Three Priorities

The weekly operations meeting began at 0810 in Conference Module B, twelve minutes late because Engineering had held the room past schedule to finish a maintenance briefing that ran over and five minutes early by Science standards, because Priya Chandrasekaran disliked beginning any meeting with an apology.

The room was built for twenty-two. Seventeen seats were occupied. The display wall carried the station agenda in three color bands: Science Directorate items in blue, PEC items in amber, CCA items in white. Maren took her usual place halfway down the table where she could see both the wall display and the reflection of the room in the darkened viewport beyond it. Tomás Herrera sat to Priya's right, broad-shouldered, expression neutral, engineering slate already open. Daniela Mossberger from the CCA sat two seats farther down with a legal pad she never seemed to use for anything except making other people nervous. Kael Oduya had no official reason to be in this meeting beyond communications support and every practical reason; they sat near the door, tablet idle, watching the room's timing more than its content. Jun arrived last, apologizing softly to no one in particular, and folded himself into a seat with the distracted care of someone whose body had shown up before the rest of him.

Priya looked once around the table. "Let's begin."

Item one cleared quickly: surface resupply manifest, no objections. Item two did not.

Amber filled the display. PEC REQUEST: Vent Cluster 5 output increase, fifteen percent, effective next cycle.

Tomás stood without flourish and brought up the thermal projection. "Seasonal surface demand is running above forecast. Northern cascade draw is up nine percent week over week. PEC wants compensation from Cluster 5 because it's the least disruptive place to get it."

Priya did not look at the graph first. She looked at Tomás. "Least disruptive by which model?"

"Grid stability." He tapped the display. A clean series of load curves appeared. "Cluster 5 can absorb the increase within current tap tolerance bands."

"Current thermal tolerance bands," Priya said. "The adjacent biosequestration zone is already running elevated metabolic output. Increased extraction changes the local thermal envelope."

Tomás nodded once. "By 0.6 degrees at the colony boundary, by our estimate."

"Your estimate," Priya said, "built on substrate assumptions that predate the last expansion."

Maren watched the exchange the way she watched stress propagate through coupled systems. Neither of them was posturing. Tomás was reading infrastructure. Priya was reading biology. Both models were locally correct. The room, as usual, was optimized to pretend local correctness added up to global clarity.

Jun raised a hand halfway, not enough to interrupt. Priya saw it.

"Jun?"

He pulled up a small waveform on his tablet and sent it to the wall. "Background microseismic activity near Cluster 5 has been a little noisy this quarter. Probably nothing. But if extraction goes up, I'd like a closer look at the baseline before and after."

The waveform appeared for four seconds. Maren saw it: low-amplitude chatter with a faint regularity she could not place at this distance.

Tomás said, "Does noisy mean unstable?"

Jun frowned at the screen. "Noisy means noisy."

A few people smiled. Jun did not.

Priya said, "Can you quantify the additional compute time for comparison modeling?"

Jun hesitated. "Twelve hours for a first pass. More if you want separation from station operational noise."

Tomás checked his slate. "Those cycles don't exist this week."

There it was. Not refusal. Arithmetic.

Priya let the silence hold just long enough for everyone to register that the constraint was structural, not personal. Then she said, "Noted. We move to item three. Cluster 5 output request remains pending final allocation review."

Amber stayed on the wall long enough for everyone to understand that pending was not denial. Then blue replaced it.

Item five was Daniela's.

CCA REQUEST: Baseline recalibration of seismological data. Seventy-two hours dedicated compute.

Daniela spoke in the same calm tone she used whether requesting coffee or regulatory compliance. "The regional modeling team has identified a discrepancy between CDO's baseline and two other stations in the network. We need recalibration to rule out instrument drift."

Jun sat up a little straighter. "The array passed last month."

"This is not an accusation of error," Daniela said. "It is a certification requirement."

Priya folded her hands. "The quarterly carbon credit verification run starts tonight. Science has a hard reporting deadline."

Daniela nodded. "I am aware."

"What you are requesting," Priya said, "would displace the verification run."

"What I am requesting," Daniela said, "is confirmation that the primary warning instrumentation for the Pacific Northwest coast is producing trustworthy data."

No one in the room argued with the sentence. That was what made the conflict efficient. All three institutions had learned to speak in morally unassailable summaries while fighting over the same finite machinery.

Jun said, "I can run the recalibration overnight in segments."

Tomás shook his head before Priya could answer. "No overnight segments. Compute is saturated with tap optimization after the Cluster 5 demand notice."

Priya turned to him. "Before the request is approved?"

Tomás met her look. "Demand on the grid exists whether we approve the wording or not."

Maren made no note. She didn't need to. The trilemma was writing itself in large, readable letters: energy output, scientific verification, safety certification. Pick one first and the other two accrued cost.

Item seven arrived eighteen minutes later than scheduled and exactly when Maren expected it to: after enough time had been spent proving the room's scarcity.

Blue-white split across the wall. Cross-System Anomaly Flag: Vent Cluster 7.

Priya looked at Maren. "You filed this at 0416. Walk us through it."

Maren sent her trend reconstruction to the main display. Eighteen months of spectral data arranged without smoothing. The slope was visible from the back of the room.

"Cluster 7 metabolic output is elevated in the positive direction relative to carbon fixation baseline," she said. "That isn't the issue. Secondary mineral deposit composition is drifting from the expected profile. Iron-associated peaks have been increasing gradually since shortly after the expansion event fourteen months ago."

She overlaid the pre- and post-expansion signatures. The room leaned toward the wall by degrees.

"Individually, all samples are within accepted variance bands. Collectively, the drift is directional. I've cross-referenced that drift against nearby geothermal performance. No major anomalies. Some minor thermal efficiency deviations once seasonal compensation is stripped out. Not actionable alone. Potentially relevant in combination."

Priya said, "Your working hypothesis?"

"Possible interaction between engineered colonies and native chemosynthetic species producing altered mineral output. If the deposits are changing composition, local substrate chemistry may also be changing. If substrate chemistry is changing, upper basalt porosity and thermal conductivity may eventually shift. That could affect tap performance and, depending on scale, seismological baseline quality."

Daniela looked up from her unused pad. "Could?"

"Could," Maren said. "I don't have the substrate survey data to reduce that uncertainty."

Tomás leaned back in his chair and studied the slope. "And the recommendation?"

"Expanded spectral monitoring and an ROV geological survey of substrate adjacent to Cluster 7."

Tomás exhaled through his nose, not dismissive, just calculating. "I've got three tap inspections already queued this week, and if Cluster 5 gets approved they're all higher priority than an open-ended geology survey. Can it wait two weeks?"

"Maybe," Maren said. "I don't know if the drift is linear, plateauing, or preceding a threshold behavior. Two weeks gets us more data on the spectral side and no data on the substrate side."

Priya said, "I can authorize additional spectral sampling from the existing biolab schedule."

Daniela said, "If there's any chance this touches substrate behavior, the CCA needs to know whether it affects baseline interpretation."

Maren resisted the urge to answer too quickly. "I don't know yet."

The room accepted that answer the way institutions accept uncertainty when uncertainty does not yet force expenditure.

Priya nodded. "Additional biological sampling authorized. Geological survey pending resource review."

Pending. The station's favorite way of saying later.

Tomás marked something on his slate. "Send me the stripped thermal microdeviation set."

"I already attached it to the flag."

He gave her a brief look that translated to yes, and I will read it when I have air.

Daniela said, "Please forward any updated assessment directly to my office."

"I will."

Priya moved to item eight.

Just like that, Maren's anomaly became one more line in the station's managed backlog. Not ignored. Not accepted. Processed.

For the remaining thirty-four minutes, the meeting turned through personnel rotations, supply chain revisions, and a minor communications bandwidth dispute between lab and engineering. Maren contributed when asked and spent the rest of the time tracking the invisible subtraction. If compute went to CCA recalibration, bioverification slipped. If ROV time stayed with engineering, Cluster 7 remained a model built on inference. If Priya protected the verification deadline, Daniela escalated. None of it required malice. Optimization was enough.

When the meeting adjourned, chairs scraped back in disciplined impatience. Tomás was intercepted immediately by two engineering leads. Priya by a biolab supervisor with carbon credit spreadsheets in hand. Daniela by no one, which was its own institutional fact. Kael remained seated for three seconds after everyone else rose, watching the dispersal pattern, then slid their tablet shut.

Maren was halfway into the corridor curve when she heard quick steps behind her.

"Dr. Leath."

Jun never called her Maren in public when he was carrying data.

She slowed. The corridor lights were in day-cycle brightness now, clean and shadowless against the curved pressure wall. Jun held out his handheld display without preamble.

"This is the noise near Cluster 5 I mentioned."

She took the display. The waveform was clearer than it had been in the meeting: low-amplitude background activity with a periodic structure nested inside it, regular enough to bother the eye even before the math touched it.

"How long?" she asked.

"I noticed it six days ago. I thought it was array contamination. It isn't."

"You're sure?"

"I checked station machinery cross-talk, tap cycling signatures, comms bleed, and software artifacts." A small pause. "Then I checked them again."

Maren enlarged the section around the repeating interval. Tectonic background should have scattered. This had rhythm.

"What could produce this frequency?" she asked.

Jun looked past her shoulder for a moment, not evasive, just searching a library only he could hear. "I don't know."

It was the least reassuring answer available and, coming from Jun, the most useful. He did not protect anyone with false specificity.

"Send me the raw waveforms," she said. "All six days."

"I already routed them to your private workspace."

Of course he had.

She handed back the display. "Keep monitoring. Don't brief it yet."

He blinked once. "Because we're not sure?"

"Because we're not sure and because if you put an unexplained periodic signal in the room we just left, three institutions will turn it into three different emergencies before we've identified the variable."

Jun considered that, then nodded. "Understood."

He started to leave, then stopped. "Do you think it's related to Cluster 7?"

Maren looked down the empty curve of corridor toward the operations hub. "I think the station is generating too many small things that point in one direction."

"Which direction?"

"I don't know yet."

That answer seemed to settle him more than certainty would have. It also settled something in her. A second data point was no longer a possibility. It existed. The spectral slope at Cluster 7. The periodic noise near Cluster 5. Maybe connected. Maybe not. But the shape of the week had changed. There was now enough wrongness distributed across enough systems to justify a larger model.

Jun left for seismology. Maren continued around the ring toward the hub, passing Communications where Kael stood half inside the doorway speaking quietly to a surface tech through a wall display. They glanced at her once, took in her pace, Jun's retreating back, and the fact that she was already thinking six variables ahead. Their expression altered by less than a degree. They knew something had moved.

Back at Cross-Systems Station Three, Maren opened the raw waveform files Jun had sent. Six days of microseismic data populated the lower band of her display. She placed the Cluster 7 spectral slope above it and did nothing for eleven seconds except look.

No conclusion. Not yet.

Just two anomalies in a station designed to let every department believe its problems ended at the edge of its feed.

She opened a new workspace and labeled it Cluster 7 / Cluster 5 correlation test.

At the top of the display, the water reclamation ticket still sat closed with its routine reply. Within normal variance.

Maren dragged it into the corner rather than deleting it.

Then she began building the next layer of the model.