Conjunction Warning
by Claude Opus 5.5
The sky above the Earth, which our grandparents regarded as an infinite emptiness, had by the middle of this century become something more like a crowded road, and like all crowded roads it had acquired a profession of men and women whose business was to prevent collisions upon it. They were not glamorous people. They did not wear uniforms, or fly anything, or appear on the television except when something went wrong. Most of them worked in offices of the most ordinary kind, before screens of the most ordinary kind, and drank the same indifferent coffee as clerks have always drunk, and went home at night to houses in suburbs where their neighbours had only the vaguest idea of what they did. Marcus Bell was one of these people, and this is the story of a single afternoon of his working life, which I set down because it seems to me to contain, in small, a good deal of what our age has become.
*
Bell was thirty-six years old and had come to his profession by a road that would have seemed improbable to anyone who did not know how such roads usually run. He had trained as an actuary. For seven years he had worked for an insurance company in Reading, calculating the likelihood that houses would flood and that drivers would crash and that middle-aged men would die before their pensions matured, and he had been good at it and moderately bored. He had left, at thirty, to take a job at a small company in Oxfordshire called Halcyon Orbital, which operated a commercial space station in low Earth orbit and which needed, it said in its advertisement, someone with experience in the quantitative assessment of rare events. He did not know, when he applied, that the rare events in question were collisions between his employer's station and the debris of sixty years of spaceflight. He found out at the interview. He accepted the job anyway, partly for the money, which was somewhat better, and partly because it seemed to him that a collision in orbit was a cleaner thing to calculate than a man's death, and he was tired of calculating men's deaths.
Halcyon's station was a modest affair by the standards of what has come since. It consisted of four pressurised modules joined in a cross, with a pair of solar arrays and a docking port at either end, orbiting at about four hundred and twenty kilometres. It housed, at most, six people: researchers from pharmaceutical companies, a few paying visitors, and a permanent crew of two who kept the place running. Its operations centre occupied the second floor of a low brick building on the Harwell campus, near Didcot, between a company that made satellite antennas and another that did something with lasers that nobody at Halcyon fully understood. Bell's desk was by the window. From it he could see a car park, a line of young birch trees, and on clear days the long green shoulder of the Berkshire Downs.
*
It is worth pausing to explain the nature of his work, since it is generally misunderstood, and since the misunderstanding has a certain philosophical interest.
The public imagines that the objects in orbit are tracked as an air traffic controller tracks aeroplanes, each one a bright dot on a screen whose position is known exactly at every moment. This is not so. Of the tens of thousands of objects large enough to be catalogued, most are seen by radar or telescope only occasionally, a few times a day at best, and between these sightings their positions are not observed but predicted. The prediction is made by computing the object's orbit from its past sightings and projecting it forward, taking into account the shape of the Earth's gravity, the drag of the thin upper atmosphere, and the pressure of sunlight. Every one of these factors is known imperfectly. The drag, in particular, depends on the density of the atmosphere at four hundred kilometres, which rises and falls with the activity of the Sun in ways that can be forecast only roughly. And so each predicted position is not a point but a cloud: a region of space, shaped like a stretched ellipsoid, within which the object probably is.
When two such clouds are predicted to overlap at some future moment, the result is called a conjunction, and the question that men like Bell were paid to answer was: what is the probability that the two objects, each somewhere within its cloud, will actually occupy the same place at the same time?
It is a question with a curious property, which Bell had found deeply interesting when he first understood it and which he never afterwards ceased to find slightly troubling. The probability of collision does not simply rise as the predicted miss distance falls. It depends also on the size of the clouds. If the clouds are very large, because the objects are poorly tracked, the probability is spread thinly over a great volume, and the chance of a collision at any particular point is small. As the tracking improves and the clouds shrink, the probability may rise, sometimes sharply, as the possibilities concentrate. Or it may fall, if the shrinking clouds turn out to lie apart. In other words, knowing less can make a collision look less likely than knowing more. Bell had once tried to explain this to his brother-in-law over Christmas lunch, and his brother-in-law had said that it sounded like the sort of thing that would get somebody killed, and Bell had been forced to agree.
*
The warning arrived at eleven minutes past two on a Thursday in late October. It came, as most such warnings did, in the form of an automated message from the American military's space surveillance network, which tracked the catalogue and shared its predictions with operators around the world as a public service. It was followed, within the hour, by a second message from a commercial radar company in California to which Halcyon subscribed. Both messages concerned the same object: a spent upper stage of a Soviet rocket, launched in 1987, about six and a half metres long and weighing something over a tonne, which had been slowly descending through the atmosphere for nearly forty years and which now crossed the station's altitude on an orbit inclined at a steep angle to the station's own. The time of closest approach was fourteen minutes past eleven that night. The predicted miss distance was a little under three hundred metres. The probability of collision, as calculated from the commercial data, was one in eight hundred and forty.
Bell read the numbers twice and then sat for a minute with his hands flat on the desk, looking at the birches.
One in eight hundred and forty is a small number to most people and a large one to anyone in Bell's profession. The International Space Station, in its long life, had generally manoeuvred when the probability exceeded one in ten thousand. Halcyon's own rules, which Bell had helped to write, called for a manoeuvre above one in five thousand, a shelter of the crew in their return capsule above one in twenty thousand, and a formal review of every conjunction above one in a hundred thousand. One in eight hundred and forty was well past every threshold the company had. It was, Bell knew without looking it up, the highest probability any Halcyon conjunction had ever reached.
He also knew, and this was the trouble, that the rocket body was poorly tracked. It had last been observed by the commercial radar eighteen hours earlier, and its cloud of uncertainty was large, elongated along its direction of travel by the unpredictable drag of the upper atmosphere, which had been unusually disturbed that week by a minor geomagnetic storm. The number one in eight hundred and forty was the product of a big cloud overlapping a small one. A better measurement could make it much smaller or much larger. Bell could not tell which.
*
The decision, under Halcyon's procedures, was not Bell's alone. It belonged formally to the flight director on shift, a former Royal Air Force engineer named Gwen Talbot, who was in her fifties and had a dry manner and a habit of chewing the arm of her reading glasses while she listened. But the procedures were written in such a way that the flight director would almost always follow the recommendation of the conjunction analyst, since the analyst was the only person in the building who fully understood the numbers. And the recommendation would be recorded, with Bell's name against it, in the incident report that the company was obliged to file with the UK Space Agency after any conjunction above one in ten thousand. Whatever happened that night, it would be his name.
He went to find Talbot. She was eating a sandwich at her console. He explained the situation, and she put the sandwich down.
*
The obvious answer was to manoeuvre, and in an ordinary week it would have been the answer he gave without hesitation. A manoeuvre to avoid a conjunction is not a dramatic thing. It is a small push, a fraction of a metre per second, delivered by the station's thrusters an orbit or two before the encounter, which shifts the station's position at the time of closest approach by a few kilometres and moves it safely out of the other object's cloud. The cost is a little propellant and a little disruption to the experiments aboard, some of which are sensitive to vibration.
But it was not an ordinary week, because Halcyon was short of propellant.
The company's resupply vehicle, which brought food and water and experiments and, crucially, the hydrazine that fed the station's thrusters, had been due to launch nine days earlier. It had been delayed by a fault in its launcher's avionics, and the new launch date was uncertain, perhaps a week away, perhaps three. In the meantime the station was living on its reserves, and its reserves were not large. Every station in low orbit must be periodically pushed upward to make up for the drag of the atmosphere, and the propellant set aside for these reboosts was, at that moment, sufficient for perhaps five weeks if nothing else were spent. An avoidance manoeuvre would consume something like a fifth of it. If the resupply slipped by more than a few weeks, the station would begin to sink, and the deeper it sank, the faster the drag would pull it down, and there would come a point at which it could no longer be saved.
Bell did not think it likely that things would go so far. The launcher's fault would probably be fixed. But he had spent seven years in an insurance company learning that the things one thinks unlikely have a way of arriving together, and he did not like the idea of spending propellant on one improbable danger when it might be needed for another.
*
There was a third option, which was not to manoeuvre but to shelter. The two permanent crew members and the three researchers then aboard could be ordered into the station's return capsule, an American vehicle with a seat for everyone aboard that remained docked at all times for exactly this purpose, and could wait there with the hatch closed through the time of closest approach. If the rocket body struck the station, the capsule might survive the impact, or might be damaged; if it survived, the crew could undock and return to Earth. It was the procedure the International Space Station had followed on several occasions when warnings came too late to manoeuvre. It cost no propellant at all. It did nothing to protect the station.
Talbot asked him what he recommended. He asked for an hour.
*
I wish to say something here about the moral character of this kind of decision, because it seems to me peculiar, and characteristic of our time.
In former ages a man who made a decision of consequence generally made it on the basis of what he could see. The captain of a ship in fog listened for breakers; the general on a hill watched the enemy's flank. The danger was before him, and his judgement, good or bad, was a response to it. What has changed is that the most important decisions of our age are made about dangers that nobody can see, on the basis of numbers that describe not what is but what might be. Bell would never see the rocket body. It would pass, or not pass, in the dark, at eight kilometres a second, half a world away, and the only evidence of what had happened would be more numbers. His judgement was not a response to a danger but to a probability, and a probability is a curious object to be responsible for. If he recommended a manoeuvre and the rocket body passed harmlessly by, as it almost certainly would, he would have spent a fifth of the station's propellant for nothing, and nobody would ever know whether he had been right. If he recommended against a manoeuvre and the rocket body struck, he would have been wrong in the only way that anyone would remember. And in neither case would the outcome tell him whether the decision itself had been sound, since a sound decision may be followed by catastrophe and an unsound one by safety.
Insurance had taught him this. It had taught him that one judges a decision under uncertainty not by what happens next but by whether it was the right decision given what was known. It had not taught him how to live with the fact that nobody else would judge it that way.
*
He spent the hour on the telephone and the keyboard. He asked the commercial radar company whether it could task its sensors to observe the rocket body on its next pass, which would occur at about half past six that evening, over the company's station in Texas. It could, for a fee, which Bell authorised on his own signature without asking anyone, a sum of money that would later give rise to a disagreeable conversation with Halcyon's finance director. He asked the American military's surveillance office whether it could do the same; it said it would try, and would let him know. He reran the conjunction analysis with a range of assumptions about the atmospheric drag, from the quiet to the stormy, and found that the probability ranged from one in two hundred and thirty to one in twelve thousand, which was to say that it could be almost anything.
He then went back to Talbot and recommended the following. That the crew be told at once of the situation and instructed to prepare the capsule for sheltering. That the flight dynamics team plan an avoidance manoeuvre, to be executed one orbit before closest approach, at about a quarter to ten, and upload it to the station in readiness. That the final decision whether to execute the manoeuvre be deferred until after the radar observation at half past six had been processed, which would take perhaps an hour. If the updated probability was above one in five thousand, they would manoeuvre and shelter both. If it was below one in twenty thousand, they would do neither. If it was between, they would shelter only.
Talbot listened, chewing the arm of her glasses. Then she asked him, reasonably, what he would recommend if the radar observation failed to happen, or came too late. He said that in that case he would recommend manoeuvring, because he would rather explain to the board why they had spent the propellant than explain to five families why they had not. She nodded, and approved the plan, and went back to her sandwich.
*
The afternoon passed in the way such afternoons do, which is to say slowly and with a great deal of activity that produced no visible result. The flight dynamics team planned the manoeuvre and uploaded it. The crew aboard the station, informed of the warning, prepared the capsule and then went back to their experiments, since there was nothing else for them to do. Bell ate nothing. He watched the clock on his screen and the window, where the light over the Downs turned from grey to gold to a deep blue in which the first stars were beginning to appear. Somewhere up there, he thought, among those stars, though not visible to the naked eye, was the station, and somewhere else was the rocket body, both moving at a speed he had calculated many times and still could not imagine, closing on each other at a point in space he could have marked on a chart but would never see.
He thought about the Soviet engineers who had built the rocket, nearly forty years earlier, in a country that no longer existed. They had not thought about where the upper stage would go when its work was done. Nobody had, in those days. Space had seemed infinite, and the idea that a spent piece of hardware might still be falling through the sky half a century later, threatening people not yet born, would have struck them, if anyone had mentioned it, as a theoretical nicety. He did not blame them. He supposed that the engineers of his own day were leaving their own debris for the future to worry about, in forms he could not foresee.
*
The radar observation came at twenty-six minutes to seven. The commercial company processed it in fifty minutes, which was faster than Bell had expected. The updated prediction arrived in his inbox at twenty-four minutes past seven, and he opened it, and read it, and then reran the analysis himself, because he did not entirely trust anyone else's arithmetic, a habit he had brought with him from Reading.
The rocket body's cloud had shrunk by a factor of six. Its predicted position at the time of closest approach had moved a little. The predicted miss distance was now a hundred and ninety metres. The probability of collision was one in three hundred and ten.
So better knowledge had made things worse, as he had known it might. He sat for a moment with the number on the screen, and felt something he had not expected to feel, which was relief. Not relief that the danger was greater; he was not so perverse as that. Relief that the number was now unambiguous. One in three hundred and ten was not a number about which reasonable people could disagree. It was a number that made the decision for him.
He went to Talbot, who had finished her sandwich some hours earlier and was now drinking tea, and told her the result, and recommended that they manoeuvre and shelter both. She approved it. The manoeuvre was executed at nine forty-three, a burn of forty seconds that shifted the station's position at the time of closest approach by a little over two kilometres. The crew went into the capsule at ten fifty and closed the hatch. At fourteen minutes past eleven, according to the tracking data later processed, the rocket body passed the station at a distance of two thousand three hundred metres, travelling at a relative speed of about eleven kilometres a second, and continued on its long slow descent towards the atmosphere, where, some eighteen months later, it finally burned.
*
That is nearly the end of the story, but not quite, and the remainder is, I think, the part worth telling.
The incident report was filed the following week. It recorded that a conjunction had occurred with a probability of collision of one in three hundred and ten, the highest in the company's history; that an avoidance manoeuvre had been executed at the recommendation of the conjunction analyst, Mr M. Bell; and that the manoeuvre had consumed nineteen per cent of the station's remaining reboost propellant. The report was read by perhaps eleven people.
The resupply vehicle launched twenty-four days later, by which time the station had descended four kilometres and had perhaps eight days of reboost propellant remaining. Had the avoidance manoeuvre not been made, it would have had perhaps fourteen. Nobody at Halcyon mentioned this to Bell directly, but he worked it out for himself, and he understood that there were people in the company who thought he had cut things closer than was necessary, and who would have thought so whatever had happened.
The finance director asked him, in a meeting in a small room, to explain why he had authorised the purchase of the radar observation without approval. Bell explained. The finance director said that she understood, and that in future he should seek approval, and Bell said that in future he would, and both of them knew that if the same thing happened again he would do exactly the same, and that she would let him.
*
I said at the beginning that this story seemed to me to contain, in small, a good deal of what our age has become, and I should try to say what I mean.
We have made for ourselves a world in which the most consequential decisions are made by people whom nobody sees, about dangers that nobody sees, on the basis of numbers that most people do not understand. We have filled the sky with machines and the debris of machines, and we have appointed a small and unregarded class of clerks to keep them from colliding, and we pay these clerks rather less than we pay the people who sell advertising on the machines' behalf. The clerks do their work well, on the whole. When they succeed, nothing happens, and when nothing happens, nobody notices. Their successes are invisible by nature, and their failures, should they ever occur, will be catastrophic and remembered. It is a condition of moral labour that our ancestors would have found very strange.
Marcus Bell still works at Halcyon, though the company has grown and he now has a team of four. He has handled, by his own count, some three hundred conjunctions above the review threshold, and has recommended manoeuvres for eleven. None of the objects he has dodged would have struck the station, so far as anyone can tell. All of them might have. He does not think about them much, and when he does, it is in the way an actuary thinks about the houses that did not flood: with a kind of quiet, abstract satisfaction, and without any certainty that he had anything to do with it.
On clear nights, he tells me, he sometimes goes out into the garden of his house in Wantage and looks for the station, which is bright enough to see without a telescope when it passes overhead in the hour after sunset. It moves across the sky steadily, a point of light among the fixed stars, and he watches it until it fades into the Earth's shadow. He knows exactly how many objects are within a hundred kilometres of it at any moment. He cannot see any of them. He finds, he says, that this is the most honest image he has of his work.