Australia, Japan, and the Americas Before the Flood Date

"That is why we won’t accept any date that contradicts the eyewitness evidence of human history recorded in the Bible.  Such contradictory dates can’t be right." — Tas Walker, Creation Ministries International, on Mungo Man, 2003¹

"Invoking a series of very different sorts of contamination processes to dismiss the consistent dating of the tracks, lake beds, and wetlands requires serendipity in the extreme" — Vance Holliday and colleagues, on the White Sands footprints, 2025²

Summary

Genesis puts the Flood about 2350 BC on the Hebrew text and about 3100 to 3300 BC on the Greek Septuagint, and has every living person descend from Noah's family.  Young-earth writers place all human settlement of Australia, Japan and the Americas after the dispersal from Babel, a century or a few centuries later.  Five famous sites test that placement.  Each is dated by a different team, and most by more than one method.

1. Every site is older than any Flood date, by a factor of 3 to 15 measured from Babel. Madjedbebe in northern Australia: first occupation about 65,000 years ago by luminescence dating, "conservatively 59.3 ka",³ and about 50,000 on its strongest critics' reading.⁴ Lake Mungo: two burials at "40 ± 2 kyr".⁵ Odai Yamamoto I in Japan: pottery residues at 13,070 to 13,780 radiocarbon years,⁶ about 15,700 to 16,500 calendar years. Monte Verde in Chile: a camp at about 14,500 years.⁷ White Sands in New Mexico: footprints between about 21,000 and 23,000 years.⁸
2. Where two methods meet, they agree. At Madjedbebe, radiocarbon on charcoal and luminescence on sand agree down to about 34,000 years.³ At Monte Verde, luminescence on the peat that seals the camp gives 13,793 ± 906 years over a camp radiocarbon-dated to about 14,500.⁷ At White Sands, radiocarbon on aquatic seeds, on land pollen and on lake mud, measured by several laboratories, and luminescence on quartz all give 21,000 to 23,000.⁹²
3. Two of these records run straight through the Flood date. The Jomon culture of Japan runs from about 16,000 to 2,500 years ago, with dated sites on both sides of every Flood date;¹⁰ a Hebrew-text Flood lands at the start of the Late Jomon and a Septuagint Flood in the Middle Jomon. Madjedbebe's upper occupation runs from about 7,100 years ago to the present.³ The other three are single horizons that bear only on how much time has passed since.
4. The young-earth writers who discuss these sites set the dates aside; none of them names a measurement that went wrong at any of these sites. Their shared model is a low level of carbon-14 in the air after the Flood, unreliable luminescence, and a dispersal from Babel by boat and land bridge during a single Ice Age of about 700 years. Tas Walker: "Such contradictory dates can’t be right."¹
5. Priced, the model needs luminescence clocks running 3 to 15 times fast on average since Babel, or 17 to 100 times fast if the excess stops with the Ice Age. Radiocarbon needs the air at 2 to 35 percent of today's carbon-14 level at these sites, then a production rate about 7 times today's for a thousand years to bring it back.
6. The one quantitative young-earth radiocarbon curve puts these people at or before Babel. Robert Brown's 1992 conversion, which reproduces its own published table here to within 5 years, dates the White Sands footprints 99 years after his Flood and the Madjedbebe charcoal 21 years after it.¹¹ Babel at Peleg's birth is 101 years after the Flood on the Hebrew text and 531 on the Septuagint.
7. Two clocks on one surface are the hardest problem. The White Sands footprints must date to Sibley's "around 4,000 years BP"¹² by two clocks at once: by radiocarbon, which needs the air at 16 percent of today's carbon-14, and by luminescence, which needs the sand's radiation dose at 5 to 51 times today's. On the model as its writers state it, low carbon-14 in the air for radiocarbon and fast decay for luminescence, the two corrections have nothing to do with each other, yet they give the same answer here and at Monte Verde and Madjedbebe.
8. The mainstream case has real weak points, and they are printed here. Lake Mungo's burials have been dated at 30,000, 42,000 to 45,000, 62,000 and 40,000 years; four methods once agreed on the wrong one. Madjedbebe is a single site in sand that termites rework. White Sands has no stone tools. The Japanese pottery dates come from one pot. These disputes are worth from a few hundred to about twenty thousand years; none brings any site within 8,000 years of a post-Babel date.

Verdict.  On the evidence, people were in Australia, Japan and the Americas thousands to tens of thousands of years before any date Genesis allows for the Flood, and the records of two of them run through that date without a break.  The young-earth account has a dispersal model but no dating model.  Its only published radiocarbon curve contradicts its dispersal model, and no young-earth writer has explained why unrelated clocks agree at these sites.

Part 1: Five sites and how each is dated

Two methods do most of the work here.  Radiocarbon measures how much carbon-14 is left in something that once lived, such as charcoal, a seed or burnt food;  it needs to know how much carbon-14 the air held when the plant grew, and the calibration curve supplies that (see Radiocarbon).  Optically stimulated luminescence (OSL) measures the radiation dose stored in buried quartz grains since sunlight last emptied them;  dividing the stored dose by the yearly dose from the surrounding sand gives the years of burial (see Luminescence Dating).  "OSL dating gives an estimate of the time since mineral grains were last exposed to sunlight."³  The two methods share no physics.  Radiocarbon depends on the chemistry of the air;  OSL depends on the uranium, thorium and potassium in the ground.

Madjedbebe, Arnhem Land, Australia

Madjedbebe is a rock shelter in Mirarr country in the Northern Territory, formerly called Malakunanja II.  Chris Clarkson's team dug twenty one-metre squares there in 2012 and 2015, plotted about 11,000 artefacts, and dated quartz grains one at a time from 56 sediment samples, four of them re-measured blind in a second laboratory.³  Their headline:  "Human occupation began around 65,000 years ago, with a distinctive stone tool assemblage including grinding stones, ground ochres, reflective additives and ground-edge hatchet heads."³

The stated uncertainty is large and stated plainly.  The lowest band of artefacts has "modelled mean start and end ages of 65.0 ± (3.7, 5.7) and 52.7 ± (2.4, 4.3) kyr";  the bigger error term includes the method's systematic error.³  The authors' own floor is "conservatively 59.3 ka".³

Radiocarbon checks the upper part.  "The 14C ages of 22 charcoal samples (15 from the northwest squares) increase progressively to around 34 calibrated kyr before present (bp) at a depth of approximately 1.6 m, but some isolated fragments have stratigraphically inconsistent ages," and the OSL ages "are also in good agreement with the 14C chronology."³³  Below 1.5 m the charcoal dissolved in the laboratory:  "16 of 40 charcoal samples submitted did not survive the chemical pretreatment, with the number of samples dissolved increasing with depth," including the sample from the lowest hearth.¹³  So the 65,000-year figure rests on luminescence alone.

The critique is serious.  James O'Connell, Jim Allen and colleagues argued in 2018 that "an age estimate of >50 ka for this site is unlikely to be valid."⁴  Their reason is the sand itself:  termites, rain splash and creep move stones down through loose sand, so artefacts end up beside older grains.  "In particular, abandoned termite galleries collapse and disappear."⁴  They point to a charcoal fragment from the 1989 dig, at the depth of the oldest artefacts, that "reported a calibrated 14C date of 12.8–13.0 ka", and to Nauwalabila, a nearby shelter where the same problem was proven.⁴  Clarkson's team answers with refitted stone flakes:  "post-depositional vertical mixing of the deposit and artefact movement is restricted to depths of approximately 10 cm."³  O'Connell's own preferred date is about 50,000 years, the date "Appropriately calibrated molecular clock studies of Sahul genetic diversity all point to".⁴  So the live dispute is 65,000 against 50,000 years.

Lake Mungo, New South Wales

Lake Mungo is a dry lake in the Willandra Lakes, famous for two burials:  Mungo I, "the first recorded cremation", and Mungo III, "the world's oldest ritual ochre burial" in the words of the team that redated them.⁵  Its dates are the most revised of any site here, and a skeptic of dating will find the history quickly, so here it is.  Radiocarbon first put Mungo I at roughly 20,000 to 26,000 years.  Mungo III was then estimated at 30,000, at 42,000 to 45,000, and in 1999 at 62,000 ± 6,000 years by four methods that agreed with each other.¹⁴⁵  In 2003 James Bowler's team published "a new series of 25 optical ages showing that both burials occurred at 40 ± 2 kyr ago and that humans were present at Lake Mungo by 50–46 kyr ago".⁵

The 1999 agreement was a false one.  Three of its methods read the same uranium in the same bone and shared one assumption about when that uranium arrived;  the site's article on concordance tells the story.  Why the fourth, a luminescence date on sediment, also read 61,000 is not settled here:  the study that dated the grave fill itself, by John Olley and colleagues in 2006, was not obtained.  The early radiocarbon dates were too young, the usual fault of radiocarbon near its limit, where a trace of modern carbon pulls an old sample young (see Radiocarbon).  What has held since 2003 is the luminescence dating of the sand units that hold the graves.  The range of every published figure, 20,000 to 62,000 years, is still 5 to 15 times the time since Babel.

Odai Yamamoto I and the first Jomon pottery, Japan

Odai Yamamoto I is a small terrace site at the northern tip of Honshu.  In 1998 it yielded 46 sherds and 262 stone tools in a patch about ten metres across, and "The potsherds are all from the same vessel."⁶  Burnt food and soot crusted on the sherds were dated directly.  "Except for two charred wood 14C dates, 7070 ± 40 and 7710 ± 40 BP, all five charred-residue samples and one wood charcoal sample gave older 14C ages of 12,680–13,780 BP".⁶  The authors' own estimate for when the pot was used is "in the range from the average value 13,070 ± 440 BP to the oldest age 13,780 ± 170 BP."⁶  The famous "16,500 years" is the calibrated age of that single oldest crust.⁶

The authors name their own uncertainties.  The five crusts spread over 1,100 radiocarbon years on one pot.  Soot from old firewood can read "several decades or a few hundred years (but rarely more than hundreds of years) older than the period of the pottery usage".⁶  The two young charcoal pieces are maple and plum-yew, warm-climate trees, and "most likely were mixed into the loam layers at the site by later humans or natural agents."⁶  Two of the crusts measured δ¹³C of −30.5 and −29.6 per mil, a land-plant value, which argues against a large marine reservoir effect.⁶

Odai Yamamoto does not stand alone.  Charcoal from Kitahara gives five dates of 13,020 to 13,060 radiocarbon years.¹⁵  A 2004 review of the East Asian dates concludes that "In Japan, the earliest pottery can be dated to about 13,500 BP (about 16,750–15,700 cal BP)."¹⁵  On the Amur River in Russia, pottery dated by thermoluminescence, which uses the heat of firing rather than carbon, gave "13,460 ± 1460, 10,430 ± 1160, and 8580 ± 1490 yr ago", which "are very close to the calibrated dates of the charcoal 14C dates they correspond to".¹⁵

Monte Verde, southern Chile

Monte Verde is a creek-side site in southern Chile, excavated by Tom Dillehay's team.  Its upper level, MV-II, is a camp "associated with the remains of a long tent-like dwelling, the foundation of another structure, hearths, human footprints, economic plants, and wood, reed, bone, and stone artifacts."⁷  MV-II is older than any Clovis site, the culture long held to be the first in the Americas, and it has been central to the debate "For almost four decades".⁷  In Dillehay's 2015 summary:  "Previous work at Monte Verde revealed one valid human site (MV-II) dated ~14,500 cal BP".⁷

Two clocks agree on MV-II.  The camp lies on an old creek terrace and is sealed by a peat layer.  Luminescence dates the top of the terrace to about 15,000 years, and "stratum MV-5, which has an OSL assay of 13,793±906 years", seals the camp.  "The OSL dates also confirm the prior 14C dating of cultural materials in the single component occupation of the MV-II site".⁷

The 2015 paper also reports older and much less certain material:  scattered flakes and burnt patches "radiocarbon and luminescence dated between at least ~18,500 and 14,500 cal BP."⁷  Dillehay himself grants that "The character of some of these horizons may not meet the traditional criteria used by some archaeologists to define valid early sites".⁷  This article uses only MV-II.

White Sands, New Mexico

At White Sands National Park, human footprints lie in layers of lake-edge mud alongside the tracks of mammoths and giant ground sloths.  Matthew Bennett's team reported in 2021 that the prints "are stratigraphically constrained and bracketed by seed layers that yield calibrated radiocarbon ages between ~23 and 21 thousand years ago."⁸  The seeds were of Ruppia, a pondweed.

The objection was immediate and fair.  A water plant takes its carbon from the water, and groundwater that has passed through limestone carries old carbon with no carbon-14 in it, which makes a plant look older than it is.  Charles Oviatt, David Madsen and colleagues tested it:  "Ruppia specimens collected in 1947 from nearby Malpais Spring returned a radiocarbon age of ca. 7400 cal yr BP."  They asked for "independent means for dating the footprints".¹⁶

Two independent tests followed.  Jeffrey Pigati's team dated land pollen from the same layers, because "Terrestrial pollen is produced by cone-bearing and fruit-bearing plants living on land and fixing atmospheric carbon and therefore is not subject to hard-water effects."⁹  Three samples of about 75,000 conifer pollen grains each gave calibrated ages from 23.4 ± 2.5 to 22.6 ± 2.3 thousand years, and quartz from within the footprint layers gave a minimum luminescence age of more than 21.5 ± 1.9 thousand years.⁹  In 2025 Vance Holliday's team traced the footprint layers into the old lake beds and reported "26 additional radiocarbon dates on palustrine mud determined by two labs independent of the original investigations".²  At the same depth in one profile, mud read 18,051 ± 42 and seeds 18,499 ± 60 radiocarbon years;  at another depth, mud 18,551 ± 52 and seeds 18,208 ± 73.²  The seeds carry no consistent old-carbon offset.

The weaknesses are real.  The pollen ages needed blank corrections "approximately an order of magnitude larger than what is typically used in radiocarbon dating", and the luminescence ages are minimums from a single level.⁹  No stone tools have been found with the prints.  A 2025 critique of the seed dates by Dave Rachal and Robert Dello-Russo was not obtained for this article.  The live dispute is 21,000 to 23,000 years against a reservoir-corrected reading which Andrew Sibley, summarising the critics, puts at 15,000 to 13,000.¹²

The sites at a glance

Part 2: Which records run through the Flood date

A global Flood about 2350 BC, or about 3100 to 3300 BC on the Septuagint, would end every human community on earth except one family in the Near East.  So a sequence of occupation that runs through that date without a break, read on its conventional dates, shows no Flood.  The young-earth answer is that the whole sequence is younger than it looks and began after Babel.  That makes such a sequence a test of compression, not of the Flood's existence:  the whole of it must fit into the time after Babel.  A single dated horizon, by contrast, only says how far back people were in one place.

The Jomon sequence of Japan runs through both Flood dates.  The Jomon culture is defined by its pottery, which begins with sites such as Odai Yamamoto and runs without a gap to the arrival of rice farming:  "The Jomon culture of Japan (ca 16,000 ~ 2,500 cal BP)".¹⁰  Its periods, fixed by radiocarbon on pottery-phase sites, are "the Early (6,950 ~ 5,470 cal BP), Middle (5,470 ~ 4,420 cal. BP) and the Late (4,420 ~ 3,220 cal BP) Jomon periods".¹⁰  Ussher's Flood, 2348 BC, is 4,298 years before 1950, so it falls just after the start of the Late Jomon.  A Septuagint Flood of 3298 to 3386 BC falls at 5,248 to 5,336, in the Middle Jomon.  Enrico Crema's team gathered 1,433 radiocarbon dates from three regions of eastern Japan for 7,000 to 3,000 years ago.¹⁰  In Hokkaido alone the archaeological survey records 2,766 Middle Jomon sites, 1,598 Late and 1,462 Final.¹⁰  The population rose and fell, and the authors trace those swings in detail, but dated sites and the pottery tradition continue across 4,300 years ago in all three regions.

Madjedbebe's own sequence runs through it.  The site's upper band of artefacts "has a modelled start age of 7.1 ± (1.0, 1.1) kyr", and the Bayesian model's last phase ends at the present.³  A thylacine bone coated in red pigment comes from this upper deposit, and the caption dates it:  "The probable age of the thylacine specimen is 2.7–3.9 calibrated kyr bp".³  The deposit holds the same people's tools and food remains from before any Flood date to after it.

Lake Mungo, Monte Verde and White Sands are single horizons.  Each says when people were there.  None is a continuous record through 2350 BC, and none of them could show a Flood break if there were one.  They bear on a different question:  how much time the young-earth account has between Babel and the first people in each place.

This article does not use the Australian rock-art sequences.  Most painted art is dated by indirect means such as mud-wasp nests or mineral crusts over the paint, and its dates are less direct than the five sites above;  the sister article Behavioral Modernity treats dated rock art in Indonesia and Europe.

Part 3: What the young-earth writers say

Young-earth writers agree on the history and say little about the dates.  Their history is that the people of Australia, Japan and the Americas descend from Noah through the families scattered at Babel, and arrived during a single post-Flood Ice Age.

When.  Andrew Snelling of Answers in Genesis puts the Flood "about 4,350 years ago" and Babel "just over 4,000 years ago", after which "all human migrations and cultures have occurred".¹⁷  Michael Oard puts Babel "about 100 to 300 years after the Flood" and the Ice Age at "roughly 700 years".¹⁸¹⁹  Tas Walker dates "the dispersion from the Tower of Babel (c 2,200 BC)".²⁰  Kurt Wise concludes that "both Lower and Upper Pleistocene sediments are post-Babel in age", which puts every site in this article after Babel.²¹  Marcus Ross, Peter Brummel and Todd Wood write that "the dispersal of humans most likely occurs post-Babel".²²  The longest Flood date any biblical text allows is 3386 BC, on Chris Hardy and Robert Carter's count.²³

Australia.  On Mungo Man, Walker writes:  "After the Flood, and after the confusion of languages at the Tower of Babel (Genesis 11), their ancestors migrated to Australia."¹  A 1993 Creation article holds that the Aboriginal people "have been here only a few thousand years—certainly less than four thousand."²⁴  On Madjedbebe's 65,000 years, Creation magazine wrote that "the Bible teaches that after the Flood about 4,500 years ago, people groups travelled out from Babel in the Middle East and eventually dispersed to all habitable areas of the earth."²⁵  Peter Line, the movement's most careful reader of the Australian fossils, notes that "The evolution-based dating methods give inconsistent dates for some of the Australian fossil remains, depending on the method used."²⁶  How people crossed to Australia is left open.  Oard, writing on animals, now judges that "the idea of a land bridge between Sunda and Sahul is unlikely", since the gap from Timor to Australia is about 150 km of sea.²⁷  So the first Australians came by boat, which no one disputes.

Japan.  No young-earth treatment of Odai Yamamoto or the Jomon sequence was found.  The nearest is Walker's 2009 article on early pottery in a Chinese cave, dated 18,300 to 15,430 years:  "If we leave the dates to one side, we discover that the data can be interpreted quite nicely within the biblical framework of history.  The early settlement of China began after the dispersion from the Tower of Babel (c 2,200 BC), when the first immigrants migrated eastward across Asia from the Middle East."²⁰

The Americas.  Oard draws "the route of animals and man into the Americas from Eurasia" over the Bering land bridge and down the ice-free corridor, adding that "It is possible man could have taken a coastal route".²⁸  He now thinks the bridge was raised by uplift early in the Ice Age rather than exposed late by low sea level.²⁸  David Smith of Answers in Genesis welcomes the boat route:  "Based on Genesis, we know that descendants of the great shipbuilder Noah must have settled in the Americas soon after the Flood", and Monte Verde, "older than the Clovis site, provides strong evidence that the first settlers moved rapidly by boat."²⁹  On White Sands, Andrew Sibley wrote in the Journal of Creation in March 2023, before the pollen and luminescence tests were published:  "Creation scientists would interpret the human footprint data as that of the post-Babel dispersion of people from the Middle East during the later stages of the Ice Age, with a timeframe of around 4,000 years BP."¹²

How the dates are handled.  Three approaches appear.

1. Set them aside. Walker on Mungo: "Such contradictory dates can’t be right."¹ Walker on China: "If we leave the dates to one side".²⁰ Creation on Madjedbebe calls OSL "a method based on unprovable assumptions."²⁵ Ross, Brummel and Wood concede that "none of the dating methods have a widely accepted method of recalibrating to a creationist timescale."²²
2. Radiocarbon: too little carbon-14 in the air after the Flood. Walker: "the carbon-14 method does not account for the disruption of the carbon balance during the Flood some 4,500 years ago," so early post-Flood samples "give ages that are too old by tens of thousands of years."¹ The 1993 Creation article: "A radiocarbon ‘age’ of say 30,000 years may thus actually represent a true age of only a few thousand years."²⁴ Snelling's chart says radiocarbon dates "are unreliable before 1000 BC because of a stronger magnetic field and rapid buildup of radiocarbon after Noah's flood".¹⁷ Wise puts "Flood carbon" at 40,000 to 60,000 radiocarbon years and has radiocarbon agree with tree rings by 2,000 to 3,000 years ago.²¹ Sibley adds the specific objection at White Sands, the reservoir effect in water plants.¹²
3. Luminescence: fast decay soon after the Flood. Snelling's chart files luminescence ages of "Early Human Carvings & Pottery" under "Centuries soon after the flood (~2250 BC–present)", and states that "radioactive elements decayed rapidly during and soon after the flood".¹⁷¹⁷

The one quantitative curve.  The only young-earth paper found for this article that turns radiocarbon ages into calendar dates with a published equation and table is Robert H. Brown's, in the Creation Research Society Quarterly in 1992.  Brown credits earlier attempts by Whitelaw in 1970, Hefferlin in 1972 and Hanson in 1976.¹¹  He assumes the air's carbon-14 started low after the Flood and rose smoothly toward today's level.  "In my judgment the most suitable value for F is 5350", that is, a Flood 5,350 years before 1950, or 3400 BC.  He sets the starting level from coal's radiocarbon age:  "The fraction of equilibrium activity represented by age 37,650 is 0.011", 1.1 percent of today's.¹¹  He warns that his formula is "highly questionable" above 34,000 radiocarbon years.¹¹  He tested it on a sloth cave:  "Converting 40,000 and 20,000 to 5348 and 5271 years real-time gives a 77 year interval".¹¹  Brown says nothing about Babel.  Part 4 does the comparison he did not.

Part 4: The compression, priced

The young-earth inputs below are the writers' own:  a Flood 4,350 years ago, Babel 100 to 300 years later, an Ice Age of 700 years, and the Septuagint's longest Flood date for the generous case.  The site ages are those in Part 1.  The arithmetic is in a short script beside this article, research/compression.ts, with its output saved.³⁰  The Behavioral Modernity article ran the same arithmetic on African and European sites and its numbers agree.

Luminescence: how fast the clocks must have run

An OSL clock starts when people bury sand, or when sand buries what people left.  On the young-earth account that happened after Babel, so all of the stored dose built up since then.  Two cases bracket the need.  Spread over all 4,250 years since an early Babel, the dose rate must have averaged the first number below, in multiples of today's.  If the excess stopped when the Ice Age ended, 3,650 years ago, it must have been squeezed into 600 years, which gives the second.

These speed-ups must also follow a schedule.  The Madjedbebe sand must gain 65,000 years of dose while the White Sands mud gains 21,500 and the Monte Verde peat 13,800, each according to how long before the end of the speed-up it was buried.  A decay rate that changed everywhere at once would do that automatically.  But then it changed in every human body and every hearth stone too, in the centuries when the young-earth account has people walking to Chile.  The heat and radiation that would release has been priced elsewhere on this site, in Behavioral Modernity and The Oldest Homo sapiens, and no young-earth writer has proposed it for the post-Babel period.

Radiocarbon: how little carbon-14 the air must have held

The young-earth explanation for old radiocarbon ages is not faster decay but less carbon-14 in the air after the Flood.  That can be computed.  For a sample of a given radiocarbon age to be only as old as Babel, the air's ratio of carbon-14 to ordinary carbon at the time must have been this fraction of today's:

Three things follow.  First, the Lake Mungo and Madjedbebe levels, 2 to 4 percent, are close to the "Flood carbon" Kurt Wise assigns to the Flood itself, 40,000 to 60,000 radiocarbon years,²¹ which works out to 0.1 to 1.2 percent of today's air.  People in Australia would have been living in air holding barely more carbon-14 than the Flood left.  Second, the order is fixed.  Air whose carbon-14 only rises puts Mungo first, then Madjedbebe's upper charcoal, White Sands, Odai Yamamoto and Monte Verde, in the same order the conventional dates give.  Third, the rise costs production.  Taking the air from the White Sands level, 16 percent at Sibley's 4,000 years, to 95 percent by 1000 BC, where Snelling says radiocarbon becomes reliable, needs carbon-14 made at about 7 times today's rate for a thousand years in today's carbon reservoir;  in 500 years, about 14 times.  Snelling's "stronger magnetic field" pushes the wrong way.  A stronger field shields the atmosphere and makes less carbon-14, and the measured record shows the reverse:  air with "values 1.70 times higher" than today's during the Laschamp event, when the field was weak.³¹

Robert Brown's curve against Babel

Brown's equation can be solved for any radiocarbon age.  Rebuilt here, it reproduces his printed table:  5,200 years before 1950 gives 15,470 radiocarbon years, as he prints, and 5,300 gives 23,287 against his 23,290.¹¹  On Brown's own curve, with his Flood at 3400 BC:

Genesis 11 counts 101 years from the Flood to Peleg's birth on the Hebrew text, and 531 on the Septuagint, which adds a generation named Cainan.  Genesis 10:25 says of Peleg that "in his days was the earth divided", and Oard's range for Babel, 100 to 300 years after the Flood, starts at that point.  On Brown's curve a Babel 101 years after the Flood falls at 18,244 radiocarbon years:  the White Sands footprints were made at the moment of Babel, 12,000 km from Babylon, and the Australian dates come 80 to 93 years before it.  A Babel 300 years after the Flood falls at 10,962 radiocarbon years, so all five sites are older than Babel.  On a Septuagint Babel, 531 years after, everything older than 7,850 radiocarbon years is pre-Babel.  Brown's curve and the post-Babel dispersal cannot both be right.

A young-earth writer could move Brown's Flood, but moving it later makes things worse.  Brown's equation stretches its time axis over the gap between the Flood and 4,000 years before 1950, when he requires radiocarbon to agree with history.  With a Hebrew-text Flood of 2348 BC that gap is 298 years, and the same equation puts White Sands 19 years after the Flood and Monte Verde 46.  To put White Sands after Babel, the curve has to stay near Flood-carbon levels for a century or more and then rise much faster than Brown's, which makes the production problem above worse.  No such curve has been published.

Two clocks on one surface

The hardest problem is not the size of either correction but their agreement.  At White Sands the same layers are dated by radiocarbon and by luminescence.  For Sibley's "around 4,000 years BP" to be right, the air must have held 16 percent of today's carbon-14 when the pondweed grew, and the sand's radiation dose must have averaged 5.4 times today's ever since, or 51 times today's if it returned to normal with the end of the Ice Age.  One is a fact about the chemistry of the atmosphere, the other about the uranium, thorium and potassium in a New Mexico lake bed.  Monte Verde poses the same pair:  air at about 34 percent, and the sealing peat's dose rate at 3.4 times today's.  Madjedbebe poses it again to 34,000 years.  No young-earth writer has proposed a cause that would move both clocks by the same amount at three sites on three continents.  The one cause that could, a speed-up of every kind of radioactive decay by the same factor, is not the model the young-earth writers give, and objection 4 below explains why it fits badly with their own decay research.  The mainstream reading needs no coincidence:  both clocks are right.

The Jomon sequence

The whole Jomon must fit after Babel.  Snelling's chart treats radiocarbon as reliable from 1000 BC, so the conventional span from 16,500 to 2,950 years ago, 13,550 years of pottery phases, must fit between Babel and 1000 BC.  With Babel 100 years after a Flood 4,350 years ago, that is 1,224 real years, 11 conventional years to each real one;  with Babel 300 years after, 13 to one;  with Hardy and Carter's earliest Flood and a Septuagint Babel, about 7 to one.  The compression must also be uneven:  little in the Late and Final Jomon, where radiocarbon is nearly right, and heavy in the Incipient and Initial Jomon.  And the calibration that turns Jomon radiocarbon ages into calendar years at 16,500 rests in part on a Japanese lake, Suigetsu, whose counted layers the young-earth side has disputed;  the site's Lake Varves article weighs that dispute and finds the count running short, not long.

Part 5: Where the mainstream case is weaker than it looks

A skeptic of dating will find these points, so they are set out here, with what each one is worth.

1. Lake Mungo's dates moved by a factor of two. Mungo III has been put at 30,000, 42,000 to 45,000, 62,000 and 40,000 years. In 1999 Alan Thorne's team combined uranium-series and spin-resonance dating of the bones into "62,000+/-6000 years" and reported it "in close agreement with OSL age estimates on the sediment into which the skeleton was buried of 61,000+/-2000 years."¹⁴ Four results agreed and were 55 percent too old. Walker is right that this is a warning against treating agreement as proof.¹ What it is worth: the error ran between 40,000 and 62,000 years. It was never within a factor of nine of a post-Babel date.
2. Madjedbebe is one site in loose sand. O'Connell's team is right that sand-sheet shelters let artefacts sink, that Nauwalabila nearby was overdated for that reason, and that "substantiation of this antiquity would require evidence from another site in Sahul situated in a more secure geomorphic setting."⁴ What it is worth: their alternative is about 50,000 years.
3. The White Sands seeds can carry old carbon, and the site has no tools. The 1947 spring seeds read about 7,400 years too old.¹⁶ The pollen ages carry about 2,400 years of uncertainty and the luminescence ages are minimums.⁹ What it is worth: even the critics' reading leaves people in New Mexico three to four times longer ago than Sibley's 4,000 years, and the land pollen, the lake mud and the quartz all agree with the seeds.
4. The first Japanese pottery is dated on one pot. Its five crusts spread over 1,100 radiocarbon years, old firewood can add "a few hundred years", and two charcoal pieces in the same layer are intruders 6,000 years younger.⁶⁶ What it is worth: the authors' own lower figure, 13,070 radiocarbon years, is still about 15,700 calendar years, and other sites give the same age.¹⁵
5. Monte Verde's older layers are thin. Dillehay concedes that some "may not meet the traditional criteria" for a valid site.⁷ This article uses only MV-II, which has a dwelling, hearths, footprints and two clocks.
6. Luminescence has a famous failure. At Jinmium, also in northern Australia, thermoluminescence dates of more than 100,000 years were later cut by a factor of more than ten, because quartz flaking off the shelter wall was buried without ever seeing daylight, and even the intermediate figure of 22,000 years was more than twice too old; the Luminescence Dating article tells the story and sets beside it Madjedbebe's own earlier dates, which single-grain dating confirmed. What it is worth: Jinmium was caught by the same single-grain method used at Madjedbebe, and it was caught by mainstream scientists.

These points are worth from a few hundred years at Odai Yamamoto to about 22,000 at Lake Mungo.  The young-earth shortfall at these sites runs from about 10,000 years at Monte Verde to about 60,000 at Madjedbebe.  None of the live disputes brings any site within 8,000 years of a post-Babel date, even with the Septuagint's longest chronology.

Objections and responses

1. "Lake Mungo proves that methods can agree and still be wrong." It does, and Part 5 prints it. But Mungo's error was a choice between 40,000 and 62,000 years, made by a second, larger set of luminescence dates on the grave's own sand units. It shows that dating can be off by half. It does not show that dating can be off by a factor of ten in the young direction.
2. "Radiocarbon on water plants is unreliable, as the 1947 Malpais Spring seeds show." Sibley and Oviatt are right about the seeds, and Oviatt's team asked for independent dating.¹⁶ It was done: land pollen, which takes its carbon from the air, and lake mud from two laboratories gave the same ages as the seeds, and luminescence, which uses no carbon at all, agreed.⁹² Sibley's article was written before any of that was published.¹²
3. "Archaeologists keep the dates that fit and drop the rest." The 1993 Creation article quotes an archaeologist saying so. At these sites the record runs the other way. Mungo's 62,000 years was dropped for a younger figure. Madjedbebe's artefacts below the 65,000-year band, in sand dated as old as 87,000 years, were set aside by its own excavators as having sunk.⁴ The White Sands team handed its dating to independent laboratories. And the dates they did publish are the ones that kept the oldest site in Australia and the oldest prints in America under open dispute.
4. "Radioactive decay ran fast soon after the Flood." If only uranium, thorium and potassium sped up, the luminescence clocks would move and radiocarbon would not, and radiocarbon would need its separate correction, the low-carbon air, matched to the first by chance. That is the model as Snelling's chart gives it: fast decay for luminescence, a changing atmosphere for radiocarbon.¹⁷ If carbon-14 sped up too, by exactly the same factor at the same time, the two clocks would move together for anything buried before the speed-up, and their agreement would be expected. That version would answer Part 4's two-clock problem. No young-earth writer quoted here proposes it; every one explains radiocarbon by the air. And it runs against the young-earth decay project's own finding, that "Alpha-decay methods generally seem to give older 'ages' than β-decay methods", which means different speed-ups for different kinds of decay.³² Carbon-14 and potassium-40 are beta emitters; much of a sediment's dose comes from the uranium and thorium chains. It would also have to happen while people were living at these sites, at 3 to 100 times today's radiation.
5. "The air had little carbon-14 after the Flood, so radiocarbon ages are inflated." This is the young-earth side's strongest card, and it is a real mechanism. Priced, it needs the air at 2 to 35 percent of today's level at these sites and then about 7 times today's carbon-14 production for a thousand years. Its one published curve, Brown's, puts the White Sands footprints at Babel and the Australian charcoal before it. The tree-ring and cave records that calibrate radiocarbon, discussed in the site's Radiocarbon article, are a separate problem it also has to answer.
6. "The decline between the Middle and Late Jomon, about 4,400 years ago, is the Flood." No young-earth writer has said so, and it cannot work. The same pottery tradition and the same regions carry on through it, with 1,598 Late Jomon sites in Hokkaido alone.¹⁰ A Flood there would make Japan's Incipient to Middle Jomon a pre-Flood culture with no survivors whose pottery style the post-Flood Japanese then resumed. The young-earth writers put all of it after the Flood.
7. "The Septuagint gives another thousand years." It helps a little. Hardy and Carter's earliest Flood, 3386 BC, with a Septuagint Babel 531 years later, gives about 4,800 years since Babel. That cuts the Madjedbebe factor from 15 to 13 and the Jomon compression from 11 to about 7. On Brown's curve it makes things worse, because Babel moves later than every site.
8. "Every dating method rests on assumptions." True, and each assumption here was tested. The hard-water assumption at White Sands was tested by dating pollen and mud. The bleaching assumption at Madjedbebe was tested by dating grains one at a time and in a second laboratory. The old-wood assumption at Odai Yamamoto was named by the authors and bounded. An objection has to name which measurement failed and by how much. None of the young-earth sources quoted here does.
9. A skeptic's objection: "Why give the young-earth side a paragraph at all?" Because the reader should see the case at its strongest. Its history, people dispersing from one family by boat and land bridge, fits the genetic and archaeological picture of fast, recent spread better than many skeptics assume. Its trouble is entirely the clock.

Issues and unanswered questions

1. Olley et al. 2006, the single-grain dating of the Lake Mungo grave fill, was not obtained. It is the study that should settle Mungo III directly.
2. A 2025 critique of the White Sands seed dates by Dave Rachal and Robert Dello-Russo, in the Journal of Archaeological Science volume 179, was found only as a title. It may bear on the seed ages, but the pollen, mud and luminescence results do not depend on the seeds.
3. The 1997 Monte Verde site visit by David Meltzer and colleagues, which brought MV-II into the mainstream, was not read.
4. No young-earth treatment of the Jomon sequence or of Odai Yamamoto was found in searches of creation.com, answersingenesis.org and icr.org. The nearest is Walker's 2009 article on Chinese cave pottery. If one exists, this article should quote it.
5. The decay-rate speed-up and its heat are priced in sister articles for sites in Africa and Europe; this article does not compute the dose to the people living at Madjedbebe or Monte Verde during a 15-fold to 100-fold increase.
6. Brown's curve is the only quantitative one found. Whitelaw, Hefferlin and Hanson, whom Brown cites, and later writers may have published curves that place Babel differently. A curve that put these five sites after Babel would need to hold the air near Flood-carbon levels for a century or more and then raise it faster than Brown's; no such curve was found.
7. The Australian rock-art record, such as Kimberley paintings dated by wasp nests and mineral crusts, also runs through the Flood date and was left out here because its dating is indirect.

Verdict

People were in northern Australia by 50,000 to 65,000 years ago, at Lake Mungo by 40,000, in New Mexico by 21,000 to 23,000, in Japan making pottery by about 16,000, and camping in southern Chile by 14,500.  Four of the five rest on two or more dating methods, the Japanese pottery on radiocarbon from two kinds of sample and two other sites, and where methods meet they agree.  The live disputes, Madjedbebe's 65,000 against 50,000 and White Sands against a reservoir effect, move the dates by thousands of years;  none comes within a factor of three of a post-Babel arrival.  The Jomon sequence and Madjedbebe's upper layers run through both biblical Flood dates without a break.

The young-earth account has a history for these peoples but no dating model for them.  Its writers set the dates aside;  none names a failed measurement at any of these sites.  Its one quantitative radiocarbon curve, Brown's, places the White Sands footprints at Babel and the Australian charcoal before it.  And at White Sands, Monte Verde and Madjedbebe, the model as its writers state it needs two unrelated corrections, one to the air and one to the ground, to agree by chance.  On the evidence, the first peoples of Australia, Japan and the Americas were there long before any date Genesis allows for the Flood.

Sources