How climate changes faster than the cause

The discovery that climate can change in just a few years can be seen as the culmination of a progressive ratcheting up of the pace of events as succeeding generations of researchers found better methods and tools that offered more accurate visions of the past. Certainly, change that might have been thought in the 1950s to take thousands of years was found in the 1990s to have occurred in a decade or less. But it is not just that things happen more quickly than expected. A climate that is subject to abrupt change is fundamentally different, more variable, and less predictable, posing questions that lead to different, more difficult explanations of causes and effects. Most researchers in the 1950s would have said that, aside from the far-distant ice ages, climate really doesn't change much at all.

The 1950s researchers were seeing climate from the perspective of a sailor observing a far shore. The contours of the distant land are smooth. In the grand scheme of things, on the planetary scale, the climate of Earth seems to have been relatively stable for the last 4 billion years at least. Various internal feedback processes keep temperatures within the narrow range that permits water to remain liquid—not so cold that it all freezes or so warm that it evaporates into space. The abrupt change pattern is the signature of the climate processes that keep this liquid window open. Seen for the first time on the human scale, the wiggles of climate change are much bigger, of course, more sudden, and not nearly so benign. Researchers in the 1990s were like sailors negotiating shallow coastal waters and making a landing. From the perspective of their launches pitching in the surf, the shore is lined with rocky crags and cliffs of abrupt change.

Recognizing the new detail and the new timescale of this more eventful history forces scientists across an uneasy theoretical divide. Left behind is the venerable idea that climate is a ponderous system anchored by stable, gradually evolving features—by the orbital motions of the planet, by the oceans and their long thermal memories, and by the ice sheets and their geologically paced advances and retreats. Such a system is complicated, to be sure, but not insurmountably so, and given time, its behavior is predictable. It is infinitely more congenial to analysis and simulation than a system driven by the dynamics of chaos. Abrupt change means that, like the weather itself, climate sometimes behaves in ways that defy prediction. Processes in the atmosphere, in the ocean, and on the land are known to interact with one another, and even though scientists think they know all of the parts and all of the important processes, still they cannot be sure of the outcome of these interactions from one time to the next.

Trying to capture these more elaborate and subtle but less predictable properties of their subject, climate scientists find themselves adopting a vocabulary that seems far removed from the mechanisms of climate and weather. It is the language used by computer modelers and mathematicians who study characteristics of chaotic and complex systems. Concepts with names such as nonlinearities, feedbacks, critical thresholds, and multiple equilibria are their stock in trade. Their formulations are designed to simulate the behavior of economic systems that provoke stock market crashes, biological systems that prompt mass extinctions, tectonic systems that cause earthquakes, and climate systems that change abruptly. In the vernacular, the climate system is nonlinear, which means its output is not always proportional to its input—that, occasionally, unexpectedly, tiny changes in initial conditions provoke huge responses. It is chockablock with feedbacks, loops of self-perpetuating physical transactions, operating on their own timescales, that amplify or impede other processes. This constant cross talk of positive and negative feedbacks is said to be balanced, more or less, at various critical thresholds in the system. Forced across such a threshold, by whatever external or internal triggering mechanism, important variables begin gyrating or flickering, and the system suddenly lurches into a significantly different semistable mode of operation, a new equilibrium. All of these variables, all of these timescales, make for a system that is full of surprises. Scientists use an enigmatic phrase to describe how such chaotic systems violate one of the central principles of linear physics. The whole is more than the sum of its parts.

Nonlinear characteristics of the atmosphere and the ocean have been part of the literature since the early 1960s. In fact, it was a meteorologist who first described the physical principles of chaos. In 1961, in his office at the Massachusetts Institute of Technology, Edward N. Lorenz was using his little Royal McBee computer to execute a rudimentary numerical model of the atmosphere. The slightest changes in initial conditions, Lorenz discovered, inevitably led to major changes in the output of the model as the computer performed repetitions of his equations. As Lorenz and others developed the theory, they laid the groundwork for the way scientists in a variety of fields study the behavior of dynamic systems.

At the same time, just down the way at Woods Hole Oceanographic Institution, oceanographer Henry Stommel was developing the idea that the North Atlantic had two different patterns of circulation and that it was capable of abrupt transitions between these equilibrium states. At the time, this concept seemed to conform to the prevailing theories about climate changing between ice ages and warm periods in concert with long-term changes in the geometry of Earth's orbit of the Sun. It was Stommel who also introduced the concept of thermohaline circulation that linked the motion of currents to differences in water density, a process driven by changes in temperature and salinity. It would be many years, not until the mid-1990s, before climate scientists knew what to make of these ideas.

An abrupt climate change is the very definition of a nonlinear event. In 2002, a report by a special committee of the National Research Council (NRC), chaired by Richard Alley, so defined it: "Technically, an abrupt climate change occurs when the climate system is forced to cross some threshold, triggering a transition to a new state at a rate determined by the climate system itself and faster than the cause." Moreover, the cause of such an abrupt climate change may be "undetectably small."

Change faster than cause and cause undetectably small—these are the footprints that Lorenz followed through the curious behavior of his computer simulation. This discovery had fairly obvious and discouraging implications for weather forecasters, who in the 1950s had dreamed of the day when the usefulness and accuracy of their predictions would extend far into the future. Now they know that, although ingenious numerical models executed by the most powerful computers generate excellent simulations of typical atmospheric changes, capturing the nonlinear motions of thermodynamics, still there are practical limits to their weather forecasting skill. Because the system is chaotic, a typical three-day forecast is not nearly as reliable for the third day as for the second.

When they come to simulating climate, powerful computer models seem to encounter the same sorts of problems with chaotic behavior in the system. Throw all of the interactions and feedbacks into a model that couples the intricate physics of the atmosphere and the ocean, and the output fairly accurately reproduces the typical workings of the system. But that is not always the way the world works. "There have not yet been any successful simulations of the pattern and magnitude of the Younger Dryas or of recurrent Dansgaard/Oeschger events with coupled ocean-atmosphere general circulation models," Alley and colleagues wrote in the NRC study. Nobody is sure why this is so, but something important seems to be missing. Theorists point to problems with models, and modelers point out that the theory could be wrong. Summarizing their work in a 2003 review in Science, the NRC team described a consistent "mismatch" between the computers' idea of abrupt climate change and the real record of the last 100,000 years. Across the board, the real record shows changes that were greater and more widespread than the computers reproduce.

Is it possible that the computer models have it right and climate scientists have it wrong, that they are misinterpreting their data? Not likely, say the scientists. The record of the past is too clear, the cross-checks are too numerous to take such doubts very seriously anymore. That argument collapsed when the big Greenland projects reported their polar ice findings from Summit. Either some natural processes are missing from the models, the NRC team concluded, or the computers systematically "underestimate the size and extent of climate response to threshold crossings."

The difficulties of simulating past abrupt climate changes lead some researchers to question whether predicting future climate is even a realistic goal. Participants in a 2001 workshop at Duke University on nonlinearity in the environment posed the question: Given the nearly certain occurrence of sudden transitions between climate states, is "prediction" per se achievable? The group seemed impressed most by "a relatively poor understanding" among researchers of the nonlinear character of the climate system and the mechanisms that drive it to rapid, episodic change. "Abrupt climate change is believed to be the result of instabilities, threshold crossings and other types of nonlinear behavior of the global climate system, but neither the physical mechanisms involved nor the nature of the nonlinearities themselves are well understood," wrote theorist José A. Rial, of the University of North Carolina's Chapel Hill Wave Propagation Laboratory, and colleagues in the journal Climatic Change in 2004. Citing several examples of nonlinearities, the group was led "to an inevitable conclusion: since the climate system was complex, occasionally chaotic, dominated by abrupt changes and driven by competing feedbacks with largely unknown thresholds, climate prediction is difficult, if not impracticable." Whether prediction will ever be accurate or not, it looks, for better or worse, as though computer modeling and the study of abrupt climate change are wedded to one another. It is a weird marriage of extremes. In the antiseptic, environmentally controlled surroundings of computer laboratories, mathematicians manipulate ice core and ocean sediment data that have been gathered in some of the coldest, wettest, most difficult heavy-lifting fieldwork in science. And yet, in the study of climate, it is hard to imagine one line of research without the other. Weather forecasters don't have long to wait to find out if they've got the system right, but without computer simulations, climate researchers really have no way to test their ideas. More than that, however, although the evidence for abrupt change now seems obvious, there is nothing obvious about the explanations for it. The old idea of a stable, slowly evolving climate was so widespread through the twentieth century—and died such a slow death—because it seemed to make the most sense. Abrupt change, in contrast, is so counterintuitive and so elusive that it is like a concert being played at a pitch beyond the range of human hearing. The chaotic nature of the climate system might never have been recognized were it not for the mathematical genius of Edward Lorenz and his little Royal McBee computer. Now as then, it is the so-called simple computer models or models of "intermediate complexity" that are most useful for testing the validity of one idea or another.

Another instance of the merging of meticulous data analysis and computer modeling has led to interesting evidence that the curious nonlinear behavior known as "stochastic resonance" may account for a 1,500-year cycle of abrupt change detected in the record. It may explain the timing of the two dozen Dansgaard-Oeschger events that dominate the last 100,000 years of climate.

The North Atlantic's nonlinear thermohaline circulation—its density-dependent formation of deep water in the far north, its abrupt switching between multiple modes of equilibrium—remains a central feature of the leading rapid change scenarios, although not all of the prominent events during the last ice age are likely to yield to the same explanation. After all, the most widely observed and carefully studied rapid change—the Younger Dryas, 11,400 years ago—was a sudden plunge back toward frigid cold conditions that came at a time when Earth was warming out of the last ice age. In contrast, the most common episodes of abrupt change in the past 110,000 years, the Dansgaard-Oeschger events, were periods of sudden warming that came with some regularity throughout the ice age.

To explain the Younger Dryas, theorists can point to specific places, times, and events. About 14,000 years ago, as Earth began warming from the last ice age and the great Laurentide ice sheet began shrinking, a vast meltwater lake began to form at the edge of the ice in southeastern Canada. The Great Lakes are remnants of this ice sheet melting. Sediments in the Gulf of Mexico attest to this meltwater lake draining for centuries into the Mississippi River. Then, about 11,400 years ago, so the story goes, an ice dam gave way and an enormous flood of fresh water suddenly poured through the St. Lawrence River into the North Atlantic. This rapid freshening so altered the density balance of the ocean that far northern surface water no longer sank to the bottom when it chilled. This halt in deep-water formation caused a shutdown of the ocean thermohaline conveyor, closing off the flow of warm water from the Tropics into the North Atlantic and plunging the region back toward ice age conditions. A cold, dry, windy regime held sway for 1,300 years before warming resumed. The Younger Dryas cold was transported to other regions by circulation changes in other oceans and broadcast by large-scale and persistent atmospheric responses, paleoclimatologists believe, leaving a distinct climate signal over much of the globe.

But what of the main features of ice age climate history? What caused the Dansgaard-Oeschger warmings and the supercold Heinrich episodes? A group of leading abrupt-change specialists pondered these questions at an American Geophysical Union conference in 1998 in Snowbird, Utah. The stabs of extreme cold—the six Heinrich events during this glacial period—seem closely tied to the D-O cycle, coming as an exaggerated cooling phase at the end of a bundle of several warming events, each separated by periods of progressively colder temperatures. Evidence from the ocean sediments, showing the Canadian origin of the ice-rafted debris layer, pointed to the dynamics of the Laurentide ice sheet as their cause. Most researchers have accepted the idea proposed by glaciologist Douglas R. MacAyeal that the ice sheet over North America underwent a binge-purge cycle, becoming unstable as it built up over thousands of years. MacAyeal proposed that Heinrich events "were caused by free oscillations in the flow of the Laurentide ice sheet which arose because the floor of Hudson Bay and Hudson Strait is covered with soft, unconsolidated sediment that forms a slippery lubricant when thawed." The great ice sheet "periodically disgorged icebergs in brief but violent episodes which occurred approximately every 7,000 years," MacAyeal said, when geothermal heat trapped under the overburdened glacier thawed its base. These icebergs pushed the spreading sea ice to exceptionally low latitudes, amplifying the cold conditions, completely shutting down the conveyor, and halting formation of deep water in the North Atlantic.

As Alley summarized the consensus view in an AGU monograph, the Dansgaard-Oeschger events, while clearly related to ocean circulation, were a bigger puzzle. The big climate swings have come not when conditions such as temperatures and ice extent were high, and forces such as carbon dioxide in the atmosphere or radiation from the Sun were extreme, but rather when conditions were relatively moderate and CO2 and solar forces were changing rapidly. Maybe the swings came when forces would push the climate system "in the gap between two modes." Alley offered an alternative explanation—that "the stability of the system may be sensitive to the rate of change of forcing variables, such that climate response is analogous to that of a drunken human: when left alone, it sits; when forced to move, it staggers with abrupt changes in direction."

What could provoke the sudden warmings in the midst of an ice age? In his lectures around the country, Alley found that his ideas attracted other scientists to his line of thought. The Penn State glaciologist soon found himself teamed with two new partners, geophysicist Sidhar Anadakrishnan and physicist Peter Jung, to test a line of thought that stochastic resonance might account for what seemed to be a 1,500-year period between most D-O events. The idea behind stochastic resonance is that a feeble periodic signal is amplified by the presence of stochastic—that is, random—"noise" in the climate system, that the weak signal and the noise resonate and produce a signal capable of triggering an abrupt change.

Alley performed a frequency analysis of the oxygen isotope signals of the U.S. GISP2 and European GRIP ice cores. If stochastic resonance played a role in the 1,500-year cycle, the researchers reasoned, a certain signature would emerge from frequency analysis of the climate record. If the cycle were strictly periodic—strictly linear—all transitions would land on the 1,500-year mark on a histogram, a chart depicting frequency distributions. If it were no more than random noise at work, events would be scattered over the histogram. What emerged on the histogram was just the pattern that stochastic resonance would leave. Almost all events landed at multiples of 1,500 years— most at the 1,500-year frequency, fewer at 3,000, and fewer still at 4,500. While other researchers confirmed that the North Atlantic's conveyor system was potentially subject to stochastic resonance, they estimated that only a relatively powerful signal would be able to switch it on and off.

While Alley and coworkers were massaging their data, Stefan Rahmstorf and Andrey Ganopolski were doing some interesting computer modeling at the Potsdam Institute for Climate Impact Research in Germany. Their "intermediate complexity" model simulates a more sensitive ice age North Atlantic than other models of the ocean. The warm mode of their model Atlantic, when warm water is flowing northward, cooling, and sinking to the depths, is an "excitable" state that is less stable than the glacial mode. As the German researchers put the case in a 2001 monograph, "we have shown that the 'cold' mode is the only stable mode of the glacial Atlantic ocean circulation." Rather than completely switching off deep-water formation, their model shifts the critical overturning zone farther south. Alley's team noticed that this ice age ocean is more likely to respond to stochastic resonance. Alley and Rahmstorf brought their two lines of research together in an article in the American Geophysical Union journal Eos. Combining random noise with a very weak 1,500-year climate cycle in this model produces Dansgaard-Oeschger events that "are very similar to those recorded in Greenland ice and other paleo-climatic archives," they wrote. And the recurrence times closely resembled the pattern developed by Alley and colleagues. "This shows how the Atlantic ocean currents can act like a threshold amplifier, turning a feeble signal into dramatic climate swings."

More work was going to have to be done before anyone could confidently claim that stochastic resonance explains abrupt climate change, they wrote. "But for the first time, data analysis has shown strong hints and model simulations have provided a possible quantitative mechanism for stochastic resonance causing some of the most abrupt climate shifts known." The concept raised interesting questions. What is the source of the 1,500-year cycle? Could natural or human-caused noise affect the system in the future?

Several researchers have proposed possible causes of a weak periodic signal. As early as 1990, Wally Broecker had suggested that abrupt climate events were triggered by oscillations in the salt content of the Atlantic that periodically shifted the density balance of the ocean and switched the circulation conveyor on and off. In 1997, the U.S. ice core project leader at Summit, Paul A. Mayewski, analyzing the chemical properties of the GISP2 ice core, described a 1,450-year cycle of rapid climate change events that involved "massive reorganizations of atmospheric circulation." The glaciologist suggested that solar variation ice sheet-atmosphere feedbacks may help explain the timing and amplitude of rapid climate changes. That same year, Gerard Bond at Lamont-Doherty, analyzing North Atlantic sediment cores, found a long cycle of abrupt change on a timescale of 1,470 years that ran through the ice age and—at lower magnitude— continued to punctuate the supposedly stable climate of the Holocene, the last 11,000 years. In 2001, Bond and colleagues also proposed that these cycles, marked by the spread of drift ice in the North Atlantic, were tied to solar variation. In 2000, Charles D. Keeling and Timothy P. Whorf suggested that a 1,800-year tidal cycle was responsible for abrupt climate change. According to them, "variations in the strength of oceanic tides cause periodic cooling of surface ocean water by modulating the intensity of vertical mixing that brings to the surface colder water from below."

Apart from the enormous fact of it—the data in the ice cores and the ocean sediments that plainly show climate changing dramatically in just a few years and the increasing evidence of its global reach—almost no important issues about the mechanisms of abrupt climate change are entirely settled. The problem is new; the science is young. Still scientists are asking themselves questions as basic as that posed in 2003 in the journal Science by Wally Broecker: "Does the Trigger for Abrupt Climate Change Reside in the Ocean or in the Atmosphere?" Focusing on the Younger Dryas, the most thoroughly documented of the abrupt changes, Broecker reviewed data suggesting that the North Atlantic's circulation was responsible and other data favoring an atmospheric mechanism centered in the equatorial Pacific. The oceanographer found in favor of the ocean, again, and the scenario of change driven by reorganization of the ocean conveyor in the North Atlantic that he first proposed in the mid-1980s. Still, the state of the science is such that few conclusions can yet be stated with much conviction, even by a theorist who is famous for doing just that. As Broecker would observe, "we are still a long way from understanding how our climate system accomplished the large and abrupt changes so richly recorded in ice and sediment."

The idea that the oceans play a central role in the climate of a planet so dominated by the presence of water is not really in dispute. At every timescale, the atmosphere and the ocean are engaged in intimate conversation—the atmosphere picking up water vapor and temperature changes and discharging precipitation and current-bending winds, the oceans storing and transporting heat from the Tropics and cold from the poles. Coming from Broecker, the question of a trigger or mechanism for change in the ocean or the atmosphere is recognition of an ongoing controversy among leading theorists and modelers about where in the world abrupt climate change originates. The scenario that invokes the thermohaline circulation and its mode shifts in the North Atlantic was the first and still is the most thoroughly developed line of thought about the causes and the timing of abrupt changes through the past 100,000 years.

Broecker would be the first to acknowledge that the visual icon of this reasoning, the Great Ocean Conveyor of continuous strips of cold and warm currents wending their ways through the world's oceans, is a cartoon rendering of how the oceans really work. First drafted for a popular magazine article in 1987, the picture would forever offend physical oceanographers for its over-simplification. "Had I known that this conveyor belt diagram would become a logo for global change research," he would recall, "I probably would have worried a bit more about the details."

More fundamental are differences in evaluating the role of thermohaline circulation—the THC—the currents that respond to density differences that are themselves a function of temperature and saline properties of water at different depths. In a 2000 commentary in the Proceedings of the National Academy of Sciences, the German theorist Jochem Marotzke wrote: "The THC's role in abrupt climate change is not comfortably established—on the contrary, it poses major scientific challenges and thus provides a powerful focus for climate research." Elsewhere, Marotzke observed that "first-order advances are needed" to fill critical gaps in the basic concepts of how climate changes. "Our current knowledge is insufficient to determine the likelihood of abrupt climate change, or even to answer some simple questions with confidence," he wrote. "For example, why was the THC in glacial times apparently subject to large instabilities, but in the present interglacial has been considerably more stable? Did the giant ice sheets dramatically amplify a climate signal that has persisted into the interglacial, but at a suppressed amplitude? Or does the relative quiescence of the present interglacial primarily result from a change in the atmospheric forcing applied to the ocean, or perhaps a change in ocean processes, the nature of which may have altered because of decreased sea ice coverage or a higher sea level?" Oceanographer Carl Wunsch of the Massachusetts Institute of Technology argues that the THC concept has been overused and overstated. In a 2002 article in Science, Wunsch maintained that the ocean's motions are "sustained primarily by the wind, and secondarily by tidal forcing." In both models and the real ocean, the buoyancy of surface water strongly influences the flows of heat and salt, "because the fluid must become dense enough to sink, but these boundary conditions do not actually drive the circulation."

The following year in Nature, modeler Stefan Rahmstorf noted that the THC "can be defined as currents driven by fluxes of heat and freshwater across the sea surface and subsequent interior mixing of heat and salt." This process is "clearly distinct" from the turbulent mixing caused by winds and tides, although because they work on the same circulation, the two mechanisms cannot be separated from one another in ocean measurements. In any case, heat-transporting ocean currents are critically important to the climate of regions surrounding the North Atlantic. "Changes in these currents are our best explanation for the abrupt and marked climate swings that occurred over the north Atlantic many times during the most recent glacial period, as shown by Greenland's ice cores and by deep-sea sediments," Rahmstorf wrote.

Changes in North Atlantic ocean circulation may be the best explanation, but there are other explanations for abrupt change—for the triggers, amplifying feedbacks, and "globalizer" processes that transmit its impacts across the planet. The gaps in knowledge still are large enough to accommodate more than one school of thought. Theorists and modelers are like frustrated cooks trying to work up a delicate soufflé without the benefit of a satisfactory recipe. Try after try, they can't get the dish to come out they way they want. Perhaps the problem is more than a matter of tweaking the proportions of this or that ingredient. Maybe it's not a soufflé recipe at all.

What bothers many scientists about the North Atlantic THC scenario of abrupt change is not so much what it proposes—that changes originate in the North Atlantic—but what it proposes to do without—the tropical Pacific. In the scheme of things, one might think it would be difficult to account for such powerful environmental episodes without involving the world's largest ocean or the Tropics, where sunlight deposits most of the planet's energy. Nevertheless, theorists were much influenced by 1970s research that seemed to show that while the rest of the world slipped into the last ice age, the Tropics cooled only a mere 1.8°F. Is the formation of North Atlantic deep water really the key to abrupt climate change? Or is it just the explanation most readily at hand that climate scientists happened to find?

Being a human enterprise, science is not always such an orderly business. Would our current understanding of natural climate variation be the same if Henri Bader had not persuaded the U.S. National Committee for the International Geophysical Year to support an expedition to Greenland in the 1950s? What if stable isotope technology had not been available in the 1960s to tease the temperature of ancient snowfall out of the polar ice? Climate science seems especially dependent on technology, its direction subject to the whims of contingency. The Swiss climate modeler Thomas F. Stocker touched on this issue in Science in 1998 as he evaluated prevailing thinking about the causes of abrupt change. "One of the pressing questions is where the centers of activity responsible for these abrupt and millennial-scale climate changes are located," he observed. "The region of the high northern latitudes, especially the North Atlantic, has been the classic focus of this research. However, it may be argued that the first institutions of paleoclimatic research have been located around the North Atlantic, and a certain bias cannot be excluded."

The same sort of bias probably could be observed in most lines of research. Lecturing to undergraduates, Harvard University geneticist Richard Lewinton tells the old story of the drunkard whom a passerby finds crawling around on his hands and knees under the light of a street lamp. He's looking for his car keys, he says.

"So, you lost your keys down here somewhere?" asks a passerby. "No, I dropped them across the street," says the drunkard, "but this is where the light is good."

Maybe it is just a matter of proportion. Later research found much more intense cooling of the Tropics during the last ice age—locally as much as 7.2°F. And while no one can yet describe the details, no one really argues that the Tropics have nothing to do with abrupt climate change. In 1997, for instance, even as he was describing thermohaline circulation as "the Achilles heel of the climate system," Wally Broecker was ruminating about the climate-altering power of water vapor, the major greenhouse gas that is primarily a product of the Tropics. As Alley put the case in a commentary in Nature in 1998, "The idea that changes at high latitudes can affect widespread regions extending across the Equator is unpopular with many workers who do not believe that the small, energy-starved polar 'tail' can wag the large, energy-rich tropical 'dog.'"

In 1999, speaking for the tropical dog, geophysicist Raymond T. Pierrehumbert wrote in the Proceedings of the National Academy of Sciences: "It is not certain that the millennial scale fluctuations observed in different parts of the globe are all related to each other, but it is certain that such fluctuations are observed just about everywhere, so one should at least entertain the possibility that the NADW [North Atlantic deep water] picture alone cannot account for everything that is going on." It doesn't explain climate fluctuations in different regions, for instance, or how the impact of changes in North Atlantic deep-water turnover is spread around the globe.

Pierrehumbert was a member of the National Research Council's Committee on Abrupt Climate Change whose 2002 report described the tropical Pacific, where the waxing and waning of the El Nino-Southern Oscillation spreads well-known patterns of deluge and drought around the world on a three- to five-year cycle, as "a natural candidate to be a 'globalizer' of climate influences." Maybe it is more than that. The atmosphere's response to the tropical Pacific's main features—its wind-driven eastern "cold tongue" and western "warm pool" and the line of intense convection known as the intertropical convergence zone—brings to Pierrehumbert's mind the same sort of dynamical behavior that Broecker envisions in the ocean currents of the North Atlantic.

"The formation of convecting air is tippy in precisely the same sense and for precisely the same reason that formation of North Atlantic deep water is tippy," Pierrehumbert wrote. The tropical atmosphere is a powerhouse of storminess, but the air itself—like North Atlantic surface water—is delicately balanced, rising in response to a subtle interplay of moisture and temperature changes. As Pierrehumbert put it, "Tropical convection creates 'top air' in much the same way oceanic convection creates 'bottom water.'"

"The tropical atmosphere-ocean system offers a rich palette of possible amplifiers and switches that could in principle lead to abrupt climate change," the NRC committee wrote. Foremost among them is water vapor, the greatest greenhouse gas, what Pierrehumbert called "the premier atmospheric feedback amplifying the response to what would otherwise be minor climate forcings." Moreover, the circulation of the Tropics produces "boundary layer clouds" that can have an albedo—that is, reflectivity— "approaching that of ice and exert a potent cooling effect when they form As a nexus for climate change, the tropical Pacific also has a clear advantage over the North Atlantic, in that the global atmosphere is exquisitely sensitive to tropical Pacific sea surface temperature anomalies," wrote Pierrehumbert.

Variations in the surface salinity of the tropical and subtropical Atlantic during the past 136,000 years have been linked to the waxing and waning of the North Atlantic's conveyor by University of California researchers who analyzed two sediment cores from the Caribbean Sea. Cold periods bring relatively dry conditions to the tropical Atlantic, so the waters of the Caribbean become increasingly salty. How much the North Atlantic's circulation picks up during the next warm period may depend on how much extra saltiness the Caribbean has to deliver. Reporting these results in Nature in 2004, Matthew W. Schmidt and Howard J. Spero surmised that the tropical rainfall cycle "has a direct role in regulating rapid climate change."

Could El Niño trigger abrupt change? In modern climate, at least, nothing more dramatically rearranges the hydrologic cycle of the Tropics from one year to the next. In Germany, Mojib Latif tested the effects of warming ocean temperatures and was able to show a link between the salinity of the tropical Atlantic and El Niño-like changes in air-sea interactions in the tropical Pacific. Moving the center of heavy precipitation into the central tropical Pacific creates a warm, dry regime of descending air over northeast Brazil that increases evaporation and salinity in the surface of the tropical Atlantic. Larry C. Peterson studied sedimentary records from the Cariaco Basin off northern Venezuela and found evidence that "supports the notion that tropical feedbacks played an important role in modulating global climate" during the last ice age. In 2003, David W. Lea, Peterson, and colleagues compared the timing of sea surface temperatures recorded in Cariaco Basin sediments and suggested an "active tropical feedback" between the western tropical Atlantic and Greenland air temperatures during the Younger Dryas.

Modeling work by Amy C. Clement and coworkers points to the El Niño-Southern Oscillation as a possible trigger for abrupt climate change. In response to subtle, constant alterations in solar radiation, this unstable tropical Pacific ocean-atmosphere cycle can lock "in phase" with the seasonal cycle, producing several centuries of La Niña conditions. Their experiments suggest that such a locked-in La Niña could have triggered the Younger Dryas about 11,000 years ago.

James P. Kennett, investigating the Santa Barbara Basin, has developed an entirely different line of thought about abrupt change mechanisms. He challenges the common view that variations in the concentration of methane gas in ice cores and other climate archives are primarily signals of the waxing and waning of low-latitude wetlands as temperatures and sea levels rose and fell. This provocative "Clathrate Gun Hypothesis" argues that the powerful greenhouse gas itself was the trigger mechanism for abrupt warmings. Crystalline gas hydrates embedded in sediments along the continental shelves, sensitive to changes in temperature and pressure, occasionally became destabilized and produced bubbles that rose to the surface and released massive amounts of methane gas into the atmosphere.

Methane-laced ocean sediments represent an enormous reservoir of carbon—by some estimates, at least twice the amount of carbon found in all known fossil fuels. Warren T. Wood has shown that hydrates in some continental slopes may be more unstable than previously thought. In 2003, Kai-Uwe Hinrichs, examining the Santa Barbara Basin sediments, studied fossil remnants of bacteria that would have flourished only under high concentrations of methane. "It is evident from our data that substantial quantities of methane were rapidly mobilized at least several times over the past 60,000 years," Hinrichs reported in Science. Hinrichs's research supports the Clathrate Gun Hypothesis, although it doesn't say how much methane escaped from the ocean. "But one thing is for sure," he said, "our results clearly show that relatively minor environmental changes can have a major impact on sensitive coastal regions with yet unknown consequences for climate and [for life on Earth]."

0 0

Post a comment