Part 2 — Dating Distinctions

The second of four parts. Part 1 sets out the nineteen lines of evidence and the timeline the dating methods are tested against here.

One question organizes both sections that follow: was this material disturbed by the event, or was it not?

Where the answer is yes, a real measurement is converted into an age through a calibration the event violates, and the result is too old in a direction that can be calculated. Where the answer is no, the method reads undisturbed material and reads it true — and the framework says so plainly rather than quietly declining to mention it. Section 4 takes the first half, method by method. Section 5 takes the second.

A framework that only attacked dating methods would be special pleading. The test of one that is not is whether it can say in advance which methods it expects to be wrong and which it expects to be right — and then be held to both lists.

4. Calibration-Dependent Methods

The evidence inventory produces dates in the thousands of years. The conventional geological chronology produces dates in the millions. The disagreement is not subtle — it spans three to four orders of magnitude. Both frameworks cannot be right, and a project setting out nineteen mutually constraining lines consistent with a recent date owes the reader a direct account of why the established chronology gives a fundamentally different answer.

The explanation is not that dating is broken.

The Common Structure

A large family of dating methods shares one architecture: a real, reproducible measurement is converted to an age through a calibration, and the calibration — extended into deep time — carries the very assumption in dispute. This is true of potassium-argon and argon-argon, radiocarbon, luminescence, cosmogenic-nuclide exposure dating, thermochronology, and the molecular clock. It is worth stating plainly what is and is not claimed for all of them, because the honest position is narrower than either side's slogan, and it is the same position for every method in this group.

The measurement is sound. Fission-track densities, argon ratios, nuclide inventories, SNP counts — these are physically present and reproducibly measured, and nothing here disputes them.

The recent, historically-anchored end of each calibration is sound. Where a method is tied to a genuinely known reference — a dated eruption, a documented founding, a tree-ring-anchored horizon — it works, and the framework grants this without qualification.

What the framework contests is the single component these methods share: the extrapolation of a calibrated relationship deep into the past on the assumption that present conditions and rates held throughout. That extrapolation cannot, by its nature, independently establish the timescale it presumes — because the timescale is built into the calibration. This is the sense in which the conventional chronology buys long time with slow rate: the great ages are not measured directly but derived by dividing an observed quantity by an assumed slow rate, so the deep time is a consequence of the rate premise, not an independent finding. A method calibrated on that premise cannot then be cited as independent confirmation of it.

None of this makes a method "broken." Each has a specific, calculable vulnerability under the conditions the catastrophe model specifies — and each returns its true answer wherever those conditions do not apply. The physics of radioactive decay in particular is among the most precisely measured phenomena in science. Decay constants for potassium-40, uranium-238, uranium-235, rubidium-87, and carbon-14 are known to high precision and are not in dispute — not by this project, not by any serious framework. The atoms decay at the rates they decay. Nothing here requires or proposes any change to nuclear physics. The vulnerability is never in the physics, and never in the rates. It is in the conditions the calculation assumes — the state of the system at time zero, and the environmental conditions that prevailed while the clock ran.

Every age of this kind takes the same general form: measure a ratio, divide by a known rate, and read off an age — but only if the assumed initial condition matches the actual one. Where the assumption deviates from reality, the age is wrong. The known-age evidence developed method by method below shows that the error, where it occurs, is not random but has a consistent direction — a claim earned there, on the data, rather than asserted here.

The scope of the contest is important. The initial-condition vulnerability is specific to methods that must assume a fixed starting point. Methods built to avoid that assumption — isochron and uranium-lead among them — are treated separately and are not contested here; they belong with the methods that comport with the model, taken up in the following section.

Before a single date is touched, it is worth settling what the compression does and does not reach. If the event was violent enough to reset the clocks, how is anything still in order? The event touched what it touched and left the rest.

Where the event did not rework the material — the standing pre-existing column, wherever the disruption did not reach — relative order is preserved, and only the absolute ages compress. Everything stratigraphically older remains older. The geological column was built by relative methods: superposition, cross-cutting relationships, faunal succession, established before radiometric dating existed. Compressing absolute ages leaves that sequence untouched.

Where the event did rework the material, the arrangement is the event's: hydraulically sorted, mixed provenance, old fragments in young beds. That is not a contradiction. It is the event's signature. Disturbed and undisturbed are fixed by the event's geography — failure sutures, opening basins, paths of catastrophic transport — not chosen at need. That map is not drawn here.

K-Ar and Ar-Ar

The potassium-argon system dates a rock by the ratio of radiogenic argon-40 to the potassium still present. The argon-argon variant measures both quantities on one sample. Both methods assume the rock contained negligible argon-40 at formation. In a subaerial eruption that assumption is often reasonable. It is not reasonable where argon cannot escape, or where ambient argon was elevated, or where the rock quenched too fast to degas.

The case that bears on this project's timeline is submarine basalt — the ocean floor's own rock. The excess-argon mechanism is seen most plainly first in subaerial flows of documented age.

The known-age evidence. Dalrymple's studies of historic Hawaiian flows report the excess; it is not in dispute for those samples. Every case that follows comes from the conventional literature. Hualalai, Hawaii (erupted 1800–1801) returned K-Ar whole-rock ages of roughly 1.4 million years (Dalrymple 1969). Historically dated Mt. Etna flows returned roughly 0.14 to 0.35 million years (Dalrymple 1969). Sunset Crater, Arizona (1064–1065 AD) returned about 0.27 million years. The 1959 Kilauea Iki flow returned roughly 1.7 to 8.5 million years (Krummenacher 1970). The 1915 Lassen Peak eruption returned about 0.11 million years (Dalrymple 1969). In every case the error runs one direction: too old. Excess argon can only add apparent age.

The conventional explanation — argon-40 trapped at formation — is almost certainly correct as a mechanism. Nor is the claim that all subaerial flows fail. Excess appears under specific conditions and is absent under others. Vesuvius 79 AD sanidine returns the correct calendar age. The question the failures force: do the conditions that produce excess argon also characterize the unknown-age samples on which the geological timescale depends?

Submarine basalt. Dalrymple and Moore (1968) dated fresh pillow basalt from the submarine flanks of Kilauea and Mauna Loa — rock essentially zero-age. Excess radiogenic argon increased with water depth and was highest in the rapidly quenched glassy rims. Apparent ages from hundreds of thousands to tens of millions of years have been reported for these and related zero-age Hawaiian submarine basalts (Dalrymple and Moore 1968; Noble and Naughton 1968), confirmed under modern protocols (Guillou et al. 2017). Many submarine basalts are simply unsuitable for K-Ar, because they trap mantle-derived argon at formation.

The age grid of the ocean floor is not primarily a radiometric age — it is a magnetic-pattern age read through an assumed spreading rate, which the Trigger Standalone contests. Radiometric dating still enters as the young-end anchors of the reversal timescale and in the rarer cases where submarine basalt is dated directly. When K-Ar returns million-year ages on rock of known zero age, under the same formation conditions that characterize the new ocean floor, the radiometric contribution to that timescale is called into question.

Dalrymple and Moore's strongest finding was that the excess peaked in the glassy rims — fast cooling against seawater traps argon regardless of depth. The new basins form that way: molten basalt quenched against flooding seawater (Trigger), during mantle degassing that elevated ambient argon. Rapid quench and elevated ambient argon both trap excess at the surface. Pressure adds to it at depth; the argument does not rest on pressure.

Where the method works. Ar-Ar on Vesuvius 79 AD sanidine is noted without qualification. Sanidine crystallizes hot and degasses. The ocean floor, formed from fine-grained or glassy submarine basalt, resembles the Kilauea pillows more than the Vesuvius sanidine.

The recalibration. Three specified conditions violate negligible-initial-argon during and after the event: mantle degassing as the shell fails; hydrothermal circulation through fresh crust; rapid subaqueous quench. Highest excess for rock forming during the event; decaying toward correct ages as ambient argon falls. The qualitative prediction is stated. The full compression curve is a future target.

Radiocarbon

Radiocarbon dates organic material by the ¹⁴C/¹²C ratio. The method assumes the atmospheric ratio at death was approximately the modern pre-industrial value, or a tree-ring-calibrated value. The measurement is sound and the recent calibration is sound. What is contested is the assumed atmospheric ratio during and after the event.

Three specified conditions perturb that ratio. Dominant: injection of ¹⁴C-dead carbon — mantle CO₂ and ocean overturn. Secondary: warm-ocean suppression of CO₂ uptake. The hot rift basins are small; the large remnant ocean warms only a few degrees, so net global suppression is modest. Third: if the field was weakened, ¹⁴C production rose and partially counteracted dilution. Direction of the bias is stated; the full carbon-cycle curve is a future target.

An organism dying when the ratio is most depressed appears far older than one dying after re-equilibration. A conventional 40,000-year date may be a few thousand years of true time if the initial ratio was low enough.

Residual ¹⁴C. If the event falls inside the window, material conventionally dated to millions of years formed during or shortly before it. At the central 5,786 years — about one half-life — any material that contained ¹⁴C at formation should retain roughly half of it. Across 4,725 to 7,200 years the surviving fraction runs from 57 percent down to 42 percent, all detectable by AMS (practical limit ~0.1–0.5 pMC). Coal dated 37–318 million years typically shows 0.16–0.33 pMC (RATE span 0.10–0.46; Baumgardner et al.). Mesozoic fossils show 0.61–5.7 pMC (Miller et al., AGU 2014 conference abstract). A Late Cretaceous mosasaur humerus returned 4.68 ± 0.1 pMC (Lindgren et al. 2011); its authors attribute that to bacteria. That is the competing explanation the source-variation test is built to separate. Diamonds at 0.008–0.31 pMC (RATE) are not leaned on at the low end, which sits at instrument background (Taylor and Southon 0.005–0.03 pMC). Deep-mantle carbon never saw cosmic-ray production. The load-bearing measurements are those above detection. At millions of years the expected content is zero, not “approximately zero.”

Contamination is real. The model predicts a combination: genuine residual from formation in the window, diluted by dead carbon. Variation should track source, not only laboratory protocol. Test 7 is that distinction.

Luminescence (OSL and TL)

Optically stimulated and thermoluminescence dating measure the time since a mineral grain — usually quartz or feldspar — was last exposed to sunlight or heated above roughly 400°C. After that reset the grain accumulates a radiation dose from its surroundings; the measured dose divided by the measured dose rate gives the age. Two assumptions are critical: that the grain was fully reset before burial, and that the dose rate has been roughly constant since. The luminescence case is necessarily site-by-site rather than blanket — whether a deposit is affected depends on its specific transport and burial history — and what follows identifies the conditions under which bias is expected, not a uniform correction.

On reset completeness: the Deposition Series moves massive volumes of material by wind-driven surge and flood pulses, often in sediment-laden water or dense aerial suspension. Many grains never see sunlight in transit; they arrive carrying a residual signal from their prior context, which the method reads as extra accumulated dose — an age too old. On dose-rate constancy: the early post-event environment carried elevated background radiation from fresh volcanic deposits, exposed crystalline basement stripped of cover, and radioactive minerals redistributed by catastrophic hydraulics. Where the early dose rate ran well above the modern measured rate, using the modern rate overestimates the age — the grain accumulated its dose faster than assumed, so less time was required. Which sites were violent enough to leave grains incompletely reset follows the surge history, not a plate-speed anemometer. Some sites may need large corrections; others — slow deposition under clear skies with normal background — little or none.

Cosmogenic-Nuclide Exposure Dating

Cosmogenic exposure dating measures the inventory of nuclides — beryllium-10, aluminum-26, chlorine-36 — that accumulate in rock while it sits near the surface exposed to cosmic rays, and converts that inventory to an age using an assumed production rate. Three things must be held apart.

The measurement is real. The inventories are physically present and reproducibly measured.

Most exposure ages are not contested. An old surface that accumulated its inventory before the event carries it intact, and mechanical disturbance does not reset it: cosmic-ray exposure records time near the surface, not location. Only a change in shielding depth matters. Most long exposure ages are therefore pre-event inheritance the model expects and does not contest.

The contested component is narrower. Production rates are calibrated against surfaces of independently known age, an assumed constancy of cosmic-ray flux, and an assumed geomagnetic-field history. The recent, historically-anchored end is real. The deep-time extrapolation inherits a field history the framework contests: a disturbed, at times weak field would have raised the flux, and thus the true production rate, above the calibrated value during that window. Where the calibrated rate is too low, the computed age is too high. That does not independently settle the age of the disputed window.

Surfaces the model identifies as freshly cut by the event — canyon walls cut in the post-fill slot — are the one local, positive case: those inventories should start at the event. Inheritance and shielding cut both ways, so that case is a measurement that can go against the model, not a claim that cosmogenic dating is broken.

Thermochronology (AFT, AHe, ZHe)

Low-temperature thermochronometers — apatite fission-track, apatite and zircon (U-Th)/He — record the time since a mineral cooled through a characteristic temperature window. The measurements are real. Converting them to an age requires a kinetic model and an assumed thermal history.

For old, undisturbed crust that cooled slowly and monotonically, those ages stand. For rock that took the event's rapid heat-and-burial path — a history the standard kinetic models are not built to expect — the same inventory interpreted on a monotonic multi-million-year cooling path is the product of that assumed history. The data do not independently demand the slow path. For material the event did not disturb, thermochronology's ages stand, and the model expects them to.

Ice Cores

Ice cores are the one annual-layer record whose chronology is claimed to run continuously through the event window and to read that window as slow — which is why, alone among the layer-counting records, ice belongs here rather than among the methods the model agrees with.

A polar ice sheet is dated in its upper reaches by counting seasonal cycles in dust, chemistry, and isotopes. In Greenland that counting extends to roughly sixty thousand years (GICC05 to 60.2 ka before 2000 CE). Below that, and in low-accumulation Antarctic cores, the published timescale is ice-flow modelling, not a count. The layers that exist are real. What is calibration-dependent is the conversion of every countable band into one calendar year, and the assumption that the whole sheet is a long pre-event plus post-event stack.

On this model the ice is entirely post-event. The pre-event world held no perennial ice (Appendix A.1). The moisture engine starts when the new rifts exist and the thermal gradient is steep. The whole record, not a recent cap, belongs to the window.

A modern counted year is several signals moving together. Under the post-event engine those signals are supplied together by the same evaporation from the same hot basins under the same aerosol load. Independence of the certifying proxies is reduced. Counted as annual, the record over-reads the window. That is the claim about the conversion. It is not a published count of bands per year at the base.

Appendix A is the supply budget: volume target defined, moisture adequate by one to two orders of magnitude, required efficiency in the neighborhood of the modern polar-to-global ratio. Accumulation rate is front-loaded — rising while the rifts widen and still boil, then decaying as they cool (A.4). That curve is a rate history, not a year-count.

The same engine predicts the character of the deep ice. High-supply fall is thick, wet, and chemically mixed — dust, salt, and volcanic load arriving with nearly continuous snowfall. Visual banding may be poor or absent. After the engine relaxes, ordinary seasonal banding can appear. Loss of annual resolution at depth is already reported in the literature. Convention reads that as thinning and diffusion over long time. This model reads it as a deposit that was never annual. Both mechanisms can produce a smeared-looking base. The smear is therefore a forced consequence of this engine, not a patch to an observable, and not by itself a clock. A survey that can tell diffused years from never-annual chemistry would discriminate. That survey is Test 15, not a result here.

Volcanic aerosol rides with the opening. It is not the ice clock.

The Molecular Clock — The German Shepherd Test

Paper 2 used the German Shepherd, founded in 1899. Dog10K SNPs compared to wolf genomes by standard molecular-clock methodology overestimate the true divergence by a factor of 336 to 2,297, depending on the published rate.

The mutation rate is not wrong. 99.95 percent of the observed differences are ancestral variants sorted into the breed, not new mutations since 1899. The clock counts all differences as accumulated mutations. The error runs the same direction as excess argon: too old. The same logic applies wherever ancestral sorting dominates a molecular-clock calibration.

The Study That Has Not Been Done

Known-age volcanic failures and the German Shepherd test are each given a local explanation in the conventional literature. What has not been done is a cross-method error-distribution study on all available known-age samples, then applied to unknown-age samples. Small random errors: conventional ages survive. Large systematic excess: compression. The study is Test 5. Its absence is a datum.

Across every method in this group the pattern is one pattern. The measurement is precise. The physics or genetics is correct. The vulnerability lives in the conditions — the state of the system at time zero, and the environment while the clock ran. Where the post-event setting differed from what the calibrations assume — and the Trigger, the Diaspora Series, and the Deposition Series specify how: plate velocity as a function of time, dead-carbon and argon injection, a rift moisture-and-heat supply curve that drives weather and ice — the calculated ages are biased too old. These are not vague claims that “conditions were different.” They are named mechanisms. Full numerical curves for each method remain future work. The next section turns to the methods that do not carry this vulnerability.

5. Model Predictions in Agreement

The previous section dealt with the methods whose deep-time calibration carries the disputed premise. This section is the other half. The organizing question is the same: was its record disturbed by the event, or not?

Uranium-Lead and Isochron Methods

These methods are not recalibrated. Isochron methods solve for the initial daughter ratio. Uranium-lead uses two chains and is typically applied to zircon, which rejects lead at crystallization. They are built to avoid the initial-condition vulnerability of Section 4.

Applied to old, undisturbed crystalline rock they return its true age. The model does not claim undisturbed material is young. The remnant Pacific lithosphere should read old by every robust method applied to it. Old ages on remnant lithosphere are a positive expectation, not a concession.

A further prediction: old material inside young deposits. Catastrophic transport carries pre-existing rock. Dating a clast returns the clast's age, not the deposit's. A spread of old ages on reworked fragments is the transport fingerprint.

Paleointensity and Remnant Directional Records

Pre-event remnant lithosphere carries the pre-event field and is not compressed. The disturbed-field regime of Section 6 and Appendix B belongs to the new floor after the lid exists. The Pacific is the control.

Layer-Counting Records: Trees, Speleothems, Coral, Varves

These records belong in agreement, not among open million-year gaps.

Where a record accumulated without the event touching it, the increments are real and the annual reading is sound. A tree-ring or speleothem spanning several thousand years is mostly — often entirely — pre-event. A reef read at tens of thousands of years is mostly pre-event structure; only the youngest growth can intersect the window. That is a small correction on the young end, not a compression of millions into thousands.

Varves are the same rule with one honest qualification. Long sequences are overwhelmingly pre-event. Where they intersect the event, deposition being the event's specialty, more than one couplet in a year is possible. That impact is owned and bounded. Genuinely young sequences whose counts fall inside the window are the live case, not the deep Green River. New lakes that begin only a few thousand years ago are anticipated: ice retreat and reshaped drainage produce late-onset basins.

Undisturbed Material, as a General Principle

The framework contests disturbed material and agrees with undisturbed material. That is the whole of it. Collapse the distinction and the project looks like an attack on dating. Hold it, and robust methods on untouched rock, true ages on transported clasts, and pre-event field history in remnant crust are confirmations, not problems.

Salt and Inland Salinity Distribution

Salt retained in closed lows, flushed from drained lowlands, absent from highland lakes that never held seawater. Consistent with the model. Also consistent with ordinary endorheic geology. Agreement, not discrimination.

Taken together, the two method sections are two halves of one boundary. Where the event disturbed a record, its conventional age is inflated in a calculable direction. Where the event left a record untouched, it reads true. The model is not at war with the dating record. It is at odds with one specific assumption — the initial and boundary conditions during a defined window — and in agreement with everything that assumption does not touch.

← Part 1 · Series · Part 3 →

© 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.