one another’s results, pro- posed ideas, tore them apart and then jointly proposed better ones. One of the justifications for these new cores, locat- ed 30 kilometers apart, was to verify and learn more about the 40-year change in climate, an event observed in both cores. The records stored in these cores were more detailed than before and showed that within a 20-year peri- od at the summit of Greenland, where ice is thickest, the amount of snow de- posited each year doubled, average an- Figure 2. Different equipment is used for cor- nual surface temperature increased by ing ice depending on the depth of the sample 5 to 10 degrees Celsius and wind to be taken. At left is the rig for collecting speeds increased. The same ice cores samples up to 3,000 meters in depth. At right also showed that the spatial extent of is an ice-core drill for depths of less than 200 decreased, atmospheric-circula- meters. Analysis of ice cores for isotopes, such as oxygen-18 and deuterium, trapped gases, tion patterns changed, and the size of dust particles and thickness of layering can the world’s wetlands increased. Many reveal much about the climate of the past. of these shifts in parameters, including at least a 4-degree Celsius increase in the average annual air temperature, happened in less than 10 years. These changes were not restricted to Green- land; the global nature of many of these ice-core records showed that low-lati- tude, continental-scale regions rapidly got warmer and wetter. The most dra- matic change occurred 11,700 years ago. But we also found comparable anomalies every several thousand years during the Wisconsin ice age (see Figure 6). Further, Antarctic ice cores also show comparable climate transi- tions at these times.

Climate, from the Bottom Down One can also learn a lot about what controls climate by studying sediments on the ocean floor. These sediments contain the decayed remains of ocean organisms and inorganic material from the erosion of rocks. Ocean organisms assimilate chemical compounds from the water as they grow, and the com- pounds they incorporate are partially Figure 3. Richard Alley of Pennsylvania State University and Mark Meier of the University determined by the environment in of Colorado count annual layers (light and dark bands about a centimeter apart) in an ice which they live. Thus the decayed re- core. (Photograph courtesy of Ken Abbott/National Ice Core Laboratory.) mains of the organisms that fall to the © 1999 Sigma Xi, The Scientific Research Society. Reproduction 322 American Scientist, Volume 87 with permission only. Contact [email protected]. ocean floor contain a record of what suring the ratio of different types of chemical compounds were available alkenones we can determine what the and the temperature of the water in surface water temperature was when which they lived. the phytoplankton were living. For example, consider an ocean-sed- By collecting cores of the ocean sedi- iment core collected at Bermuda Rise, a ments at different locations, we can de- place where ocean currents deposit a termine a lot about how the ocean cir- lot of sediment. The oxygen-18/oxy- culated water and heat in the past. The gen-16 ratio of seawater varies through rapid climate changes recorded in the time depending on how much water ice cores encouraged a search for ocean is locked in ice sheets and how much sediment records with high time reso- water is in the ocean (see Figure 7). The lution. In the past few years locations near surface–dwelling foraminiferan have been identified in the ocean Globigerinoides ruber uses seawater to where sediment accumulates rapidly, make its shell. By measuring the oxy- and the sediment cores from these lo- gen isotopic composition of the shells cations have comparable time resolu- recovered from an ocean core, we can tion to the ice cores. Coring projects off determine how much water was the coast of Bermuda by Konrad locked up in ice sheets when the Hughen, Julian Sachs and Scott Lehman foraminiferan was living. Likewise, the with the University of Colorado, in con- bottom-dwelling foraminiferan Nutal- junction with Lloyd Keigwin of Woods lides umbonifera incorporates cadmium Hole Oceanographic Institution and Ed and calcium in its shell. By measuring Boyle of the Massachusetts Institute of the ratio of cadmium to calcium in the Technology, found the same rapid shells recovered from an ocean core, changes in climate as were recorded in we can tell where the bottom water the ice cores. Other groups have found came from when the foraminiferan similar records near Santa Barbara, was living. High values of the cadmi- California and off the coast of India. um-to-calcium ratio indicate that the Paleoclimatic evidence worldwide water near the bottom came to the shows that a global change in climate Bermuda Rise from the south, whereas took place 11,700 years ago, and in the Figure 4. Ocean-bottom sediments are col- lected using a 10-meter-long piston core bar- a low ratio indicates that the bottom North Atlantic a large part of the rel. Analysis of the chemical makeup of the water came from the north. change took less than 20 years. It was a shells of , the mineralogy of the Ocean sediments also contain ground- few thousand years before the comple- sediments, and the types of organic com- up rock, which is transported and de- tion of the transition from ice age to pounds made and deposited by phytoplank- posited by ocean currents, just as wind warm period; still, in just a 20-year pe- ton indicate what the seawater temperature, carries airborne dust to be deposited riod the climate of a large part of the salinity and source areas were when the sed- on ice sheets. The mineralogy of the earth changed significantly. There was iments were deposited. (Photograph cour- ground-up rock can be used to identify no warning. A threshold was crossed, tesy of Anne Jennings.) where it came from. For example, a and the climate in much of the world layer of hematite-rich sediments in shifted abruptly from cold to warm. ocean cores near Bermuda indicates This was not a small perturbation; our that ocean currents were transporting civilization has never experienced a cli- material from the east coast of Canada mate change of this magnitude or to Bermuda when the sediments in the speed. To get an idea of what hap- layer were deposited. pened, imagine that over a 20-year pe- To determine what the temperature riod the weather at your home became of the ocean surface was in the past we that typical of a place 400 to 600 miles can use organic compounds made by farther south. What might be the phytoplankton. Phytoplankton live mechanism for so rapid and large a cli- near the ocean surface where there is mate change? light for photosynthesis. Some phyto- plankton produce compounds know as Figure 5. Ocean core shows a rapid change alkenones, which are straight chains of from sediments transported from coastal carbon atoms. Along these chains of regions by (light-colored material) carbon there can be two or three double to sediments composed of the remains of bonds. The number of double bonds organisms that lived in the ocean (dark-col- ored material). Melting icebergs at this time depends on the water temperature. The also released large amounts of freshwater at double bonds are thought to keep the the ocean surface, which reduced seawater cell membrane pliable in cold water. density and slowed convection-based cur- When the phytoplankton die, the alke- rents, precipitating a major and rapid climate nones fall to the bottom and become in- change. (Photograph courtesy of Bedford corporated into the sediment. By mea- Institute of Oceanography.) © 1999 Sigma Xi, The Scientific Research Society. Reproduction with permission only. Contact [email protected]. 1999 July–August 323 Holocene Wisconsin ice age Eemian MIS5e North Atlantic, and a large volume of Ð400 surface water and heat is drawn from the tropics to replace the sinking North Atlantic water. The heat carried north Ð450 by the northward-moving surface wa- ter warms eastern , the North Atlantic and most of Europe. deuterium (per mil) Ð500 The second mode of ocean circulation Industrial Revolution occurs when surface water sinks in 350 only one area of the North Atlantic. Less surface water sinks to the bottom, so smaller amounts of warm surface 250 water and heat are drawn north to re- (parts per place the sinking water. This mode was 2 in place during the warmest times of CO

million by volume) the Wisconsin ice age, when climate 150 was only slightly colder than current 0 20 40 60 80 100 120 140 160 conditions. In the coldest mode, no wa- thousands of years before present ter sinks in the North Atlantic; hence Figure 6. Ice-core record from Vostok, Antarctica provides a record of climate change over the no warm water is drawn north. This past 500,000 years. Shown at top is the deuterium record (a for temperature) for the was the condition during the coldest past 160,000 years, which contains the last glacial- cycle. The climate transitions portions of the Wisconsin ice age. are not as abrupt in central portions of Antarctica as they are in more coastal locations. The Each of these modes of ocean circula- atmospheric CO2 record clearly shows how human activity has increased both the concen- tion is associated with a small range of tration of CO2 and the rate of increase of CO2 to unprecedented levels. (Deuterium data from prevailing environmental conditions. J. Jouzel, Laboratoire de Modélisation du Climat et de l’Environnement, France. CO2 data, from direct observations and three different ice cores, come from C. Keeling, Scripp’s Weather anomalies such as 10-year-long Institution of Oceanography; A. Indermuhle and A. Neftal, University of Bern; and J. or wet periods, as significant Barnola et al., Laboratoire de Glaciologie et Géophysique de l’ Environnement, France.) as they may seem to human affairs, are a reflection of the small range of envi- Climate’s Control Mechanism freshwater, which makes the surface ronmental conditions associated with a Like the atmosphere, the oceans are far water more buoyant than the water un- single mode of ocean circulation. If, from static. Currents, of which the derneath it. But in the North Atlantic, however, environmental conditions are Caribbean-Atlantic Gulf Stream is just the combination of cold temperatures externally forced to be inconsistent with a small part, continually exchange wa- and evaporation makes the water dense the existing mode of ocean circulation, ter among all the oceans and between again and it sinks. the circulation will switch to a mode the surface and the depths. For the sake Wally Broecker of Columbia Univer- that is more consistent with those envi- of convenience, we shall start this jour- sity likens this circulation pattern to a ronmental conditions. An example of ney in the Gulf Stream, where water long conveyor belt that moves water, this forcing would be a change in the moves northward along the East Coast salt and heat. He was among the first to amount of solar heat reaching, and re- of the U.S. toward . Along the recognize that alterations in the path of tained at, the earth’s surface. Such a way, the water exchanges heat with the the ocean conveyor belt would change change could be the result of alterations air, warming the air and cooling the wa- climate in much the same way that either in solar output or in the way the ter in the process. Water evaporates turning off the furnace fan changes the atmosphere regulates the exchange of from the surface and leaves behind dis- temperature distribution in a house. He heat between the earth’s surface and solved salt. The combination of chilling proposed that the large oscillations in space. The transitions between different and evaporation makes surface water climate observed in the geologic record modes of ocean circulation are abrupt. denser as it moves north. In the vicinity were caused by different patterns of The ocean-sediment cores and ice cores of Iceland, the surface water becomes ocean circulation. tell us that they frequently take only denser than the water below it and several decades or less. sinks. This dense, cold water then Three Climate Modes Numerical models of ocean circula- moves south along the bottom of the Because the oceans must abide by the tion developed by Thomas Stocker of Atlantic, around the Horn of Africa constraints of geography and the laws the University of Bern and Syukuro and, still near the bottom, continues to of physics, there are only a few pat- Manabe of Princeton University show the North Pacific, where it upwells to terns in which the oceans can circulate. that each circulation mode is stable for the surface. Surface water in the North At a recent conference organized by a particular range of environmental Pacific makes room for the upwelling Robert Webb of NOAA and Peter conditions. For example, if the dis- bottom water by moving south, pass- Clark of Oregon State University, cli- charge of a river changes, altering the ing between Asia and Australia and fi- mate scientists identified three modes density of the surface water in the ad- nally catching the tail of the circulation of ocean circulation, each of which is joining ocean, the ocean-circulation pattern at the beginning of the Gulf associated with a different climate. The pattern will change only if it is unstable Stream in the Atlantic off Central Amer- current mode produces the warmest under this new set of conditions. As ica (see Figure 8). For most of its journey, conditions in the North Atlantic. Sur- long as the climate system stays within the surface water collects heat and face water sinks in two regions of the the stable-mode range, river discharge © 1999 Sigma Xi, The Scientific Research Society. Reproduction with permission only. Contact [email protected]. 324 American Scientist, Volume 87 thousands of years before present 0202535 10 15 035 40 Ð30

Ð35

Ð40 0 (per mil) 18 δ Ð45

Ð50

years before present and greenhouse-gas concentration can vary without having much influence on climate.

Ð32 11,600 11,620 11,640 11,660 11,680 11,700 11,720 11,740 11,760 11,780 Stefan Rahmstorf of the Potsdam In- Ð34 stitute for Climate Impact Research has Ð36 used numerical models to show how Ð38 surprisingly sensitive ocean circulation Ð40 can be to changes in freshwater dis- 0 (per mil)

18 Ð42 charge. His numerical models show δ Ð44 that if the climate system is near the Ð46 threshold between stable modes, a small change in the amount of freshwa- 0.30 ter entering the North Atlantic will 0.25 force a large and rapid shift to a differ- ent ocean-circulation pattern. Like a 0.20 coin on edge, which topples with only a 0.15 breath of air, an unstable pattern quick- 0.10 ly assumes a new position where it be- ice accumulation (meters per year) 0.05 comes quite stable. The climate changes 0.00 recorded by the ice and ocean-sediment cores appear to have taken place when some crucial threshold was crossed, re- 220,000 sulting in large and rapid switches—in geologic time, like the flip of a switch— 212,000 in ocean circulation. Unfortunately, no one knows yet 112,000 what caused this switch to flip. We (per milliliter) dust particles know of external forcing mechanisms, 12,000 but their time periods do not match the 1,674 1,676 1,678 1,680 1,682 record. For example, the distribution of depth (meters) reaching the earth varies according to the relative positions of Figure 7. Ice-core record from the Greenland Ice Sheet Project Two shows the oxygen-18 concentration for the past 40,000 years (top). Although the location where the water evapo- the and earth. These variations, rated from the ocean affects the isotopic concentration, higher values generally indicate called , have peri- warmer surface temperatures at the drill site. During warmer periods (red), the ocean circu- ods of tens of thousands of years and lation was similar to what it is today. Colder conditions (blue) prevailed when surface are thus too slow to explain the rapid waters traveling north in the Gulf Stream no longer sank north of Iceland, disrupting heat changes seen every couple of thousand transport out of the tropics. The coldest periods were when no surface waters sank in the years during the Wisconsin ice age. North Atlantic (dots). (Oxygen-isotope data from P. Grootes of Christian Albrechts The Milankovitch cycles define the big University, Germany.) The lower three graphs show details of three climatic parameters for picture and determine when changes the most recent rapid climate change, about 11,650 years ago. At that time, climate became could occur, but some smaller, quick- warmer (oxygen-18 concentration), wetter (annual ice accumulation) and less dusty (dust- er-acting mechanism triggered the particle concentration). The resolution for absolute age is ±200 years, but the relative ages are accurate within a few years. This change of 5 to 10 degrees Celsius within 20 years is more frequent switches during the unprecedented in recorded history and would be very disruptive if it were to repeat. (Age Wisconsin. data from D. Meese, CRREL; oxygen-18 data from J. White, University of Colorado; ice The leading idea, which has been accumulation data from R. Alley, Pennsylvania State University; dust concentration data simulated in computer models, is that from G. Zielinski, University of New Hampshire.) increased discharges of freshwater © 1999 Sigma Xi, The Scientific Research Society. Reproduction with permission only. Contact [email protected]. 1999 July–August 325 circulation would abruptly switch to a new stable mode. This would be more than just a switch in ocean circulation; it would be a switch in the way tropical heat is transported to the North Atlantic. At the least, Northern Europe and Scandi- navia would be 2 to 5 degrees colder on average than they are now, and the amount of precipitation would decrease dramatically. It would not necessarily be a rapid return to an ice age, but it might be a start in that direction. The orbital parameters of the earth are such that we are due for another ice age, and a cooling in the north Atlantic at a time Summit high time-resolution ocean core Vostok when orbital parameters favor a return warm, fresh, less dense, shallow water cold, salty, dense, deep water to a much colder climate could be the Figure 8. Ocean currents continually exchange water throughout all the oceans and from the trigger that initiates such a change. surface to the depths, much like a liquid conveyor belt. In the current mode, warm water A switch in climate from a warm pe- (red) flows northward along the East Coast of the U.S. toward Iceland. The warm water riod (like the current Holocene epoch) exchanges heat with the cooler air, becoming cooler and saltier. Near Iceland, the water to an ice age has happened before. becomes more dense (cool and salty) than the water below and sinks, flowing southward Ocean and lake cores tell us that the along the floor of the Atlantic. From there it continues around the Horn of Africa and, still warm Eemian period from about near the bottom, flows into the North Pacific, where it upwells to the surface. Surface water 131,000 to 114,000 years ago—when the in the North Pacific makes room for the upwelling bottom water by moving south, passing distribution of ice sheets was similar to between Asia and Australia and finally catching the tail of the circulation pattern at the what it is today—switched to the Wis- beginning of the Gulf Stream in the Atlantic off Central America. There are also places consin ice age in no more than 400 around Antarctica where surface water sinks to great depths; these are not shown because their location and magnitude are less well established than those in the North Atlantic. Two years, the minimum time resolution of other modes of ocean circulation exist, which correspond to different intensities of ice-age the record from these ancient sediments. climate. During most of the the Wisconsin ice age, surface water sank only south of Iceland; Unfortunately, we have yet to recover there was no sinking north of Iceland as shown here. This reduced the amount of water that an ice core that shows in sharp detail was circulated by the conveyor. During the coldest times of the last ice age (indicated by the how the Eemian Period ended. This is dots in Figure 7), the surface water did not sink in the North Atlantic, and only a small old ice. It is difficult to find a place amount of water, if any at all, was circulated by the conveyor. where it snowed enough to produce a high time-resolution record but not so glacial ice and river runoff into the Tampering with Our Stable Mode? much as to smear the record against the North Atlantic reduced the density of Human beings have made major mod- bedrock. An international project, led by the surface water enough that it could ifications to the earth’s environment in Claus Hammer of the University of not sink. This slowed the ocean con- little more than a century, increasing , has identified the most veyor, forcing it to switch to another the concentration of carbon dioxide in likely location in Greenland for this ice circulation pattern. Other emerging the atmosphere to its highest level in to be found and is collecting a core. concepts place the source of the dis- 260,000 years. Numerical models can Many ice cores tell us that 8,200 ruption of the conveyor in the tropical be used to estimate what will happen years ago the climate approached ice Pacific. Variability in the sun’s output when anthropogenic increases in the age conditions for a 400-year period be- is another possible cause of the climate atmospheric concentration of gases fore returning to conditions similar to variations, but the record of solar out- such as carbon dioxide and methane today. This excursion was most likely put is not good enough to adequately block heat from leaving the earth. The caused by the one-time draining of investigate this idea. Furthermore, the increased concentration of these gases lakes left behind by the melting of the dynamics of ocean circulation around acts like a greenhouse, and the average Canadian ice sheets. This change in Antarctica are too poorly understood temperature of the earth gets warmer. freshwater flux to the oceans was large to completely exclude the possibility But the numerical models of Stocker, but not that much different from what that they may play a role. Whatever Rahmstorf and others suggest there greenhouse-induced changes may pro- the cause may be, it is worrisome that may be surprises in the greenhouse. duce in the future. The fact that it took the phenomenon that has repeatedly When the warms place when climate conditions were triggered major changes in ocean cir- the earth, it accelerates the hydrologic similar to today demonstrates that large culation and the earth’s climate is so cycle, more water moves around in the and rapid climate switches do not hap- subtle that we have not been able to atmosphere, and rainfall increases in pen exclusively when there are exten- identify it. This emphasizes how large many places. Some models suggest sive ice sheets. It changes in the interaction of the that this will slowly decrease the salin- is ironic that greenhouse warming may oceans, atmosphere and ice sheets have ity of the North Atlantic, making the lead to rapid cooling in eastern North been triggered by small perturbations surface water less dense. Were a critical America, Europe and Scandinavia, and of the environment. density threshold to be crossed, ocean it is possible that altered ocean circula- © 1999 Sigma Xi, The Scientific Research Society. Reproduction with permission only. Contact [email protected]. 326 American Scientist, Volume 87 quickly enough. Even a short-duration cores. But when we model the future, 1 2 event such as the Dust Bowl years in we have no empirical basis to judge the 3 the 1930s had a large influence on model’s accuracy. If we take no action rapid American society. The Little Ice Age, until we are completely confident the

warm in Europe climate which caused major resettlement in models are correct, then the only use change Europe in the 15th and 16th centuries, for the models will be to explain what rapid is a more likely analogue of where we happened. Our insistence on a tested climate might be headed. model is part of the reason society is change Some have proposed that we could continuing to conduct the largest ex- 5 counterbalance the greenhouse effect periment ever done, the experiment of strength of ocean circulation 4 by manipulating the global exchanges increasing the atmospheric concentra- cold in Europe of heat and mass. Methods that have tion of greenhouse gases. freshwater inflow to North Atlantic been discussed include blocking the It will be another 20 years before the Strait of Gibraltar to change the salinity climate changes that are predicted to Figure 9. When the ocean-current conveyor of the North Atlantic, using airplane- be associated with the greenhouse ef- belt is in a stable warm state (1), such as distributed particles or large orbiting fect become large enough to be unam- exists today, it can experience relatively sunshades to shade the earth, and fer- biguously differentiated from naturally large changes in freshwater inflow (2) to the tilizing the ocean with iron to promote occurring variations in climate. As a North Atlantic without undergoing signifi- cant change. Ocean circulation is strong, and production of carbon dioxide-consum- society we have the choice of ignoring considerable heat is transported northward ing biomass. But we have a poor record the warning signs that our studies from the tropics, warming Europe. As fresh- of managing even small ecosystems have uncovered or taking some action. water inflow increases, circulation remains and lack a complete understanding of I think we should spend the next 20 relatively strong and heat transport signifi- the ocean-atmosphere interactions that years aggressively investigating our cant until a threshold is crossed (3), causing govern our climate. Intentionally ma- options. We should continue to focus circulation and transport to plummet, and nipulating climate would not only be on improving our ability to predict cli- Europe to cool by at least 2 to 5 degrees costly and imprecise; it would also be mate change. At the same time, we Celsius (4). Likewise, a large drop in fresh- impossible to benefit some regions should test the technologies and po- water inflow (5) is required to cause a return without adversely effecting others. It lices we might need to reduce green- to the warmer state. (Adapted from work by Stefan Rahmstorf, Potsdam Institute for would be a risky experiment on the house-gas emissions, implementing Climate Impact Research, Germany.) only planet we can call home. them on a small scale where there Although we do not know the criti- would be minimal economic and social tion could lead to much larger changes. cal level of greenhouse-gas concentra- disruption. I am not alone among sci- We have no experience predicting cli- tion, we do know that reducing the entists in anticipating that 20 years mate switches between stable modes, rate of greenhouse emissions would from now our society may have to so it would be wise to expect surprises. help in two ways. First, the atmospher- choose between disruptions associated ic concentration of greenhouse gases with our current approach to energy Climate and Choices would increase more slowly. Second, use and disruptions associated with Climate is the result of the exchange of numerical models by Thomas Stocker adopting an approach to energy use heat and mass between the land, ocean, and Andreas Schmittner of the Univer- that produces fewer greenhouse gases. atmosphere, ice sheets and space. As sity of Bern and others predict that the Procrastination will prevent making an long as changes to the land, ocean, at- climate threshold will occur at a higher informed decision and will increase the mosphere and ice sheets stay below the concentration of greenhouse gases if social and economic costs. thresholds I have just described, climate the concentration of greenhouse gases changes will happen slowly. But the cli- increases slowly. Slowing the rate of Bibliography mate will change rapidly if those thresh- greenhouse-gas emissions would buy Broecker, W. S. 1997. , olds are crossed. So rapidly that it would us more time to understand the conse- the Achilles heel of our climate system: Will be impossible to rearrange agricultural quences of our actions and might al- man-made CO2 upset the current balance? Science 278:1592–1588. practices quickly enough to avoid stress- low us to increase greenhouse-gas con- Hughen, K. A., et al. 1996. Rapid climate ing world food supplies. So rapidly that centrations to a higher level before changes in the tropical Atlantic region dur- many species would not be able to reaching the critical threshold. ing the last deglaciation. Nature 380:51-57. adapt, because their habitat, already It is true that computer models are Manabe, S., and R. J. Stouffer. 1995. Simulation greatly reduced by human activities, not perfect; they indicate general pat- of induced by fresh- would be eradicated. terns. And we need to improve our un- water input to the north . Na- Human ingenuity would most likely derstanding in many areas before the ture 378:165–167. Rahmstorf, S. 1997. Bifurcations of the Atlantic allow us to adapt to a rapid change in models can pinpoint thresholds. For ex- thermohaline circulation in response to climate, but we would pay a larger ample, our understanding of the details changes in the hydrological cycle. Nature price than our civilization has ever of ocean circulation is poor, and the 378:145–149. known. Imagine the economic and so- physics of cloud formation and their in- Stocker, T. F., and A. Schmittner. 1997. Influ- cial cost of moving, in a 20-year period, fluence on heat exchange is elusive. ence of CO2 emission rates on the stability most of our agricultural activities 500 When we model previous switches in of the thermohaline circulation. Nature 388:862–865. miles south of their current locations. climate, we can compare the model to Taylor, K. C., et al. 1997. The Holocene-Younger Imagine the social cost and if the results of real-world experiments Dryas transition recorded at Summit, agriculture could not be relocated recorded in ocean sediments and ice Greenland. Science 278:825–827. © 1999 Sigma Xi, The Scientific Research Society. Reproduction with permission only. Contact [email protected]. 1999 July–August 327