Part 3 — The Mechanism and the Tests

6. The Ocean Floor and the Reversal Record

The two preceding sections did two things. The first questioned the calibration-dependent methods, where a real measurement is converted to a deep-time age through an assumption the event violates. The second agreed with the robust methods, which read undisturbed material and read it true. Questioning some and agreeing with others is the whole of a critique. This section is not a critique. It is the one place where the model does not respond to the conventional dating record but adds to it — a mechanism the standard account has no equivalent for, on a piece of evidence that at first looks like a rapid emplacement's fatal contradiction and turns out to be its natural product.

The ocean floor is dated by its magnetism, not its chemistry

The conventional age of the ocean floor does not, in fact, rest primarily on radiometric dating of the ocean floor. Fresh submarine basalt is notoriously difficult to date directly — for exactly the excess-argon reason developed in the calibration-dependent section — which is why the community does not attempt it systematically. The age grid of the Atlantic is a magnetic age. Ship-track surveys map the pattern of magnetic-polarity zones frozen into the crust, identify each boundary, and convert distance from the ridge axis into age by assuming a spreading rate. The reversal timescale that calibrates this conversion is anchored, in turn, by radiometric dates — chiefly argon-argon on subaerial lavas at its young end. So the load-bearing assumption behind the 180-million-year Atlantic is not an initial-daughter assumption at all. It is the assumption of a slow, roughly constant spreading rate integrated across the entire interval.

That assumption is precisely what this model replaces. The velocity history in the evidence inventory (Section 3) has the Atlantic opening fast and then decaying toward the modern few centimeters per year — the profile derived in the Trigger paper. Appendix B uses that same profile for a different job: the rate at which cold slab is delivered into the mantle. The magnetic data is not in dispute; the pattern is real and it spans the basin edge to edge. What differs is the spreading history the pattern is read through. Read through a constant modern rate, the zones span a hundred and eighty million years. Read through the catastrophe's velocity profile, the same zones span thousands. This is not a radiometric argument, and it does not require any dating method to fail. It turns entirely on which spreading history is correct.

When the record could begin: batch cool, then surface freeze

A magnetic band cannot form in molten rock. The recording of the field is gated by a coherent lid, and that lid is the same gate the sea-level work already uses.

The heat budget of the new floor is Trigger Appendix F. Fresh mantle is a hot, buoyant pond. Quenched skin remelts back into the mix until enough heat has left that the batch sits near the solidus. That wait is the slow step. Paper 5 anchors the first defensible sea-level point at that transition: year 293–463, the fast-to-slow boiling-flux bracket on F's Phase 1→2 crossing. No sea-level claim, and no coherent magnetized plate, before that bracket.

Once the batch is cool enough to hold a freeze, the top against seawater locks fast. That is the same rapid subaqueous quench used in Section 4 for trapped excess argon. The skin takes a near-instantaneous snapshot of the field. The interior, insulated beneath, finishes more slowly and averages more of the field's motion. The layer that dominates the sea-surface magnetic anomaly is that fast-frozen upper skin. The ridge axis, where the newest crust is still forming, is the last part to close and is not the already-set floor the survey reads as older stripes.

Freeze geometry follows the same cooling: inward from the cold continental margins and downward from the ocean. Early recording is therefore margin-in. After a plate exists, further recording is only in the central crack.

What a fast surface freeze records of a disturbed field

A rapidly frozen surface does not average the field over a long cooling interval. It snapshots whatever the field is doing as that strip of crust passes the blocking window. Where the freeze is fast against the field's motion, the parcel locks a clean direction; where it freezes slowly, it averages and records something messier.

The field being snapshotted is not a normal dipole calmly reversing. Cold-slab delivery to depth holds the dynamo in a disturbed, low-dipole state for the active phase — the regime strong, equatorially weighted CMB heat-flux forcing is known in the dynamo literature to produce. In that state the field is weak and directionally restless. A weak field still records: grains cooling through their blocking temperatures lock a direction at an intensity that scales with field strength. That is how existing transitional lava and crust records were written at all.

The texture follows. Stronger, cleaner bands where the field momentarily organized; weaker, messier ones where it was collapsed. Neither a metronome of identical chrons nor undifferentiated noise: major correlatable structure with short-wavelength complexity superimposed. The conventional account reads that fine structure as paleointensity fluctuation over long time. This model reads it as a disturbed field varying while the surface froze. Stripe width is not a prediction. The field's fidgets are not computed here, and the recording geometry is not “the ridge wrote every band from day one.”

The delivery of the perturbation

Subduction geometry is mapped, not invented. What changes is the rate. The diving edge carries pre-event floor downward on the same velocity history that opens the basin. At event speeds the cold mass reaches the deep mantle in centuries — short against the window, negligible against ordinary mantle-convection time. The perturbation begins as cold material arrives.

Appendix B works the supply of thermal deficit into the mantle. At peak that supply exceeds modern total CMB heat flow by roughly five orders of magnitude; stacked adversarial haircuts still leave about three. The slab arrives still cold. What fraction of that deficit couples into the actual boundary heat flux is not modeled. The margin is wide enough that coupling would have to be extraordinarily inefficient before the published O(1) disturbance threshold was missed. Enhanced equatorial CMB cooling is what the dynamo literature finds most effective at suppressing a stable dipole; low-latitude slab delivery is that pattern. Amplitude class and geometry fall out of where the slab goes. They were not arranged to match a stripe count.

Still running

Once the cold mass has ponded at depth it warms only by conduction. Across a body tens of kilometers thick that is a long process. Over a few thousand years the insulating rind grown by diffusion is thin against the slab. The thermal deficit is therefore still present. That is not a claim that the CMB heat-flux anomaly sits at “full strength” today. Coupling of deficit into boundary flux is the unrun dynamo/mantle model. What faded quickly was the violence of delivery — plate speed on a few-hundred-year decay — not the existence of the cold mass.

The same youth shows in the floor's own heat. Observed Atlantic heat flux averages about sixty milliwatts per square meter, which Trigger Appendix F converts to a residual of roughly 4 × 10²⁷ joules still stored in the new-floor body. A floor with a hundred and eighty million years behind it has no formation heat left to hold. Two present-day quantities — a field still unsettled and heat still stored in the lid — are the live tail of one young event.

One boundary stays clean. This account concerns the event interval and the new ocean floor. The remnant Pacific carries the long pre-event field history and is not reinterpreted. A field behaving this way for reasons intrinsic to the core would mark every basin alike. The signature is specific to the new floor; the cause is specific to the new floor's formation.

What the argument establishes, and what it does not

If the Atlantic opened on that velocity profile, the field was disturbed while the new floor froze, and the new floor froze fast at the surface once the batch was cool enough to hold a lid. A disturbed field snapshotted that way writes mixed clean-and-messy structure. The mechanism is assembled from parts already on the table: F's batch-cool gate, the quench that traps argon, Appendix B's deficit budget, and the fact that weak fields still record.

What the argument does not do is compute the result. It does not claim a stripe-width pattern, a reversal count, or that a dynamo forced this way reproduces the Atlantic in detail. That simulation is Test 4, not a result here.

7. Modern Damped-Tail Observables

A violent event inside the window cannot have left the Earth already at rest. The systems the mechanism touched should still be settling, and the settling should still be measurable. The prediction is qualitative and forced. The rates below are measured, not predicted.

The Ocean Floor's Own Heat

Observed heat flux through the Atlantic floor averages about sixty milliwatts per square meter. Under the model that flux measures formation heat still stored beneath a thickening lid. Converted through that lid, the measured flux implies a residual of approximately 4 × 10²⁷ joules (Trigger Appendix F). Under the conventional age the same floor has had a hundred and eighty million years to drain, and has no formation heat left to hold.

Geomagnetic Secular Variation

The Earth's magnetic dipole has been weakening at approximately 5 to 6 percent per century since systematic measurement began in 1840 (Olson & Amit 2006). The South Atlantic Anomaly has been expanding and deepening over the same period. The north magnetic dip pole accelerated through the late twentieth century and in recent World Magnetic Model epochs has been near 36 kilometers per year (WMM/IGRF; Thébault et al. 2015). Regional patches tracked by Swarm since 2014 show a complex pattern of change.

The conventional reading is ordinary secular variation. That reading is not disputed as a description of core flow. The question is whether the present rate and pattern are a long-term steady state or the tail of the same CMB perturbation Section 6 and Appendix B bound. The Steens Basalt is an analog for style only — jerky directions, sustained low intensity — not for an exact duration, and not for the contested within-flow rate claims.

A damped return on the order of a couple of thousand years is a stated check, not a derived output. If the field reverses or strengthens abruptly, that check fails.

Himalayan Isostatic Rebound

GPS rock uplift in the Himalaya is about 2 mm/yr against a long-term thermochronologic average near 1 mm/yr (Han et al. 2024, Chomolungma sector). Han et al. attribute the disconnect to drainage capture; that process is not disputed. Under this model the extra unloading includes post-event ice on the high country: the rift moisture engine built ice after the event, then that load diminished as the engine relaxed. Volcanic aerosol rides with the opening; it is not the ice clock. No specific modern rate or ice thickness is back-calculated.

Patagonian Glacial Isostatic Adjustment

GNSS stations around the Patagonian ice fields record 18 to 41 mm/yr of uplift (Lange et al. 2014). The conventional mechanism — ice-mass loss over a low-viscosity mantle — is not disputed. What is disputed is only the timing of the load: emplaced in the post-event accumulation phase, losing mass since. Patagonia is the louder of the two rebound cases. The formal test is Section 9.

Residual Radiocarbon in Ancient Carbon Reservoirs

Listed here because, unlike the others, it is visibly decaying on a human timescale. Coal dated to 37–318 million years typically shows 0.16–0.33 pMC (RATE; broader span 0.10–0.46). Mesozoic fossils show 0.61–5.7 pMC (Miller et al., AGU 2014 conference abstract), with a mainstream mosasaur value of 4.68 pMC (Lindgren et al. 2011). Section 4 develops the dilution model and the contamination competitor. The source-variation test is Test 7. Residual ¹⁴C is not, by itself, the timeline.

The claim for this section is the pattern: the systems the mechanism touched are still moving. Each line has a conventional explanation of its own.

8. A Request for Specialized Modeling

What remains is not unfinished reasoning but unrun computation. Four full-scale models would carry the assessment past what reduced-order work can reach. Each is specified tightly enough to confirm or break the framework.

The geodynamo model

A three-dimensional simulation forced by the CMB supply history in Appendix B. Confirm: a suppressed, restless low-dipole regime the fast-freezing surface could snapshot into mixed texture. Break: the dipole rides through essentially undisturbed. No reversal tally is required. This is Test 4.

The climate model

A coupled atmosphere and moisture-transport model of the three-basin state and the rift supply curve (Appendix F; Capstone Appendix A). Confirm: enough moisture, without pre-event ice, to feed the volume budget, with volcanic load riding along. Break: cannot supply the ice. Fabric of the core is what the run returns, not a pass/fail already specified as annual bands.

The ice-sheet model

Initialized with no pre-event polar ice. Driven by the climate run. Confirm: observed volume inside the window on a front-loaded accumulation-rate history (Appendix A.4); deep ice that can be chemically loaded (dust, salt, volcanic) even if visually massive. Break: needs far longer than the window, or cannot make the ice at all. Countable basal couplets are not a required output.

The tectonic model

Three-dimensional failure and opening, of which the velocity profile is a reduced-order projection. Confirm: rapid opening that decays toward modern rates and closes the observed separation inside the load-clock window, with underthrust that delivers cold slab to depth. Break: cannot open the basin in the time allowed, or demands a spreading history incompatible with the profile.

None of these four has been run here. They are on the record so that whoever has the tools can pick them up.

9. The Discriminating Tests

Twenty-one tests. For each, both predictions are stated in advance. Every test can count against the model.

Geophysical Tests

Test 1 — Mid-Atlantic Ridge lithospheric thickness transect. Conventional: thickness follows √t with distance from the ridge. This model: basin-scale thickness is approximately uniform — crust of one age, not a 180-million-year fan. Any residual should be the opposite of half-space: a somewhat thicker, quieter lid toward the old margins that locked first, livelier toward the axis that stayed open longer. One transect. Two qualitative shapes.

Test 2 — East Pacific Rise westward transect. Conventional: continuous Pacific crust thickening westward from the EPR. This model: the EPR is where new post-event crust meets old remnant shell — a detectable step in thickness, velocity, or thermal structure. MELT's west-flank asymmetry is already on the table as a candidate signature, not as a completed proof.

Test 3 — Ocean-floor magnetic stripe texture. Conventional: wide, clean polarity bands written by a stable dipole over long time. This model: mixed clean-and-messy structure written by a disturbed field onto a fast-frozen skin, first from the margins inward, then only at the central crack. Stripe width is not predicted. No reversal count is predicted. The Steens Basalt (Mankinen et al. 1985) is a style analog — jerky directions, low intensity — not a duration assigned to the ocean floor.

Test 4 — Geodynamo simulation driven by Appendix B forcing. Conventional: no forced excursion on this timescale. This model: a disturbed, suppressed low-dipole regime for the duration of the forcing. Discriminator is the regime, not a tally.

Radiometric Tests

Test 5 — Systematic known-age recalibration study. Error distribution across published historically dated volcanics, applied to unknown-age samples. Small symmetric errors: conventional ages survive. Large systematic excess: compression.

Test 6 — Replicate Dalrymple and Moore with modern instruments. Fresh zero-age submarine pillows by K-Ar, Ar-Ar, and U-Pb. Million-year excess persisting confirms 1968 at modern precision. Excess vanishing strengthens the conventional reading.

Test 7 — Expanded residual ¹⁴C survey. Multiple materials, multiple labs, blind blanks. Contamination should track protocol and vanish with rigor. A genuine diluted signal should track carbon source and persist.

Genomic Tests

Test 8 — Population-invariant private mutational load. Equal-sized ancestry subsamples in gnomAD v4. This model: approximate parity. Out-of-Africa: higher African private load. Specified in Paper 11. Not yet run.

Test 9 — Core versus peripheral modularity boundary. Sharp step versus smooth conservation gradient.

Test 10 — Stress-induced repertoire expansion. Localized to peripheral modules versus genome-wide hypermutation.

Test 11 — Drift equation on further families. Window fixed; solve for required Ne. Ordinary sizes continue, or a family lands off the biological map.

Test 12 — Late-onset degradation clustering. Shared maintenance-subsystem failures versus unrelated causes.

Climate and Earth-System Tests

Test 13 — Ocean-floor heat-flow transect. Conventional: heat flow falls with distance from the ridge (√t). This model: approximately uniform at basin scale, hydrothermal noise allowed; any systematic residual should be higher toward the axis, not lower toward “older” crust.

Test 14 — Geomagnetic decay trajectory. Stated check: continued dipole weakening at a gradually decreasing rate, SAA evolution, no abrupt reversal or sudden strengthening. Shape of the trajectory, not a derived time constant.

Test 15 — Deep-ice character against Appendix A. Not a Paper 5 snowfall field and not a layer-count exam. Reanalysis of existing cores for a front-loaded accumulation-rate history and for chemical load (dust, salt, volcanic) in the deep ice even where visual banding is poor. Seasonal couplets appearing later, as the engine relaxes, are allowed. A record that requires a pre-event ice sheet, or that cannot be fed inside the window, counts against the model.

Archaeological Tests

Test 16 — Additional organized-from-founding sites. Organized construction from the first occupation layers. A documented experimental precursor phase counts against.

Test 17 — Transported domesticates without local wild precursors.

Test 18 — Early specialist trade before bulk commodities.

Isostatic Tests

Test 19 — Rebound rate decay. Himalayan and Patagonian rates should decline over coming decades if the load is recent post-event ice. Trajectory over decades, not a derived time constant.

Geological Tests

Test 20 — Bright Angel high-resolution weather record. Centimeter-scale section through a complete Bright Angel. This model: storm interbeds thin and fine upward as the rift supply curve decays — hot-basin evaporation falling after the boil — not as a readout of plate speed and not as an SST hurricane table. Conventional: tidal/seasonal/tectonic cyclicity without that monotonic wind-down. Field methods only.

Specimen Tests

Test 21 — Radiocarbon dating of the Devon Island rhinoceros. Conventional (~23 Ma): zero ¹⁴C. This model: detectable ¹⁴C. Apparent age near the 4,725–7,200 window places the animal before the event; an inflated age places him in the dead-carbon pulse. The specimen is in a collection. Organic carbon is confirmed. Zero versus detectable is the test.

The Count

Twenty-one tests. Four geophysical. Three radiometric. Five genomic. Three climate and Earth-system. Three archaeological. One isostatic. One geological. One specimen.

Every test specifies both predictions. Every test can break the model. The predictions are on the record.

A full series of evidence. Twenty-one tests that could break it. That is the dare.

← Part 2 · Series · Part 4 →

© 2026 D. L. White. Licensed under CC BY-ND 4.0. https://creativecommons.org/licenses/by-nd/4.0/

AI Collaboration Disclosure: Drafting and calculations by Claude (Anthropic), with adversarial review by Grok (xAI), under the direction of D. L. White. Neither AI system endorses all conclusions as settled.