The core sample arrived at Station Tuktu on a Tuesday morning in February, flown in by helicopter from the drilling site eighty-seven kilometers northeast of the station, and Dr. Sarah Koenig received it the way she had received eleven previous cores during her three winters at the permafrost monitoring outpost: with the exhausted, functional competence of someone who had long ago stopped expecting the ice to tell her anything she wanted to hear. The core was a cylinder of frozen earth and ancient ice one meter in length and ten centimeters in diameter, extracted from a depth of fourteen meters, and it contained within its compressed layers approximately thirty-two thousand years of climate history. Sarah's job was to read that history. She was good at her job. She had published seventeen papers in the preceding six years, built her reputation on the precision of her readings and the restraint of her conclusions, and she had learned, over three winters at Tuktu, to separate the data from whatever she felt about the data, the way a surgeon separates the body on the table from whatever she knows about the person it belonged to.

She ran the core through the gas chromatograph at fourteen hundred hours on the day of its arrival. The instrument hummed and clicked and produced its output in columns of numbers on a screen that had been manufactured in 2016 and had developed, over the intervening years, a slight flicker in the lower left quadrant. Sarah read the numbers. She read them again. She printed them and held the printout under the fluorescent light and read them a third time. Then she unplugged the gas chromatograph from the power supply, waited thirty seconds, plugged it back in, recalibrated against the reference standard, and ran the same core through the same machine a second time. The numbers were identical. Both readings. Both valid. Both impossible.

The first reading indicated that the permafrost at a depth of fourteen meters was thawing at a rate of 4.7 centimeters per year, accelerating, and that the methane trapped within the frozen layers was beginning its release into the atmosphere at concentrations that, if extrapolated across the Alaskan and Siberian permafrost zones, predicted a feedback loop of catastrophic proportions within the next eighteen to thirty-six months. The second reading, from the same core, processed by the same instrument on the same afternoon, indicated that the permafrost was thermodynamically stable, that the ice lenses within the core showed no sign of degradation, and that the methane concentrations were consistent with a system in equilibrium, neither releasing nor absorbing, unchanged across the thirty-two thousand years of compressed history the core contained. Both readings were scientifically valid. Both had been produced by identical procedures on identical equipment. Both used the same calibration standards and the same computational models and the same statistical tests for significance. Both could not be true. The permafrost could not be simultaneously collapsing and holding steady. The atmosphere could not be simultaneously at risk and at peace. Sarah Koenig could not simultaneously be the scientist who documented the beginning of the end and the scientist who confirmed that, for now at least, the end was not yet beginning.

She did not tell anyone that night. She placed the printouts in separate folders, labeled them Core 14-A and Core 14-B as though they were distinct specimens, and locked them in the steel cabinet beside her bunk. She ate dinner in the station mess with the four other researchers who wintered at Tuktu. She ate rehydrated beef stew and listened to Dr. Marcus Okafor describe the problem he was having with the barometric pressure sensors on the eastern ridge and she nodded at the appropriate moments and she did not mention that her own data had split into two incompatible realities and that she had no way of determining which one was real.

Over the following week, Sarah attempted to resolve the contradiction. She sectioned the core into ten-centimeter segments and ran each segment separately through the chromatograph. Each segment produced two readings, identical in their contradiction. She shipped a portion of the core to the University of Alaska Fairbanks for independent analysis. The analysis returned with two sets of results, each valid, each irreconcilable. She checked her calibration logs, her maintenance records, her power supply voltages, her software version numbers. Everything was correct. Everything was within specification. Everything was simultaneously conclusive and inconclusive, the way a photon is simultaneously a particle and a wave until someone measures it and forces the universe to choose.

On the eighth day she told Dr. Elias Thorn. She told him because he was the only other person at Tuktu who would not immediately dismiss the contradiction as instrument error or operator fatigue or the cognitive decline that was rumored to afflict researchers who spent too many winters too far north. Elias was a glaciologist from the University of Oslo, forty-one years old, with the kind of quiet, unhurried competence that Sarah had always found more attractive than she was willing to admit, and when she showed him the two sets of data he did not ask her whether she had checked the calibration. He spread the printouts across the table in the data room and studied them for twenty-three minutes and then he looked up and asked her which version she believed.

Which version she believed. Not which version was correct. Not which version the evidence supported. Which version she believed, as though belief were a variable in the equation, as though the permafrost cared what Sarah Koenig thought about it. She said she believed the first reading. She believed the ice was melting. She believed the methane was releasing. She believed the feedback loop was beginning. She believed these things because she had spent her entire career documenting the slow violence of a warming planet and she could not, at this point in her life, accept that the violence had paused. Elias nodded. He said he believed the second reading. He said he believed the permafrost was stable because he had spent his entire career looking for evidence that the systems of the Earth were more resilient than the models predicted, and he had found that evidence often enough to trust that it existed even when he could not point to it directly.

They sat in the data room while the Arctic night pressed against the windows and the generators hummed their steady note beneath the floor. Sarah understood, with a clarity that felt almost physical, that the two readings of the core sample had become the two readings of her life. If the first reading was true, if the ice was melting and the methane was releasing and the feedback loop was beginning, then everything she had worked for was justified and everything she feared was confirmed and the relationship she had been building with Elias Thorn across three winters of shared meals and shared data and shared silence was a temporary arrangement between two people who would soon be consumed by a catastrophe too large for personal connection. If the second reading was true, if the ice was holding and the methane was stable and the system was resilient, then her career had been built on an alarm that was, if not false, at least exaggerated, and the relationship with Elias was not a temporary arrangement but a possible future, a thing that could grow without the shadow of impending collapse darkening every conversation.

She lived the following week in both realities simultaneously. She could not help it. The data had split and her life had split with it. On Monday morning she walked to the eastern ridge to help Marcus Okafor with the barometric sensors and she felt, in the reality where the ice was melting, a grief so total that the white landscape seemed already to be vanishing beneath her boots, the permafrost softening, the tundra releasing its ancient carbon, the world tilting into an irreversible cascade. On Monday afternoon she ate lunch with Elias and she felt, in the reality where the ice was stable, a cautious optimism that she had not experienced since her first winter in the Arctic, a sense that the future might actually arrive, that there might be time for things besides data and papers and the slow accumulation of evidence for a conclusion no one wanted to reach.

On Tuesday she worked on a paper draft and typed two different introductions for two different papers that would reach two different conclusions from the same core sample, and she did not delete either one. On Wednesday she walked with Elias to the weather station on the north ridge, a three-kilometer trek across the packed snow, and in one reality they walked in silence because the melting ice made conversation feel frivolous, and in the other reality Elias told her about his daughter in Oslo, about the last summer they had spent together before he came to Tuktu, about the way she had asked him whether the ice would still be there when she was his age, and Sarah listened and answered and felt the shape of a possible life forming in the space between their words.

On Thursday she received the formal report from Fairbanks. Two reports, technically, bound in a single folder with a cover letter from the department chair expressing confusion and concern and a request for additional samples. The two reports agreed on nothing but their own validity. Sarah placed the folder in the steel cabinet with the others. She had twelve core samples now, each with two readings, twelve pairs of contradictory truths stacked in a cabinet in a research station at the edge of the habitable world. She did not know whether the cabinet contained the evidence for an apocalypse or the evidence for a reprieve. She did not know whether she was a prophet of catastrophe or a technician of normalcy. She did not know whether the man in the bunk next to hers would be her partner in survival or her companion in the final months before the systems failed.

On Friday night Elias came to her bunk. The station was quiet. The generators hummed. The Arctic darkness pressed against the windows with the weight of a depth the sunlight would not touch for another month. He sat on the edge of her bunk and he did not touch her. He said he had been thinking about the two readings, about the impossibility of the data and the impossibility of their response to it. He said that he had realized something. In the reality where the ice was melting, he said, he wanted to spend whatever time remained with her, because even a catastrophe was better faced with someone who understood the machinery of its arrival. And in the reality where the ice was stable, he said, he wanted to spend whatever time was available with her, because a future that held was a future worth sharing. He said the conclusion was the same in both realities. The data did not matter. The permafrost did not matter. What mattered was that they had been given two truths and that both truths led to the same decision.

Sarah looked at him in the dim light of the bunk and she understood that he was wrong. The conclusion was not the same. The data did matter. In one reality she would love Elias Thorn in the shadow of an ending, every touch freighted with the knowledge of what was coming, every word a valediction, every moment of happiness a temporary reprieve from a grief that would eventually consume everything. In the other reality she would love him in the light of a continuing, every touch a beginning rather than an ending, every word an opening rather than a closing, every moment of happiness a down payment on a future that would actually arrive. The two loves would feel different. They would mean different things. They would be different experiences, different memories, different lives.

She told him this. She told him that she needed to know which version was true. She told him that belief was not enough, that choosing to believe one reading over the other was not the same as knowing, that she could not live the same life with two different meanings because meaning was the only thing that distinguished one day from another, one love from another, one life from a different life. Elias listened. The generators hummed. The Arctic darkness held its breath. And then Elias said something that Sarah would return to for the rest of her time at Tuktu and long after. He said that perhaps the ice knew both things. Perhaps the permafrost was simultaneously melting and stable, releasing and holding, ending and continuing. Perhaps the contradiction was not a failure of measurement but a property of the system itself. Perhaps the universe did not choose between the two possibilities because both were required for either to have meaning. Perhaps a love that did not know whether it was a farewell or a hello was the only kind of love that was honest about its own uncertainty.

Sarah did not answer. She did not know how to answer. She did not know whether Elias was offering her a philosophy or a proposal. She did not know whether she could live in superposition, whether she could hold two contradictory truths in her mind simultaneously without one collapsing into certainty under the weight of her own need for resolution. She did not know what the ice knew. She knew only that the core sample sat in the steel cabinet with its two readings, that the methane was either rising or stable, that the future was either diminishing or holding, and that Elias Thorn was sitting on the edge of her bunk in the permanent night of an Arctic February waiting for an answer that she could not give because the answer depended on which question was being asked and she could not tell the questions apart.

She took his hand. It was an answer of sorts, or a postponement of the question, or an acknowledgment that some decisions were made not by the mind but by the body, which knew things about connection that the mind could only analyze and never understand. They sat together in the dim light of the bunk while the generators hummed and the station settled into the small sounds of its own maintenance, and Sarah Koenig felt, simultaneously, the grief of the ending and the hope of the continuing, the weight of the catastrophe and the lightness of the reprieve. She did not know which feeling was true. She did not know whether truth was the right word for what she was feeling. She knew only that she was holding a man's hand at the edge of the habitable world and that the ice beneath them, thirty-two thousand years deep, was either melting or holding, and that she would spend the rest of her life learning to live without knowing which.