201: Chapter 201 Before the Main Force

September of the twentieth year.

After the 37,412th repair fragment sent by OUTER CACHE-07 passed verification, a record in the FDSO-2076A directory that had been gray for twenty years turned green.

[Outer Heliosphere—Earth-Moon System Continuous Observation Segment]

[Effective Span: Thirty-one Years and Seven Months]

[Data Completeness: 71.4%]

[Cross-system Unified Time Scale: Missing]

The certificate printed with [6 < 25/4] beside the workbench had already turned yellow, and the seventh cache verification slip of the month had been pressed beneath the transparent board.

Jiang Lin pushed the freshly brewed cup of tea away from the host computer's air vent and pulled up the five data streams from the records.

The outer heliosphere boundary, dust particle size spectrum, farside lunar observation link, Mars Joseph-Louis Lagrange Relay, and TM-7 load history were laid out in sequence across a six-meter-long horizontal screen; the five curves were each continuous, yet they lived within five mutually unrecognized sets of time.

If aligned directly according to the surviving tags, the farside lunar link would have lost contact eight years before the anomaly appeared in the outer heliosphere, while the Mars relay would have lost ninety-eight percent of its effective forwarding capacity long before the so-called main front entered the solar system.

The most absurd part was TM-7.

While the external observation systems were still reporting that everything was normal, it had already taken over the offline cache and computing load of the deep space array for three consecutive years, acting like a warehouse that began moving goods to the roof years before the flood arrived.

MPS-Memory listed 216 time-scale stitching schemes, and the scheme ranked first required the fewest modifications to the existing tags.

Jiang Lin deleted it.

The writers of the tags had been dead for over two hundred years, yet the planets were still running along their respective orbits.

Every time the high-gain antenna of the Mars relay switched targets, the records would leave behind the azimuth, elevation, and lead time; the farside lunar array had to avoid Earth's radio pollution, and its working window opened and closed repeatedly following lunar libration; the probes further out used stellar sensors to maintain their attitude, and the incomplete quaternions still retained the direction of the sky.

These data, which were originally responsible for telling the machine where to look, had now become a more reliable clock than the system date.

Jiang Lin used orbital inversion to reconstruct the planetary positions, first forcing the five sets of time scales into the same set of celestial geometric constraints, and then searching the records for solar high-energy particle events jointly observed by multiple systems.

A major solar eruption would leave burst noise in the near-Earth link, cross the Mars orbit hours later, and take even longer to reach the outer probes.

The intensity order and pulse intervals of each eruption were different, and when dozens of them were linked together, they formed a string of notches that was hard to mistake.

The first set of time-scale corrections took him forty-seven days.

When the five curves finally shared the same zero point, an attitude correction of the Mars relay was still fourteen months earlier than the corresponding particle event.

Following these fourteen months to the local calendar field after the relay switch, Jiang Lin found that the existing decoder expanded the length of a Martian year according to Earth years; the longer the record, the further forward the date slipped.

The second set of corrections compressed the error to twenty-one days.

The outer heliosphere records still did not match, because the probe there, after the master clock failed, had used a millisecond pulsar to maintain its local frequency, but the archive dump program had mistaken the pulse count for seconds.

Jiang Lin re-solved the counting multiplier, and the final set of teeth of the solar event also locked into the corresponding position.

[Cross-system Relative Time Scale: Reconstructed]

[Inner Solar System Alignment Error: ±11.2 min]

[Outer Heliosphere Alignment Error: ±3.7 d]

The five curves on the horizontal screen stretched out in the same direction.

Jiang Lin first called upon the Dark Dust String model established earlier based on catastrophe fragments, sending [Single Main Front] into the inverter as the first branch to be ruled out.

The model quickly produced a set of fine parameters, and just as quickly began accumulating debts everywhere.

If the farside lunar link's disconnection time was used to fit the main front's velocity, the outer heliosphere boundary would have started compressing at least seventeen years too late; if the boundary compression was allowed to happen first, the particles encountered by the Mars relay would have been too large too early; if the particle size spectrum was adjusted to explain the Mars data, the total mass of the main front would push subsequent observations up by two orders of magnitude.

Every time one curve was fixed, another would rear up from the horizontal screen.

Those incomplete entries back then had placed [Main Front Particle Size Spectrum] and [Heliosphere Boundary Compression] in the same directory layer, so the single wavefront had always remained in the first place among the candidate branches, yet it had never crossed the evidentiary boundary to become a conclusion.

After the unified time scale was fixed, three sets of mutually incompatible residuals simultaneously crossed the preset threshold, and this candidate branch completed its sole task: being strictly ruled out by the data.

Jiang Lin closed the [Single Front] branch and switched to the second pre-listed two-component model, allowing the incoming flow to possess two overlapping components in velocity, particle size, and charge-to-mass ratio.

The first curve immediately flattened from a steep wall into a long slope.

The density at the very front of the slope was very low, and what it carried was mainly plasma and charged fine dust; after entering the heliosphere boundary, they were jointly filtered by the magnetic field, radiation pressure, and solar wind, and the amount reaching the inner solar system was not large.

However, the effect of this layer of material on the boundary occurred long before the increase in visible dust.

The external dynamic pressure continued to rise from a fixed sky direction, the windward side of the heliosphere shrank inward, the deflection pattern of small charged particles changed accordingly, and the periodic particle size fluctuations originally caused by solar activity gradually developed an additional bias that could not be reversed with the flipping of the solar magnetic field.

Only afterward did the high-density component emerge from behind the long slope.

It contained more particles with larger inertia, a steeper particle size spectrum, and a more concentrated coming direction—it was precisely the main front independently named in the old index.

Jiang Lin labeled the two components as [Thin Precursor Layer] and [High-Density Main Front] respectively, and re-fitted all observations.

The compression of the outer heliosphere boundary fell at the very front of the precursor layer, the dust spectrum change was located in the middle of the long slope, and the damage to the Mars relay and farside lunar link appeared sequentially along the incoming direction; for the first time, five conflicting records simultaneously remained within the error band.

The ninety-eight percent blockage of the Mars relay also found its origin.

What appeared earliest were attitude maintenance interruptions caused by phase scintillation, charged surface discharges, and tiny impacts; in order to preserve the link, the relay continuously increased the frequency of error correction, retransmission, and pointing correction, and the backup channels dropped out one by one amidst years of accumulated damage, finally leaving only two percent of effective forwarding capacity.

The scientific broadband was shut down first, routine mapping subsequently yielded its power, and the navigation and emergency narrowband persisted until the very end; each downgrade over the seven years retained a different authorization signature.

The farside lunar observation link took over part of the forwarding tasks in the direction of Mars, and after its own bit error rate exceeded the upper limit, it pushed the unprocessed raw data back to the ground.

As for TM-7, it could not even be considered an observation device.

After the nodes on Mars and the farside of the moon began to drop offline, a large amount of raw data, model states, and engineering materials that needed to be preserved flowed back to the ground, causing the transient thermal load of the underground computing array to rise ahead of schedule.

The task undertaken by the outer-ring phase-change cooling system increasingly resembled trying to catch the leaks of an entire roof with a single basin.

Jiang Lin moved the TM-7 load curve to the very bottom, where that gray line, which had been rising since the eighteenth year, clung closely to the cumulative area of deep space node failures.

Jiang Lin strung FDSO, the farside lunar link, the Mars relay, the Sparks of Civilization, and TM-7 into the same causal chain, changing the directory name from [Catastrophe-Related Fragments] to [Deep Space System Retreat Sequence].

In the fourth month of the twenty-second year of the current cycle, the inverter produced the first complete timeline version.

[T0: TM-7 enters frozen preservation]

[Precursor layer first discernible: T0 - 23.8 years]

[Cross-source stability confirmation: T0 - 19.6 years]

[Mars relay continuous degradation: T0 - 11.3 years]

[Farside lunar link lost contact: T0 - 4.9 years]

[Main front high-density peak expected arrival: T0 + 6.7 years]

Jiang Lin magnified the last item, checking the observation gap, the upper limit of the velocity distribution, and the final positions of the outer probes.

On the final verifiable cross-section, the high-density peak of the main front was still outside the orbit of Neptune.

Six years and seven months was not an exact countdown, and even the most conservative error range could not push it before T0.

When TM-7 was frozen, that wall the civilization had waited for over twenty years was still on the way.

Inside the control room, three unmanned operation shift groups continued handing over task capsules along the long-term testing route, and the one-hundred-thousand-hour counter on the wall had long since crossed 470,000.

Jiang Lin dragged the red line corresponding to the main front peak to the right of T0, leaving a black background gap of more than six years in the middle of the horizontal screen.

The old model originally attributed all destructions to beneath that red line; after the revision, most of the failure records remained to the left of the red line.

He reopened FDSO's cataclysmic causality catalog and split [Dark Dust String caused civilization destruction] into two items.

[Foreign dust-plasma structure enters the solar system: Evidence chain closed]

[Systemic collapse of civilization occurs before the arrival of the main front: Reason unknown]

The second item remained in the red state.

Jiang Lin had no one left to question, and OUTER CACHE-07 would not explain for people from more than two hundred years ago why they had squandered a twenty-year early warning window.

Jiang Lin crossed out [Search for main front] in the reality translation task and changed it to [Search for precursor layer].

The main front was far too distant; by the time it brightened enough for ordinary astronomical observations to directly identify it, more than half of the choices left for the Earth-Moon system would have already vanished.

The precursor layer was much lower in density, but it would leave traces first from the boundary of the solar system, the screening methods of charged dust, and the noise floor of deep-space links.

When these traces appeared individually, they closely resembled old problems: solar activity, instrument aging, local debris clouds, attitude control faults, or the exhaustion of communication equipment lifespans.

To recognize them as the same incoming flow, the observation system had to retain a sufficiently distant spatial baseline and allow bad news that originally belonged to the operations and maintenance department to enter the scientific data.

In the twenty-third year of the current cycle, Jiang Lin compressed the FDSO inversion instrument down to the conditions usable in 2022.

He retrieved the plasma and magnetic field records of Voyager, the all-sky celestial high-energy neutral atom maps of the interstellar boundary explorer IBEX, the solar wind and dust counts of New Horizons, and the sixteen years of archived interstellar dust data from Ulysses from the offline database.

The design goals, orbits, sensitivities, and sampling times of these detectors did not match one another, and piecing them together looked very much like assembling a set of measuring tools from four secondhand markets; their biggest commonality was that none of them would fill in the missing column for Jiang Lin.

The first round of inversion predictably reported two candidate precursor layers.

The stronger item fell near the change in the direction of Ulysses's late-stage dust influx, while the weaker item came from the increase in plasma density after Voyager 1 crossed the heliopause.

Jiang Lin sent the two candidates respectively into spatial propagation testing.

The former changed with the solar magnetic field polarity and the detector observation geometry, and upon switching to another set of particle size response functions, the directional offset shrank back into the error band; the latter only existed on one outward baseline, while the other baseline lacked evidence of co-directional propagation.

Both alerts were withdrawn.

[2022 Public Data: No precursor layer meeting the joint criteria was found]

[Conclusion boundary: Sparse observations, incoming flows lower than existing sensitivity are not ruled out]

These two lines of text on the terminal were not startling enough, making them just right to bring back to reality.

If an early warning method only knew how to call normal solar activity the end of the world, its greatest use would probably be to expand the headline database for news websites.

Jiang Lin added three hard gates to the inversion instrument.

External disturbances had to propagate from the outside inward, in a direction opposite to the outward propagation of solar flares; the three types of channels—plasma pressure, dust particle size, and deep-space links—had to point to the same incoming direction cone; the anomalies also had to span multiple observation cycles, and local instrument drifts and one-off debris clouds were not allowed to enter long-term alerts.

He then downsampled the FDSO precursor layer data, deleting high-precision payloads that future systems possessed but 2022 lacked, retaining only the time, direction, and coarse particle size information that existing instruments could provide.

The precursor layer disappeared in the first two experiments.

For the third time, he retained two outer baselines separated far enough apart, and added the phase noise and attitude correction records of the deep-space links into the observables, allowing the inversion instrument to find the incoming direction once again.

When seventy percent of the data was randomly excised, it still compressed the precursor layer into a 12.4-degree sky cone, with the velocity median error lower than nine percent, and the arrival interval of the main front peak covering the archival true value.

If any outer baseline were further deleted, the incoming direction cone would expand to half the sky, and the early warning time would also lose its constraint.

The minimum observation network thus stopped at three layers.

The near-Earth layer retained all-sky heliopause boundary imaging and solar activity baselines; the outer planet layer was configured with dust-plasma combined payloads capable of distinguishing incoming directions and coarse particle sizes; and at least two independent baselines separated by a sufficiently large distance were retained outside the heliopause.

Phase scintillation, error-correcting retransmissions, surface discharges, and micro-impact attitude corrections of all long-distance links entered long-term archiving, no longer serving merely as operational garbage periodically cleaned up by the communications department.

This network did not need to see the main front itself clearly.

As long as the precursor layer maintained the minimum intensity in the FDSO records, it would cross the joint criteria at least twenty-one years before the arrival of the high-density peak.

From the twenty-fourth to the thirty-second year of the current cycle, Jiang Lin threw the model into nine rounds of long-term blind tests isolated from one another.

Each round of blind testing mixed in complete solar activity cycles, instrument aging, detector orbit changes, local debris clouds, and natural degradation of deep-space links. Half of the data was generated by MPS-Memory according to failure records, and the labels were not revealed until the end of the experiment.

The first round reported seven alerts: four came from solar magnetic field reversals, two from single dust counter response drift, and the remaining one was caused by a centralized upgrade of the communication link in the direction of Mars.

Jiang Lin withdrew the alerts one by one, incorporating operational behaviors that would cause the model to misjudge into the baseline.

After the fifth round, false alarms dropped to zero for the first time, but missed alarms appeared in succession.

A weak precursor layer with 82% of its data excised passed between the two outer baselines, and the joint criteria consistently lacked consistency in the final item—phase noise.

Jiang Lin retained this missed alarm, keeping the threshold at its original value, while changing the minimum continuous observation duration from a fixed window to a propagation window that varied with baseline distance.

In the seventh round, the same group of signals was re-injected, and the three-layer network crossed the threshold twenty-two years and three months before the arrival of the high-density peak of the main front, with the false alarm set still empty.

Over the next decade, the model entered long-term operation alongside the unmanned operation shift groups of the Outpost.

Every six months, Jiang Lin extracted a public data adapter, deliberately replacing timestamp formats, missing fields, and instrument response curves, and then handed it over to another offline verifier to re-establish the baseline from the raw materials.

The adapter had broken seventeen times, but the inversion main chain had never changed its conclusion boundary even once because of it.

In the winter of the forty-second year of the current cycle, fine ash fell continuously outside the Outpost for seventeen days.

Jiang Lin was already sixty years old, and those size-8 characters at the bottom of the horizontal screen required stepping back half a step to see clearly; he simply rearranged all proof dependencies and false alarm boundaries into size-12 characters, at the cost of the same page expanding from three screens to five screens.

The machine had no objections to this, only recording an extra 267 megabytes in the storage requirements column.

In the second month of the forty-third year of the current cycle, he spent four months completing the final blind test, sealing the precursor layer model, public data adapter, synthetic injector, minimum observation network, and false alarm list into the same long-term research package.

[Star-Tide precursor layer inversion model: Closed]

[2022 Public data baseline: Completed]

[Minimum observation network: Three layers / Four types of observables]

[Conservative effective early warning window: ≥ 21 years]

[Reality co-orbit assumption: Prohibited]

[Publication status: Sealed]

Jiang Lin left the FDSO raw data in the permanent isolation zone, adding only the algorithmic structure, false alarm boundaries, and the 2022 public data inventory to the regression candidates.

The five curves on the horizontal screen were archived, while the unified time axis remained on the control room wall.

The far left was the blue line of the tenuous precursor layer squeezing the heliopause boundary for the first time; the middle was the gray line of the Mars and lunar backside links successively extinguishing; and the right side was the peak of the main front still rushing over.

The black line representing the freeze-preservation of TM-7 stopped six years and seven months before the main front.