Eielson Air Force Base, Alaska, 2024

Dr. Sarah Chen was thirty-four years old and had spent the last four years of her life measuring the temperature of air that did not exist. This was, she knew, a reductio ad absurdum of her job description, which officially read Director of the Sub-Arctic Atmospheric Monitoring Station, Eielson Research Division, but which in the informal language of grant proposals and conference panels meant a woman who lived in a metal box at the edge of the tree line in interior Alaska and sent data about temperature and humidity and carbon dioxide concentration to people in offices in Washington and Geneva who would read the numbers and file them and write reports that would be read by people who would sit on committees and decide whether the numbers justified the expense of doing something about the numbers.

The station was a collection of six modular buildings clustered around a weather radar dish that had been installed in 2003 and updated in 2015 and was, as of 2024, simultaneously the most sophisticated instrument in the region and the most obsolete. The radar could detect precipitation particles down to a tenth of a millimeter in diameter and scan a radius of two hundred kilometers every six minutes. It produced approximately four terabytes of data per day, of which Sarah and her two assistants, a climatologist named Priya Sharma and a meteorological engineer named Erik Lindstrom, processed perhaps five percent. The rest was archived and largely ignored, because the people who controlled the funding — the National Science Foundation and the Department of Energy and a private foundation called the Aurora Trust that had been established by a technology entrepreneur with an interest in climate data — wanted summaries, not raw data, and summaries were easier to sell to the people who wrote the checks.

Sarah Chen arrived at the station in May 2020, three months after the pandemic had turned the continental United States into a series of isolated bubbles and travel had become, for most people, a theoretical concept. She had come from NOAA's Boulder laboratory, where she had worked on atmospheric modeling for six years, and she had accepted the Eielson assignment because it offered two things that her previous position had not: complete intellectual autonomy and isolation. She wanted to work on a problem that she had been thinking about since her doctoral studies at MIT, a problem that required a vantage point that only interior Alaska could provide, and she wanted to work on it without the institutional pressures that had made her previous years at NOAA a slow compression of her curiosity into the shapes that funding agencies preferred.

The problem was simple. Or rather, the question was simple. Over the previous decade, the atmospheric data collected at Eielson and at a network of thirteen other sub-arctic stations had shown a pattern that did not fit any of the prevailing climate models. The temperature fluctuations were too large, too rapid, too localized. The standard models predicted a gradual warming with seasonal amplification — warmer winters, longer summers, a slow northward shift in the tree line. The data showed something else: violent oscillations, periods of intense warming followed by periods of intense cooling, patterns that looked almost periodic but were not, that looked almost random but were not, that looked, to anyone who had spent enough time staring at the graphs, like the signature of a system that was oscillating between two states without settling into either one.

Sarah Chen developed two hypotheses. She developed them simultaneously, in the way that a quantum system exists in superposition — both states occupied at once, neither collapsed into the other by the act of observation. She did not choose between them. She did not present them as competing theories. She presented them as coexisting explanations for the same phenomenon, and she asked the data to decide. The data did not decide.

Hypothesis One: The oscillations were natural. They were the signature of a complex atmospheric system that had always been capable of this kind of behavior but had never been observed at this resolution because the monitoring network had been insufficient. The sub-arctic atmosphere was a boundary layer between the Arctic vortex and the mid-latitude jet stream, and boundary layers are, by their nature, unstable. The oscillations represented the natural variability of a system that was being squeezed between two larger systems whose own behaviors were changing in ways that amplified the instability at the boundary. This was not — she emphasized this point in her grant proposals and her papers and her presentations — evidence that climate change was not happening. It was evidence that climate change was more complex, more volatile, more unpredictable, than the models suggested. The direction of the change was still warming. The rate and pattern were the question.

Hypothesis Two: The oscillations were artificial. They were a measurement artifact — not an error in the instruments, which had been calibrated regularly and maintained by Erik Lindstrom with a devotion that bordered on the religious, but an error in the methodology. The act of measuring the sub-arctic atmosphere at this resolution, with this technology, in this geographic configuration, was itself introducing a systematic bias that created the appearance of oscillation where none existed. The radar was scattering signals off ice crystals that should have been filtered out. The temperature sensors were being affected by the heat radiating from the station buildings. The data was not showing natural variability. It was showing the imprint of the measuring apparatus on the system being measured.

Both hypotheses were supported by the data. Both hypotheses were consistent with the existing literature. Both hypotheses had been reviewed by colleagues and found plausible. And both hypotheses could not be true at the same time. One said the oscillations were real. The other said they were not. One said the boundary layer was more volatile than we thought. The other said our instruments were lying to us.

Sarah Chen held both hypotheses in her mind for four years without collapsing the superposition. This was not a philosophical position — she was not a postmodernist who believed that all explanations were equally valid or that truth was a social construct. She was a scientist, and she believed in truth, and she believed that the truth about the sub-arctic atmosphere was a specific thing that existed independently of her beliefs about it. But she also believed in intellectual honesty, and intellectual honesty required her to acknowledge that the data supported two contradictory explanations and that she did not have, and might never have, the information necessary to choose between them.

The isolation of Eielson supported this position. There were no colleagues to perform the social function of observation that collapses the quantum state. There was Priya, who was a brilliant climatologist and a good friend but who believed, firmly and with data to support her belief, in Hypothesis One. There was Erik, who believed, equally firmly, in Hypothesis Two, because the instruments were his children and he could not accept that they were lying. But neither of them could prove their position, and Sarah did not need them to. She needed only the data, and the data was indifferent.

The station operated in a landscape that reinforced Sarah's willingness to hold contradictory ideas simultaneously. Alaska is a place of extreme and coexisting realities. In the summer, the sun does not set for fifty-seven days, and the landscape is bathed in perpetual golden light that makes everything look like a photograph of itself. In the winter, the sun does not rise for fifty-seven days, and the landscape exists in blue darkness broken only by the aurora borealis and the orange glow of the station's exterior lights. Both states are real. Both states occupy the same space. Neither cancels the other out. The landscape does not choose between light and dark. It contains both, in a superposition that the locals accept without question and the visitors find disorienting.

Sarah had come to Alaska from San Francisco when she was twenty-two and had spent six years before that in Cambridge, Massachusetts, and before that in Seattle, Washington, and before that in a small house in the Oregon suburbs where she had learned, from her father, a Tibetan Buddhist monk who had left Tibet in 1980 and found a monastery in Oregon, that the mind is capable of holding multiple contradictory ideas without discomfort if it has been trained to do so. Her father had taught her meditation when she was seven, and she had continued the practice in Boulder and in Eielson and in the metal buildings at the edge of the tree line, and the practice had given her a capacity — or a habit, or a weakness, depending on the day — for occupying two positions at once without collapsing into one.

The winter of 2023-2024 was the most extreme in the station's history. The temperature dropped to minus forty-six degrees Fahrenheit in November and did not rise above minus thirty for eleven weeks. The aurora was visible every night, a curtain of green and purple light that stretched from horizon to horizon and pulsed in patterns that Priya said were related to solar activity and Erik said were beautiful and Sarah said were, in the language of her two hypotheses, either a real phenomenon or an artifact of the human visual system adapting to darkness.

During those eleven weeks, the data continued to flow. Four terabytes per day from the radar alone, plus the temperature and humidity and greenhouse gas measurements, plus the manual observations that Sarah and Priya and Erik recorded in leather-bound notebooks that Erik had ordered from a craftsman in Fairbanks because he did not trust digital storage. The data showed the oscillations — violent temperature swings of ten, fifteen, twenty degrees in the course of a single day, followed by periods of stability that lasted hours and then broken again by another swing, as if the atmosphere were a system that could not make up its mind about what temperature to be.

Sarah analyzed the data with both hypotheses active. When she looked through the lens of Hypothesis One, the oscillations were the signature of a boundary layer in crisis — a system that was being pushed beyond its capacity for stable equilibrium by the competing forces of the Arctic vortex and the jet stream, both of which were changing in response to the warming Arctic. The oscillations were a warning, not of instability for its own sake, but of a system that was approaching a threshold — a point at which the oscillations would stop being oscillations and become a shift, a regime change, a new state that the models did not predict because the models are written by people who assume that the future will resemble the past in its broad patterns, and the sub-arctic atmosphere was demonstrating that the future could be patterned differently.

When she looked through the lens of Hypothesis Two, the oscillations were a reminder of the limits of human knowledge — not the romantic, poetic limits of knowledge, but the specific, technical, instrument-specific limits of a monitoring network that had been designed for a different climate than the one that existed. The radar had been calibrated in 2015 based on the atmospheric conditions of 2015. The temperature sensors had been installed based on thermal models that assumed a gradual warming trend. Neither had been designed for the violence of the oscillations that they were now recording. The data was not lying, but it was not telling the whole truth either. It was telling a truth that was filtered through the assumptions of the instruments that produced it, and those assumptions were wrong, and the oscillations were the imprint of those wrong assumptions on the raw numbers.

Both explanations were true. Both explanations were false. Sarah Chen held them in her mind the way she held the meditation — not as beliefs to be acted on but as states to be observed, occupied, released, and occupied again. The superposition was not a failure of science. It was the most honest position that science allowed, given the data. To collapse the state prematurely — to choose one hypothesis and publish a paper that declared the oscillations real or artificial — would have been, in her view, less honest than maintaining the ambiguity and letting the data accumulate and hoping that, eventually, the evidence would favor one explanation and the superposition would collapse on its own terms.

The outside world was less patient. The Aurora Trust, which funded sixty percent of the station's operations, sent a representative in January 2024 to evaluate the research and recommend continued funding. His name was Gregory Vaughan, and he was a former military officer who had moved into climate funding because he believed that climate change was a security threat and that data was the weapon that would address it. He spent three days at the station, reviewed the data, spoke with Sarah and Priya and Erik, and asked Sarah directly: Which hypothesis do you believe?

She told him that she believed both. He did not understand. She explained. He did not understand better. He wrote a report that said the research was intellectually sophisticated but practically ambiguous, and that the Aurora Trust needed to know whether the oscillations were a real threat or a measurement artifact before they could determine the appropriate response. If they were real, the station's funding should be increased to expand the monitoring network. If they were artificial, the station's instrumentation should be upgraded and the data collection methodology revised. Ambiguity, he wrote, was not a funding strategy.

Sarah read the report and felt, for the first time in four years, the pressure to collapse the superposition. Not from Vaughan — he was doing his job, which was to allocate resources based on actionable information — but from herself. She was a scientist, and scientists are trained to seek clarity, and the clarity that she had avoided for four years was calling to her, in the voice of a former military officer and the voice of her own training and the voice of the data itself, which was accumulating and accumulating and accumulating and might, if she continued to hold both hypotheses, accumulate to a point where it would make the choice obvious.

She did not make the choice. She wrote a response to Vaughan that said the Aurora Trust should fund both directions simultaneously — expanded monitoring to test Hypothesis One, and instrument upgrade to test Hypothesis Two — because the cost of choosing wrong was higher than the cost of pursuing both. Vaughan declined. He reduced the Trust's funding by thirty percent and told her that the remaining funding was contingent on a clear research direction within six months.

Sarah Chen sat in the station on a night in February 2024 when the temperature was minus forty-one degrees and the aurora was a green curtain that stretched from east to west and pulsed in time with nothing she could measure, and she looked at the data and she looked at the two hypotheses and she felt the superposition holding. Not because she was stubborn. Not because she was afraid of choosing wrong. But because the superposition was, in its own way, a kind of truth — not the truth of one explanation or the other, but the truth that the sub-arctic atmosphere could not be described by one explanation or the other, that it occupied a space that was larger than either hypothesis, that the oscillations were either real or artificial or both or neither, and that the most honest thing she could do was to keep measuring, keep observing, keep holding the state open, until the data itself collapsed it or until she died, whichever came first.

She kept measuring. The aurora pulsed. The temperature swung. The data accumulated. And in the metal buildings at the edge of the tree line in interior Alaska, a woman who was thirty-five years old in 2024 and who had come to Alaska from the east coast and to the east coast from the west coast and to the west coast from Oregon and to Oregon from a monastery in the high country where her father had taught her to hold contradictory ideas without discomfort, kept measuring the temperature of air that did not exist and calling the measurements data and calling the data truth and calling the truth a superposition that she would not collapse because the act of collapsing it would be an act of violence against the complexity of the thing she was trying to understand.

The line between real and artificial did not exist. The line between the atmosphere and the instrument that measured it did not exist. The line between the observer and the observed did not exist. And Sarah Chen, holding both hypotheses in her mind like two states of a quantum system that refuses to collapse, occupied the space where the lines did not exist and called it, not truth exactly, but the closest thing to truth that a human mind could reach without reducing the world to something it could manage.

She called it work.