by Sarah Dry
7
OLD ICE
One rainy Saturday in June 1952, Willi Dansgaard stood in his back garden in central Copenhagen. He set a beer bottle on the ground and put a funnel in its mouth. The rain was thick and persistent, the product of an epic storm front that stretched all the way to Wales, one thousand kilometers away.1 In a few hours, he stepped again into the back garden. The bottle was nearly full. He poured its contents into a small container, sealed the container, and replaced the empty bottle on the back lawn, like any amateur meteorologist collecting rainwater. He re-entered the house and wrote the date and time on his sample, setting it next to the others he’d already taken. The containers crowded his kitchen table.
He kept it up throughout the weekend, even waking in the night to collect the rainwater. The rain, making up an unusually well-developed warm front followed by a sharply delineated cold front, continued. At some point, he ran out of sealed containers and started collecting water in pitchers and pots from the kitchen. He kept it up through the early hours on Monday morning and then loaded the whole collection into his vehicle and drove carefully to his laboratory.
A mass spectrometer, the instrument that made the collection of all this rainwater something more than an exercise in amateur weather-watching, was waiting for him. It had been his since he arrived back in the city after a few early years of adventure in Greenland. He’d fallen in love with the harsh beauty of the vast island, but he had been offered a salaried research post at the university in Copenhagen that he couldn’t refuse. The task he’d been assigned was to use the chamber of the mass spectrometer to test the utility of the so-called stable isotopes for medical or biological uses. Radium and its highly unstable cousins had been used for nearly fifty years to treat cancer, helping burn away sick cells or track the progress of disease, but they were damaging to the healthy tissues of the body too. Dansgaard’s job was to investigate whether something useful could be done with the less protean isotopes of oxygen and nitrogen. These non-radioactive elements were free of the dangers that accompanied radiation treatment.
Dansgaard had little luck researching the use of stable isotopes in medical applications, but the machine was his to use more or less as he pleased. He knew that the oxygen in rainwater existed in the form of several isotopes, variants distinguished by the number of neutrons in their nucleus. The powerhouse 16O, the isotopic bully, was vastly more prevalent than the heavier 17O and 18O versions. Generated in the belly of stars, at the tail end of the process of helium fusion, 16O makes up more than 99.7 percent of all the oxygen molecules in the earth’s atmosphere. That leaves about one molecule of 18O for every three hundred molecules of 16O. The difference of a few neutrons between these isotopes doesn’t matter chemically. But it does mean that 18O, which has two more neutrons than 16O, weighs ever so slightly more.
On this basis, the isotopes of oxygen could be separated in the mass spectrometer. He’d done as much. With the machine, it was possible to tease apart the scarcely before separated strands of oxygen with relative ease. He knew that water evaporates in warm temperatures and condenses in cooler temperatures. One form of this condensation is rain. He also knew that 18O, being heavier, was about ten percent more likely to condense than 16O. The converse was also true: 16O, being lighter, was ten percent more likely to evaporate than 18O. Storms were made up of weather fronts, alternating masses of warm and cold air. Putting the two ideas together, Dansgaard wondered whether rainwater had a distinctive isotopic composition, like a fingerprint. Could it change, from rainfall to rainfall, or even during the course of a single storm?
The bottles sloshing in his car were snapshots across time of the isotopic profile of the storm that had just dumped so many inches of rain on Copenhagen. He put the samples through his mass spectrometer, and the results were more than good. It was as though he’d put a stethoscope to the storm and listened to the heartbeat of isotopic oxygen pulsing within it. The proportion of 18O in the rainwater he collected rose as the warm front passed over Copenhagen, as the “heavy” water condensed as rain.2
It was all there waiting to be measured and understood, a secret code to the life history of the storm, its changing temperature profile captured in an isotopic progression. He knew immediately that he now had to get inside a cloud. That was the next step, to understand the relationship between the atmosphere at the center of a cloud, that which lay at its top edge and that which was below it. He imagined cumulus clouds as giant condensers in the sky, separating water vapor, sending warm air upward, cold 18O-dense water below, and finally warm 16O-dense air to evaporate at the top.
A buddy with connections to the Royal Danish Air Force got him and his wife, Inge, in a plane. Inge insisted on going along for the ride. She did not, she said, wish to be a young widow. So up they went into the violent air currents inside a cumulus cloud. Using cold traps cooled by dry ice, Willi tried to collect cloud droplets as they tossed about. The samples were too small to quantify the effect, but the results confirmed his theory. Dansgaard looked for a way to gather larger samples, samples that would speak more loudly. He thought of weather in three dimensions: not simply what fell to earth, or what the thermometers measured at a set of points, but masses of gas and liquids traveling together, clouds occupying space the way an actor fills the stage. Clouds had form, depth, density. They moved and changed. And they were freckled with the stable isotopes of oxygen.
The whole water cycle was his to explore. Where would he go next? His mind flew to the rivers, which were liquid averages of the rain that fell across vast distances. He was leaving behind the temperature profile of a single storm to think, instead, about what weather adds up to when it is averaged: to climate. Would rivers bear witness to the regions from which their water was gathered, presenting an isotopically determinable average temperature for the region? His mind raced outward, searching for a way to grasp the whole earth with his new insight, his new tool. What he needed was a global network for rainfall collection. And that, thanks to the offices (literally) of a Danish shipping company with international reach, is what he got. Another buddy with connections managed to put him in good graces with the Danish East Asiatic Company, which obligingly provided him with samples from its worldwide branches. Soon enough, he had a new collection of bottles, this time sloshing not with Copenhagen rain but with the products of storms ranging from the tropics to the Arctic.
He was now equipped to continue his stepwise progression through the waters of the globe. He was working outward and upward. Would the relationship between the isotopic ratios and temperature that he had observed in his backyard samples hold true for water from other parts of the globe? At first, it didn’t look good. Things got messy in the tropics, and it was hard to discern a strong correlation between the δ-value (a measure of the isotopic ratios) and the temperature of the water. But all was not lost. Samples collected closer to the poles, and, in particular, at high altitudes in the Arctic, showed the same relationship he’d been so happy to find in the Danish rain. Thanks to the Danish East Asiatic Company, he had proved that the isotopic technique could in principle be used to measure temperature at the time that rain or snow fell.
FIG. 7.1. A schematic diagram showing the excess 18O in rainwater samples collected during the June 21–23, 1952 storm. During the passage of the warm front, the rain became colder and isotopically heavier. This was the beginning of Dansgaard’s idea for an isotopic thermometer. From Willi Dansgaard, “The Abundance of 18O in Atmospheric Water and Water Vapour,” Tellus 5 (1953): 461–469. Copyright the Author 1953.
It was a good result, but Dansgaard was not satisfied. In the paper he published on the subject in 1954, Dansgaard wondered whether the correlation between isotopic ratios and temperatures in cold places like the Arctic also held true for the distant past of the earth. “On the supposition that the character of the circulatory processes, in all essentials, have not varied over a long period of time,” he reasoned, the isotopic
technique could offer the possibility of “determining climatic changes over a period of time of several hundred years in the past.”3 This was the first inkling of what he later called “maybe the only really good idea” he ever had.4 Maybe, thought Dansgaard, he could use his mass spectrometer to measure not only today’s rain, but yesterday’s snow, and many, many yesterdays before that. He thought he might be able to go back in time a few hundred years because the only way to access ice at depth was to measure it at the margins of the glaciers, where it was tens of meters thick. He ended his paper with the statement: “an investigation will be undertaken as soon as an opportunity offers.”
* * *
Dansgaard’s idea depended on new technology and a newly developed understanding of atomic isotopes, but the problem to which it would eventually contribute with stunning results was of a piece with old questions about the role of ice in the earth’s history. Thanks to the work of men like Agassiz, Forbes, and Tyndall, the ice age theory gradually grew in acceptance through the second half of the nineteenth century. Much of this had to do not with the convincing power of the theory itself, but with the accumulation of what amounted to an avalanche of evidence in its favor. Mapping, in particular, played a critical role in turning the ice age theory into an ice age consensus by bringing together countless observations of glaciers, ice sheets, and the moraines they left behind. By the 1870s, enough evidence had been plotted on maps to convince the most skeptical that erratic and uncontrolled drifting icebergs could no longer be considered a plausible mechanism for the degree and nature of the evidence compiled.5 What James Geikie called the “prejudices” of British geologists in favor of icebergs (in a letter to John Muir) had been overcome by the sheer mass of evidence in favor of ice sheets. In Britain, the culmination of this mapping mode arrived in 1914 with the publication of William Bourke Wright’s comprehensive The Quaternary Ice Age, which indicated the flow lines of glaciers believed to have moved across Britain, re-creating the movement of the ice based on the moraines and striations it had left behind.
FIG. 7.2. Map of Europe showing largest extent of glaciation, from William Bourke Wright, The Quaternary Ice Age (London: Macmillan, 1937).
FIG. 7.3. Two views of a glaciated boulder showing the different directions of the striations. From William Bourke Wright, The Quaternary Ice Age (London: Macmillan, 1937).
Maps like Wright’s helped convince the remaining skeptical geologists that not just one but several ice ages had in fact occurred. “There is a growing body of evidence that several distinct glaciations were involved,” explained Wright.6 But it still remained unclear precisely what had caused the earth’s climate to flip-flop between ice ages. James Croll had made many assumptions about the complicated interactions between different aspects of the earth’s climate system, with ice at its poles, water vapor that created cooling fogs or reflective clouds, and ocean currents that meandered across the globe transporting heat all playing a role. While Croll and his supporters were convinced that the laws of physics meant that such a system had to exist, the idea that the oceans, ice, and atmosphere interacted at global scales was both new and almost impossible to prove. Geologists who were used to basing their grand theories on extensive field observations and synthesizing maps were unconvinced. And so it was that Croll’s theory remained marginal to most discussions of ice age theory in the decades following his death.
One man took up Croll’s astronomically based theory with an almost revolutionary zeal. Milutin Milankovitch was a Serbian engineer who ended up as a prisoner of war during World War I. He used his four years of imprisonment to begin to work out detailed calculations of the changing orientation of the earth and the sun over hundreds of thousands of years. Three main cycles, corresponding to the tilt and wobble of the earth on its axis and slight changes in the shape of its orbit around the sun, which changed over long time cycles ranging from 23,000 to 41,000 to 100,000 years, determined how the sun’s radiation hit the earth. While they did not alter the total amount of sunlight that reached the earth, they did affect the distribution of that sunlight throughout the year and across the globe in ways that affected the earth’s climate dramatically (as Croll had understood). Interestingly, Milankovitch shared Croll’s belief that changes in the earth’s orbit were the cause of the ice ages, but he thought that increased summer sunlight in the northern latitudes where snow and ice accumulate and then melt each year, rather than increased winter cold, was the main reason for the imbalance.
Like Croll, the Serbian struggled to elicit mainstream recognition for his work. But also like Croll, he received support from a few eminent scientists. One important group of researchers called themselves climatologists. Concerned with mapping as well, these scientists sought to establish not the past history of the earth but the current geographical distribution of different climates. Once content to focus on the local or regional scale, climatologists had by the turn of the century become increasingly confident in their ability to map climate differences around the globe. Place, rather than time, was their main concern. These intellectual descendants of Alexander von Humboldt were not concerned, as the geologists were, to explain change over time. More attuned to the stability of regional climates than to any long-term changes, they were inclined to stay focused on the climates of the present. Cleveland Abbe, chief scientist of the U.S. Army Signal Office, for example, considered that “the true problem for the climatologist to settle during the present century is not whether the climate has lately changed, but where our present climate is, what its well-defined features are, and how they can be most clearly expressed in numbers.”7
The concerns of climatologists were shared by—and partly a response to—nations seeking to consolidate or increase their territory. Mapping had always been a tool of government. Where to plant which crops, how many settlers a region could reasonably accommodate, how much rainfall fell in which regions were climatological matters of central importance to state powers. Since climatological regions did not always follow national boundaries, one nation’s drive to map its territory naturally spilled over into others—prompting international cooperation even as individual nations sought to better manage their own territory. Climatology gradually grew in both prestige and ambition. By the early decades of the twentieth century, climatologists were increasingly determined not merely to characterize distinct regional climates but to create a global climatology that could synthesize local information. Given the new global orientation of the field, it makes sense that Milankovitch’s work on global climate changes would catch the eye of perhaps the most accomplished climatologist alive, Wladimir Köppen. Köppen, too, took a global view of climate. In 1884 he produced one of the first maps of climate on a global scale, plotting regions of similar temperature and precipitation, flora and fauna across the globe. Köppen and his son-in-law, the meteorologist and Arctic explorer Alfred Wegener, were both impressed with Milankovitch’s work.
Soon after his first paper came out in 1920, Milankovitch received a postcard from Köppen praising his publication, which he cherished, he later said, “like a relic.” The product of Milankovitch’s decades of toil was the first time line showing how summer sunlight had varied over the past 600,000 years. While Croll had made qualitative guesses about how radiation might have changed, Milankovitch had managed to generate a curve giving specific dates for astronomical oscillations that would have affected how much sunlight hit the earth during the summer. Köppen and Wegener realized that with Milankovitch’s curve, they had an independent marker of past climate change. Geologists had already generated their own crude time line of past ice ages based exclusively on geological traces—on the moraines, striations, and other features that had first been noticed back in the 1830s and 1840s. There was rough agreement between these curves when they were aligned, giving both geologists and physicists more confidence that changes on the earth might correspond to astronomical cycles. It was now possible to make retrospective predictions about w
hen ice ages and warmer periods may have occurred, and to do so on the basis of a theory with an impeccable physical basis. The long arm of Newton stretched across the twentieth century, demonstrating that something as seemingly contingent as the spread of snow and ice across the earth was intimately related to the celestial mechanics of the earth and the sun.
It took some time for geologists to come around to Milankovitch’s theory. During the 1930s, as he refined and extended his calculations, geologists sought to match the curves of past ice ages generated by increasingly detailed geological evidence to the ever more detailed astronomical curves that Milankovitch produced. The concordance was convincing, even if there were significant divergences between the curves as well. Over time, something important shifted in the relationship between Milankovitch’s theory and the field-based work of the geologists. No longer was the geological record being used to test Milankovitch’s theory. Now the theory was used to verify—or ground-truth—the geological record. “In this manner,” wrote a triumphant Milankovitch, “the ice age was given a calendar.”8
Milankovitch had reached into the heavens to produce a calendar for changes on the earth. His theory leveraged the predictive power of astronomy to make inferences about the physics of the earth. In a sense, this was the culmination of ambitions laid out by men such as John Herschel and Norman Lockyer, the “cosmical physicists” who hoped to simultaneously draw the connections between Earth and the solar system and to raise the physics of terrestrial phenomena to the same status that positional astronomy had once held. Milankovitch’s reliance on Newtonian calculations of orbits rather than on the physical traces of magnetism or radiation marked him out as different from Lockyer and Herschel. And while Milankovitch’s theory had been corroborated by geological evidence on earth, it was still a rough-grained picture of the history of the ice ages. It was possible, and indeed likely, that other changes relating to the so-called secondary factors which Croll had initially identified—factors such as the circulation of the world’s oceans and changes in atmospheric phenomena—occurred on much shorter timescales than the cycles Milankovitch had identified. To find out what these were would require new tools for looking into the past. The heavens had provided what they could. The time had come to look back to the earth and to seek new methods, which went beyond the descriptive mapping into which the geologists had put so much effort, to bring the past into focus.