Megasequences and the Flood Sequence

"Sequences should be considered as rock units, assemblages of formations and groups.  They are simply the strata which are included between objective, recognizable horizons, and are without specific time significance since their limits do not coincide with time lines and may include rocks of different ages in various areas." — Laurence Sloss, William Krumbein and Edward Dapples, GSA Memoir 39, 1949, pp. 110–111, defining the unit¹

"The surfaces of unconformity are rarely isochronous in cratonic stratigraphy since such surfaces typically involve truncation and loss of record in the subjacent units and time-consuming progressive onlap of overlying strata." — Laurence Sloss, twenty-five years later²

"Water levels seem to have risen and dropped simultaneously across both continents." — Tim Clarey, Institute for Creation Research, 2015³

Summary

In 1963 Laurence Sloss cut the North American rock record into six great slabs.  Each slab is bounded top and bottom by an erosion surface that can be traced across most of the continent.  He named them after Native American peoples and places:  Sauk, Tippecanoe, Kaskaskia, Absaroka, Zuni, Tejas.  The scheme is standard and taught in textbooks.

Tim Clarey of the Institute for Creation Research has spent a decade turning those six slabs into a calendar for Noah's Flood.  On his model the first three were laid down in the first forty days.  The Absaroka and the Zuni followed between Day 40 and Day 150.  The Tejas was laid down as the water receded.⁴  It is the most data-rich young-earth model of the rock record anyone has built, and it deserves a real test.

The verdict here is that the framework is genuine and the calendar is not.  The megasequences are real rock units.  They are not a set of moments.  The day-count needs them to be moments.  The man who defined them said in print, in 1949 and again in 1988, that they are not.

Three findings go to the young-earth side first.  They are true, and the case depends on them.

1. The Phanerozoic record is not an eroded remnant. The standard old-earth answer to "why is there so little Cambrian rock?" is that most of it was eroded away. Shanan Peters and Jon Husson measured it. Shallow-marine sediment on continental crust, about 75% of all surviving sedimentary volume, "does not exhibit any steady decrease in quantity with increasing age."⁵ A later review of the field⁶ says Phanerozoic rock quantity carries "little discernible overprint of erosion."⁶ This is Clarey's own argument against fellow creationist Andrew Snelling, and the best available data confirm it. How much rock was laid on the continent really did change through the Phanerozoic. Clarey is asking the right question.
2. At the median site, three of the five internal boundaries show no measurable gap. A Macrostrat run built for this article finds the median hiatus at the Sauk/Tippecanoe, Tippecanoe/Kaskaskia and Kaskaskia/Absaroka boundaries to be zero within the resolution of the data.⁷ Clarey reads this as water that never fully drained between packages. In the middle of the continent the rock does run across those contacts.
3. The Sauk really is lopsided between continents. Clarey and Davis Werner, in a 2017 Answers Research Journal paper,⁸ measured the first package at over 12 million square kilometres in North America, just under 9 million in Africa, and about 1.45 million in South America.⁸ A single ocean rising over a static world does not do that. Whatever the cause, early flooding was uneven.

Then the findings that decide it.

4. Sloss defined a sequence as a rock unit, and said it is not a time unit. The founding sentence, from Sloss, Krumbein and Dapples in 1949, is the first pull quote above: sequences are "without specific time significance".¹ In 1988 Sloss called the bounding surfaces "rarely isochronous" and gave the reason. The top of the lower package keeps being cut while the base of the upper one keeps onlapping.² For the Zuni/Tejas contact, Clarey's Day 150, Sloss puts the start of the gap anywhere from the late Late Cretaceous to the late early Paleocene.²
5. The boundaries are measurably diachronous inside one continent. Among Macrostrat columns holding rock on both sides of a boundary, the 10th-to-90th-percentile spread of the truncation level runs from 40 million years at the Zuni/Tejas contact to 118 million at the Absaroka/Zuni contact. Between 29 and 75 columns per boundary have the upper package sitting directly on Precambrian basement with the lower one absent.⁷ ICR's popular claim is that water levels "seem to have risen and dropped simultaneously across both continents."³ They did not rise and drop simultaneously across one.
6. The count of six is a convention. Sloss named four first and added two later.⁹ Harry Wheeler argued in 1963 for eight. Sloss rejected him on a judgement about causes, without a measurement.¹⁰ Stephen Meyers and Shanan Peters ran the spectral analysis Sloss could not, and propose "a quantitative redefinition that includes nine distinct units and two mega-sequences."¹¹ A Flood calendar keyed to the number six is keyed to a 1963 editorial decision.
7. Clarey's data source is Peters's data source, and it holds thousands of boundaries. Clarey states that "about 500 columns from the USA" came out of the AAPG COSUNA correlation charts.¹² Peters compiled the same charts and found 4,173 gap-bound rock packages in the United States.¹³ The Macrostrat run for this article gives 7,239 for North America, a median of seven per column, with 53.3% of columns holding more than six.⁷ The six named surfaces are six of thousands.
8. Clarey contradicts himself on the one internal signature he offers as evidence. In 2015 he reported "very few 'ideal' uniformitarian megasequences."¹⁴ In 2023 he answered the erosion objection by appealing to "the consistent internal stratigraphy of each megasequence."⁴ Macrostrat sides with the 2015 Clarey. The basal unit is sandstone in 68.8% of Sauk columns but only 36.5% of Tippecanoe and 32.4% of Kaskaskia columns.⁷ In 2025 he conceded: "there are few ideal megasequences (other than the Sauk)."¹²
9. His continental sediment volumes are about half the published figures. Clarey's North American volumes sum to 53.6 million cubic kilometres. Peters and Husson put roughly 110 million cubic kilometres of Phanerozoic sediment on North America.⁵ This project's independent Macrostrat computation gave 111.3 million, within about 1% of them.¹⁵ Two compilations agree with each other and disagree with Clarey by a factor of two. The database that would settle it has never been released.
10. One clean prediction fails, and nobody on the young-earth side has answered it. Peters and Noel Heim measured where inside a rock package fossil species first and last appear. Last appearances pile into the top 20% of a package. First appearances are spread evenly. The asymmetry "is not consistent with an unconformity-related sampling bias."¹⁶ On a rising-water model both endpoints are burial artifacts of the same water level and should be symmetric.

What this article does not repeat.  The base of the Sauk is treated in The Great Unconformity, and the surfaces between the packages in Flat Gaps.  Clarey's areas, volumes and flooding curve were tested against Macrostrat in Fossil Order and the Flood, which also credits the one thing his model gets right:  the marine-to-terrestrial order of fossils up the column is real.  This article is about the packages, their boundaries, and their count.

Part 1: What the six packages are, and what defines them

Sloss started with a mapping problem.  Working in Montana in the late 1930s he kept finding regional erosion surfaces that did not line up with the fossil-based divisions of the geologic column.⁹  With William Krumbein and Edward Dapples at Northwestern he turned those surfaces into mapping units, and presented them at a Geological Society of America symposium in 1948.¹  The definition they published the next year is the one quoted at the top of this page.  A sequence is an assemblage of formations lying between two recognizable erosion surfaces, and it carries no time meaning.¹

He restated the thesis in 1964, in a Kansas Geological Survey bulletin that is still free to read.¹⁰

"The writer (1963) has recently reiterated the thesis that the stratigraphic record of the North American craton (late Precambrian to the present) is punctuated by six interregional, cratonwide unconformities.  The unconformities are recognizable surfaces which subdivide the cratonic stratigraphic column into six vertically successive groupings of rocks, each identified by a geographic term in the manner of other lithostratigraphic units."¹⁰

A megasequence is a lithostratigraphic unit.  It is defined by a surface you can walk up to and put your hand on.  No age and no fossil enters the definition.  That is why Clarey likes the framework, and he is right to.  A young-earth geologist who rejects the radiometric timescale can still accept a unit defined by rock.

Two passages of Sloss 1963 go to Clarey's side, and they are printed here.  On the shape of a sequence, Sloss described the transgressive-regressive motif Clarey's model runs on, and said the regressive half is usually the missing half.  Each sequence begins "at the cratonic margins and in the basins of greatest subsiding tendencies, gradually spreading to the more stable areas of the cratonic interior, and ultimately lapping up on the margins of the Canadian Shield.  The transgressive phase, buried and protected by a cover of younger strata, is commonly well preserved.  The closing regressive phase of each sequence is typically poorly preserved".¹⁷  On time, the 1963 paper is less severe than the 1949 one:  "Although sequences have a greater time-stratigraphic significance than classical rock units … there is no implication in the sequence concept of an attempt to establish a North American, as opposed to a Western European, time scale."¹⁷  Sloss thought his surfaces carried more time meaning than an ordinary formation.  What he denied was that they were a chronology.

The six, with the systems each spans in Clarey's own usage:

The system spans are Clarey's own.  The third column converts them to the standard scale, which is how the tests below were set up.⁷  The day numbers are from his 2023 paper.⁴

The six were not found at once, and this matters later.  Robert Dott, who interviewed Sloss in 1996, records that "at first, four sequences were so named, which approximately spanned all of the Paleozoic systems.  Later, as they became more acquainted with the younger record toward the margins of the craton, two more sequences were named."⁹

The idea was unpopular when it was new, and the objection was about time.  At the 1948 symposium Philip King asked why the authors would burden their technique with "the wholly unwarranted 'operational units,' which are newly named and which have no relation to time-stratigraphic units?"  Dott adds:  "This was a common reaction."⁹

The modern textbook agrees about what the units are.  Octavian Catuneanu's Principles of Sequence Stratigraphy¹⁸ calls Sloss's sequences "informal mappable units that could be used for stratigraphic correlation and the subdivision of the rock record into genetically-related packages of strata."¹⁸  On the surfaces themselves the textbook is blunt:

"Strictly speaking, sequence stratigraphic surfaces commonly are not true time lines but in fact are to some degree time transgressive, or diachronous.  However, because true time lines are not commonly observed, the geoscientist is relegated to using these surfaces as proxies for time lines, being pragmatic and accepting the notion that in most instances, within the confines of most study areas they are at least very close to being time lines and therefore, are fundamentally useful."¹⁸

"Within the confines of most study areas" is the phrase that matters.  Treating a sequence boundary as a time line is a local convenience, allowed inside one basin.  A Flood-day assignment needs it to hold across a planet.

Sloss thought the cause was tectonic.  His own reading was that the craton moved vertically relative to base level.  Dott, who knew him, writes that "to his grave, Sloss remained assured that tectonics, presumably plate tectonics, somehow controlled the sequences."⁹  When his own students took the sequence idea to Exxon and rebuilt it around global sea level, Sloss mocked them.  They "find that they sleep well when they place their faith in eustatic sea levels and dream pleasant dreams when glacial controls on eustatics can be invoked."⁹

And he found no cycle in the six.  This part of Sloss 1964 is rarely quoted.  He asked whether the six repeat a pattern and answered no:  "although the observed sedimentary associations of the several sequences serve to illustrate their similarities and differences, there is no readily detectable orderly succession in the occurrence of major associations."¹⁰  Four of the six fit a model of slow crustal warping.  The Absaroka and the Tejas do not fit it, and Sloss attributed those two to mountain-building instead.  Of that pair he wrote that there is "no evidence of cratonwide synchronism."¹⁰

So the framework Clarey inherits comes with three properties its author insisted on.  The units are rock units and carry no time meaning.  The boundaries are not synchronous.  And two of the six he traced to mountain-building rather than to water.

Part 2: What Clarey's model claims

Clarey's programme is called the Column Project, and it has grown in public.  In March 2015 it held "over 500 drilling-core and outcrop samples from across North America."¹⁴  By February 2017, with South America added,¹⁹ it was "over 1,500 compiled stratigraphic columns from around the world."¹⁹  The 2023 flagship paper⁴ reports "nearly 3000 stratigraphic columns" from five continents.⁴  The 2025 reply reports coverage figures for six.¹²

The method is stated plainly, which deserves credit.  Lithology, boundaries and coordinates go into RockWorks, a commercial geology package, which draws thickness maps.  Then, in his own words, "we then trimmed the computer-generated isopach maps to match the extent of" each megasequence.⁴  Those maps give six areas and six volumes per continent.  They were tested against Macrostrat in Fossil Order and the Flood and are not re-tested here.

The claim under test here is the correlation.  Clarey needs the six surfaces to be the same six surfaces everywhere on Earth, and to mark the same six moments.  Here is how he gets there:

"Although Sloss (1963) initially defined his megasequences across only the interior of North America, oil industry geologists quickly extended these megasequence boundaries to the offshore regions surrounding North America and to adjacent continents using well logs, seismic data and outcrops … These same Sloss-megasequence boundaries were correlated to at least three other continents based on seismic data and oil well drilling results (Sloss 1972;  Soares et al. 1978;  Hubbard 1988;  Van Wagoner et al. 1990).  In fact, very similar megasequence boundaries were identified and correlated to erosional events in North America, the Russian Platform, and Brazil (Soares et al. 1978)."⁴

That paragraph carries the whole model, and it is a citation rather than a measurement.  Four papers support it.  Two of them, Sloss 1972 and Soares and colleagues 1978, could not be obtained for this article.  The Issues section below says how far the search got.

The day-count is equally explicit:

"The earliest three megasequences (Sauk, Tippecanoe and Kaskaskia) seem to have inundated only shallow marine environments as the fossils within these megasequences are almost exclusively marine … We interpret that these megasequences were deposited in the first 40 Days of the Flood … As the water rose higher, floating the Ark (on or after Day 40) and flooding portions of the dry land, the first massive coal seams appear and the first land animal fossils appear in great numbers. … depositing the Absaroka and Zuni megasequences between Days 40-150 of the Flood, until the water covered the highest hills"⁴

In the popular ICR writing the synchrony claim is much stronger than in the technical papers.  On Africa and North America³:  "The two continents' strata match up;  they independently record the same levels of the Flood at the same time and in many cases contain even the same type of sediments."³  And:  "Water levels seem to have risen and dropped simultaneously across both continents."³  Day 150, the high-water mark, is pinned to the top of the Zuni because the Zuni shows the greatest extent of sediment.³

Two points of fairness before the tests.

Clarey does not smuggle in deep time, and he says so.  "Acceptance of the validity of the geologic column does not mean we advocate for deep time nor acceptance of the geologic timescale so commonly used in conventional geology."⁴  He uses the column as a stacking order only.

His best argument is spatial.  The megasequences, he writes, are "confined to nearly the same identical locations across each of the five continents, and stack uniformly one on top of the other.  This is the general rule for all five continents.  Random erosion would not leave this consistent of a megasequence pattern across five continents."⁴  That is a good argument.  After the fossil succession it is the strongest thing in the programme, and it is answered below.

A second young-earth model makes the boundaries by draining the continents.  John Baumgardner and Evan Navarro, at the 2023 International Conference on Creationism, proposed that God cooled the Flood's new ocean floor in pulses rather than steadily.  Each pulse lowers the average sea bottom 700 m within a day, drains the continents, and cuts one of the erosion surfaces between megasequences.²⁰  A 2025 abstract revises the pulses to 600 m, one at the base of each of the first five megasequences.²¹  The paper's own rates put the sea below its pre-Flood shoreline for about 43 of the Flood's first 80 days, draining every continent four times.  That is the opposite of Clarey's reading of the eastern United States, where the floodwaters "may never have drained" between the Sauk and the Tippecanoe.  The model, the heat each pulse would have to remove, and its schedule are audited in The Heat Problem.

Part 3: Are the boundaries the same moment?

What Sloss said

Twenty-five years after naming the six, Sloss wrote the chapter on cratonic evolution for the Geological Society of America's Sedimentary Cover volume.²  It is his mature statement.  One sentence on page 25 answers the question this article is asking:

"The surfaces of unconformity are rarely isochronous in cratonic stratigraphy since such surfaces typically involve truncation and loss of record in the subjacent units and time-consuming progressive onlap of overlying strata."²

The reason is mechanical.  A sequence boundary has two halves.  Below it, erosion cuts down for different lengths of time in different places.  Above it, the next sea laps further inland.  It reaches the basin centre long before it reaches the craton margin.  Neither half happens everywhere at once.  So the surface between them cannot be a single instant.

Sloss also said something that helps Clarey.  On the same page he argues that all the rock below an interregional unconformity is everywhere older than the turn from offlap to onlap.  All the rock above it is everywhere younger.  So the surfaces "place unambiguous bounds on time slices for analysis of cratonic evolution."²  The boundaries are real, ordered surfaces with a consistent direction.  They cannot supply a day number.

Sloss is explicit about how ragged the Flood model's most important surface is.  The Zuni/Tejas contact is Clarey's Day 150.  Sloss writes that "the time of initiation of the pre-Tejas hiatal (nondepositional) interval ranges from late Late Cretaceous to late early Paleocene, and the severity of pre-Tejas erosion varies from stripping of all Phanerozoic strata from many uplifted blocks to trivial degradation at the Atlantic and Gulf margins and at the axes of deeply subsident Laramide basins."²  On the conventional scale that spread is about fifteen million years.  On Clarey's scale it is a surface that started forming at different times in different places.  A single Flood day cannot do that.

By 1988 Sloss was also no longer working with only six.  He had added subsequences with Roman numerals.  He had also folded in "intervening significant interregional unconformities" from Peter Vail's and Bilal Haq's seismic work.²

What the rock says, measured for this article

Sloss's statement is a judgement from a career of mapping.  It can be turned into a number.

The method.  Macrostrat is an open, keyless, CC-licensed database of stratigraphic columns.  Its North America project holds 24,141 rock units in 949 columns covering 24.81 million square kilometres.  Each unit carries a top age, a bottom age, a lithology and a gap-bound package identifier.  A Deno script written for this article assigns every unit to one of the six megasequences using Clarey's own system-level definitions.  Then, for each of the five internal boundaries, it takes every column holding rock in both the megasequence below and the megasequence above.  For each such column it records two numbers:  the age of the youngest top in the lower package, and the age of the oldest base in the upper one.  A synchronous surface would give nearly the same first number in every column.  The script and its raw output are published with this article and re-run in about a minute.⁷

Here is what it returns.

Read the median-hiatus column first, because it favours Clarey.  At three of the five boundaries it is negative.  That means the rock above and the rock below overlap in age within the resolution of the database.  In the median North American column there is no measurable break at the Sauk/Tippecanoe, Tippecanoe/Kaskaskia or Kaskaskia/Absaroka contact.  Clarey saw the same thing in 2015 from the raw columns.  In the eastern United States the lowest Tippecanoe rock is limestone sitting on Sauk limestone.  That "suggests that the floodwaters may never have drained off these areas between the Sauk and Tippecanoe Megasequence cycles."¹⁴  The measurement agrees with him.

Now read the spread column, which does not.  At each boundary the truncation level varies from column to column by 40 to 118 million conventional years between the 10th and 90th percentiles.  That is the middle 80% of the data.  A few outliers cannot cause it.  At the Absaroka/Zuni boundary the top of the lower package sits at 156 Ma in one tenth of columns and at 274 Ma in another tenth.  And in 29 to 75 columns per boundary the upper package rests directly on Precambrian basement, with the lower one missing altogether.

A Flood day is one day everywhere.  Suppose the top of the Kaskaskia were laid down at a single moment and then cut by a single event.  The level of the cut would then be similar from place to place.  It is not.  The claim that water levels "seem to have risen and dropped simultaneously across both continents"³ fails inside one continent, on data whose quality Clarey has never disputed.

The limits of this measurement.  Macrostrat's unit ages are interpolated from assigned chronostratigraphic intervals.  So a unit with loose age control smears across a boundary.  That produces the negative median hiatuses.  They mean "no resolvable break", and they do not prove continuity.  The smearing also inflates the spread slightly.  Macrostrat is onshore-dominated, so Clarey's offshore Tejas argument is outside its reach.  And a column counts toward a megasequence if it holds even one unit of that age, which is coarse.  None of that can manufacture an 80-million-year spread out of a synchronous surface.  A single global event recorded at 400 sites does not scatter its own erosion level across a seventh of the Phanerozoic through interpolation error.

The circularity that young-earth geologists raised first

The sharpest published statement of the correlation problem comes from three young-earth geologists.  Carl Froede, Jerry Akridge and John Reed wrote in 2015²² that Sloss's transcontinental unconformities "cannot be physically traced across the continent but are projected by connecting 'correlative conformities'.  These correlative conformities are aligned on the basis of their age, as dated by the geologic timescale."²²

That is a real problem for the Flood model.  Suppose two exposures of the same boundary a thousand kilometres apart are tied together because their fossils give them the same age.  Then the correlation is a product of the timescale.  A model that rejects the timescale cannot use that correlation as independent evidence.

Clarey has an answer.  In the 2025 reply he describes how the Grand Canyon correlation was made:  "we made correlations of the various rock types (found in the same approximate order) from column to column across southwestern USA and Mexico. … The rock correlation had little to do with fossil content, deep time, and evolutionary theory.  It was merely the physical comparison of similar patterns of rocks."¹²  That is a genuine reply.  It defeats the charge at the scale he describes:  one region traced column by column.

It does not reach the scale the model needs.  Walking a limestone from Texas to Arizona is one thing.  Saying that the surface under the Tejas in Angola is the same surface as the one under the Tejas in Wyoming is another.  No chain of adjacent columns crosses the Atlantic.  For the intercontinental step Clarey cites four papers, and no traverse.  The Issues section below takes that up.

He also concedes a version of the problem in one case.  Asked why the whole Tippecanoe is missing under the Grand Canyon, he writes:  "We infer that the tsunami waves of the Flood simply did not reach this point during deposition of the Tippecanoe, for some unknown reason."¹²  An entire package absent for an unknown reason, in the most-studied section on the continent, is a large concession.  The model's evidence is that the packages appear in the same places everywhere.

Part 4: Six surfaces out of thousands

The same charts hold four thousand boundaries

Clarey states where his American data came from.¹²  "The COSUNA stratigraphic columns were created by oil industry geologists and state geologists who used oil well and outcrop data. … We ultimately used about 500 columns from the USA."¹²  COSUNA is the American Association of Petroleum Geologists' Correlation of Stratigraphic Units of North America project.  It is a set of published correlation charts.

Shanan Peters compiled the same charts, at the same resolution, and counted every break in them.¹³  His result:  "the COSUNA charts were assembled at a temporal resolution of approximately 1–3 million years (m.yr.) in the Phanerozoic and contain 4173 gap-bound rock packages."¹³  The compilation he built from them became Macrostrat.  That is the database this article and its predecessor test Clarey against.  Clarey's data source and the database that contradicts him are the same body of observations.

Macrostrat's current North American holdings give 7,239 gap-bound packages.  That is a median of seven per column, a mean of 7.63, a maximum of 25, and 506 of 949 columns holding more than six.⁷

That number carries less weight than it looks.  Seven or eight depositional packages at one site over 371 days is one break every fifty days.  A violent year of water can absorb that.  Clarey's own answer is that "each megasequence is composed of hundreds or thousands of individual beds generated by individual waves."¹²  The count of beds is beside the point.  The record presents thousands of hiatus-bound packages, and six of them have names.  Nothing in the data marks those six as different in kind from the rest.  Nothing shows continuous rock between them.

The biggest break in North America is also missing from the six.  Peters again:  "The largest increase in the rate of sediment truncation occurs at the end-Permian, which marks a clear and fundamental temporal discontinuity in the sedimentary record of North America.  Smaller discontinuities occur at the end-Ordovician, the end-Triassic, and the end-Cretaceous."¹³  The end-Permian sits in the middle of Clarey's Absaroka.  If cratonwide erosion surfaces record stages of the Flood, the largest one on the continent records nothing at all.

Six is a number somebody chose

Four counts of the same rock exist in the literature.

• Four. What Sloss first named, covering only the Paleozoic.⁹
• Six. What he settled on by 1963, after learning the younger record near the craton margins.
• Eight. Harry Wheeler's division, published the same year. Sloss printed a note in 1964 acknowledging it: "It is Wheeler's thesis that the writer's Tippecanoe and Kaskaskia sequence should each be divided into two sequences."¹⁰ He rejected it on the ground that Wheeler was confusing local stabilisation with cratonwide uplift. That is a judgement about causes, with no measurement behind it.
• Nine, plus two. Meyers and Peters ran a spectral analysis on the area of North America covered by sedimentary rock through the Phanerozoic. They found a 56-million-year rhythm.¹¹ Their conclusion: "The newly identified 56 Myr cycle in sedimentation delineates most of the cratonic sequences that have previously been identified qualitatively in North America, but here we propose a quantitative redefinition that includes nine distinct units and two mega-sequences."¹¹

A model that assigns Day 40, Day 90 and Day 150 to particular surfaces needs those surfaces to be the natural joints of the record.  Wheeler found eight joints.  Meyers and Peters found nine.  The day numbers move if the count moves, and the count has never been settled.

One reply to head off:  "even the experts disagree, so nobody knows."  The disagreement is about how finely to slice the Phanerozoic.  All four counts agree it is hundreds of millions of years long.  Wheeler and Meyers quarrel with Sloss over where to draw lines inside 539 million years.  None of them doubts that the record is old.

The 56-million-year rhythm is also a problem in its own right.  Meyers and Peters attribute it to tectonics or mantle convection.¹¹  They are careful about it.  The signal accounts for about 14% of the variance.  The longer-term mode in their data is statistically indistinguishable from red noise.  So the rhythm should not be oversold.  It is still there.  The next calculation is this article's own.  Neither Meyers and Peters nor Clarey has done it.  Put the 538.8 million years of the Phanerozoic onto the 371 days of Clarey's Flood year.  Oard uses the same 371 days when he checks Clarey's arithmetic.²³  One day then buys 1.45 million years.  A 56-million-year period becomes 38.6 days.  It repeats 9.6 times inside the Flood year.  Catastrophic plate tectonics, the mechanism ICR proposes, is monotonic.  Seafloor is created and then cools.  That gives one rise and one fall.  Nothing in the young-earth literature produces a repeating beat, and no creationist paper has attempted it.  The scaling is crude.  A defender is free to say the mapping from years to days is not linear.  Saying so costs the model the day numbers, which are the thing under test.

Peters also states, in the paper that counted the 4,173 packages, that "Sloss's and other similar descriptions of the geologic record are primarily qualitative synopses of temporal and spatial patterns in the rock record."¹³  Clarey's areas, volumes and mean thicknesses are therefore his own construction.  They cannot inherit Sloss's authority.  They stand or fall on a database that has never been published.

Part 5: Where the model argues against itself

The ideal megasequence

Clarey's 2023 answer to the charge that erosion has removed much of the record is this:

"the consistent internal stratigraphy of each megasequence testifies against significant erosion.  Megasequences often start out with sandstone followed by shale and then carbonate rock. … Vast erosion in between each megasequence cycle would have likely destroyed this systematic signature in many locations, if not totally."⁴

Eight years earlier he had reported the opposite from his own columns.¹⁴  "Careful comparison of the rock types in each column with the sequence boundaries show very few 'ideal' uniformitarian megasequences."¹⁴  And in more detail:  "Some megasequences start with limestone at the bottom and sandstone at the top—a complete reversal of the secular story!  Others start with shale or even salt at the base, and some have very little sandstone in the sequence at all."¹⁴  On one package he was specific:  "The Kaskaskia Megasequence is also mostly limestone at its base with virtually no sandstone across its entire expanse."¹⁴

Macrostrat settles which Clarey was right.  Take the oldest unit inside each megasequence in each column.  It is sand in 68.8% of Sauk columns, 36.5% of Tippecanoe, 32.4% of Kaskaskia, 53.6% of Absaroka, 45.3% of Zuni and 68.6% of Tejas.⁷  The Sauk is a strong case and the Tejas matches it.  The Tippecanoe and the Kaskaskia begin with carbonate more often than with sand.  The lithology classifier lumps quartzite with sandstone.  If anything, that favours the sand reading.

Michael Oard is a young-earth meteorologist and the leading advocate of a late-Cenozoic Flood boundary.  He pressed the same point from South America²³:  "Basal carbonate is a major problem for megasequences because it deviates from the ideal theoretical transgressive–regressive sequence.  If erosion occurred, the upper carbonate should be eroded preferentially over the lower clastics.  So, if carbonate forms the basal Sauk in many areas, either the ideal sequence is not globally predictive or the boundaries are not correct."²³

Clarey conceded in reply:  "we agree that there are few ideal megasequences (other than the Sauk) as envisioned by uniformitarian geologists.  Instead, the chaotic nature of the stratigraphic column is more the 'rule'."¹²  That concession retires the sand-shale-carbonate signature as evidence.  With it goes half of the 2023 argument against erosion.  What is left is the flatness of the boundaries and their consistent locations.  That is a narrower case than the 2023 paper made.

"Most locations"

The 2023 paper states:  "These earliest three megasequences stack one on top of the another in most locations."⁴  Macrostrat gives the numbers.  411 of 949 North American columns hold all three.  That is 43.3% of columns and 33.0% of the continental area.  Only 119 columns, 12.5%, hold all six.⁷  The claim reaches a majority only if the denominator is columns holding at least one of the three.  There it is 57.2%.  "Most locations" is true under one reading and false under two others.  The paper does not say which it means.

The factor of two

Clarey's North American volumes across the six packages sum to 53.6 million cubic kilometres.  Peters and Husson, from 1,013 Macrostrat columns holding 22,282 sedimentary units,⁵ report "a large volume of Phanerozoic sedimentary rock (~110 × 10⁶ km³)" above the Great Unconformity in North America and the circum-Caribbean.⁵  This project's own Macrostrat computation, done independently for the fossil-order article, gave 111.3 million.  That is within about 1% of Peters and Husson.¹⁵

Two compilations built by different people for different purposes agree with each other and disagree with Clarey by a factor of two.  That does not prove he is wrong.  Volume from isopach maps and volume from unit thicknesses are different operations.  Nobody quotes the uncertainty on either.  It does mean the disagreement cannot be resolved.  The Column Project database has never been released.  Froede's group asked for it in 2015 and Oard again in 2025.

Part 6: The prediction nobody has answered

Peters and Heim asked a question that turns out to be a clean test of any rising-water model.¹⁶  Inside a hiatus-bound rock package, where do fossil species first appear and where do they last appear?  They crossed Macrostrat packages against the Paleobiology Database and found:

"LADs are more common than expected in the top ~20% of package durations, whereas FADs are more uniformly distributed.  This result is not consistent with an unconformity-related sampling bias."¹⁶

LAD is a last appearance datum, FAD a first appearance datum.  Last appearances bunch at the tops of packages.  First appearances do not bunch at the bottoms.

On Clarey's model both endpoints are the same kind of thing.  A group first appears in the record when its elevation belt is flooded.  It last appears when that belt is buried out.  Both are hydraulic events driven by one water level.  So both should crowd the package boundaries in mirror image.  The measurement says they do not.  The authors rule out the sampling explanation that would rescue the symmetry.

The mainstream reading is that extinction and marine regression run on similar timescales, while origination and transgression do not.  A Flood advocate can reply that regressive burial concentrates last occurrences.  That reply is reasonable.  It then has to explain why transgressive burial does not equally concentrate first occurrences.  Nothing in the young-earth literature attempts it.  Tomkins and Clarey's 2023 fossil paper does not cite Peters and Heim.²⁴  Both databases are open and the analysis is free to re-run.

Objections and responses

"Sloss's caution about time lines is a caution about biostratigraphic dating.  The Flood model rejects that dating anyway."  Half right, and the right half matters.  Sloss's 1949 point was that a rock unit need not line up with the fossil-defined divisions.  A creationist can shrug at that.  The 1988 statement is about geometry.  A truncation surface cuts for different lengths of time in different places.  An onlapping base arrives in different places at different times.²  That holds whether the interval is fifty million years or fifty hours.  A defender who rejects Macrostrat's ages still has to explain one thing.  At the same "moment", the upper package rests on bare Precambrian in dozens of columns and on a full lower package in hundreds of others.

"The packages are real and continent-wide.  That is all the model needs."  It needs more.  Real, continent-wide packages give an ordered stack of rock.  Nobody disputes that.  Clarey's model needs Day 40, Day 90 and Day 150 to be surfaces.  It needs the surface in Angola to be the same surface as the one in Wyoming.  Order is compatible with any duration.  The day numbers are the claim under test.  They need synchrony, the one property Sloss denied his own surfaces.

"Macrostrat's ages are radiometric and biostratigraphic, so testing a Flood model against them is circular."  This is the strongest creationist reply and it is partly correct.  Every age in the table in Part 3 comes from the conventional timescale.  Two things blunt it.  First, the test does not use the ages as ages.  It uses them as a common ruler to ask whether one surface is at the same level everywhere.  Compress the whole scale by a factor of half a billion, and the spread compresses with it.  The ratio of scatter to package duration does not change.  A surface whose cut level varies across a fifth of the Phanerozoic is a ragged surface on any scale.  Second, the megasequence assignment itself is lithostratigraphic.  Macrostrat's 24,141 units partition into Clarey's six intervals without difficulty.  That is a point in his favour.  The fossil-order article reported it as such.

"The counts disagree because correlating anything to a published chart is loose.  Andrew Miall showed random sections match the Exxon chart most of the time."  He did,²⁵ and the number is worth having:  "77% correlation with the Exxon Chart can be achieved between four random number-generated sequences."²⁵  A chart with more than a hundred third-order cycles will find a partner for almost any section.  So a match to it proves little.  That is a real result, and it goes against a common anti-creationist move.  It does not touch this article's counts.  Macrostrat's 7,239 packages are contacts observed in columns and recorded there, with no chart involved.  Meyers and Peters's nine units come out of a spectral analysis of how much of the continent held sediment through time, again with no chart.  Miall's target is the Exxon eustatic curve.  This article does not use it.

"Sequence stratigraphy is ordinary mainstream geology.  Why does it matter that a creationist uses it?"  Because this use is unusual.  Petroleum geologists correlate sequence boundaries inside a basin.  That is where Catuneanu's "within the confines of most study areas" licence applies.¹⁸  Clarey extends five surfaces across six continents and dates each in a 371-day calendar.  The framework being mainstream answers the charge that megasequences are invented.  It says nothing about the calendar.

"You conceded that the Phanerozoic record is not an eroded remnant.  Doesn't that hand Clarey the argument?"  It hands him a real question and removes a bad answer.  Suppose the thin early Paleozoic is thin because little was deposited.  Then continental flooding really did change through the Phanerozoic, and that needs explaining.  The mainstream explanation is the supercontinent cycle.  It predicts two flooding maxima, Late Ordovician and Late Cretaceous.  That two-peaked shape is what Macrostrat's marine coverage shows.  Clarey rejects it by name.  The concession costs the old-earth side its standard reply.  It does not buy the Flood calendar.

Issues and unanswered questions

Sloss 1963 is quotable but not readable.  The founding paper, GSA Bulletin 74:93–114, is paywalled at GeoScienceWorld, absent from archive.org's periodical holdings, and returns 403 from HathiTrust.  Google Books holds a scan of volume 74.  Its search-inside endpoint returns the printed page of each exact phrase.  That is enough to confirm a quotation and pin its page, and no more.  So the article quotes Sloss 1963 only twice and paraphrases nothing from it.¹⁷  The rest is carried by three other texts.  Sloss's own 1964 restatement in Kansas Geological Survey Bulletin 169 is open.  His 1988 chapter was read paragraph by paragraph through archive.org's search-inside service.  The 1949 definition was read in the printed memoir itself.

The intercontinental claim is left open, and here is how far it got.  Clarey's global correlation rests on four citations.  Two carry the weight:  Sloss 1972, on the North American craton and the Russian Platform, and Soares, Landim and Fúlfaro 1978, on the Brazilian intracratonic basins.⁴  Sloss 1972 is a 24th International Geological Congress proceedings paper.  It is print-only and not indexed by Crossref or OpenAlex.  It was not obtained in any form.  For Soares and colleagues 1978 the full paper is still closed, but the publisher's own abstract was read.²⁶  It does not name Sloss's six.  It divides the Brazilian record into seven units of its own, lettered Alpha through Zeta, and splits Clarey's Absaroka interval across two of them.  On the intercontinental step it says the cycles "are synchronous on the Brazilian craton and are correlated with cratons of other continents," and describes the Devonian rise and fall as running "in a pattern similar to that of the North American craton and the Russian platform."²⁶  Synchronous within Brazil;  similar in pattern between continents.  That is the weaker of the two readings, and it is the one the abstract states in its own words.  An abstract is not a paper, and the body may argue for more.  So the question stays open.  But nothing yet seen in Clarey's own cited source claims that a boundary in Brazil is the same moment as a boundary in Wyoming.  This article's case against synchrony is therefore measured inside North America and left open between continents.

Miall's randomisation critique was checked and is now used, in the objections above.  The 1992 Geology paper itself is still closed.  Miall restates the result in his own textbook, and that sentence was read at its printed page.²⁵

The Column Project database is still unpublished, ten years after it was first asked for.  Every quantitative disagreement above would be settled in an afternoon if it were released.  The request is made without insinuation.  The model may be right.  The only way anyone can find out is the data.

Verdict

Sloss's six packages are real.  So are the erosion surfaces between them.  And the old-earth habit of explaining away the thin early record as erosion is wrong.  Clarey saw that before most of his critics did.

What fails is the calendar.  The man who defined the units said they are not time units, and that the surfaces are rarely isochronous.  Measured against 949 North American columns, each of the five internal boundaries cuts to a level varying by 40 to 118 million conventional years.  The upper package sits on bare basement in dozens of places and on continuous rock in hundreds.  North America holds 7,239 gap-bound packages, and six of them have names.  The count of six has been disputed since the year it was published.  The current quantitative alternative is nine.  The one internal signature offered as evidence has been withdrawn by its own author.

Megasequences are a real framework, and a young-earth geologist is entitled to use it.  They are not a clock.  They will not tell anyone what happened on Day 150.

Sources