The concourse was quiet.
It was the measured, heavy quiet of people doing their jobs. Outpost Four slid forty meters along the ring, absorbing a tidal surge from the ice below. No alarms sounded. The drift system handled the tension perfectly, bleeding the kinetic load sideways without a shudder.
Above the holotable, the ocean was no longer a guess. Three million probes were active. The mesh was only partially deployed, waiting on the next Forge drop to cover the southern hemisphere, but the resolution in the north was already staggering.
"We can see it now," Dr. Voss said.
She stood at the primary console, her posture rigid but grounded. The frantic edge that had gripped her yesterday was gone. She was operating in command mode again, watching the neat, synchronized data streams feed into the station's mainframe.
"Bulk flow is holding at ninety-eight percent accuracy," a systems tech reported from the secondary banks. "The drift compensators are anticipating the macro-currents perfectly."
"And the micro-currents?" Voss asked.
"That's where we're catching drag."
The tech highlighted a series of deep-water boundaries on the projection.
"Prediction drift is consistent across runs," he said. "The macro-waves are hitting their marks, but at the transition points—where the cold down-welling meets the thermal plumes—we're missing the timing. It's small. Fractions of a second, but it's compounding."
Voss frowned, leaning over the table. "Are the pressure waves arriving early or late?"
"Both. It depends on the cluster."
"It's calibration," an engineer suggested from across the room. "The new probes are still settling into the thermal gradients. The hardware needs time to acclimate to the ambient cold."
"If it were noise, it wouldn't repeat like this," another engineer countered, pulling up a secondary display. "Look at the telemetry. It's not random. The errors are clustering in specific, repeatable regions. We're not wrong everywhere. Just where it matters."
"Magnetospheric bounce?" Voss asked.
"Checked it. The induction shielding on the probes is holding perfectly."
A fluid-dynamics specialist tapped her screen, pulling up a raw acoustic return from Sector Four. "There's a localized drag inconsistency. Something is physically altering the acoustic bounce in these specific boundary layers."
Daniel looked at the readout. He felt the friction in his Lace. "Density irregularities."
"More than that," the specialist said. "There's a conductivity variance. The water isn't uniform down there. We're getting particulate scattering. It's behaving like a suspension film—something heavy and complex altering the local viscosity."
She shook her head, staring at the numbers. "Something in the water is changing how it moves. We're not modeling the medium correctly. It's altering the local flow response."
"Then add the viscosity variables to the simulation," Voss ordered. "Update the medium parameters."
Across the room, six engineers with newly flashed Mark III Lace upgrades went to work. They took the raw particulate data, the conductivity variances, and the density films, and fed them directly into the fluid-dynamic algorithms. They spun up massive, multi-threaded simulations, throwing sheer computational brute force at the boundary behavior.
The holotable fragmented.
Instead of generating one clean predictive model, the projection split into four overlapping, conflicting flow-paths.
"What's happening?" Voss demanded.
"The compute is saturated," an engineer said, aggressively rubbing his temples as his upgraded Lace struggled to manage the load. "When we introduce the new density variables, the math branches. All models are mathematically valid based on the telemetry. None of them converge."
"We can simulate it," another tech said, his voice tight with cognitive strain. "We just can't collapse it into a single outcome."
Daniel stepped up to the holotable.
He didn't look at the branching models. He reached into the haptic interface and started collapsing the data. He stripped out the temperature decimals. He muted the salinity readings. He filtered the particulate scattering and the conductivity spikes, paring the simulation down to the dominant gradient.
"What are you doing?" the fluid-dynamics specialist asked, her hands hovering over her locked board. "You're throwing out the medium."
"You're using too much data," Daniel said. His eyes were half-focused, tracking the deep, heavy blue lines of the bulk flow beneath the noise. "You're trying to calculate every suspended particle in the boundary. The system doesn't move like that."
"It's a chaotic medium, Daniel. You have to account for the variables."
"Watch where it goes," Daniel said quietly. "Not what it is."
A massive pressure wave rolled into the northern grid on the live feed. It was approaching one of the problematic boundary layers where the density film hung thickest.
The engineers' multi-threaded models on the secondary screens began spinning wildly, offering a half-dozen possible deflections, choking on the viscosity math.
Daniel didn't look at the models. He watched the shape of the incoming current. He felt the gradient in his own mind, the physical, undeniable necessity of the pressure.
He pointed to a blank, unremarkable coordinate on the holographic grid.
"It's going to shift here," Daniel said. "You'll see it before the model updates."
The engineers stopped typing. They watched the live telemetry.
The wave hit the boundary. The particulate film dragged at it. The automated models stuttered, caught in an infinite loop of variables, unable to decide which valid mathematical outcome would occur.
But the live telemetry updated a fraction of a second later.
The pressure wave sheared exactly where Daniel's finger was pointing, rolling cleanly into the new vector. It wasn't guesswork. It was simply the only place the water could go.
Before Voss could speak, two voices rang out from opposite sides of the concourse.
"It sheared at grid 4-A," a junior tech called out.
"I've got the shift at 4-A," a senior analyst echoed simultaneously.
They looked up from their consoles, locking eyes across the room.
"You saw that too?" the junior tech asked, surprised.
"We all flagged the same region," the analyst said, looking back at her board. "The model missed it, but the flow trajectory made it obvious."
Voss stood motionless.
She looked at the holotable, where the automated models were still trying to computationally resolve an event that had already happened in reality. She looked at the six engineers massaging their temples, overwhelmed by the data limit of their high-end Laces.
Then she looked at Daniel, and the two analysts across the room who had caught the shift manually, without coordinating.
The frantic panic of the prior days was completely absent. What replaced it was a cold, operational clarity.
"This isn't a modeling failure," Voss said. Her voice carried easily through the quiet hum of the room.
The engineers turned to her.
"The telemetry is accurate," Voss continued, stepping slowly around the projection. "The probes are working exactly as intended. The data is sufficient." She looked at the fragmented projection, and the humans surrounding it. "This is a coverage problem."
Daniel dropped his hand from the interface. The Old Instinct hummed in quiet agreement. The Forge had given them the ocean. The probes had given them the numbers. But the machine couldn't hold the shape of it.
"This doesn't resolve with more data," Daniel said softly.
Voss stared at the vast, dark blue sphere of the Europan sea, realizing exactly what the bottleneck truly was.
"We don't need better models," Voss said. "We need more minds in the water."