Debris

Chapter 1: The Night Before

The calibration file was 1.7 megabytes. A whisper of data. Marta could have left it alone.

She sat in the pre-flight quarantine suite at Johnson Space Center, her laptop balanced on a pillow because the desk lamp had burned out three days ago and she'd never reported it. The room smelled of recycled air and the faint chemical tang of the disinfectant they used on every surface. Outside, Houston was dark and wet, rain streaking the windows in slow diagonals.

It was 2:14 a.m.

Launch was in nine hours.

The debris-tracking array was finished. Tested. Certified. Signed off by three independent review boards and one very tired program manager named Hal Pemberton who had shaken her hand that afternoon and said, "Marta, it's good. Go to bed."

She had nodded.

She had not gone to bed.

Because the array was good, but good was the word people used when they meant adequate, and adequate was the word people used when they meant we hope this works, and hope was the word people used when they didn't have enough data.

Marta had spent three years building something better than hope.

The Orbital Debris Tracking and Prediction System—ODTPS, which everyone called "the lattice" because the sensor arrangement looked like a crystalline web when deployed—was designed to solve one specific problem. Existing tracking systems could detect objects larger than ten centimeters in low Earth orbit. They could model trajectories, predict conjunctions, issue collision warnings.

What they could not do was see the small things fast enough.

A fleck of paint traveling at seven kilometres per second carries the kinetic energy of a bowling ball dropped from the top of a building. A bolt carries the energy of a hand grenade. A fragment the size of a marble can punch through the pressure hull of a space station like a bullet through paper.

These fragments numbered in the hundreds of thousands. They moved in clouds and streams and lonely arcs, invisible to ground radar, too small for optical tracking, too fast for the statistical models that tried to predict where they would be.

Marta's lattice was supposed to change that.

Twenty-four external sensors mounted on Kepler Station's truss structure, each one a hybrid of millimeter-wave radar, LIDAR, and passive infrared detection. Together they would create a real-time map of the debris environment within a two-hundred-kilometre sphere around the station, updating sixty times per second, feeding into a predictive algorithm that could model cascade events before they happened.

The algorithm was Marta's. Every line of it. Every coefficient, every weighting function, every probabilistic branch.

It was the best work she had ever done.

It was the work she had started the week after they buried Tomek.

She did not think about that. She thought about the calibration file.

The file governed how the sensors interpreted signal return times. A fraction of a microsecond's difference in how the system measured the bounce-back of its own radar pulses would cascade through the prediction model, shifting estimated positions, velocities, threat assessments. The calibration had been set during ground testing at the Goddard facility, using a debris simulation chamber that could reproduce orbital-velocity fragment behavior within a controlled environment.

But the simulation chamber was not space.

Marta knew this. Everyone knew this. The calibration would be refined once the array was operational, using real orbital data. That was the plan. That had always been the plan.

But she had been reviewing the ground-test data one more time—just one more time, she told herself, the way she always told herself—and she had noticed something.

Sensor band seven.

Its return-time baseline was 0.003 microseconds faster than the other twenty-three sensors. Not outside tolerance. Not flagged by quality control. But consistently, persistently, measurably faster.

It could be nothing.

It could be a manufacturing variation in the radar transceiver.

It could be a tiny systematic bias that would compound over millions of measurements and shift the prediction model's accuracy by a fraction of a percent.

A fraction of a percent.

Tomek had died because a debris warning came four seconds too late.

Four seconds.

Marta opened the calibration file.

She adjusted sensor band seven's baseline to match the mean of the other twenty-three. A simple correction. Obvious. Defensible. She would document it in the morning, file the deviation report, explain her reasoning.

Then she noticed that the adjustment changed the cross-correlation matrix for the entire array. Bands six and eight now showed a slight mismatch with band seven's new baseline. Not large. But visible.

She adjusted those too.

The cascade of small corrections took forty minutes. Each one was logical. Each one was small. Each one was made at 2:30, then 2:45, then 3:00 in the morning by a woman who had not slept properly in three years and who needed this system to be perfect because perfection was the only currency that could pay for what she had lost.

When she finished, the calibration file was 1.7 megabytes of revised data.

She uploaded it to the flight software package.

She closed her laptop.

She lay in the dark and listened to the rain and did not sleep.

At 5:00 a.m., she got up, showered, put on the blue flight suit they had fitted for her six months ago, and walked to the van that would take her to the launch pad.

In the van, Commander Buck Dorr was already seated, his broad frame filling the bench, his silver-streaked hair cropped close in the way of men who had worn helmets for decades. He was fifty-six years old, built like a retired linebacker who had taken up distance running, and he had the easy, weathered confidence of someone who had been to space four times and come home every time.

"Morning, Doc," he said. His voice was deep, unhurried, the drawl of a man who had grown up in west Texas and never seen a reason to talk faster than the wind.

"Morning, Commander."

"You sleep?"

"Some."

Buck studied her for a moment. He had pale blue eyes that missed very little and forgave most of what they saw.

"Liar," he said. But he smiled when he said it.

Flight engineer Luis Iglesias climbed in behind Marta, all nervous energy and careful courtesy. He was thirty-three, slight, dark-eyed, born in Guadalajara and raised in Pasadena, an MIT-trained systems engineer on his second flight. He carried a small notebook everywhere because he believed that anything worth remembering was worth writing down, and he wrote in a precise, tiny hand that could fit an entire checklist on a Post-it note.

"Beautiful day for it," Luis said, even though it was still raining.

"Rain stops at T-minus-forty," Buck said. "Weather office promised."

"You trust weather offices?" Marta asked.

"I trust promises," Buck said. "People try harder when they've promised."

The van rolled toward the pad.

Marta watched the rain streak the windows and thought about the calibration file.

She should mention it. She should tell someone. She should—

The file was within tolerance. Every adjustment was defensible. She would document it after they reached orbit, when there was time, when the system was live and she could validate against real data.

It was fine.

It had to be fine.

The rain stopped at T-minus-thirty-eight.

Buck caught her eye and winked.

"Close enough," he said.

They launched into a sky that was clearing from the west, the clouds pulling apart like torn fabric, and the ride was rough and bright and deafening, and for eight minutes and thirty-two seconds Marta felt the full weight of the Earth trying to hold her down and failing.

When the engines cut and silence flooded the cabin and the straps went slack and her body lifted gently against the harness, she looked out the window at the thin blue line of atmosphere curving beneath them and felt, for the first time in three years, something that was not grief.

It felt like possibility.

It felt like the beginning of something that could make the loss mean something.

She did not know it was the beginning of the end.