1 Nine Tracklets
JPL, Pasadena · Wednesday, 22 March – Wednesday, 29 March 2028.
| tracklets | 9 Catalina 3 | Pan-STARRS 1 2 | ATLAS Mauna Loa 2 | NEO Surveyor (comm.) 2 |
| arc | 2028-02-18.11 to 2028-03-21.13 UTC 32.0 d |
| V | 17.9 +/- 0.3 (ground surveys only) |
| elong | 61.4 deg falling |
| orbit | a 1.0003 au e 0.071 i 9.2 deg |
| region | Earth co-orbital, trailing (L5) |
| fit grav | rms 0.71" chi2/dof 3.8 |
| fit +NG | rms 0.29" chi2/dof 0.9 |
| A1 = (+0.1 +/- 0.7) A2 = (-2.0 +/- 0.8) A3 = (0.0 +/- 0.6) e-9 au/d^2 | |
| class | TROJAN? NONGRAV? |
The question marks were hers. Dana Okafor had written the classifier to hedge when it didn’t know, and at four in the morning, on cores nobody had assigned it, it had hedged twice.
She read the report standing up, coat still on, while the marine layer sat on the parking structure and the towers on Mount Wilson came and went in it. Then she sat down and read it again.
The linker was a side project with no charge number. Every night it went through the Minor Planet Center’s file of isolated tracklets, the short strings of detections no survey had tied to anything, and tried to tie them to each other. The production linkers assumed gravity alone, because gravity alone was right nearly every time and each extra term multiplied the false links. Hers freed the three Marsden terms, the nongravitational accelerations everyone fit for comets, before it gave up on a link. In two years that had caught four dark comets, all already known, and about three hundred false linkages she’d had to kill by hand. False linkages were three tracklets from one survey, glued together by one bad night. This was nine from four.
She took her coat off.
By ten the fog had burned back to the foothills and she had the Catalina images open.
The first detection was from the eighteenth of February, in the twilight sweep the Mount Lemmon telescopes ran for comets near the Sun, where the sky was bright, the window was short and nobody else bothered to look. That was why nobody had linked it. The rest was the orbit.
She ran it gravity-only and plotted the residuals against time. They didn’t scatter. They bent, low at both ends and high in the middle, a smooth arc along the track and nothing across it. She freed the Marsden terms and the bend went away. A1 and A3 came back zero. A2 came back where the overnight run had put it, −2.0 ± 0.8, negative. Something was braking it.
Every first-year mechanic knew the sign convention, and her gut still wanted it to go the other way: brake something in orbit and it drops inward and pulls ahead. This one was about four hundred kilometers in front of where gravity put it, a little over half an arcsecond, which was the bend.
Two and a half sigma on thirty-two days at sixty degrees elongation was how people humiliated themselves in Icarus.
She started trying to kill it.
She dropped each tracklet in turn and refit, nine times, and A2 wandered between −1.4 and −2.6 and never crossed zero. She threw out the two Surveyor tracklets, since a spacecraft still in commissioning was the likeliest place for a timing error, and the error bar grew and the value stayed. She fit each survey against the others in case one station’s clock was off. She split the tracklets every way that left three on a side, in case it was two faint rocks sharing a month of sky. One orbit, one bend.
V of 17.9 at that distance and phase put the absolute magnitude near 16: two kilometers at a typical albedo, more if it was dark. The Yarkovsky push on two kilometers of rock was five orders of magnitude too small. Whatever was slowing it down, it wasn’t sunlight.
Which left outgassing, and outgassing left the dark comets.
There were a dozen or so in the literature now, all small, all spinning fast, all neatly explained by steam nobody could see. It was a perfectly good explanation, and it had the advantage that nobody could ever check it. None of them was two kilometers. A body this size braking this hard would be shedding hundreds of kilograms a second, and hundreds of kilograms a second of anything made a coma.
The two Surveyor tracklets had come in as a pipeline test during commissioning. She pulled the frames behind them from the mission archive and stacked nineteen of them on the object’s motion, the stars smearing into lines.
In the center sat one round point, small and sharp, with flat noise around it out to the edges. No coma. No tail. Surveyor had been built to find dark asteroids by their heat, and it had seen this one by its heat and nothing else. The short band ran brighter than she’d have guessed, but thermal models weren’t her field, and the commissioning flags covered a multitude of sins.
Down the hall somebody was laughing on a call about the travel budget.
At five she started the email to Glen.
He was her branch lead, and this one was worth a press release before anyone said a word about residuals. Nobody had ever found an Earth Trojan at L5. The two at L4 had turned up by accident. People had searched L5 on purpose and come back with nothing, and theorists had written papers about why.
She wrote it properly. The linkage, the orbit, the size, and then a paragraph on the residuals: the jackknife, the stack, the Yarkovsky arithmetic, and her opinion of dark comets this big, which was not polite. She read the paragraph, put the cursor at the end of it, and held down delete until the message was four lines long.
Linker joined nine orphaned tracklets from Catalina, PS1, ATLAS and Surveyor commissioning into one object, 32-day arc. Orbit is co-orbital, trailing, near L5. H around 16, so about 2 km. Gravity-only residuals show some structure, probably astrometry at low elongation; I’ll ask Catalina for follow-up.
Probably was the only word in it she didn’t believe. She sent it anyway. Glen loved an anomaly almost as much as a discovery, and a negative A2 at two and a half sigma would be on a slide by Thursday.
His reply came while she was packing her bag.
L5!! If real that’s a paper.
Then, a minute later: Who else has the Surveyor data? Let’s talk tomorrow before anybody at MPC gets cute.
Catalina got it on the twenty-sixth and the twenty-eighth, a little fainter each time, with a note that the second night had been lousy. Dana had both in the fit by eight on Wednesday morning, the twenty-ninth.
Thirty-nine days, eleven tracklets. Gravity-only was worse. With the Marsden terms free, A2 came back −1.8 ± 0.6.
Three sigma.
She ran the jackknife again, then with the bad night deweighted, then with it thrown out. Three sigma or a hair under, every way she asked.
Glen stopped at her door at six with his bike helmet already on.
“Catalina came through?”
“Two nights. The second one’s ugly.”
“Still Trojan?”
“Still Trojan. Libration’s small.”
“First L5 Earth Trojan.” He said it to the doorframe, trying it out. “First one. Ours. You know how many papers explained why L5 was empty?”
“Three,” Dana said. “I refereed two of them.”
He laughed and tapped the frame twice. “I talked to the Surveyor project scientist about data rights. They want to be on it, I said fine, but the linkage is ours and the paper comes out of this branch. And we hold the designation until we’ve got a draft. I don’t want to read about this on somebody’s blog.”
“It’s still there,” she said.
“What is?”
“The structure in the residuals. Three sigma with the new nights.”
“At sixty degrees from the Sun.” He was smiling, not unkindly. “Dana. Every astrometrist in the world goes soft down there. Get me a longer arc and it washes out. They always wash out.”
“Sometimes they don’t.”
“Sometimes they don’t, and then it’s dark comet number whatever, and that’s a second paper.” He was already turning to go. “Either way it’s a good spring. Write the orbit up clean. Leave the weird stuff for later.”
She listened to the stairwell door, then the bike rack.
The floor emptied by seven.
Cutting the arc into sub-arcs was statistically obscene. A thirteen-day arc didn’t constrain a nongravitational term at all, and if a postdoc had brought her the plot she was about to make, she’d have sent them back to their desk with Bevington’s chapter on error analysis.
She did it anyway. Three pieces, early, middle and late, each fit for A2 with the orbit from the full arc held fixed.
−1.9. −1.8. −1.6. Each ±1.4.
On a plot the error bars overlapped so completely that the three points made one fat stripe with three ticks in it. Consistent with constant. Consistent with zero, on any one of them.
She drew the line through them anyway. Three points, one degree of freedom. It rose toward zero and crossed it in late August, and she computed the date rather than trust the axis: August twenty-first, give or take. Move the last point by a tenth of its error bar and the date moved two months. Three noisy points could be made to point at Christmas.
She should have closed the window. Instead she put her line in for A2, propagated the orbit through the summer, and plotted the object’s drift relative to L5, averaged over the kilometer-a-second swing its orbit made every year.
The drift came down through April and May, flattened through July, and settled onto zero in the third week of August, closer to L5 than her error ellipse.
She put the two curves on one plot. The brake tapering, the drift tapering. They reached zero in the same week.
She looked at the error bars instead. She had built a curve out of a stripe and a coincidence out of a curve, and anyone who did that in public deserved what came after.
At eight-thirty she wrote the version for the shared drive. Gravity-only orbit, eleven tracklets, thirty-nine days, the residual statistics stated plainly, with a note that the low-elongation astrometry was being re-reduced. No Marsden terms, no sub-arcs, nothing relative to L5. It was a clean, correct, publishable orbit for a two-kilometer Earth Trojan with untidy residuals, and it was what Glen had asked for. She sent him the link: Orbit for the draft. Residuals still under review.
The braked fits, the jackknife, the stack and the three segments she moved to her laptop’s own disk, outside the synced directory. She checked the sync settings twice.
Then she took a yellow sticky note off the pad by her monitor and wrote 8/21 on it in pencil. She folded it over on itself, sticky side in, so it held shut, and put it in the inside pocket of her bag, behind her badge.