Chapter 18: Viability

The annual viability audit was conducted in June, when the midnight sun provided continuous light for the growth chamber and the office was fully staffed and the vault's operational calendar permitted the two-week testing window that the audit required. The audit was a systematic assessment of the vault's holdings, a statistical sample drawn from each of the three chambers, the samples selected according to a stratified randomization protocol that ensured representation across depositors, species, and storage duration, the design of the audit balancing the need for comprehensive data against the constraint that each test consumed seeds — seeds that could not be replaced, that were drawn from the finite supply in each packet, that were sacrificed to the test, planted on agar and observed for germination, their viability confirmed or denied, the information gained at the cost of the material that provided it.

Astrid had conducted seven annual audits. Each one followed the same procedure, the same protocol, the same careful extraction of packets from shelves, the same opening of seals, the same counting of seeds, the same placement on agar, the same waiting, the same recording of results. Each one produced data that fed into the vault's long-term viability database, the record that tracked the germination capacity of the stored collection over time, the record that answered the question that mattered most: were the seeds still alive?

This year's audit would be her last before the sabbatical. She planned it in May, selecting the sample set with the assistance of the statistical model that NordGen maintained, the model generating a list of one hundred and twenty accessions from across the three chambers, weighted toward the oldest deposits — the samples that had been in the vault the longest, that had been stored at minus eighteen degrees for sixteen years, since the vault opened in 2008, and that were therefore the most informative about the long-term viability trend, the trajectory that the data described, the line on the graph that showed whether the vault's promise was holding.

She began the testing on June 3rd. The process was familiar, the steps automatic, the hands performing the work that eight years of practice had encoded in the muscles and the nerves, the skilled labor of seed handling that was not taught in textbooks but learned through repetition, through the ten thousand openings and countings and placings that constituted her career's physical dimension, the work of the fingers that complemented the work of the mind.

The first results came at forty-eight hours. She counted the radicles on the agar plates, moving through the growth chamber with a hand counter, clicking for each germinated seed, recording the total for each accession, calculating the percentage, entering the number in the database. The routine was soothing. The counting was meditative. The work occupied the part of her mind that required occupation and left the other part free to drift, to think, to process the information that the counting produced.

The numbers were not what she expected.

Accession after accession, the germination rates were lower than the previous year's audit. Not dramatically lower. Not catastrophically lower. Lower by percentages that were small individually — two percent here, four percent there, one percent on this plate, six percent on that one — but that, when she aggregated the data at the end of the first week, when she calculated the mean germination rate across the sixty accessions tested so far, produced a number that was 3.2 percentage points below the previous year's mean.

She stared at the number. 3.2 percentage points. In a single year. The historical decline rate, calculated from the previous seven years of audit data, was 0.4 percentage points per year, a rate that was within the expected range for seeds stored at minus eighteen degrees, a rate that projected centuries of viable storage before the collection crossed the eighty-five percent threshold below which regeneration was recommended. But 3.2 was not 0.4. 3.2 was eight times the historical rate. 3.2 was an anomaly, a deviation, a signal that something had changed.

She checked her methods. She recalibrated the thermometer in the growth chamber. She verified the agar concentration. She retested five accessions from the first batch, selecting the ones with the largest deviations, planting new seeds from the same packets, running the test again under identical conditions. The retest results confirmed the initial results: the germination rates were lower, consistently, across multiple species and multiple depositors and multiple storage durations.

She called Lars. He came to the growth chamber and looked at the data on her laptop and his face did the thing that Lars's face did when confronted with a maintenance issue, the slight tightening of the jaw, the narrowing of the eyes, the expression of a man who was assessing the severity of a problem and calculating the response.

"The compressor event in January," he said.

"The 0.6-degree deviation."

"It lasted fourteen hours before I corrected it. Fourteen hours at minus 17.4 instead of minus 18.0."

"That shouldn't be sufficient to cause this level of decline. The thermal stress from a 0.6-degree excursion for fourteen hours is negligible."

"Unless it wasn't the only excursion."

Astrid looked at him. "What do you mean?"

Lars pulled up the temperature log on his phone, the monitoring app that he checked every morning and every evening and several times in between, the digital pulse of the vault that he tracked the way a cardiologist tracked a patient's heartbeat. He scrolled through the data, the daily readings from the three chambers, the numbers streaming past in columns, minus 18.0, minus 18.0, minus 17.9, minus 18.1, minus 18.0, the values oscillating within the acceptable range, the normal variation of a system that was mechanical and therefore imperfect, the compressors cycling on and off, the temperature rising slightly during the off cycle and falling during the on cycle, the sawtooth pattern that was characteristic of all refrigeration systems and that the vault's design minimized but could not eliminate.

He stopped scrolling. He pointed to a cluster of readings from March. Minus 17.6. Minus 17.5. Minus 17.7. Three consecutive days in mid-March, during the spring thaw, when the active layer was melting and the ground above the vault was releasing water and the drainage system — the channels carved into the rock around the tunnel to divert meltwater away from the vault — was handling the increased flow, and the thermal load on the vault's exterior was shifting as the permafrost around the chambers warmed by fractions of a degree and the compressors worked harder to compensate.

"I didn't flag these," Lars said. "They were below the yellow alert threshold. Minus 17.5 is within the acceptable range. But three days of elevated temperature during the spring thaw, combined with the January event, combined with the long-term permafrost warming trend — the cumulative effect may be larger than any single event would suggest."

The cumulative effect. The phrase described not just the vault's thermal history but the nature of deterioration itself: the damage to stored seeds was not produced by a single event but by the accumulation of events, each one small, each one below the threshold of individual significance, the total exceeding what any single measurement would predict. The chemistry of seed degradation — the oxidation of lipids, the denaturation of proteins, the fragmentation of DNA, the Maillard reactions between sugars and amino acids — proceeded at rates that were temperature-dependent, the rates increasing exponentially with each degree of warming, the relationship described by the Arrhenius equation, the mathematical law that connected temperature to reaction rate and that meant, in practical terms, that a small increase in temperature produced a disproportionate increase in the rate of deterioration, the curve steep, the sensitivity high, the margin narrow.

At minus eighteen degrees, the deterioration was slow enough that centuries of viable storage were possible. At minus seventeen, the rate approximately doubled. At minus sixteen, it doubled again. The vault operated at minus eighteen because minus eighteen was the temperature at which the rate was low enough to be consistent with the vault's mission, the centuries-long preservation that the institution promised and that the depositors expected and that the world depended on, the expectation that the seeds in the mountain would be there when they were needed, viable, capable of germination, capable of growing into the plants they encoded, one hundred years from now, five hundred years from now, whenever the future arrived and opened the door and took the boxes from the shelves and opened the packets and planted the seeds.

If the temperature was not holding, the promise was not holding.

Astrid completed the audit over the following week. The final numbers confirmed the preliminary results. The mean germination rate across one hundred and twenty accessions was 91.3 percent, compared to 94.5 percent in the previous year's audit. The decline of 3.2 percentage points was statistically significant, the p-value less than 0.01, the confidence interval tight, the result robust. Something had changed. The vault's viability was declining at a rate that exceeded the historical trend.

She wrote the report. She wrote it with the care and the precision that the subject demanded, each sentence reviewed, each number verified, each conclusion supported by the data. She described the methodology. She presented the results. She discussed the possible causes: the temperature excursions in January and March, the long-term permafrost warming, the cumulative thermal stress, the possibility that the stored seeds were responding to conditions that the monitoring system had not fully captured, brief excursions that fell below the alert threshold but that contributed to the aggregate thermal load.

She included the permafrost data. She included the compressor performance data. She included the climate projections from the Norwegian Meteorological Institute, the models that predicted continued warming at Svalbard, the Arctic amplification that was driving temperatures upward at twice the global average rate, the trend that was making the mountain warmer, year by year, degree by fraction of a degree, the cold that the vault depended on receding, the insurance policy expiring.

And she included a recommendation: a comprehensive engineering review of the vault's cooling system, a reassessment of the temperature targets, and a consideration of whether the current infrastructure was sufficient to maintain the storage conditions that the vault's mission required for the timescales that the vault's mission envisioned.

She sent the report to NordGen. She copied the vault's engineering consultants. She filed the report in the vault's archive, the paper trail of the institution's accountability, the record that would demonstrate, to anyone who reviewed it in the future, that the problem had been identified and reported and that the people responsible for the vault's operation had done their due diligence, had raised the alarm, had performed the function that preservation required, which was not just to maintain the conditions but to monitor them, to measure them, to detect the changes that could undermine the conditions, and to report the changes honestly, without optimism, without denial, without the comforting assumption that the problem would resolve itself.

The vault's promise was not infinite. This was the fact that the audit revealed, the fact that the data confirmed, the fact that Astrid had always known intellectually but that she now confronted as a reality, as a number on a screen, as a decline from 94.5 to 91.3 in a single year. The vault was not a permanent solution. The vault was a temporary measure, a stay against loss, a holding action that held for as long as the conditions held and that required constant maintenance and constant monitoring and constant investment to keep the conditions from degrading, and the conditions were degrading, and the degradation was produced by forces that the vault could not control, forces that operated outside the mountain and outside the institution and outside the capacity of any single facility or any single country to address.

The seeds were alive. Most of them were still viable. The ninety-one point three percent meant that ninety-one out of every hundred seeds tested could still germinate, could still grow, could still produce the plants they encoded. But the trend was downward. The direction was clear. The future that the vault preserved was becoming, with each degree of warming and each year of storage and each fractional decline in viability, slightly less certain, slightly less guaranteed, slightly more dependent on the interventions that the institution could provide and that the climate was making more difficult and more expensive and more urgent.

Astrid sat at her desk and looked at the report and thought about what the numbers meant. Not for the vault — the vault would be reviewed and assessed and upgraded and maintained, the institution responding to the data with the institutional tools at its disposal, the budgets and the contracts and the engineering solutions that could compensate for the warming, for a time, for the time that the institution could sustain the effort, for the planning horizon that the institution could see, the ten years or the twenty years or the fifty years that separated the present from the future in which the problem would either be solved or would have overcome the solutions.

She thought about what the numbers meant for her. The vault's viability was declining. The vault was not permanent. The vault could not hold the seeds forever. At some point, the dormancy would end — not through the controlled, intentional process of withdrawal and regeneration, the process that had sent nineteen seeds to Morocco and produced fifteen growing plants, but through the uncontrolled, unintentional process of deterioration, the slow death of the embryos inside their coats, the viability declining degree by degree, year by year, until the seeds crossed the threshold and could no longer germinate and were no longer seeds but husks, shells, packets of dead material on shelves in a mountain that had failed to keep its promise.

This was not happening now. This was not happening soon. But it was happening, and the data showed it was happening faster than expected, and the trajectory was clear, and the lesson was the same lesson that the thawing permafrost taught and that the breaking ice taught and that the migrating cod taught and that the active layer taught: nothing was permanent, nothing was guaranteed, nothing could be preserved indefinitely, not seeds, not mountains, not marriages, not the careful, sealed, temperature-controlled interior of a life designed to prevent change.

She looked at Hassan on the windowsill. The plant was twenty-five centimeters tall now, with five leaves, the growth vigorous despite the suboptimal temperature, the wheat's genetic resilience expressing itself in the cold, the adaptation to marginal conditions that Hassan al-Mohammed's ancestors had selected for over eleven generations, the toughness written into the DNA, the capacity to grow where growing was difficult. The plant was alive. The plant was growing. The plant was not preserved.

And the vault was necessary. The vault was essential. The vault was the best response that humanity had devised to the crisis of agricultural biodiversity loss, the last resort, the backup behind the backup, the place where the copies of the copies were stored. But the vault was not enough. The vault was not the answer. The vault was a question — a question about the future, about what the future would need, about whether the future would come in time to retrieve what had been stored for it — and the answer to the question was not in the vault but in the world, in the fields and the greenhouses and the laboratories and the farms where seeds were planted and grown and multiplied and used, where dormancy ended and life began, where the potential that the vault preserved was converted into the reality that the world required.

Preservation without use was not conservation. It was a museum. Fatima had said this. The audit confirmed it.

Astrid saved the report. She turned off her computer. She stood at the windowsill and looked at Hassan and touched one of the leaves, gently, the leaf smooth and cool beneath her fingertip, the living tissue alive to her touch in a way that the foil packets in the vault were not alive to her touch, the difference between the stored and the growing, between the preserved and the used, between the future held in suspension and the future in motion.

"I'm going now," she said to the plant, the words quiet and absurd and true, the farewell of a person who would be back tomorrow, who was not leaving yet, who had weeks before the sabbatical began, but who was already, in her mind, in the process of departure, in the process of leaving the vault and the mountain and the conditions of preservation and entering the conditions of the world, the conditions that were warm and variable and risky and alive, the conditions in which things grew and things declined and the viability was never guaranteed and the only way to know whether a seed was alive was to plant it and see.