The Age of the Solar System

"Extrapolation of the current rate of lunar recession of ∼3.8 cm/y back in time places the Moon at a close position to Earth as recently as 1.5 billion years ago." — Margriet Lantink and colleagues, PNAS, 2022¹

"Thus, until new evidence is forthcoming, creationists should not continue to use the dust on the moon as evidence against an old age for the moon and the solar system." — Andrew Snelling and David Rush, Journal of Creation, 1993²

"The best estimates from observational surveys indicate that this source may fall short by more than two orders of magnitude the estimates obtained from theoretical models." — Kathryn Volk and Renu Malhotra on the supply of short-period comets, 2008³

Summary

The standard age of the solar system is 4,567.30 ± 0.16 million years.⁴ That number comes from lead-isotope dating of the oldest solid grains in meteorites, and it is confirmed to within its own error bar by a completely separate method that never touches a rock: the Sun's internal sound waves.⁵

Young-earth writers answer with a list of things in the solar system that look too young to be that old.  Comets burn away.  The Moon is moving out.  Saturn's rings are eroding.  Io is too hot, Enceladus is too active, Titan has methane that sunlight should have destroyed.  Planetary magnetic fields are decaying.  The list is old, it is repeated on every creationist website, and almost none of it has been audited by anyone friendly to it.

This article audits it.  The verdict is mostly against.

The observations are usually right.  Short-period comets really do burn out in thousands of years — the number in the mainstream literature is 4,000 to 40,000 years.⁶ Saturn's rings really may be younger than the dinosaurs.⁷ Titan really would lose its methane in about ten million years without resupply.⁸ A creationist saying these things is repeating what specialists say.

The inference is usually broken, and it breaks the same way every time.  Each argument measures how fast something is destroyed, assumes nothing replaces it, and runs the clock backwards.  That is the same structural error found in the natural arches and erosion rate arguments, and it fails for the same reason: a destruction rate constrains a loss term, not an age.

Three of the arguments are now dead by direct measurement, not by argument.

1. The Moon's recession was killed by reading the Earth-Moon distance straight out of dated rock. Sedimentary cycles put the Moon at 340,900 ± 2,600 km 1.4 billion years ago and 321,800 ± 6,500 km 2.46 billion years ago.^{9 1} The measured recession rate in the deep past was slower than today, not faster — the exact opposite of what the young-earth calculation assumes.
2. Io's heat was killed when the tidal energy input was measured directly from a century of moon positions, and came out matching the heat output.¹⁰
3. Enceladus's heat was killed the same way, when Saturn's tidal dissipation was measured astrometrically and turned out about ten times stronger than theory had assumed.¹¹

Two arguments survive as real open problems, and both are stated most forcefully in the mainstream literature rather than the creationist literature: the source population for short-period comets falls short of what the models need,³ and Titan's methane requires an episodic resupply nobody has observed.⁸ Neither gets anyone to 6,000 years.  Both are worth watching.

One prediction genuinely landed.  In 1984 Russell Humphreys published, in advance, that Uranus and Neptune would be found to have magnetic dipole moments "of the order of 10²⁴ J/T."¹² Voyager 2 measured about 3 × 10²⁴ and about 2 × 10²⁴.^{13 14} That is a dated, published, correct prediction, and this article says so plainly.  It also shows why the theory behind it cannot be tested by that success.

The movement has already retracted two items on this list itself.  Moon dust was abandoned by creationist researchers in 1993, in print, at length.² The solar neutrino argument was abandoned in 2002.¹⁵ Both retractions are creditable and both are still quoted at us by people who have not read them.

The scorecard

What this article does

It states each argument in the strongest form its own advocates give it, traces every number to a primary source, and does the arithmetic in the open.  Where a figure is computed here rather than quoted, the computation is shown so it can be checked.

It does not cover how the solar system formed.  The nebular hypothesis has real unsolved problems — the metre-size barrier, giant-planet migration, the Sun's angular momentum — and creationist writers press them hard.  Those are arguments about mechanism, not about age, and a mechanism can be unknown while a clock still reads what it reads.  They belong in a separate article.

It also does not cover radiometric dating on Earth, distant starlight, or the Genesis text.  Links to companion pages are at the bottom.

Every source it rests on is catalogued, with what each one measured and which way it cuts, in a companion page: the literature.

Part 1 — What the 4.567-billion-year number actually is

Almost every young-solar-system argument is offered as a rebuttal.  It is worth knowing what is being rebutted, because the target is not what most readers assume.

Nobody dates the Earth directly

There is no rock on Earth as old as the Earth.  Plate tectonics, erosion and impacts recycle the crust, and the oldest surviving mineral grains — zircons from Jack Hills, Australia — are about 4.4 billion years old, younger than the planet.  So the age of the solar system is measured on meteorites, which are pieces of bodies that never got hot enough or big enough to erase their own history.

Clair Patterson did this first.  In 1956 he measured lead isotopes in three stony and two iron meteorites, using the troilite from the Canyon Diablo iron meteorite to fix the isotopic composition of primordial lead, and got 4.55 ± 0.07 × 10⁹ years.¹⁶ That number has moved by less than two per cent in seventy years, and the error bar has shrunk by a factor of three.

The modern number, and what it is the age of

The best-constrained value comes from lead-lead dating of calcium-aluminium-rich inclusions — CAIs, the white refractory blobs in carbonaceous chondrites, which are the first solids known to have condensed in the disk.  James Connelly and colleagues dated four of them individually in two laboratories:

"CAIs define a brief formation interval corresponding to an age of 4567.30 ± 0.16 million years (My), whereas chondrule ages range from 4567.32 ± 0.42 to 4564.71 ± 0.30 My."⁴

An error of 0.16 million years on 4,567 million is a precision of 35 parts per million.  What is being dated is the condensation of the first solids, so "the age of the solar system" strictly means "the age of the oldest rock we can find from it."

The part that is hard to argue with is the concordance

A single date can be wrong.  The reason planetary scientists treat 4.567 Ga as settled is that mutually independent clocks agree.

The long-lived systems have wildly different half-lives — ²³⁸U 4.47 billion years, ²³⁵U 0.70, ⁸⁷Rb 49, ¹⁴⁷Sm 106, ¹⁸⁷Re 41.6.  The short-lived ones — ²⁶Al (0.72 Myr), ⁶⁰Fe (2.6 Myr), ¹⁸²Hf (8.9 Myr), ¹²⁹I (15.7 Myr) — are extinct in nature, which is itself a constraint: they are gone, and the relative order they impose on meteorite formation matches the order the long-lived systems give.  Connelly's lead-lead ages agreed with the aluminium-magnesium relative chronology on the same objects, which is why the paper could resolve a three-million-year sequence.

This matters for the young-earth reply, and the reply is discussed in Part 5.

The Sun, dated without a rock

The strongest single check is that the meteorites are not needed at all.

Helioseismology reads the Sun's interior from the frequencies of its acoustic oscillations.  Because the Sun's core composition changes as hydrogen fuses, the frequency pattern is a clock.  Alfio Bonanno, Helmut Schlattl and Lucio Paternò fitted solar p-mode frequencies and got a seismic solar age of 4.57 ± 0.11 billion years — a number that involves no meteorite, no decay constant, and no assumption about initial lead.⁵

Two methods with nothing physical in common agree to within one part in forty.  Any young-solar-system model has to explain that coincidence, and none of the arguments below attempts it.

Part 2 — The shape of every young-age argument

Almost all of them have this form:

X is being destroyed at rate R.  There is a finite amount of X.  Divide, and X cannot be older than T.  T is much less than 4.5 billion years.  Therefore the solar system is young.

Write the population as a balance:

dN/dt = S − μN

Measuring the destruction gives μ.  It says nothing whatever about S, the source term.  Setting S = 0 is a choice, not a measurement, and every one of these arguments makes that choice silently.

The honest version of the conclusion is therefore not "the solar system is young."  It is: either the solar system is young, or something replaces X.  That is a genuine and useful result — it forces the mainstream to name a source and then go look for it — and in several of the cases below the source has since been found and photographed.  In one case it has not.

There is a second, subtler problem.  Several of these clocks return answers in the range of 10⁷ to 10⁸ years.  A young-earth reader treats that as a win, because 100 million is much less than 4,500 million.  But 100 million years is also 17,000 times longer than 6,000 years.  An argument that establishes "Saturn's rings are 100 million years old" has not moved one per cent of the distance to the biblical timescale; it has simply shown that a subsystem is younger than the whole.  Rings, moons and comets are objects that form and are destroyed inside an old solar system.  Nobody in planetary science thinks otherwise.

Part 3 — The arguments

Comets and the Oort cloud

This is the flagship, and it is the strongest one.  It is item 2 of Russell Humphreys' Evidence for a Young World:

"Yet each time a comet orbits close to the sun, it loses so much of its material that it could not survive much longer than about 100,000 years.  Many comets have typical ages of 10,000 years."¹⁷

The fullest technical version is Danny Faulkner's 1997 paper in the Journal of Creation.¹⁸ Its structure: comets are consumed; the Oort cloud has never been seen; the Kuiper belt does not resupply fast enough; therefore comets are young, therefore the solar system is young.

The observation is correct and uncontroversial.  The 2022 Comets III review chapter puts the physical lifetime of Jupiter-family comets, fitted to their observed inclination distribution, at

"≈ 4000–40,000 years (≈ 500-5000 orbits), which is a small fraction of the dynamical lifetimes."⁶

So a comet dies from wear long before its orbit throws it out.  The creationist number is the mainstream number.

Two of the three sub-claims have been overtaken by discovery.  Faulkner wrote in 1997, five years after the Kuiper belt's first object was found.  Since then:

• The Minor Planet Center's lists carried 4,089 trans-Neptunian objects and 2,238 Centaurs and scattered-disk objects when checked on 9 August 2026.¹⁹ The Centaurs are the intermediate stage — objects caught in the act of migrating inward from beyond Neptune toward Jupiter-family orbits. The resupply route is not inferred, it is populated with catalogued objects.
• The source of short-period comets was reidentified in 1997 as the scattered disk, not the classical Kuiper belt.²⁰ Modern simulations that track the whole journey reproduce the observed orbital distribution of comets well.²¹

Faulkner's argument that the Kuiper belt objects seen so far are far larger than comet nuclei was correct in 1997 and is no longer decisive: Bernstein and colleagues pushed Hubble three magnitudes fainter and found objects down to about 25 km diameter.²²

One sub-claim about the Oort cloud is still true, and less damaging than it sounds.  Nobody has imaged an Oort cloud object in place.  Carl Sagan's line, which Faulkner quotes, still stands: there is no direct observational evidence for the cloud.  But the cloud was not proposed from imagination.  Jan Oort inferred it in 1950 from a measured distribution: the reciprocal semi-major axes of long-period comets pile up in a narrow spike at 1/a < 10⁻⁴ au⁻¹, meaning those comets are arriving on their first pass from something bound at tens of thousands of astronomical units.  The spike is the evidence, and it is data, not theory.  Calling the Oort cloud unobserved is like calling a planet unobserved because you have only seen the wobble it puts on its star.

The part of the argument that is genuinely open is the arithmetic, and it belongs to the mainstream.  Kathryn Volk and Renu Malhotra opened their 2008 paper this way:

"However, the best estimates from observational surveys indicate that this source may fall short by more than two orders of magnitude the estimates obtained from theoretical models of the dynamical evolution of Kuiper belt objects into JFCs."³

They considered and could not cleanly resolve it, floating a change in the size distribution below the detection limit, tidal break-up of larger bodies, a temporary excess of comets above the long-term average, or additional source populations.  By 2017 the gap had narrowed a great deal — Nesvorný and colleagues' end-to-end simulation reproduces the orbital distribution and comes within a factor of a few on the numbers.²¹ It has not closed.

Verdict:  the observation holds, the inference fails, and a real problem remains.  The comet supply is a live quantitative question in planetary dynamics, and creationists were pointing at it before most apologists had heard of the scattered disk.  But "we cannot yet count the reservoir accurately" is a long way from "there is no reservoir," and the reservoir now has thousands of catalogued members and a photographed transit route.

A related argument has already collapsed the same way.  Interplanetary dust spirals into the Sun under the Poynting-Robertson effect, so the zodiacal cloud has a short lifetime and was once offered as a young-solar-system clock.  Its source term has since been identified and measured: 85 to 95 per cent of the zodiacal cloud's mid-infrared emission comes from particles shed by Jupiter-family comets.²³ The dust is comet debris, and it is replaced as fast as it is lost.

One more thing should be said, because it is rarely said.  Comet arguments are self-limiting.  Even granting everything, a comet argument dates comets.  Rosetta measured 67P/Churyumov-Gerasimenko losing about 10.5 × 10⁹ kg per orbit — roughly 0.1 per cent of its mass.²⁴ On a 6.45-year orbit, that is a lifetime of a few thousand years, right inside the 4,000-to-40,000-year range the literature already gives.  The measurement settles that comets are short-lived objects.  It says nothing whatever about the age of Jupiter.

The Moon's recession

The clearest version is by Don DeYoung and Jason Lisle for Answers in Genesis:

"Currently, the moon is moving outward from the earth by 3.82 cm/yr (1.5 in/yr).  However, this recession is highly nonlinear and would have been greater in the past.  If one assumes unlimited extrapolation back in time, gravity theory shows the moon in direct physical contact with earth about 1.55 billion years ago."²⁵

The physics is right and the arithmetic is right.  Lunar laser ranging measures the Moon's semi-major axis increasing at about 3.83 cm per year.²⁶ The tidal torque scales as r⁻⁶, so dr/dt ∝ r^−11/2, so r^13/2 grows linearly with time and the time to move out from zero is

t = (2/13) × a / (da/dt)

Putting in a = 384,400 km and 3.83 cm/yr gives 1.54 billion years.  DeYoung and Lisle's 1.55 Ga is correct to their inputs.  Anybody can check it.

Everyone, including the mainstream, agrees this is a real constraint.  Here is the opening of a 2022 PNAS paper, written by people with no interest in Genesis:

"Extrapolation of the current rate of lunar recession of ∼3.8 cm/y back in time places the Moon at a close position to Earth as recently as 1.5 billion years ago (Ga).  This timing is in direct conflict with the formation age of the Moon of around 4.5 Ga, based on radioisotopic dating of lunar rocks, implying that, over the course of Earth's history, the average tidal dissipation rate must have been much lower than present day."¹

That is the young-earth argument, stated by its opponents, in a mainstream journal, in 2022.  The disagreement is entirely about the last clause.

The mainstream answer used to look like special pleading.  The standard reply since Kirk Hansen's 1982 modelling is that today's tidal dissipation is anomalously high because the present ocean basins happen to be near resonance with the semidiurnal tide, and that for most of Earth's history the coupling was weaker.  A skeptic is entitled to notice that this is a model built to produce the required answer, with continental configurations as free parameters.  Hugh Ross cites Hansen to close the case; that is too quick.

The argument was settled by measurement, and the measurement is the interesting part.  Milankovitch cycles — the wobbles in Earth's orbit and spin axis — leave a rhythm in sediment.  The ratio of the precession period to the eccentricity period depends on how fast Earth spins and how far away the Moon is.  So a well-preserved rhythmic sediment sequence is a ruler for the Earth-Moon distance at the time it was laid down.

Two such rulers have now been read:

Both are bracketed by high-precision uranium-lead dates on the same rocks.^{9 1}

Now compute the average recession rate over each interval, which is arithmetic anyone can do:

• 2.46 Ga → 1.4 Ga: (340,900 − 321,800) km over 1.06 billion years = 1.80 cm/yr
• 1.4 Ga → today: (384,400 − 340,900) km over 1.4 billion years = 3.11 cm/yr
• today: 3.83 cm/yr

The recession was slower in the past, and it has been speeding up.  That is the reverse of the assumption the young-earth calculation is built on.

Run the young-earth model forward against the same data and it fails outright.  Under dr/dt ∝ r^−11/2 anchored on today's rate, the Moon reaches 340,900 km only 0.84 billion years ago — but the rock says 1.4.  It reaches 321,800 km only 1.06 billion years ago — but the rock says 2.46.  The model runs the clock more than twice too fast at 2.46 Ga, and the error grows with age, exactly as a resonance-limited dissipation history predicts.

This is worth being precise about, because it is the single cleanest refutation in the whole subject.  The mainstream did not answer the Moon argument with a model.  It answered with two measurements of where the Moon actually was, taken from dated rock, using a method that does not care what anyone believes about tides.  Meyers and Malinverno say so:

"The TimeOptMCMC reconstructed Earth-Moon distance from the Xiamaling Formation (340,900 ± 2,600 km, 2σ) is consistent with that derived from ocean models that imply smaller torques, reduced tidal dissipation, and slower lunar retreat rates in the distant past."⁹

Verdict:  fails.  Not on interpretation — on data.  This argument should be retired, and the people repeating it should be told why.

One honest caveat in its favour.  Two data points spanning 2.46 to 1.4 Ga do not fix the whole history, and the interval before 2.46 Ga is still unconstrained by cyclostratigraphy.  A young-earth advocate can point out that the deep Archaean remains model-dependent.  That is true, and it does not help, because the two points that exist already break the calculation.

Saturn's rings

Here the young side wins the sub-argument and it does not matter.

Cassini's Grand Finale measured the ring mass from the spacecraft's own acceleration: (1.54 ± 0.49) × 10¹⁹ kg, about 40 per cent of the mass of the moon Mimas.²⁷ Ring material is falling into Saturn at 432 to 2,870 kg per second, which empties the rings in about 292 million years.²⁸ And micrometeoroid dust should darken clean ice on a measurable schedule; Sascha Kempf and colleagues measured the infall and concluded the rings are no more than a few hundred million years old.⁷

So the mainstream position really is "Saturn's rings are younger than the dinosaurs," and creationists have been quoting NASA saying so since 2002.  They are quoting it accurately.

And the argument still does not reach.  292 million years is 49,000 times 6,000 years.  To use the ring clock for a young solar system you would have to argue the loss rate was once tens of thousands of times faster, which is the mirror image of the assumption the same writers reject when a mainstream scientist makes it.

The honest complication is that the young-ring result is now contested.  In December 2024 Ryuki Hyodo, Hidenori Genda and Gustavo Madeira showed in Nature Geoscience that micrometeoroid impactors are vaporised on contact with ring particles, forming charged nanoparticles and ions that are then swept out of the ring system rather than staying to darken it.²⁹ If that is right, clean rings are not young rings, and the pollution clock reads too fast.  Separately, Aurélien Crida and colleagues had argued in 2019 that ring mass does not date rings either, because massive rings spread faster and any initial mass converges toward the observed one after about four billion years.³⁰

Verdict:  the observation is real, the mainstream concedes far more than a skeptic expects, and the inference is off by four to five orders of magnitude.  It is also a good example of why quoting a mainstream surprise is risky: the surprise may not survive the next paper.

Io, Enceladus, and the heat arguments

These are the "too hot to be old" family.  Wayne Spencer's Creation article for CMI states the Io version:

"However, the amount of heat that Io gives off, that scientists have observed with infrared telescopes and spacecraft, is about 10 times the amount of heat generated by all known sources of heat, including tidal dissipation."³¹

The structure is: the moon radiates more heat than any long-lived source can supply, therefore it is running down primordial heat, therefore it is young.

Io was answered by a direct measurement of the energy input.  Tidal heating is not a free parameter.  If Jupiter is dissipating energy into Io, the whole Galilean system's orbits must be changing at a rate that a long enough baseline of telescope positions can detect.  Valéry Lainey and colleagues used astrometry from 1891 to 2007 to measure exactly that, and reported in Nature that the tidal dissipation is strong enough to supply the observed surface heat flow, with Io close to thermal equilibrium.¹⁰

That is not a model fit.  It is the input side of the energy budget, measured independently of the output side, and the two balance.

Enceladus was answered the same way, and the answer was a genuine surprise.  Enceladus's south polar terrain radiates 15.8 ± 3.1 GW,³² against roughly 0.3 GW of radiogenic heating.  The steady-state tidal calculations of the 2000s gave about a gigawatt, an order of magnitude short — which is precisely what the creationist articles quote.  Then Lainey and colleagues measured Saturn's tidal dissipation from a century of satellite astrometry and found Q = 1,682 ± 540, about ten times stronger than theory expected, "as a consequence, eccentricity equilibrium for Enceladus can now account for the huge heat emitted from Enceladus' south pole."¹¹

It got stronger still.  In 2020 the same group measured Titan migrating outward from Saturn at 11.3 ± 2.0 cm per year — a hundred times faster than expected — which points to resonance locking, a mechanism in which tidal heating stays roughly constant over time instead of declining.³³

Note what happened there, because it cuts both ways.  The creationist writers were right that the textbook tidal budget did not close.  The reason it did not close was that the textbook value of Saturn's dissipation was wrong by an order of magnitude, and the correction runs in the direction that keeps moons hot for billions of years rather than the direction that makes them young.

Something real does survive.  Matija Ćuk, Luke Dones and David Nesvorný argued in 2016 that Saturn's mid-sized moons cannot be primordial and probably re-accreted from a disk about 100 million years ago.³⁴ Mimas appears to have gained an ocean recently.  Io's heat output may have been badly underestimated: a 2026 reanalysis of Juno infrared data suggests the volcanic paterae radiate far more than the long-standing figures, and Juno's microwave radiometer has found tidal heating warming the subsurface.  So the outer solar system is more active and more youthful-looking than the 1990s picture allowed.  None of this is a 6,000-year result; it is a statement that moons and rings turn over.

Verdict:  fails, and the failure is instructive.  This family of arguments was defeated by measurements of the energy source that did not exist when the arguments were written.

Titan's methane

Titan's atmosphere is about five per cent methane.  Ultraviolet light destroys methane and it does not come back on its own.  Robert Lorenz and colleagues, after Cassini's radar mapped the polar lakes:

"Put another way, there is an order of magnitude less liquid in the lakes than there is methane in the atmosphere, and photochemical models predict that inventory to be depleted in ~10 Myr."⁸

Ten million years, on a four-and-a-half-billion-year moon.  This is a real anomaly and creationists are right to press it.

The mainstream answer is episodic outgassing, and it is not observed.  Gabriel Tobie, Jonathan Lunine and Christophe Sotin proposed in Nature in 2006 that methane is stored as clathrate hydrate in Titan's interior and released in episodes driven by the moon's thermal evolution.³⁵ There is one independent piece of support: the Huygens probe's mass spectrometer detected radiogenic argon-40 in Titan's atmosphere, which can only have come from potassium-40 decay in the rocky interior, so the interior does vent to the surface.³⁶ But no cryovolcanic methane eruption has been observed, and Titan's surface is short of the ethane that a long history of methane destruction should have left behind.

Verdict:  partly survives.  The observation is solid, the timescale is genuinely short, and the proposed source is a hypothesis rather than an observation.  What it does not do is date Titan: an episodic source that fires every few tens of millions of years leaves no trace of how many times it has fired.  A young-earth reader gets "Titan's atmosphere is not in steady state," which is what the papers say, and no further.

Note the parallel with the argon itself.  The best evidence that Titan's interior is connected to its atmosphere is a decay product — argon-40 from potassium-40, with a 1.25-billion-year half-life.

Planetary magnetic fields: scoring Humphreys

This deserves detailed treatment, because it is the one place in solar-system creationism where a published, dated, numerical prediction was made in advance and came out roughly right.

The theory.  Russell Humphreys proposed in 1984 that God created the planets as spheres of water with the hydrogen nuclear spins aligned, producing an initial magnetic moment, which then decayed as ordinary resistive current decay in the body's core.¹² The initial moment is

M_c = 0.9425 · k · m

where m is the body's mass in kilograms and k is the fraction of spins aligned, between 0 and 1.  Humphreys uses k = 0.25 as a default.  The present moment follows from exponential decay over 6,000 years with a decay time set by the body's core size and conductivity.

The prediction, verbatim, from the 1984 conclusions:

"The upcoming Voyager 2 encounters with Uranus and Neptune should show planetary magnetic moments less than the k = 1.0 limit: 8.2 x 10²⁵ J/T for Uranus and 9.7 x 10²⁵ J/T for Neptune."¹²

And, in the body of the paper, an estimate rather than a bound:

"In that case, according to equation (3), (4), and (30), the present dipole moment of Uranus would be of the order of 10²⁴ J/T."¹²

with the same order of magnitude given for Neptune.

The result.  Voyager 2 reached Uranus in January 1986 and Neptune in August 1989.  Converting the published dipole moments to SI gives about 3.8 × 10²⁴ J/T for Uranus and about 2.0 × 10²⁴ J/T for Neptune.^{13 14} Humphreys' own reporting gives 3.0 × 10²⁴ and 1.5 × 10²⁴, slightly lower in both cases.³⁷

Score it honestly.

In his favour.  The estimate was published thirteen months before the Uranus encounter and five years before Neptune's, in a numbered journal, with the numbers in the text.  It was right to within a factor of four, twice.  Nobody should pretend this did not happen, and Christians who cite it are citing something real.

Against.  Four things.

1. The formal prediction was a bound two orders of magnitude above the answer. A prediction of "less than 8.2 × 10²⁵" is satisfied by 3 × 10²⁴, and also by 10²⁰, and also by zero. The number that actually landed — "of the order of 10²⁴" — was an estimate resting on an assumed core conductivity, and it is presented in the paper as such.
2. The scaling does the work, and the scaling is nearly theory-free. M_c ∝ mass. Any hypothesis in which a planet's field tracks its bulk gives the same answer for a body of Uranus's mass, and the mainstream had its own such scaling laws — Russell's, and Desch and Kaiser's radiometric Bode's law, published for Uranus in Nature in the very same year, 1984.³⁸ Humphreys' claim that "many evolutionists had predicted that Uranus would have a much smaller field, or none at all"³⁷ appears without a citation to any such published prediction.
3. The theory has one free parameter per body and therefore cannot fail. The decay time is fitted after the fact, separately, for every object. Working from the numbers in his own paper, the fitted decay times are under 220 years for Mars, under 364 for the Moon, 813 for Mercury, 5,300 for Saturn, 19,000 for the Sun, and more than 41,000 for Jupiter — a spread of more than two orders of magnitude, each value chosen to reconcile a fixed creation moment with a measured present one. (The Mars and Moon figures are computed here from his tabulated values; the rest are his.) Expressed as present moment per kilogram, the solar system's bodies span twelve orders of magnitude, from Jupiter at 0.8 J/T per kg down to Mars below 3 × 10⁻¹³. A model that accommodates a twelve-order-of-magnitude spread by adjusting one parameter per object is not being tested by any single agreement.
4. A prediction it did make sharply has not held up. See Mercury, next.

Verdict:  a real hit, on a theory that the hit cannot confirm.  The right thing for a creationist writer to say is "Humphreys got the order of magnitude right in advance, and here is why that does not settle the question."  The right thing for a skeptic to say is "yes, he did publish it first."

The evidence that actually bears against the theory is paleomagnetic, and it is decisive.  Humphreys' 1984 paper treats the magnetisation of Apollo lunar samples as snapshots of a field decaying within the last 6,000 years, quoting sample 62235 at 6.3 × 10²¹ J/T and sample 15498 at 1.1 × 10²⁰ J/T.¹² But those samples are dated, and modern work reads a coherent history out of them: a lunar core dynamo running from about 4.2 to 3.56 billion years ago with mean surface palaeointensities near 78 µT, followed by a decline by a factor of roughly 18 to 27 between 3.56 and 3.19 Ga.³⁹ MESSENGER found the same kind of record on Mercury, magnetised crust implying a dynamo operating at least 3.7 to 3.9 billion years ago.⁴⁰ Under Humphreys' model, that entire rise-and-fall has to be compressed into a few thousand years and the radiometric dates on the samples have to be wrong by a factor of about 600,000.

Mercury's "fading" field

This is the sharpest test Humphreys' theory offers, and it is worth walking through because the reporting on both sides is loose.

His 1984 prediction was specific and dated:

"Mercury's decay rate is so rapid that some future probe could detect it fairly soon.  In 1990 the planet's magnetic moment should be 1.8 percent smaller than its 1975 value."¹²

Later restated as a 4 to 6 per cent decrease by 2010.⁴¹ In 2012, after MESSENGER entered orbit, he published his result:

"My result: Mercury's dipole is 7.8 (± 0.8)% weaker in 2011 than it was in 1975."⁴¹

Three things need saying.

First, 7.8 per cent is outside his own predicted 4 to 6 per cent.  Humphreys says so in the article — the field came down "possibly by more than my prediction."  A test that overshoots its own interval by more than two standard deviations is not a confirmation.

Second, the comparison is his own reanalysis, not a published measurement of change.  The MESSENGER team did not report a decline.  Humphreys re-fitted the 1975 Mariner 10 data assuming the near-zero tilt and northward offset that MESSENGER established in 2011, and compared that to the MESSENGER value.  He explains why he did this, and the reasoning is not unreasonable — it is an apples-to-apples correction.  But it is a reprocessing of forty-year-old flyby data by a party with a stake in the answer, and it has not been replicated.

Third, the 1975 baseline will not support a 7.8 per cent inference.  Mariner 10 made three brief flybys and could not separate the dipole from higher-order terms.  Norman Ness's own review of the Mariner magnetic data lists fifteen different published models with centred dipole moments ranging from 136 to 350 nT·R_M³ — a spread of a factor of 2.6.⁴² MESSENGER's orbital coverage finally resolved the dipole-quadrupole trade-off that had "plagued earlier Mariner 10 flyby analysis."  Reading a 7.8 per cent secular change across a baseline whose model-dependent spread is 160 per cent is not possible.  Independent work combining both datasets to look for secular variation exists,⁴³ and no established decline has come out of it.

Verdict:  not established.  The prediction was admirable in form — numerical, dated, falsifiable — and that is more than most apologetics offers.  The data cannot yet test it, and the one attempt to test it missed its own target.

The faint young Sun

Wayne Spencer lists this as a heat problem for long ages:

"At that time in the Sun's life cycle it would have given off about 30% less light energy than today, and the Earth would have been much colder.  Earth would likely have looked like a ball of ice."³¹

The problem is real and it is not solved.  Carl Sagan and George Mullen raised it in 1972.  Stellar structure requires the young Sun to have been about 70 per cent as luminous, yet the geological record shows liquid water and sediments throughout the Archaean.  The standard answer is a stronger greenhouse — more carbon dioxide, some methane — and geochemical proxies keep failing to supply enough carbon dioxide to close the gap.  Whether the Proterozoic can be kept ice-free at all remains open.

But it is not an age argument.  It is an argument about early climate, and it presupposes the long timescale it is offered against: without 3.8 billion years of stellar evolution there is no faint young Sun to have a paradox about.  Cited as a young-earth evidence, it is a category error.  Cited as "here is an unsolved problem in the standard picture," it is fair and accurate.

The same applies to the composition results from the Genesis solar-wind sample mission — the Sun's oxygen and nitrogen isotopes differ from Earth's — which Spencer also lists.  That is a constraint on how the planets formed, not a clock.

Part 4 — Arguments the movement has already retracted

Two items still circulating were retired by creationist researchers themselves.  Both retractions are more careful than the arguments they replaced, and both deserve to be better known.

Moon dust

The claim was that meteoritic dust should have piled metres deep on the Moon over billions of years, and did not.  Andrew Snelling and David Rush spent more than seven years on it and published the audit in 1993.  Their conclusion:

"It thus appears that the amount of meteoritic dust and meteorite debris in the lunar regolith and surface dust layer, even taking into account the postulated early intense meteorite and meteoritic dust bombardment, does not contradict the evolutionists' multi-billion year timescale (while not proving it).  Unfortunately, attempted counter-responses by creationists have so far failed because of spurious arguments or faulty calculations.  Thus, until new evidence is forthcoming, creationists should not continue to use the dust on the moon as evidence against an old age for the moon and the solar system."²

And their acknowledgements, which are the most honest sentence in the literature reviewed for this article:

"We, of course, take responsibility for the conclusions, which unfortunately are not as encouraging or complimentary for us young earth creationists as we would have liked."²

Creation Ministries International carries the retraction on its standing list of arguments creationists should not use, along with a 2014 editorial note that new NASA data showing a tenfold higher influx "seems to raise the possibility of at least a partial revival," while cautioning that any reassessment "would need to be similarly thorough and careful."⁴⁴

Missing solar neutrinos

The claim was that the Sun's neutrino deficit showed it was powered partly by gravitational collapse rather than fusion, limiting its age.  CMI's list:

"However, a new experiment was able to detect the 'missing' flavours, which seems to provide conclusive evidence for oscillation.  This means that neutrinos must have a very tiny rest mass after all—experimental data must take precedence over theory.  Therefore creationists should no longer invoke the missing neutrino problem to deny that fusion is the primary source of energy for the sun.  So it cannot be used as a young-age indicator—nor an old-age indicator for that matter."¹⁵

That is the correct response to a falsified prediction, and it was written in the same decade the falsification landed.

Part 5 — The one move that would rescue the young position, and what it costs

Every argument above is a side-skirmish.  The 4.567-billion-year figure rests on radiometric decay, and the only way to make the solar system young is to make decay faster in the past.  That is the RATE programme's accelerated-decay hypothesis, and it has been argued at length elsewhere on this site.  Three points bear directly on the solar system.

Meteorites are the hardest case for accelerated decay, and RATE conceded it.  Terrestrial discordances can be blamed on open systems and heating.  Meteorites are closed systems with no water, no metamorphism and no plate tectonics, and their chronometers agree.  RATE's own second volume acknowledges that meteorites were "not adequately addressed" and that measurements of meteorites show a high degree of concordance.

The concordance is the problem, not the dates.  Alpha decay and beta decay have unrelated physics; alpha rates depend exponentially on the Coulomb barrier, beta rates do not.  To speed up decay by a factor of 10⁹ and still leave uranium-lead, rubidium-strontium, samarium-neodymium, rhenium-osmium and hafnium-tungsten in agreement, every one of those rates must be scaled by very nearly the same factor by a mechanism that has no reason to treat them alike.

The heat is fatal.  Compressing 4.5 billion years of crustal decay into the Flood year releases enough energy to raise average crustal rock by roughly 10⁵ K — several times the energy needed to vaporise it.  RATE says so in its own words: the heat "would have melted the crustal rocks many times over."  This is worked through in full on the companion pages.

So the solar-system arguments cannot be evaluated in isolation.  Even if every one of them worked, the meteorite chronometers would still have to be explained, and the explanation on offer boils the Earth.

Objections and responses

"You are just assuming uniformitarianism."

No.  The Moon section is the counter-example.  The mainstream case there does not assume the recession rate was constant — it assumes nothing about the rate, and reads two past Earth-Moon distances directly out of dated rock.  The young-earth calculation is the one that assumes constancy, in the form of an unchanging tidal response.

"The Oort cloud has never been seen.  That is not science."

The Oort cloud is inferred from the measured energy distribution of incoming long-period comets, which piles up exactly where a distant bound reservoir predicts.  Inference from a measured distribution is ordinary science.  Where the criticism does bite is on population counts: the number of objects required by dynamical models has not been reconciled with the number surveys find, and that is stated most bluntly in the mainstream literature itself.

"Mainstream scientists keep having to invent rescue devices — the Oort cloud, resonance locking, episodic outgassing."

Some of these were invented to solve a problem, and that is not a criticism by itself; the question is whether they were later checked against something independent.  Resonance locking was checked, by measuring Titan's migration rate.  Tidal heating of Io was checked, by measuring a century of orbital positions.  Episodic outgassing on Titan has not been checked, and this article says so.  The test is whether the rescue makes a new prediction that could fail.

"You accept the radiometric dates, which is the whole question."

The article's strongest arguments do not depend on them.  Lunar laser ranging, cyclostratigraphic Earth-Moon distances, Cassini's ring mass, and the astrometric measurement of tidal dissipation are all independent of decay constants.  The helioseismic solar age is independent of decay constants.  A young-earth model has to defeat these too, and none of the sources reviewed here attempts it.

"These arguments only claim to set maximum ages, and any maximum below 4.5 billion years is a win."

DeYoung and Lisle say exactly this: "all these age estimates are an upper limit."²⁵ It is a fair framing, but the upper limits have to be limits on the right thing.  A maximum age for Saturn's rings is not a maximum age for Saturn, any more than the maximum age of a snowbank is a maximum age for the mountain.  Of the arguments audited here, only the comet argument and the magnetic-field argument even try to bound the age of the whole system, and both bound a subsystem in practice.

"Creationists made a successful prediction with the magnetic fields."

One, in 1984, to within a factor of three, and it is documented here.  That prediction should be conceded by skeptics and it should not be oversold by creationists, because the theory behind it fits one free parameter per planet and the same theory's sharp prediction — Mercury's decay — has not been confirmed.

Verdict

Against the argument, with two live exceptions and one genuine prediction.

The solar system's measured age rests on lead-isotope chronometry of meteorites, confirmed independently by helioseismology.  No argument reviewed here touches either.  What the arguments touch are objects — comets, rings, small moons, atmospheres — whose own lifetimes are short, and the mainstream literature agrees that those lifetimes are short.  Turning a short object lifetime into a short system age requires setting a source term to zero, and in most of these cases the source has since been found.

Three arguments are now refuted by direct measurement: lunar recession, Io's heat, Enceladus's heat.  Two are retracted by creationists themselves: moon dust and solar neutrinos.  Four fail on arithmetic while resting on true observations: Saturn's rings, Mercury's field, planetary field decay generally, and the comet argument in its usual form.  One is a category error: the faint young Sun.

Two things remain worth watching, and a Christian writing on this subject should be pressing on these rather than on the Moon:

1. The short-period comet source population. The observed reservoir has not been reconciled with the number the models require. The gap has narrowed from two orders of magnitude to a factor of a few in fifteen years, which suggests it will close, but it has not closed.
2. Titan's methane. The atmosphere is demonstrably not in steady state on a ten-million-year timescale, and the proposed resupply has never been observed.

And one thing should be conceded: Humphreys published a numerical prediction of the Uranus and Neptune magnetic moments before the measurements existed, and he was right to within a factor of three.

Companion pages

• The literature — every source consulted, catalogued with what it measured and which way it cuts.
• Erosion rates and the age of the Earth — the same "no source term" error, on continents
• Natural arches, rock bridges, and the age of the Earth — the same error again, worked in full
• RATE and helium diffusion in zircon — the accelerated-decay hypothesis and its heat problem
• The hydroplate theory's heat problem — an alternative creationist route around the same obstacle

Sources