Deep-Sea Sediment

"can the oceans supply enough nutrients, in some 4500 years, to explain the observed sediments?" — Larry Vardiman, Creation Research Society Quarterly, 1996¹

"the carbonate flux to the floor of the ocean is in the order of 25 to 250 mg m⁻² day⁻¹ which is several thousand times slower than the potential figures given above." — Ariel Roth, Origins, 1985²

"Because the LR04 stack is constructed by graphic correlation, its stratigraphic features are essentially independent of any timescale." — Lisiecki and Raymo, Paleoceanography, 2005³

Summary

The deep ocean floor carries about 400 m of sediment on average.  About sixty per cent of it by area is the shells of plankton.  The young-earth question is whether that pile could have formed in the 4,500 years since a global Flood.  The only quantitative young-earth model says yes:  Larry Vardiman's 1996 ICR monograph, condensed in the Creation Research Society Quarterly, has deposition start enormous and decay exponentially to today's rate.  The verdict of this article is that the measured rates, the sediment's own record, and the creationist literature itself all rule the model out.  The strongest evidence comes from measurements that contain no assumed age at all, so the usual creationist reply, that the rates are circular, does not reach them.

Three findings favour the young-earth side, and each comes from mainstream sources.  They are stated first so nobody can say they were buried.

1. Jake Hebert's 2017 recalculation of the "Pacemaker of the Ice Ages" paper is real technical work, and no direct mainstream reply to it was found. Hays, Imbrie and Shackleton dated one of their cores with a magnetic reversal then put at 700,000 years. The mainstream age is now 780,000. Hebert re-ran their spectral analysis with the new age and reports that the obliquity and precession peaks moved off the predicted frequencies.⁴ This article answers it only indirectly.
2. The ages on the standard deep-sea isotope stack really are tuned, and its authors say so. Lisiecki and Raymo's LR04 stack aligns 57 cores to each other by shape, then assigns ages by fitting to an insolation-driven ice model.³ Only the succession of stages is independent of any timescale. And Huybers and Wunsch, who built an untuned chronology, could not reject the null hypothesis that the 100,000-year cycle is unrelated to orbital eccentricity or precession.⁵ The 100-kyr cycle is a live mainstream problem.
3. Manganese nodules really are almost absent below the top few metres of sediment. Glasby's 1978 survey of 370 DSDP cores found nodules deeper than 200 m in only ten of them, and called their scarcity at depth "the most striking feature of the data."⁶ This is a real pattern with no complete mainstream explanation, and the creationist literature is right to press it.

None of these three rescues the model.  Four findings decide the case.

1. The measured rate is now known without any age model, and it is a few centimetres per thousand years. Thorium-230 is made in seawater from dissolved uranium at a rate fixed by two decay constants, and it sticks to falling particles. How much of it a gram of sediment carries tells you how fast that gram arrived. No stratigraphy, no tuning, no reversal and no assumed age enters the calculation.⁷ Costa and 33 colleagues applied it to over 1,000 records from every ocean in 2020 and got a global late-Holocene mass flux of 1.48 to 1.68 g per cm² per thousand years.⁸ Vardiman's model, by his own arithmetic, needs 0.75 to 3.5 m of sediment on the whole deep ocean floor in its first year, which is about 30,000 to 140,000 times that flux.
2. Vardiman's model puts a magnetic reversal in the Iron Age. His own re-plotted core, V28-238, carries the Brunhes-Matuyama reversal at 12.0 m depth. On his own age equation that layer was laid down 2,700 to 3,300 years before present, which is roughly 1300 to 700 BC. ICR's current position, stated by Hebert in 2014, is that every reversal happened inside the Flood year.⁹ Both cannot be true. And in exactly that window, Vaknin and colleagues measured the field in 20 dated burnt destruction layers across Israel and found it strong, normal, and drifting by three or four degrees at a time.¹⁰ No reversal happened.
3. The founding Milankovitch result was not tuned, and there is a radiocarbon date in Vardiman's own core. Hays, Imbrie and Shackleton's 1976 chronology for core RC11-120 had two control points, the core top and a 127,000-year stage boundary taken from radiometric dating of a Caribbean core and checked against uranium-dated coral terraces. Between them, age was a straight line in depth.¹¹ The same page reports a carbon-14 age of 9,400 ± 600 years at 36 to 39 cm in that core, which is 3.8 to 4.1 cm per thousand years, measured.¹¹
4. The creationist literature contradicts itself by three to four orders of magnitude and has not noticed. Kenneth Patrick's 2010 nodule paper computes "only tens of centimetres in a total of 5,000 years" for the post-Flood deep ocean and accepts it.⁶ Vardiman's model lays down 322 m in 4,500 years. Andrew Snelling's 2009 reference work caps post-Flood ocean productivity at "5 to 10 times greater than at present."¹² Vardiman needs tens of thousands of times. None of the three cites the others' number.

Vardiman is treated fairly here, and one limit must be stated at the outset.  His monograph could not be obtained.  It is out of print, not on archive.org, and reaches this article only through Michael Oard's 1996 review and Vardiman's own condensation in the Creation Research Society Quarterly.  Every number attributed to him below comes from those two documents, and the monograph's five data appendices, which Oard says are its real substance, have not been seen.

Part 1: What is on the sea floor

How much.  The best current inventory is GlobSed, a 2019 global grid of sediment thickness built from seismic surveys and drill holes by Straume and eight colleagues.  It puts the total marine sediment volume at about 3.37 × 10⁸ km³.  Most of that is piled against the continents:  the margins hold about 1.43 × 10⁸ km³ on 4.69 × 10⁷ km², a mean of 3,044 m.  The deep ocean, more than 200 km from the continent-ocean boundary, holds about 1.13 × 10⁸ km³ on 2.79 × 10⁸ km², a mean of 404 m.¹³  That 404 m is the pelagic column this article is about.  It is the same figure the creationist geologist Ariel Roth used in 1985, when he wrote that "An average depth of about 0.4 km may be generous for the floor of the oceans and is a few percent of what was conceived earlier."²  Nobody on either side now claims kilometres of ooze, so the old creationist point that the column is thinner than once thought is spent.

Vardiman's own inventory came from the Deep Sea Drilling Project.  Of 624 DSDP sites, only 186 reached sediment of Cretaceous age or older, and at those 186 the mean thickness above the Cretaceous-Tertiary boundary was 322 m, with a standard deviation of 273 m.¹  Oard's review reports the monograph's global average as 600 m of wet sediment, or 300 m once the pore water is removed.¹⁴  These are the same order as GlobSed.

What kind.  Diesing's 2020 machine-learning map of sea-floor lithology below 500 m, trained on 10,438 samples left after duplicates were dropped from an original 14,400, gives the split by area.  Calcareous ooze, the shells of foraminifera and coccolithophores, covers 121 × 10⁶ km², or 36.8 %.  Clay covers 102 × 10⁶ km², or 31.0 %.¹⁵  Radiolarian ooze is mapped at 14.2 %, diatom ooze at 8.5 %, and land-derived sediment at 9.5 %.  The radiolarian figure should be read as an upper bound:  the model's overall accuracy is about 59 %, radiolarian ooze had only 99 training samples, and Diesing notes that his map spreads it wider than the classical hand-drawn maps do.¹⁵  Roth's 1985 partition was simpler and is close enough:  half red clay, half carbonate ooze, and the carbonate half split evenly between coccoliths and foraminifera.²

Thickness follows the age of the crust beneath it.  This is the finding in GlobSed that matters most here.  Straume and colleagues fit thickness across the whole ocean, excluding crust older than 82 Myr and latitudes beyond 72°, and found it rises with the square root of the crust's magnetic-anomaly age and with latitude:  "The results presented in Figures 9 and 10 agree with previous findings that sediment thickness increases with age of the oceanic lithosphere (e.g. Olson et al., 2016).  In addition, our analysis confirms that sediment thickness is also latitude-dependent, showing an increase along equator and towards high latitudes."¹³  Their world-ocean fit is Z = √τ · (53.02 − 2.46·φ + 0.045·φ²) metres, with τ the crustal age in Ma and φ the absolute latitude in degrees.  Thickness is the square root of crustal age times a latitude factor.  That factor has its minimum near 27°, which is the subtropical desert belt of the ocean where plankton are scarce.  Thickness follows biological productivity and crustal age together, which is what a slowly raining column on a spreading floor predicts.  Vardiman's model has no crustal-age term at all.  In catastrophic plate tectonics, the sibling model most young-earth geologists now hold, the whole ocean floor formed within the Flood year, so crust labelled 5 Myr and crust labelled 80 Myr differ in real age by weeks.  Snelling's 2009 reference work says as much:  "it seems likely that most of the Tertiary sea floor was formed during the run-off stage during the year of the Flood."¹²  Then the sediment on top has no reason to track √τ, and it does.  The plate-tectonic side of this is argued in Catastrophic Plate Tectonics, which covers DSDP Leg 3 and the thickness-versus-age test;  it is not repeated here.

Part 2: How fast it accumulates, measured

Everyone in this dispute accepts the same present-day rate.  Vardiman wrote in 1996 that "The rate appears to vary between about 1 cm/1,000 years to about 10 cm/1,000 years, depending on the investigator and location on the earth," and set his model's modern rate at 2 cm per thousand years.¹  Snelling in 2009 accepts "2 to 10 centimeters (0.78 to 3.93 inches) per thousand years" for coccolith ooze today.¹²  Hebert in 2014 wrote that "it can take a thousand years for a few centimeters of sediment to be deposited on the ocean floor."⁹  The argument is entirely about whether the past rate was different.  So the question is how the rate is measured, and whether any measurement is free of the assumptions the creationist side rejects.

Method 1:  thorium-230, which needs no age at all.  Seawater holds dissolved uranium-234.  As Francois, Frank, Rutgers van der Loeff and Bacon put it in their 2004 methods paper, "While residing in seawater, ²³⁴U decays and produces ²³⁰Th at a uniform and exactly known rate (β230) throughout the water column".⁷  That rate is the decay constant of thorium-230 times the seawater activity of uranium-234:  0.0267 disintegrations per minute per cubic metre per year.  Thorium is intensely particle-reactive, so almost all of it is scavenged onto falling particles and reaches the floor beneath the water that made it.  Over a 4,000 m water column, the floor receives about 107 dpm of thorium-230 per square metre per year, whatever else is falling.  If a lot of sediment falls, each gram carries little thorium;  if little sediment falls, each gram carries a lot.  Measure the excess thorium-230 in a gram of sediment and divide it into the known production, and the vertical mass flux comes out directly.  The flux calculation itself uses no stratigraphy, no orbital cycle, no magnetic reversal and no assumed age:  "In many instances, this problem can be resolved, at least for sediments deposited during the late Quaternary, by normalizing to the flux of ²³⁰Th scavenged from seawater, which is nearly constant and equivalent to the known rate of production of ²³⁰Th from the decay of dissolved ²³⁴U."⁷

In 2020 Costa, Hayes, Anderson, Pavia and 30 co-authors gathered every published record and applied the method at two time slices:  "We synthesized over 1000 sedimentary records of ²³⁰Th from across the global ocean at two time slices, the late Holocene (0–5,000 years ago, or 0–5 ka) and the Last Glacial Maximum (18.5–23.5 ka), and investigated the spatial structure of ²³⁰Th-normalized mass fluxes.  On a global scale, sedimentary mass fluxes were significantly higher during the Last Glacial Maximum (1.79–2.17 g/cm²kyr, 95% confidence) relative to the Holocene (1.48–1.68 g/cm²kyr, 95% confidence)."⁸  At a dry bulk density of 0.6 g/cm³, typical for surface ooze, 1.5 g per cm² per thousand years is about 2.5 cm per thousand years.  The slowest places on earth are the centres of the Pacific gyres, below 0.5 g/cm²/kyr;  the fastest pelagic sites, on Pacific margins and in the Bering Sea, reach about 8.⁸  One caveat belongs here.  An age model does enter the synthesis at two points:  it picks which centimetres of a core count as late Holocene, and it sets the correction for thorium-230 decay since burial.  Costa and colleagues say so:  "All data are presented using the age models in the original publications", derived from radiocarbon or oxygen-isotope stratigraphy.⁸  But they bound how much it matters:  a 1,000-year age error changes the flux by less than 1 %.⁸  Thorium-230's half-life is 75,584 years, so over a 5,000-year window the decay correction is a few per cent.  Costa and colleagues list the method's other limits themselves:  it misbehaves near hydrothermal ridges and on steep margins, and it depends on an assumed lithogenic uranium-to-thorium ratio.  Their conclusion is that "²³⁰Th normalization is a robust tool for determining sediment mass accumulation rates in the majority of pelagic marine settings (>1,000 m water depth)."⁸  None of the listed limits is a factor-of-ten effect, let alone a factor of ten thousand.

Method 2:  radiocarbon in Vardiman's own core.  Hays, Imbrie and Shackleton's 1976 paper reports, on the same page as their age-control table, that "In RC11-120, carbon-14 dating at the 36- to 39-cm level yields an age of 9400 ± 600 years (35)."¹¹  That is 3.8 to 4.1 cm per thousand years.  RC11-120 is one of the two cores Vardiman re-plots on his compressed time axis.  The date uses no orbital theory, no reversal and no isotope stratigraphy.

Method 3:  the core-to-core rates the mainstream literature actually uses.  Table 2 of the same 1976 paper lists twelve sedimentation rates for its two cores under three age models.  All twelve lie between 2.70 and 4.14 cm per thousand years.¹¹  Lisiecki and Raymo report that across their 57 sites the standard deviation of a site's sedimentation rate averages 41 % of its mean.³  Rates vary from core to core by tens of per cent, not by four orders of magnitude.

Method 4:  sediment traps.  Roth reports the trap fluxes he knew in 1985 as 25 to 250 mg of carbonate per square metre per day, and says they are "several thousand times slower than the potential figures given above."²  At 2 g/cm³ that is 0.46 to 4.6 cm per thousand years, the same band as every other method.

Sediment focusing is real, and it is bounded.  Bottom currents move sediment sideways, and Francois and colleagues note that where bottom circulation is vigorous, "the accumulation rates of laterally redistributed sediments are often many times larger than the fluxes of material sinking from the overlying surface waters".⁷  A creationist could read that as a mainstream admission that measured rates overstate true rain.  But the whole point of thorium-230 normalisation is to strip focusing out, since the thorium moves with the sediment.  Costa and colleagues measured focusing factors explicitly and found them above 1 in some regions and below 1 in others.  Sideways transport does not create sediment;  the global inventory is unchanged.  They also note that their sites are biased "toward regions with rapidly accumulating sediments", which if anything makes the global mean an overestimate.⁸

Put the four methods together and the deep ocean floor accumulates at roughly 1 to 4 cm per thousand years, with extremes from 0.5 to about 13.  At 3 cm per thousand years, 404 m takes about 13 million years.  At 1.5 cm it takes 27 million.  Those are the numbers the young-earth model has to beat.

Part 3: Vardiman's model, and what it needs

The model, stated fairly.  Vardiman's abstract puts it in one sentence:  "The model of deposition presented in this paper differs from the conventional model primarily in the rate of deposition, which is asserted to have peaked at an enormous level during and after the biblical Flood and is presumed to have fallen at an exponential rate to the present low level."¹  He takes the Flood/post-Flood boundary at the Cretaceous-Tertiary boundary, and the column above it, 322 m on average, as post-Flood.  He assumes the deposition rate decays as dy/dt = A·e^(−t/τ), and fixes A and τ so that the column reaches 322 m in 4,500 years while the rate at the end equals today's 2 × 10⁻⁵ m per year.  He reports the result:  "The solution to this transcendental equation gives a value for τ of 373 years."¹  He then re-plots published oxygen-isotope curves from DSDP sites 277, 279 and 281 and from cores RC11-120 and V28-238 on that compressed axis, reads off an ocean cooling of 15 to 20 °C in the first thousand years after the Flood, and relabels the oscillations:  "If the young-earth age model proposed by this work is valid, the conventional correlation between sea-floor sediments and the orbital parameters is completely false.  The periods illustrated in Figures 4 and 5 are on the order of 100 years and 700 years."¹  He is candid about his reason for distrusting the modern rate.  He assumes the published figures are too low because they were produced by people who needed millions of years:  "it is likely that the average accumulation rates quoted are biased to small values."¹

The arithmetic does not close.  His equation for the column height is H = r·τ·(e^(4500/τ) − 1).  Put in his own r = 2 × 10⁻⁵ m/yr and H = 322 m and solve for τ.  The root is 427 years, not 373.  At τ = 373 the equation gives 1,295 m, four times his column.  Read the other way, τ = 373 with H = 322 m needs a Flood 3,981 years ago, not 4,500.  He says he solved the equation graphically, and the error is the size of a graphical error.  It is not harmless, because the model's initial rate is A = r·e^(4500/τ).  At τ = 427 that is 0.75 m per year;  at his τ = 373 it is 3.47 m per year.  Either way, the model demands that the entire deep ocean floor received between 0.75 and 3.5 m of sediment in the first year after the Flood, against 0.02 mm today.  That is a ratio of 38,000 to 170,000 over the present rate.  Half the column, 161 m, is down within 296 years at the corrected τ = 427, and within 50 years at his stated τ = 373.  The ratio is Vardiman's own arithmetic, and it is what the rest of this article tests.

Against the thorium-230 flux the comparison is direct.  Costa's 1.5 g/cm²/kyr over the 2.79 × 10⁸ km² of deep ocean is 4.2 × 10¹⁵ g per year.  Vardiman's 3.47 m/yr at 0.6 g/cm³ over the same area is 5.8 × 10²⁰ g per year.  The ratio is 1.4 × 10⁵;  at the corrected τ it is 3 × 10⁴.  The number comes out the same through Vardiman's linear rates or through the measured mass flux.

Where the productivity is supposed to come from.  Vardiman cites one source for it, Roth's 1985 paper in Origins, and the whole later creationist literature on this topic descends from Roth.  Roth's argument has four parts, stated on his first page:  "1) the layers of shells are not kilometers thick as has been reported, but probably at best an average of 0.2 km;  2) the biological potential of production is so great that this quantity of shells could probably be produced in much less than 2000 years;  3) a worldwide flood as described in Scripture could provide the nutrients necessary for such production;  4) caution is warranted because of the poor data that are currently available."²  The productivity figure is a cell-division calculation:  "if one assumes a high concentration of foraminifera of 100 l⁻¹ as has been reported (see Berger 1969), a doubling time of 3.65 days (Berger 1976, pp. 273, 299) and an average of 10,000 forams g⁻¹ of carbonate (Berger 1976, p. 298), the top 200 m of the ocean would produce 20 g carbonate cm⁻² y⁻¹ or (at an average sediment density of 2 g/cm³) 100 m in 1000 years."²  His coccolith figure, 54 cm of carbonate per year, comes from "13×10⁶ coccolithophores per liter as reported for Oslo Fjord (Black & Bukry 1979)" applied to the top 100 m of the entire world ocean.²  Snelling repeats it in 2009 as 55 cm per year, enough to make 100 m of coccolith ooze "in less than 200 years."¹²

Roth then says what those numbers are.  "It must be emphasized that the high rates given above are optimum and do not appear at all to represent average present-day rates.  The figures given represent the biological potential of these organisms."²  And he names the limit he cannot remove:  "If one assumes a balanced steady-state model, it does appear that at present the slow input of calcium carbonate into the oceans from rivers, etc., may be a major limiting factor in carbonate skeletal production and preservation in the ocean."²  His answer is that carbonate is "dissolved and recycled".  Recycling turns a stock over;  it does not enlarge it.

The carbonate budget will not stretch.  Cartapanis, Galbraith, Bianchi and Jaccard's 2018 review gives the modern numbers:  "the modern CaCO₃ burial in deep-sea environments has been estimated to be between 100 and 150 PgC kyr⁻¹ (Catubig et al., 1998; Sarmiento and Gruber, 2006; Dunne et al., 2012)."¹⁶  That is 0.8 to 1.3 × 10¹³ mol of calcium carbonate per year buried in the deep sea.  Milliman's 1993 budget, relayed by the same paper, put total carbonate burial at 360 Pg C per thousand years, half of it in shallow water.¹⁶  Only a fraction of what plankton make survives to burial:  "The fraction of pelagic CaCO₃ exported from the surface ocean that is eventually buried in sediments is estimated to vary between 0.6 and 0.06".¹⁶  So gross carbonate export today is at most 17 times burial, under 2 × 10¹⁴ mol per year.  Roth's optimum, 20 g/cm²/yr over the ocean, is 7.2 × 10¹⁷ mol per year, about 4,000 times the entire modern export.  Vardiman's τ = 373 model, at 1.5 g/cm³ and 50 % carbonate, needs 7.3 × 10¹⁸ mol per year, about 40,000 times it.  Dissolution makes this worse, not better.  It removes 40 to 94 % of what is exported, so to bury 322 m the model must grow 1.7 to 17 times that much shell.  A textbook value for the ocean's dissolved calcium, about 1.4 × 10¹⁹ mol, would be exhausted by Vardiman's rate in under two years and by Roth's in about twenty;  that inventory has not yet been checked against a primary source, and is flagged below.

Vardiman names the problem and does not solve it.  His conclusions end with a question:  "The accumulation of hundreds of meters of sediment, on the average, and kilometers of sediment in some locations, such as the Arctic Ocean, require very high productivity following the Flood.  Although the potential for high productivity has been suggested by Roth (1985), can the oceans supply enough nutrients, in some 4500 years, to explain the observed sediments?"¹  He also doubts his own boundary:  "It may be that the Cretaceous/Tertiary boundary is too deep in the geologic column."¹  Thirty years later no creationist publication has answered the question.

What the friendliest reviewer withheld.  Oard's 1996 review is the only published creationist review of the monograph, and it declines three of the four load-bearing claims.  On the boundary:  "the Cretaceous-Tertiary boundary in the ocean is a uniformitarian interpretation based on microfossils. ...  Since these organisms were alive before and during the Flood, and probably after the Flood, we cannot simply borrow their relative time-frame for the Flood and say that the ocean floors are late Flood."¹⁴  On the temperatures:  "the magnitude of the derived temperature change is questionable."¹⁴  On the 100-year and 700-year periods:  "At this stage of the investigation, it is likely too early to speculate on the time-frame of oxygen isotope oscillations found at the tops of cores.  These fluctuations have been correlated to repeating ice ages by uniformitarian scientists.  The physical meaning of these oscillations still needs to be explained within the creationist paradigm."¹⁴  Oard offers a mechanism, rapid post-Flood cooling driving vigorous overturning and upwelling, and then concedes:  "However, the details of these processes still need to be worked out."¹⁴  They have not been.

Part 4: What the column records

A column of ooze is not a blank pile.  It carries a global sequence of oxygen-isotope stages, magnetic reversals, and a surface layer of manganese nodules.  A model of its age has to place each of those.

The isotope stages

Shackleton and Opdyke's 1973 study of equatorial Pacific core V28-238 is the founding record.  Its abstract:  "Core Vema 28-238 preserves an excellent oxygen isotope and magnetic stratigraphy and is shown to contain undisturbed sediments deposited continuously through the past 870,000 yr.  Detailed correlation with sequences described by Emiliani in the Caribbean and Atlantic Ocean is demonstrated.  The boundaries of 22 stages representing alternating times of high and low Northern Hemisphere ice volume are recognized and dated."¹⁷  The core is 16.09 m long;  its 22 stages occupy the top 13.4 m.  Their age model was a uniform rate of 1.71 cm per thousand years, fixed by the Brunhes-Matuyama reversal at 12.0 m.¹⁷

Lisiecki and Raymo's 2005 LR04 stack extends this to 57 cores from every ocean, 38,000 measurements, 232 stage boundaries over 5.3 million years.  Two things about it must be printed together.  First, the ages are tuned, and the authors say so.  Each core is aligned to the others, then "Each alignment is evaluated by eye and adjusted, if necessary, by changing the sedimentation rate penalties or adding tie points until a good alignment is achieved which agrees reasonably well with previously published age estimates (e.g., those derived from paleomagnetic reversals or biostratigraphic data)."³  The ice-model time constant of 15,000 years "is selected to maximize agreement with the independent age estimates used for the last 135 kyr of the stack."³  Hebert's charge that the ages are circular has a real target.  Second, the succession of stages is not tuned:  "Because the LR04 stack is constructed by graphic correlation, its stratigraphic features are essentially independent of any timescale. ...  However, almost any age model could be applied to the LR04 stack."³  Graphic correlation matches shape to shape.  Fifty-seven cores in separate oceans carry the same ordered sequence of 232 oscillations at the right relative depths before any age is assigned.  Tuning cannot manufacture that.  A young-earth model has to explain it, and none has tried.

On Vardiman's own equation, V28-238's 22 stages fall as follows.  Their published age model lists 21 stage boundaries by depth.  The layer at 2.20 m, the 5-6 boundary, comes out 2,640 years old;  the layer at 13.40 m, the 21-22 boundary and the deepest in the table, comes out 3,314 years old.  So the sixteen boundaries below the 5-6, sixteen global swings of ice volume, are squeezed into 674 years, one every 42 years.  The five youngest stages take the remaining 2,640 years.  Vardiman's "100 years and 700 years" are these numbers read off the plot.  No physical oscillator with those periods, acting on global ice volume, synchronously, has been proposed by anyone.

The founding Milankovitch result was not tuned

Hays, Imbrie and Shackleton's 1976 "Pacemaker" paper is the one Vardiman calls "completely false" and Hebert calls circular.  Neither describes its chronology.  Its Table 2 lists every age-control point in every model.  The SIMPLEX model has two:  "SIMPLEX models assume uniform sedimentation rates through each core and fix the core top of RC11-120 at zero, the 6-5 stage boundary at 127,000 years (cores RC11-120 and E49-18), and the 12-11 stage boundary at 440,000 years (core E49-18; see Table 2).  It is important to note that although the SIMPLEX model presents each variable as a function of age, the age estimated is an exact linear function of depth in the core."¹¹  The 127,000-year point is footnoted to Broecker and van Donk, and the authors say it is "within the range of several radiometric dating techniques" with "an analytical error estimated at ± 6000 years", constrained by "the terrace coral ages from Barbados, New Guinea, and Hawaii".¹¹  Hebert says the same thing in his own words:  the 127,000-year figure came from "Protactinium-thorium dating applied to the V12-122 Caribbean core".⁴  Either way it is a radiometric date, not an insolation curve.  Depth was proportional to age by assumption, and the climate signal still came out periodic at 23,000, 42,000 and about 100,000 years, matching three separate astronomical periods computed from celestial mechanics.  The paper also reports an independent check on its deep tie point:  "Extinction of S. universus occurred globally on this stage boundary (44), so that the estimates of 400,000 years for the age of this extinction (46) in the North Pacific and Antarctic constitute independent determinations for the age of the 12-11 boundary."¹¹  Orbital tuning came later.  The founding paper is the untuned one.  Milankovitch theory itself is argued in Orbital Cycles in Rock.

Huybers and Wunsch took the circularity worry seriously enough to build a chronology with no tuning:  "A new chronology of glaciation, spanning the last 780,000 years, is estimated from 21 marine sediment cores using depth as a proxy for time. ...  Depth-derived ages are estimated to be accurate to within ±9000 years, and within this uncertainty are consistent with the orbitally tuned age estimates."¹⁸  They refused even minimal tuning because it "tend[s] to align the abrupt glacial terminations with a particular phase of the assumed forcing".⁵  Their 2007 extension identifies 104 isotope events over two million years on the untuned chronology.  Two concessions belong here.  The depth-derived model uses six geomagnetic tie points whose ages "do contain limited orbital assumptions", as Huybers states plainly.¹⁹  Six tie points over two million years cannot generate 104 events.  Huybers answers the point himself in the next sentence, and the answer is worth printing:  the Brunhes-Matuyama reversal "is radiometrically dated to sufficient accuracy (±2 Ka) (Singer and Pringle, 1996) that it is essentially independent of orbital assumptions, and thus so is the agemodel between the present and 0.78 Ma."¹⁹  And the result is not a clean win for Milankovitch:  "The null hypothesis that glacial terminations are independent of obliquity is rejected at the 5% significance level.  In contrast, for eccentricity and precession, the corresponding null-hypotheses are not rejected."⁵  That dispute is between 41,000-year and 100,000-year pacing.  It is not between 100,000 years and 42.

The magnetic reversal

V28-238 carries the Brunhes-Matuyama reversal at 12.0 m.  Hebert's 2014 paper accepts the reversal as real and states ICR's position on when it happened:  "Both creation and uniformitarian scientists generally agree that earth's magnetic field has 'flipped' multiple times.  Creation scientists believe these magnetic reversals occurred rapidly during the Genesis Flood (Humphreys 1990), while uniformitarian scientists generally believe that these reversals occurred slowly over thousands of years."⁹  Humphreys' rapid-reversal model is treated in Magnetic Field Decay.  What matters here is the date.  Vardiman's Eq. 13 gives the age of any layer from its depth.  For the 12.0 m layer, with his H = 322 m and τ = 373 years, the reversal is 3,273 years before present.  With the corrected τ = 427, it is 3,095.  With the 760 m he uses for the Kennett sites, it is 2,729 to 2,953.  Every plausible choice puts a full reversal of the earth's field between about 1300 and 700 BC, 1,200 to 1,800 years after the Flood.  That is the Late Bronze Age and the Iron Age.  It is not the Flood year, and ICR publishes both claims.

The field in that window has been measured on dated material.  Vaknin, Shaar, Lipschits and 20 colleagues recovered its direction and intensity from fired mud brick in 20 burnt destruction layers across Israel, 10th to 6th centuries BC, several dated to the year by Assyrian and biblical records:  "During the period in question, the geomagnetic field in this region was extremely anomalous with rapid changes and high-intensity values, including spikes of more than twice the intensity of today's field."¹⁰  The direction changes they report in prose are small.  Beth-Shean, destroyed before about 880 BC, differs by "∼4°" of inclination from Gath, Rehov IV and Tevet V, destroyed about fifty years later;¹⁰ and near the end of the sequence, "There is a difference of ∼3° in inclination between the destruction of Batash and that of Lachish II", about 600 BC and 586 BC.¹⁰  A polarity reversal is a collapse of the dipole to a few per cent of normal followed by a 180° flip.  Nothing like it is in this record, and the record is dense enough that it could not be missed.  Each datum is a burnt floor whose date comes from a text or a radiocarbon series, not from magnetism.  No orbital tuning, ice core or deep-sea stratigraphy is anywhere in the chain.

The manganese nodules

Here the creationist side has real data.  Patrick's 2010 Journal of Creation paper quotes Glasby's 1978 survey of DSDP cores:  "The most striking feature of the data is the extreme paucity of nodules in the cores."⁶  Of 370 cores, only ten held nodules deeper than 200 m.⁶  Somayajulu's 2000 review confirms that nodules occur "mainly on the top (~ 50 cm) of the oceanic sedimentary column".²⁰  The mainstream explanation is that burrowing animals and bottom currents keep nodules at the surface, which the site's Bioturbation article takes up.  It is not a complete explanation, and the creationist literature is right to press the point.

But the two creationist versions of the argument are incompatible, and the measured growth rates fit neither.  Somayajulu's review of every dating method finds "growth rates in the range of < 1 mm to ~ 100 mm/m.y."²⁰  Hebert's 2015 ICR piece answers with lakes:  "manganese nodules have consistently been observed growing in lakes and man-made reservoirs, as well as on debris fragments from World Wars I and II, at rates hundreds of thousands of times faster than these calculated rates."²¹  Patrick, in the creationist literature itself, had already refused that move five years earlier:  "it just won't do to apply, say, those rates of >1 mm/yr directly to MNs growing at the ocean seafloor.  To do so would see MNs grow to diameters of greater than 3 kilometres in only a few million years!"⁶  A lake bottom pumps reduced manganese up from anoxic mud into oxic water;  the abyss is oxic, manganese-poor, and slow.  They are different chemical systems.  Patrick then draws the conclusion the nodule data actually support in his framework:  "These extremely conservative rates consign a blanket of sediment over the deep ocean floor of only tens of centimetres in a total of 5,000 years—the timeframe assumed since the end of the global Flood.  True, this is short of the few metres or so depth that characterize most buried MNs, but it's close."⁶  Nodules survive at the top because almost nothing has fallen since the Flood.  That is Vardiman's model upside down:  it makes the 322 m of Tertiary ooze beneath the nodules a Flood-year deposit, with 22 ordered isotope stages and a magnetic reversal inside one year.  Hebert 2014 cites Patrick and Vardiman in the same sentence as if they agreed.⁹

Part 5: The young-earth literature has three answers, and they disagree

Set the published young-earth figures side by side.  Patrick 2010:  tens of centimetres of post-Flood ooze in 5,000 years.  Vardiman 1996:  322 m in 4,500 years, starting at 0.75 to 3.5 m per year.  Snelling 2009, citing Tappan 1982:  "oceanic productivity 5 to 10 times greater than at present could be supported by the available sunlight if the nutrients (especially nitrogen), were available".¹²  Snelling's ceiling is the highest productivity multiplier the creationist literature itself defends, and Vardiman needs about ten thousand times it.  Patrick's column is about a thousand times thinner than Vardiman's.  These were published in the two main creationist technical journals and the standard creationist reference work within fourteen years of each other, and none cites the others' number.

Snelling's 1,100-page Earth's Catastrophic Past has a chapter titled "Chalk and Diatomite Beds, and Deep-Sea Sediments."  It is six pages, almost all of it about chalk on land, and it contains no estimate of deep-sea thickness, no accumulation model, no mention of Vardiman's curve, and no treatment of isotope stages or reversals in cores.  He cites Vardiman's monograph once, for post-Flood ocean cooling, not for sedimentation.¹²  He also limits his own productivity argument:  "while these calculations certainly show that the quantities of calcareous oozes on today's ocean floors are easily producible in the timespan since the Flood, they are insufficient to show how these chalk beds were produced during the Flood itself."¹²  Chalk is argued in Chalk and Limestone.

Nobody on the young-earth side has published a quantitative sea-floor sedimentation model since 1996.  Hebert's 2015 Acts & Facts article promised one:  "While I am not ready to publish my full results, I can say I am making some very interesting discoveries ..."²²  Eleven years on, the results have not appeared.

Objections and responses

"The accumulation rates are circular.  They come from orbital tuning, and the reversals and fossil datums used to check them are tuned too."  This is Hebert's 2014 argument:  "Thus we see that 'fossil events' and 'paleomagnetic reversals' are not genuine independent checks on the orbital tuning method.  Rather, both the evolutionary timescale and the astronomical theory are assumed to be true, and these assumptions are then used as criteria by which dating methods are judged to be 'correct' or 'incorrect.'"⁹  It is partly right about LR04's ages.  It does not touch the rate.  Thorium-230 normalisation contains no timescale;  its clock is two decay constants and one seawater concentration.  The radiocarbon date in RC11-120 contains no tuning.  Hays, Imbrie and Shackleton's chronology was linear in depth with a coral-dated anchor.  Huybers and Wunsch built a chronology from depth alone and matched the tuned one to ±9,000 years.  Four independent routes give a few centimetres per thousand years, and no creationist source found here engages thorium-230 normalisation as a way of measuring flux.  Vardiman, Oard and Hebert's 2014 and 2015 pieces do not mention thorium-230 at all;  Hebert 2017 mentions it only as a dating method, paired with protactinium-231.⁴

"A warm post-Flood ocean, with vigorous upwelling, had far higher productivity."  This is Roth's, Oard's and Snelling's answer, and the warm-ocean model is argued in Ice Age.  Three things stand against it here.  Roth himself says his figures are "optimum" cell-division potentials, and that measured fluxes are "several thousand times slower".  Snelling's own ceiling from the available sunlight is 5 to 10 times present.  And the carbonate budget is finite:  the model needs 40,000 times the modern gross carbonate export, and dissolution, which the model invokes, only raises what must be grown.  A warm ocean also dissolves less carbonate at depth only if it is supersaturated at 4,000 m, and a supersaturated deep ocean leaves no lysocline and no carbonate compensation depth.  The modern sea floor has both, mapped.

"The reversals happened in the Flood year, as Humphreys showed."  Then the 12 m of sediment above the Brunhes-Matuyama reversal in V28-238, and the 310 m of Vardiman's column below it, are not post-Flood, and his exponential-decay model did not make them.  The two ICR positions take opposite horns.  Either the reversal is Flood-year sediment, and Vardiman's post-Flood curve is wrong about what it re-plots, or the reversal is post-Flood on Vardiman's curve, and it happened while Iron Age Israel was burning and left no trace in twenty dated floors.  No creationist source has addressed this.

"Radiocarbon in post-Flood cores is inflated;  the biosphere's carbon-14 was low after the Flood, so the dates read too old."  Grant it for argument's sake.  The 9,400-year date at 37 cm would have to be wrong by a factor of about 10,000 to make the top 37 cm a single post-Flood year on Vardiman's curve.  And the radiocarbon date is one route of four.  Thorium-230 uses no carbon.  If every carbon-14 date in every core were discarded, the measured flux would not change.

"Hebert's recalculation shows the 1976 paper was wrong, and nobody in the mainstream has answered him."  The second half is true as far as this article could find.  Hebert reports:  "I have recently reperformed the Pacemaker frequency domain calculations, using the same method as the paper's authors, but taking into account this revision to the age of the B–M reversal boundary, as well as the inclusion of the previously excluded data from the second core.  These changes dramatically weaken, if not completely invalidate, the original argument for Milankovitch climate forcing presented in that paper."⁴  He adds that "even after multiple extensive internet searches, I have been unable to find a single, solitary candid acknowledgment in the secular literature of this serious problem with the Pacemaker results."²³  Three checks against the 1976 text change what the recalculation shows.  Hebert says the chronology came from "an assumed age of 700 ka for the Brunhes–Matuyama (B–M) magnetic reversal boundary as well as oxygen isotope (δ¹⁸O) data from the V28-238 western Pacific sediment core";²³ but Table 2 footnotes the 127,000-year anchor to Broecker and van Donk's coral dates, and only the deeper points of E49-18 to V28-238.  The 400,000-year S. universus datum is an independent check he does not mention.  And Hebert concedes that "the RC11-120 results are apparently not statistically distinguishable from the background noise",⁴ so his revision works through one core.  His revised age for the stage 6-5 boundary is 143,000 years.  Uranium-thorium dates on corals and cave deposits, which never touch a sediment core, would test that directly;  this article has not run that check and says so below.  Either way, the recalculation concerns whether three spectral peaks land on three orbital frequencies at ages near 700,000 years.  It says nothing about whether 12 m of ooze took 700,000 years or 1,200.

"The 100,000-year cycle is unexplained, so the whole orbital story is broken."  Huybers and Wunsch could not reject the eccentricity or precession nulls, and that is a live dispute.  Their own answer is that terminations skip one or two 41,000-year obliquity beats.  Oard relays Hebert's result as settled and calls LR04 a stack of "57 Milankovitch-tuned deep-sea cores",²⁴ which is right about its ages and wrong about its stage succession.  Say how far the dispute is from a Flood year.  The mainstream argument is whether the pacing is 41,000 or 100,000 years.  Vardiman's is 42.

"Measured local rates are inflated by sediment focusing."  Focusing is real and Costa's synthesis measures it.  It moves sediment sideways, both ways;  the thorium moves with it;  the global inventory does not change.

"This was settled by DSDP in 1970.  Why engage a 1996 ICR monograph at all?"  Because it is the only quantitative young-earth model of the pelagic column, and ICR still cites it approvingly.⁹  DSDP Leg 3 showed that sediment age at the basement tracks distance from the ridge, which is a spreading argument and belongs to Catastrophic Plate Tectonics.  The rate question is different, and the decisive answer to it, an accumulation rate with no age model in it, came from thorium-230 work between 2004 and 2020.  The 1970 answer was right;  the reason it was right is worth stating.

Issues and unanswered questions

• Vardiman's monograph has not been read. Only Oard's review and the CRSQ condensation were obtained. Oard says the monograph's five appendices of thickness and distribution data are its substance. If they contain a different treatment of the equation, or a different H, the arithmetic in Part 3 should be re-run. Buying the print copy would settle it.
• Milliman 1993 and Shackleton and Opdyke 1973 were not obtained in full. Milliman's budget is cited through Cartapanis 2018, which quotes it verbatim. Shackleton and Opdyke's data tables were obtained from PANGAEA, and their abstract from OpenAlex; no body prose from the paper is quoted here. The PANGAEA comment prints the reversal depth as "1.20 m", which is a transcription slip: their 1.71 cm/kyr rate times 700,000 years is 12.0 m, and their table's base at 13.40 m is dated 782,000 years.
• The ocean's calcium and carbon inventories used in Part 3 are textbook values and have not been checked against Broecker and Peng, Sarmiento and Gruber, or GLODAP. The conclusion does not depend on them; the carbonate export figures from Cartapanis do the work.
• Vardiman's model makes one falsifiable prediction nobody has tested: accumulation rate should fall exponentially with time, so the bottom of a 322 m column should have accumulated about 10⁵ times faster than the top. Costa's supporting table holds thorium-normalised fluxes for over 1,000 records. Plotting flux against depth in any one abyssal core would settle it. It has not been done here.
• Hebert 2017 has no direct mainstream reply that this article could find. The answer given here is indirect. A uranium-thorium age for Termination II, from corals or from Devils Hole, would test his 143,000-year revision by a method that never touches a sediment core, and has not been fetched.
• The light budget has not been computed. Whether the sun delivers enough energy to the ocean surface to fix carbon at ten thousand times the modern rate is a one-page calculation, and it may be the cleanest test of all. Snelling's "5 to 10 times" from Tappan is the only creationist figure, and it was not checked against Tappan.
• Diesing's lithology map is about 59 % accurate overall. The calcareous and clay areas are the robust part; the radiolarian figure is not.
• Ash layers and burrows in deep-sea cores were not worked up. Neither side's literature on this topic discusses them.

Verdict

The deep ocean floor carries about 400 m of sediment, about sixty per cent of it by area the shells of plankton.  It accumulates at a few centimetres per thousand years by four independent measurements, one of which, thorium-230 normalisation, contains no stratigraphy, no tuning and no assumed age.  The only quantitative young-earth model, Vardiman's, needs 38,000 to 170,000 times that rate in its first year, by its own arithmetic, which also does not close.  It requires the ocean to produce carbonate at 40,000 times its entire modern export.  It puts a full reversal of the earth's magnetic field between 1300 and 700 BC, when twenty dated burnt floors in Israel record a strong, normal, slowly drifting field.  It squeezes sixteen global swings of ice volume into 674 years.  And it contradicts the other two published young-earth figures by three to four orders of magnitude, which nobody on that side has reconciled or tried to.

The young-earth literature is right about three things.  The ages on the isotope stack are tuned.  The 100,000-year cycle is unexplained.  Manganese nodules are scarce below the top metre.  None of these bears on the rate, and the rate is what decides.  Vardiman asked in 1996 whether the oceans could supply the nutrients.  The measured answer is no, by four or five powers of ten.

Sources