Tree-Ring Dating

"The global signature of the two cosmogenic excursions independently validates the precise dating of well-replicated tree-ring chronologies from five continents." — Ulf Büntgen and 45 co-authors, Nature Communications, 2018¹

"multiple growth rings are extremely rare in the bristlecone pine, especially in the White Mountain area, according to Ferguson" — Walter Lammerts, Creation Research Society Quarterly, 1983, opening his case that they are not²

"these long dendrochronologies appear to consist of true annual rings for at least the past 9000 years.  This is an extremely important result which, in my opinion, warrants critical scrutiny by all recent-creationist scientists." — Gerald Aardsma, Institute for Creation Research, 1993³

Summary

A tree in a temperate climate grows fast in spring and slowly in late summer, and the two kinds of wood look different.  Count the boundaries and you count the years.  That is dendrochronology, and it is the oldest and best-anchored of the counting clocks.

It is also the one where the young-earth objection has the most respectable pedigree.  Trees really do sometimes lay down two rings in a year, and really do sometimes skip a year.  The question is not whether that happens but how often, in which direction, and whether it accumulates.

Ten conclusions, stated first.

1. The continuous, absolutely dated record is 12,410 years long, and no single tree contributes more than a few centuries of it. The Hohenheim oak and pine chronology runs from AD 2002 back to 10,461 BC. It is built from more than 7,000 oaks whose mean age is 176 years, 97% of them under 300, with a mean of 108 trees overlapping in every year and 96% of the length covered by more than 20 crossdated trees.⁴
2. The strongest verification is an event that happened in one year and left a mark everywhere. In 774 CE a burst of solar energetic particles spiked atmospheric carbon-14. An international consortium measured the carbon-14 in 484 individual rings from 44 chronologies on five continents, from 42° S to 72° N, 40 to 4,000 m elevation, 13 genera and 27 species, in living trees, historic timbers, archaeological wood and subfossil logs. All of them record it in the same ring.¹ A second event in 993 CE does the same.
3. Multiple rings per year are real, and they were demonstrated experimentally by a creationist. Walter Lammerts grew bristlecone seedlings, imposed a two- or three-week drought in August, then resumed watering, and got "four growth rings instead of three."² The experiment is sound and the observation is genuine.
4. The inference Lammerts drew from it was tested quantitatively and failed. He supposed extra rings formed in about 50% of years after the Flood, which would cut Ferguson's 7,100-year bristlecone chronology to about 5,600.² Ten years later Gerald Aardsma, also at the Institute for Creation Research, pointed out that this makes a testable prediction: the century-scale wiggles in the radiocarbon calibration curve, which are solar in origin and therefore have a fixed width in years, must appear systematically wider in rings wherever trees were doubling up. He measured thirteen of them. They do not widen.³
5. Aardsma printed the conclusion anyway. "these long dendrochronologies appear to consist of true annual rings for at least the past 9000 years," and therefore "the Flood must, in fact, have occurred more than 10,000 years ago."³ He proposed gaps in the Genesis 11 genealogy to accommodate it, and acknowledged that this leaves large unanswered questions. His test is summarised accurately in the movement's own 2018 review of the tree-ring literature, which offers exactly one counter to it — that it assumes solar-cycle periods have been constant.⁵ Nobody has developed that counter into a measurement.
6. The young-earth position has moved, and its current form is not "multiple rings per year." The most technically serious creationist analysis of the bristlecone data, Woodmorappe's statistical study of the archived ring-width database, states in its abstract: "No obvious grounds exist for questioning the overall validity of the bristlecone pine (BCP) long chronologies," and "Not enough is known about processes leading to plural annual rings in BCP for further consideration."⁶ The live proposal is instead that crossmatching is real but misdated: trees that grew at the same time are strung out into a false sequence by shared disturbances rather than shared climate, and joined by fortuitous "bridge" series.⁷ A 2026 Answers Research Journal paper builds a simulation that does exactly this.⁸
7. That model makes a sharp radiocarbon prediction, and its authors have not tested it. If a chronology is a chain of contemporaneous bundles, the trees at ring 8,000 and the trees at ring 4,000 lived at nearly the same time and hold nearly the same carbon-14. Plot measured carbon-14 against ring number and the result must be a staircase. It is a smooth curve. Neither the simulation paper nor the bridge-series abstract cites Büntgen's 774 CE network, Miyake's discovery paper, or Aardsma.
8. The crossdating statistics really can lie, and dendrochronologists said so first — and then fixed it. David Yamaguchi's 1986 paper is cited on the young-earth side for returning 113 positions "significant at the 0.001 level," including dates in the future.⁹ The same paper removed the autocorrelation that caused them, eliminated 112 of the 113, and landed on the one date that matched the independent stratigraphy.¹⁰
9. Two independently built chronologies on two continents agree on the carbon-14 content of the atmosphere to eight radiocarbon years. Irish oak latewood against Californian bristlecone whole rings, 1700–1500 BC: a weighted mean difference of −8.1 ± 1.9 ¹⁴C yr.¹¹ The claim that chronologies from Ireland, England and Germany "disagree strongly" is contradicted by the primary literature, which reports the German links "confirmed by comparisons to the independent Irish oak chronology… and the Göttingen oak chronology… back to 9741 BP."⁴
10. The chronologies do get revised, and the revisions are published. When the floating Preboreal Pine Chronology was finally linked to the absolutely dated oak record, the pre-11,250 BP part shifted 70 years older; oaks damaged by cockchafer infestation were identified and discarded; and "the number of trees in parts of the chronologies had to be reduced significantly."⁴ In the same paper, the dendrochronological placement of that floating chronology came out 8 years from the position independent radiocarbon wiggle-matching had assigned it.

The thing that is actually surprising

The young-earth literature on tree rings contains four incompatible positions, and its own 2018 review says so.  Roger Sanders, surveying the whole corpus for the Eighth International Conference on Creationism, sorts the authors into those who (1) accept rings as annual but reject crossmatching with dead logs, (2) propose multiple rings per year, (3) combine the two, or (4) "accepted annual rings and cross-matched master chronologies but extended the date of the Flood prior to those chronologies via biblically acceptable gaps."⁵

So the disagreement about whether bristlecone rings are annual is not primarily between creationists and dendrochronologists.  It is among creationists.  Lammerts and Matthews say the rings are not annual;  Aardsma measured that they are;  Woodmorappe examined the ring-width database and concluded that "no more than ~1% of MWK ring boundaries are potentially tainted by doubt";  and Sanders's review ends by calling for research to "build a consensus dendrochronology calibration model," which is an admission that there is none.

Part 1: How a chronology is built

One tree

In a conifer growing where the seasons differ, spring wood has large thin-walled cells and looks pale;  late-season wood has small thick-walled cells and looks dark.  The boundary between one year's dark wood and the next year's pale wood is a ring boundary.  Ring width varies with how good the growing season was, which is what makes the next step possible.

Many trees

No living bristlecone reaches 5,000 rings, and no oak reaches 600.  A long chronology is built by crossdating:  matching the pattern of wide and narrow rings — a barcode, in effect — between a living tree and a dead one that overlapped it, then between that dead tree and an older one, and so on.

Davidson and Wolgemuth put the practical limit plainly:  "it becomes increasingly difficult at a given location to find very old wood that reliably overlaps to yield an unbroken sequence far back in time.  A gap in the record may be due, for example, to climatic changes in the past when trees did not readily grow in that area, or a time interval when most of the fallen trees fully decomposed."¹²

The published European record:

• Holocene Oak Chronology (HOC): 10,430 BP (8480 BC) to AD 2002, from more than 7,000 individual oaks. The Main river valley alone contributes "an unbroken record from 8239 BC to AD 1100."⁴
• Preboreal Pine Chronology (PPC): extends the record to 12,410 cal BP (10,461 BC) — the whole Holocene plus 820 years of the Younger Dryas.
• Replication: mean 108 trees per year; 96% of the length covered by more than 20 crossdated trees.⁴

And in North America:

• Bristlecone pine, White Mountains, California: a continuous sequence to 6700 BC, 8,681 years.
• Bristlecone pine, White Pine Range, Nevada: an independently built chronology to 3258 BC, 5,238 years, about 300 km away.¹³

Why "the oldest tree is only 4,800 years old" misses

This objection appears in almost every popular young-earth treatment, including the 2016 Answers Research Journal article:  "if the earth really is as old as uniformitarian geologists claim, then why are there no bristlecone pines having 6000, 10,000, or even 20,000 rings?"¹⁴

The answer is that a tree's lifespan is limited by biology and by the ground it stands on, not by the calendar.  Bristlecones grow on eroding dolomite slopes;  eventually the root system is undercut.  The chronology does not depend on a single old tree and never did — the Hohenheim oaks average 176 years each.  The objection would be a good one against a method that required one 12,000-year-old tree.  No such method exists.

Part 2: False rings and missing rings

This is the real technical argument, and it should be stated at its strongest.

False rings

A false ring (or intra-annual density fluctuation) forms when growth stops and restarts inside one season — typically because of drought followed by rain.  The restart produces a band of dense latewood-like cells that can look like a year boundary.

Lammerts's experiment.  Walter Lammerts germinated Colorado-strain bristlecone seed in 1978 and grew seedlings under five treatments.  Those given "a short (circa 21 days) drought stress period in August of the third growing season showed up having one more growth ring than the control seedlings, that is four growth rings instead of three.  Also seedlings which received a two week drought stress period in August of the fourth growing season showed a similar extra growth ring."²

That is a clean experimental result and it should be granted.  Lammerts also anticipated the obvious objection — that seedlings are not mature trees — and answered it from the physiology literature:  after about three years, a conifer seedling "become[s] locked into an endogenous system of plant growth," behaving like a mature tree, so his three-and-three-quarter-year-old plants should be representative.²

His inference.  Lammerts argued that the White Mountains once had a "San Francisco pattern" of rainfall — wet winter, dry summer, early autumn rains — because Owens Valley and Lake Bonneville were far wetter in the past.  Under that pattern the young forests "would have grown an extra ring per year quite often."  Assuming extra rings in about 50% of the roughly 3,000 years from the Flood to about AD 1200, Ferguson's 7,100-year chronology "would be reduced to about 5600 years."²

And he cited the difficulty of detection.  Quoting Glock and Agerter on the San Francisco rainfall pattern, "the matter awaits careful study";  and observing that in his own material "the extra ring has a very similar diameter and might well have been formed during a regular but relatively dry growing season."²

One thing Lammerts concedes at the outset, and it is the strongest sentence against him:  he opens by reporting Ferguson's finding that "multiple growth rings are extremely rare in the bristlecone pine, especially in the White Mountain area."²

Matthews's case, which is the fullest version

The most developed argument for multiplicity is Mark Matthews's 2006 paper in the Journal of Creation, and it deserves to be set out properly because it is more than a restatement of Lammerts.

• The physiological hypothesis. Multiplicity and the strip-growth habit are adaptations for conserving water: producing narrow dark bands limits evaporative loss. Matthews predicts that the trees with the most rings should be the ones in the driest microsites, and reports from LaMarche's own ecological work that they are — "Comparative aridity thus seems to be an important characteristic of the 'old-age habitat'," and "No old bristlecone pines are found in the valley bottom, which is a sheltered area with deep colluvial soil and gentle surface slope."¹⁵
• Glock's numbers. Matthews cites Glock, Studhalter and Agerter's 1960 monograph for the findings that multiplicity was "more than twice as common as annularity," that "probably very few annual increments, over the entire tree, consist of only one growth layer," that false rings "commonly do, possess outer borders indistinguishable from the borders terminating the annual increment," and that about 99% of extremely thin and partial rings were false.¹⁵
• Mirov's sentence. From The Genus Pinus: "Apparently a semblance of annual rings is formed after every rather infrequent cloudburst."¹⁵
• The preservation paradox, and it is his best point. The oldest wood in the chronology would have had to lie on the ground for more than 7,000 years without rotting — while the dead flanks of living strip-growth trees have eroded to the pith in one to two millennia, and LaMarche's own measured slope-degradation rate for the White Mountains is about 30 cm per 1,000 years. "How is it possible that seven feet (213 cm) of dolomitic surface can erode away over the course of 7,000 years, while dead wood could remain essentially in place on the surface of the ground over that same period?"¹⁵
• And the rebuttal of the annularity evidence. Matthews works through LaMarche and Harlan's four lines of evidence one by one — the absence of diffuse outer boundaries, Fritts's three-year growth study, the 1954-to-1971 recount, and the 1884 frost marker — and argues that the first fails because Glock showed false rings can be indistinguishable, and that the second, third and fourth all used trees that were not the ancient strip-growth specimens his hypothesis is about.¹⁵

Two things should be said about this.

The first is that it is a real hypothesis with a real physiological rationale and a testable prediction, and Matthews ends by listing four experiments that would settle it — dated photographs of known-age branches, dye and frost time-tags, bomb-test isotope tags, and photographic tracking of log decay.¹⁵  Twenty years on, none of them appears to have been done.

The second is that the LaMarche and Harlan recount is the item his rebuttal handles least well.  Trees sampled by Schulman in 1954 and resampled in 1971 had formed exactly 18 rings in the 17-year interval — a few 17, none more than 18.¹⁵  Matthews's answer is that these were not ancient strip-growth trees.  That is a coherent reply, and it also concedes that multiplicity has never been observed in a bristlecone under natural conditions over a known interval.

The position most creationists have actually moved to

By 2003 the technical young-earth work had shifted, and the shift is worth stating because popular articles have not caught up with it.

John Woodmorappe analysed the archived Methuselah Walk ring-width database — 285 cores — with the University of Arizona Tree-Ring Lab's own software.  His abstract:

"No obvious grounds exist for questioning the overall validity of the bristlecone pine (BCP) long chronologies, but there is 'play' in some of the crossmatches, imposing an element of flexibility in the interpretation of the data.  Not enough is known about processes leading to plural annual rings in BCP for further consideration."⁶

And in the body:

"Personal experience indicates that the rings in many BCP cores can be visually resolved without difficulty, and the resulting skeleton plots… can straightforwardly be matched against the master chronology."  A sort of the posted measurements found that "no more than ~5%" of rings from 3000–6000 BC are potentially one cell wide, and with an expert's estimate that 80% of tiny-ring boundaries can be resolved visually, "no more than ~1% of MWK ring boundaries are potentially tainted by doubt."⁶

That is a creationist, working with the actual data, setting aside both the multiplicity argument and the complacency argument.  What he proposes instead is treated in Part 4a.

Missing rings

The opposite and, in bristlecones, the commoner error.  A tree under severe stress may fail to produce a measurable ring around all or part of its circumference.  This is why cores are taken from several radii of the same tree and why chronologies are built from many trees:  a locally absent ring shows up as a mismatch against the master pattern.

The direction matters.  Missing rings make a chronology too short, not too long.  A young-earth argument needs the opposite error.

What crossdating does about both

Crossdating catches false and missing rings because both break the barcode.  A false ring inserts a boundary the neighbouring trees do not have;  a missing ring deletes one they do.  Against a master chronology with 108 contemporaneous witnesses, either shows up as a local misfit.

The limitation is honest and worth stating:  crossdating catches errors that are not shared.  A climatic event that caused every tree in a region to double up in the same year would not be caught by comparing those trees to each other.  That is precisely the scenario Lammerts proposes, and precisely the scenario Aardsma's test was designed to detect — see Part 4.

Part 3: Crossdating and its statistics

Yamaguchi's warning, and the half of it that is usually left out

The standard European crossdating program, CROS, computes a Student's t value for each candidate alignment.  In 1986 David Yamaguchi published a paper in the Tree-Ring Bulletin demonstrating what those t values are worth when the series are autocorrelated, as tree-ring series almost always are.

His test case was a 290-ring Douglas fir log from a Mount St Helens pyroclastic-flow deposit, stratigraphically bracketed to AD 1482–1668.  Run against the Douglas fir master chronology, the program returned 113 candidate bark dates significant at p = 0.001, spread over AD 1410–2240 — including 47 in the future, AD 2078–2195.  Yamaguchi's own comment:  "Clearly, 113 is too large a number of likely bark dates because less than one likely date should have arisen by chance in a time interval of this length, and only one can be the correct date."⁹

That is where the young-earth citations of this paper stop, and it is not where the paper stops.  Yamaguchi's point was not that crossdating fails;  it was that raw correlation on autocorrelated series fails, and that the autocorrelation can be removed.  Applying a "whitening" transformation to strip the repeated internal patterns out of the master sequence, he eliminated 112 of the 113 matches and was left with the AD 1647 date, at t = 5.05.  That date falls inside the independently established stratigraphic window, which is how it was checked.

This account is not a mainstream author's defence of his own field.  It is Robert Brown's, in Origins, the journal of the Adventist Geoscience Research Institute — a creationist describing the paper accurately in the course of arguing against the bristlecone chronology:

"By a mathematical technique for 'whitening' a master chronology sequence, i.e., removing the effects of correlated ring-width sequences within the master sequence (repeated patterns of variation auto-correlation), Yamaguchi was able to eliminate 112 incorrect matches and focus on the AD 1647 bark date…  These results confirmed stratigraphic placement of the volcanic eruption that buried the log within the AD 1482-1686 time range."¹⁰

Brown's own argument is the serious one, and it survives.  His charge is not that Yamaguchi's method fails but that it was not available when it mattered:  "Whitening technique was not used in the development of the Bristlecone Pine master dendrochronology that is the standard for calibrating C-14 age," and analysis of the pre-1985 master sequences "indicates that the master sequence developed by Ferguson has unique auto-correlation features, and that its use is definitely questionable."¹⁰  He also notes that even after whitening, an investigator with an unknown 290-ring log could have chosen among 66 statistically justified matches within the existing master sequence, "making his ultimate choice in accord with where he had expected, or wanted, the match to occur."¹⁰

That is a real methodological objection to how the chronology was built.  What answers it is not a better statistic but an external check on the result — which is Part 5.

What the field does about it

Two things.

It demotes the statistic.  The standard teaching, quoted by the young-earth authors themselves, is that "dating should always be performed visually prior to checking with a computer program," and that such programs should be "backups";  it is "a dangerous course to allow statistics to overrule human judgement."¹⁴

It refuses to publish weak matches.  The best worked example is the Holme-next-the-Sea timber circle in Norfolk, published in Nature in 1999.  A 168-year site chronology from a ring of 55 oak posts gave its highest correlation against the East Anglia master at t = 3.98 — which the excavators judged "statistically non-significant."  They declined to date the site dendrochronologically and submitted six contiguous 20-year blocks for high-precision radiocarbon instead.  The radiocarbon result was consistent with the preferred dendrochronological position of 2050 BC.¹⁶

For scale:  the German oak-and-pine links that carry the long chronology run at t = 7 to 9 over overlaps of hundreds of years.  A site rejected at t = 3.98 and a link accepted at t = 8 are not the same kind of evidence.

Part 4: Aardsma's test

This is the most important thing in this article, and it takes three paragraphs.

The idea

Plot radiocarbon age against tree-ring number over the whole 9,000-year record and the curve has small wiggles in it — dips where the radiocarbon age is anomalously young.  These are not noise.  They are the Sun going quiet:  fewer sunspots, weaker solar wind, more cosmic rays reaching the atmosphere, more carbon-14 made, and trees growing then take up an excess of it.  The most recent such dip is the Maunder Minimum, roughly AD 1645–1715;  the next back is the Spörer.³

These solar episodes have a characteristic duration in years.  But the horizontal axis of the calibration curve is not years — it is rings.  So if trees were laying down two rings a year, a 50-year solar minimum would occupy 100 rings and the wiggle would look twice as wide.

Multiple ring growth must therefore show up as systematic broadening of the wiggles with depth.

The measurement

Aardsma fitted a Gaussian plus a linear background to each wiggle and took the full width at half maximum.  For the five wiggles in the last 3,000 years — the interval where, as he notes, "all recent-creationist scientists of which I am aware grant the general validity of dendrochronology and radiocarbon" — the mean width is 70.5 years, splitting into a wide Spörer-like family near 100 years and a narrow Maunder-like family near 50.³

He then measured all thirteen well-defined wiggles across the full record and plotted width against ring number, with theoretical curves for a range of assumed Flood dates.  He chose his assumptions, in his own words, "in such a way as to allow the most recent Flood date possible."

The result:

"it is apparent that the measured peak width data does not support any chronologically significant multiple tree-ring growth per year.  Rather, this data seems only supportive of single ring growth per year back at least as far as 9,000 BP (c. 7000 B.C.)."³

The alternative he also closed

The other way to shorten a chronology is to mis-join it — to place two trees end to end when they actually grew at the same time.  Aardsma rejected this on two grounds.  First, the calibration curve "changes more or less monotonically with time… [it does] not suddenly repeat [itself] or show any discontinuities," which a duplicated interval would produce.  Second, "the detailed correlation of even the short-term radiocarbon structure from independent dendrochronologies on separate continents" — the same wiggles appear in North American Douglas fir and Sequoia and in Irish oak, measured by different groups using liquid scintillation and gas counting respectively.³

And what he concluded

"The extent of the presently available long dendrochronologies argues forcefully for a Flood date in excess of 10,000 years ago.  Efforts to overturn these long tree-ring chronologies and thereby allow for a more traditional Flood date have not proven successful when examined quantitatively.  By far the most hopeful suggestion was that unusual climatic conditions soon after the Flood may have given rise to multiple ring growth per year…  This suggestion has, in the present paper, been shown to be contrary to the evidence."³

His second argument, from the other direction

The peak-width test is not the only quantitative check Aardsma ran.  In an exchange with Robert Brown — whose model had post-Flood carbon-14 rising fast enough to compress the chronology — Aardsma worked out what Brown's conversion would require of the trees themselves.  As Sanders reports it:  "Brown's conversion would require a BCP living about 6000 BP to make 580 rings in 80 years, or 7 rings/year.  Other cases would require up to 20 rings/year in ring series predating 3500 BP."⁵

And in the published discussion of his 1990 conference paper, answering Russell Humphreys's proposal that a weak post-Flood magnetic field could compress the curve, Aardsma computed that fitting the necessary rings into the 1,500 years most creationists would concede requires four rings per year — and added the ecological objection that a climate wet enough to produce that many growth flushes would not be bristlecone habitat at all.⁵

What has been said against it

The movement's own 2018 review describes the test accurately, including its cross-continental control — "identical deviation peaks for Douglas fir in North America and Irish oak over the last 600 years" — and its result:  "the size classes are consistently 50 or 100 years over the time back to the earliest tree-ring dates, suggesting that growth rates have been annual throughout."

It then offers one counter, in a single sentence:  "Of course, this requires the assumption of uniformitarian conditions in sunspot activity and radiocarbon flux since the Flood."⁵

That is a fair objection in principle.  It is also the whole of the reply.  Aardsma anticipated it — "There is no theoretical motivation for assuming otherwise of which I am aware," noting that the 11-year sunspot cycle shows no drift over the centuries of direct observation — and nobody has since produced a solar model in which the quiescent episodes systematically lengthen backwards in time by exactly the factor needed to cancel the ring multiplication.  Until that exists, the test stands.

A reader is entitled to reject his resolution.  Aardsma's own answer was to date the Flood near 12,000 BC and posit large gaps in the Genesis 11 genealogy — a position most young-earth creationists do not hold, and one whose difficulties he states himself:  "where exactly do these gaps fall in this genealogy?  Secondly, what is the significance of the numerical information given in Genesis 11 if it is not to be used for direct chronological reckoning?"³

But the measurement is separable from the theology.  The peak-width test is a physical test of the multiple-ring hypothesis, designed by someone who wanted that hypothesis to be true, and it came out negative.

Part 4a: The disturbance-bundle model, which is the current young-earth proposal

Because Part 4 has often been read as the end of the argument, it needs saying that it is not.  The young-earth position moved, and the version now in the literature is different in kind.

The idea

Crossdating works because trees that lived together shared a climate.  Woodmorappe's proposal is that they can also crossmatch because they shared a disturbance — and that disturbances propagate.  Roger Sanders summarises it:

"ring correspondence is due not to climatic, synchronous perturbations but to wave-like sequential localized soil perturbations.  That is, rock and soil shifts during substrate creep over several year periods due to erosion and earthquake tremors would stress trees on a fault first, then the creep would spread to other trees over multiyear periods.  Thus, trees of the same age would have time-staggered ring patterns making them appear to be of different ages."⁵

Extended, this gives "bundles" of contemporaneous trees that crossmatch within a bundle, chained together by bridge series — individual trees that fortuitously match at the join.  Woodmorappe's 2023 conference abstract reports a demonstration:  of ten modern Finnish Scots pine series edited to include occasional doubled rings, one crossmatched with a master chronology at 2536 BC and "passed all the 'gateway' statistics."⁷

The simulation

In 2026 Marshall Jordan published a full computational version in the Answers Research Journal.  It builds virtual post-Flood trees, none older than 4,500 years, imposes disturbance patterns on those growing within 1,500 years of the Flood, and assembles them by the same procedures dendrochronologists use.  The output is a statistically robust 8,000-year chronology from trees that all lived inside 4,500 years.⁸

Three things in that paper should be quoted rather than paraphrased, because they concede more than the popular literature does:

• "all the bristlecone pine tree-ring matching statistics are robust."
• "Some have claimed that multiple rings per year invalidate the dendrochronology… Others have questioned the statistical significance of the tree-ring pattern matches" — both listed as positions the paper is setting aside.
• "By itself, this simulation does not invalidate the bristlecone pine dendrochronology, but it does offer an alternative interpretation suggesting that the 8,000 years may be illusory."⁸

This is the most sophisticated argument in the whole young-earth counting-clock literature.  It does not need false rings, missing rings, or bad statistics.  It grants all the statistics and attacks the interpretation of a match.

What it predicts, and what is found

The model has a consequence its authors do not draw, and it is decisive.

If the chronology is a chain of contemporaneous bundles, then wood at apparent ring 8,000 and wood at apparent ring 4,000 actually grew at nearly the same time.  Same time means same atmosphere, and same atmosphere means the same carbon-14.  So a plot of measured carbon-14 against ring number must show long flat treads — one tread per bundle — with drops at the joins.  A staircase.

The measurements exist and they are public:  the tree-ring section of IntCal20 is built from thousands of carbon-14 measurements on independently counted rings, and Davidson and Wolgemuth plot the raw percent-modern-carbon against ring number over the whole 14,000-year range.¹²  The curve descends smoothly.  There is no staircase, no tread, and no step at any join.

To keep the bundle model, the shared post-Flood atmosphere would have to have had a carbon-14 content that fell smoothly and monotonically in step with the apparent ring number — which is to say, the very relationship the model was constructed to deny.

And the 774 CE test again

The bundle model also has to survive the single-year marker.  If a bundle sits 3,000 apparent rings away from where it really belongs, its 774 CE carbon-14 spike sits 3,000 rings away from everyone else's.  Across 44 chronologies on five continents, it does not.

Neither the 2023 abstract nor the 2026 simulation cites Büntgen's network, Miyake's discovery paper, or Aardsma.  That is the gap in the current young-earth case:  it is a well-built model with a testable prediction, and the test has not been run by the people proposing it.

Part 5: The 774 CE marker

If Aardsma's test is the strongest internal check, this is the strongest external one, and it did not exist when he wrote.

The event

In 2012 Fusa Miyake and colleagues measured carbon-14 in single rings of Japanese cedar and found "a rapid increase of about 12‰ in the ¹⁴C content from ad 774 to 775, which is about 20 times larger than the change attributed to ordinary solar modulation."¹⁷  The physics is a burst of solar energetic particles;  the consequence is a spike in atmospheric carbon-14 that happens everywhere in the world in the same year.

The test

In 2018 the COSMIC consortium — an "unprecedented voluntary collaboration of the international tree-ring community" — gathered wood from most of the world's longest chronologies and measured the carbon-14 in 484 individual rings:

• 44 chronologies, sites on five continents, latitudes 42° S to 72° N, elevations ~40 to ~4,000 m.
• 13 genera, 27 species — 25 conifers and two broadleaf.
• Living trees, historical timbers, archaeological excavations and subfossil remains.
• 374 rings covering 770–780 CE from 27 Northern and 7 Southern Hemisphere records; 110 rings covering 990–1000 CE from 8 Northern and 2 Southern records.¹

Every record puts the excursion in the same ring.  The Northern Hemisphere sequence is low in 773, up in 774, up further in 775, small increase in 776, declining after — consistent with a boreal-summer event.  The Southern Hemisphere shows the same shape at lower amplitude, consistent with ocean-atmosphere exchange.

The authors' conclusion:

"The global signature of the two cosmogenic excursions independently validates the precise dating of well-replicated tree-ring chronologies from five continents."¹

Why this settles the crossdating objection

The crossdating objection is that a chronologist could join two sequences in the wrong place, and that a t statistic will not reliably tell.  That objection is about whether the joins are right.  Here is an external physical event, dated to a calendar year, that appears in the same ring in Siberian larch, German oak, Californian bristlecone, Japanese cedar, South American and New Zealand wood.  If any of those chronologies had a bad join anywhere between the present and 774 CE, its spike would sit in the wrong ring.  None does.

The same event is recorded as a beryllium-10 spike in five ice cores from Greenland and Antarctica.¹  When those ice chronologies were tested against it they were found to be seven years wrong — see Ice-Core Layer Counting.  The tree rings were not.

And a second, older one

Four solar-particle events are now confirmed by more than one cosmogenic isotope:  774 CE, 993 CE, 660 BCE and 7176 BCE.¹⁸  A fifth and larger candidate sits at 14,300 cal BP, where subfossil pine rings and Greenland beryllium-10 agree to "only a decade, at the most."¹⁸

Part 6: Cross-continental agreement

A separate check, and one that does not depend on any single-year event.

IntCal20 compares Irish oak latewood against North American bristlecone whole rings for 1700–1500 BC and finds "an average weighted mean difference of −8.1 ± 1.9 ¹⁴C yr."¹¹  Two chronologies, built by different teams on different continents with different species, giving the same atmospheric carbon-14 to eight radiocarbon years.

Friedrich et al. 2004 report that the German links "were confirmed by comparisons to the independent Irish oak chronology (Pilcher et al. 1984) and the Göttingen oak chronology (Leuschner 1992) back to 9741 BP."⁴

Aardsma 1993 reports that the century-scale radiocarbon wiggles are the same in North American Douglas fir and Sequoia as in Irish oak, "obtained by independent research groups using two different measurement techniques (liquid scintillation versus gas counting)," and that Minze Stuiver had reported no significant difference between the long chronologies back to 6000 BC.³

Against this stands a claim, circulated on creation.com without a citation, that "tree ring alignments from locations as close as Northern Ireland, southwest England, and southern Germany disagree strongly."¹⁹  If it has a source, that source needs producing.  As stated it is contradicted by all three items above.

Part 7: What gets revised, and how

A method that never revises anything is not being checked.  The Hohenheim paper reports its own corrections:

• The floating Preboreal Pine Chronology was re-linked, shifting "the pre-11,250 BP part of the PPC to older ages… by 70 yr."⁴
• "Oaks affected by cockchafer have been identified and discarded from the chronology" — insect defoliation distorts ring widths and corrupts matching.
• "The number of trees in parts of the chronologies had to be reduced significantly."

And in the same paper the payoff:  when the PPC was finally crossdated dendrochronologically to the absolutely dated oak record, using a 538-year overlap, "the dendrochronological crossdating resulted in a difference of only 8 yr with respect to the published ¹⁴C wiggle-match position used for IntCal98."⁴

Two independent placements of an 11,000-year-old floating sequence, eight years apart — 0.07%.  One caveat should go with it:  the wiggle-match used IntCal98, which is itself built partly from these oaks, so this is a demanding internal consistency test rather than two wholly unrelated clocks.

Objections and responses

"Trees can grow more than one ring a year." They can, it has been shown experimentally, and Lammerts's bristlecone seedling experiment is a fair demonstration.  The question is the rate and whether it accumulates.  Aardsma's peak-width test measures exactly that and finds no chronologically significant multiple growth back 9,000 years.  Neither Lammerts's paper nor any later young-earth treatment produces a competing measurement.

"False rings are indistinguishable from true annual rings." In a single core, sometimes.  Against 108 contemporaneous trees, an unshared false ring breaks the match.  Against the 774 CE spike, even a shared false ring would show:  a doubled year anywhere between now and 774 would put the spike in the wrong ring, in that chronology and not in the others.

"Crossdating is subjective — CROS returns dates in the future." Yamaguchi's result is real and important, and it is why the field dates visually first and treats the statistic as a check.  But the paper that produced the 113 matches also removed the autocorrelation causing them, eliminated 112, and got the answer the stratigraphy independently required.  The stronger form of the objection is Brown's:  that the whitening technique postdates the construction of the bristlecone master chronology, and that Ferguson's sequence has unusual autocorrelation properties.  That is an argument about how the chronology was built, and it is answered by testing the finished product against an external marker rather than by arguing about the statistic.

"The chronology depends on the accuracy of just two specimens, one living and one dead, at the join." This claim, in Snelling's wording, is repeated widely.⁵  It is a fair description of an early stage of the bristlecone work — Ferguson's original link — and it is not a description of the record as it now stands.  The Hohenheim chronology has a mean of 108 trees in every year and 96% of its length covered by more than 20 crossdated trees;  the bristlecone master sequence Brown describes rests on 81 living-wood and 118 dead-wood specimens;  and there are two independently built bristlecone chronologies 300 km apart.  A single fragile join in 1969 is not the same object as the 2020 IntCal20 tree-ring section.

"The trees crossmatched because they shared a disturbance, not a climate, so the chronology is a chain of contemporaneous bundles." This is the current and best young-earth argument, and it is treated at length in Part 4a.  Its problem is not statistical — the simulation reproducing it grants that the real matching statistics are robust — but radiocarbon.  Contemporaneous bundles must produce flat treads in a plot of measured carbon-14 against ring number, and there are none.

"The oldest chronology wood would have to have lain on the ground for 7,000 years while the mountain eroded seven feet beneath it." Matthews's preservation paradox is the best empirical point in the multiplicity literature and this article does not have a quantitative answer to it.  What can be said:  the two rates are not measuring the same thing — LaMarche's ~30 cm/1,000 yr is an average slope-degradation rate for a mountain range, not a rate for the specific bench a log lies on, and dead bristlecone wood in that climate is dense, resinous and abraded rather than rotted.  It deserves the direct test Matthews himself proposed, which is photographic tracking of decay in dated logs.

"Radiocarbon was used to place the tree-ring sequences, so the agreement is circular." Partly true and it needs care.  For the absolutely dated portion — 0 to 12,410 cal BP — placement is by ring counting and crossdating, with no radiocarbon input;  the carbon-14 measurement is then read off the independently dated ring.  For floating sequences beyond that, radiocarbon wiggle-matching is used, and there the circularity charge has real force.  The Preboreal Pine Chronology is the test case:  it was placed by radiocarbon, then later placed independently by dendrochronology, and the two agreed to 8 years.

"The oldest bristlecone has about 4,800 rings, roughly the biblical age of the earth." The coincidence is not evidence.  A tree's maximum age is set by its biology and its substrate;  bristlecones are eventually undercut by erosion of the dolomite they grow on.  The 8,681-year chronology comes from overlapping living and dead wood, and the German record — 12,410 years from oaks averaging 176 years each — shows the method does not depend on any single long-lived tree.

"The chronologies keep being revised, so they are unreliable." They are revised, and the revisions are published in the same papers that report the chronologies:  a 70-year shift here, discarded trees there.  Those are corrections of tens of years in records of ten thousand.  A young-earth reading needs a correction of about 8,000 years.

"Trees near volcanoes take up old carbon and date too old." True, and it is screened for.  IntCal20 excludes datasets on these grounds and discusses regional and altitude offsets at length, with measured magnitudes in the range of 8 to 19 radiocarbon years.¹¹

"Post-Flood climate was completely different, so present ring behaviour is no guide." This is Lammerts's actual argument and it deserves the direct answer rather than a dismissal.  It makes a prediction — extra rings, concentrated early, decaying toward the present — and that prediction was tested.  The solar wiggles do not broaden.  If post-Flood climate produced systematic double rings, the calibration curve's Maunder-like and Spörer-like features would be twice as wide in the early part of the record.  They are not.

Issues and unanswered questions

• Yamaguchi 1986 has not been obtained as a PDF. The University of Arizona repository refuses this host and no mirror served the file. The quotations used here come from research notes that reproduce the paper at length; they are internally consistent and consistent with later citations, but they have not been checked against the original. This should be fixed.
• LaMarche & Harlan 1973 — the mainstream study that tested the multiple-ring hypothesis in the field over an 18-year interval — could not be obtained. It is closed at Wiley with no abstract deposited and no repository copy. Aardsma describes it as having "demonstrated in a number of different ways including direct observation of bristlecone pine ring growth in their native habitat in the White Mountains over an 18 year interval that these trees were not growing multiple rings per year under modern climatic conditions,"³ and that description is used here rather than any quotation from the paper itself.
• Glock, Studhalter & Agerter 1960 and Mirov 1967 are cited by the 2016 Answers Research Journal article for the claims that false rings are more common than single annual increments and that rings form "after nearly every cloudburst." Neither was obtained. Both claims, if accurately reported, are about species and settings other than White Mountain bristlecone, and Ferguson's contrary finding for bristlecone specifically is quoted by Lammerts himself. They should be checked.
• Whether Lammerts's seedling result scales to mature trees in situ is the open scientific question in his paper, and he says so: "The need for further study of the effect of drying out on older trees is pointed out."
• Matthews's preservation paradox is unanswered here. Why does bristlecone deadwood survive on the surface for millennia while the dolomite beneath it erodes, and while the dead flanks of living strip-growth trees erode to the pith in one to two millennia? A published decay-rate study on dated logs would settle it either way.
• Woodmorappe's disturbance model has not been tested against radiocarbon by anybody, on either side. The prediction is unambiguous — flat treads in pMC against ring number — and the data to test it are public. Somebody should simply run it on the Jordan 2026 simulated chronology and print the result.
• Glock, Studhalter & Agerter 1960 has still not been obtained directly. Its numbers now reach this article through two creationist sources (Matthews 2006 and Sanders 2018) that agree with each other; that is better than one, and it is not the same as reading the monograph.
• Sanders 2018 names creationist works this article has not read, among them Brown 1968/1986/1990, Gladwin 1978, Setterfield 1986, Armstrong 1976, Porter 1995, Downes 2010, Woodmorappe 2001/2003b/2004/2009a/2009b, and the Aardsma–Brown exchange of 1991–92. None of them appears from Sanders's summaries to add a mechanism not already covered, but the exchange in particular would be worth reading in full.

Verdict

Tree-ring dating is the best-anchored counting clock there is, and the reason is not the counting.  It is that a physical event in 774 CE marked one ring in 44 chronologies on five continents, and every one of them has it in the same ring.

The young-earth objections are technically accurate and were, in several cases, first raised by dendrochronologists.  Trees do produce false rings.  Trees do skip rings.  Raw crossdating statistics on autocorrelated series return confident nonsense.  The whitening technique that fixes that was not available when the bristlecone chronology was built.  Chronologies do get revised.  Bristlecone deadwood really does seem to last improbably long.  All of that is in the literature and none of it is hidden.

Two things decide it.

The first is that the young-earth case has three incompatible answers and no consensus, and its own review says so.  Lammerts and Matthews say the rings are not annual.  Aardsma measured that they are, twice, and drew the conclusion.  Woodmorappe examined the archived ring-width data and concluded that the rings are readable and the crossmatching valid — and then proposed that the matches are real but misplaced.  A movement whose 2018 literature review ends by calling for research to "build a consensus dendrochronology calibration model" does not have a settled position to weigh against the mainstream one.

The second is that the strongest version of the case — the disturbance-bundle model — makes a clean prediction that the data refute.  Contemporaneous trees strung out into a false 8,000-year sequence must show flat treads in a plot of measured carbon-14 against ring number.  The plot is smooth.  That test costs nothing to run, the data are public, and nobody on the young-earth side has run it.

Further reading

• Non-Radiometric Clocks — the overview of this series
• Ice-Core Layer Counting — the 774 CE marker found the ice chronologies seven years wrong
• Lake Varves
• Radiocarbon Dating — the calibration curve these chronologies underwrite

Sources