The Persistence of Comets

Comets lose mass on every perihelion pass and die off in tens of thousands of years. After 4.6 billion the inner solar system should be empty of them — and yet we keep finding more. The mainstream answer is a reservoir nobody has ever seen.

Comets are different from the other clocks in this section. The earlier four are about rates measured on earth — the moon receding, the salt accumulating, the collagen decaying, the carbon-14 ticking down. Comets are about rates measured in the sky, on the small bodies of the outer solar system, and what those rates imply about how long the solar system can have looked the way it does today.

The argument is the simplest of the five to state. It is also, in some ways, the most rhetorically loaded, because the mainstream response to it requires positing an entity that, three quarters of a century after it was first proposed, has still not been observed.

What comets are, in one paragraph

A comet is a small body — typically a few kilometres across — composed mostly of water ice with some frozen CO₂, methane, ammonia, and embedded dust. When its orbit takes it close to the sun, the ices sublimate; the released gas and dust form the characteristic coma and the two tails (the dust tail trailing behind, the ion tail pointing away from the sun by solar-wind pressure). The Whipple “dirty snowball” model from 1950 has held up remarkably well as a description of what is going on, with detailed amendments from Deep Impact (2005) and Rosetta’s close study of 67P/Churyumov–Gerasimenko (2014–2016).¹

The relevant fact for our purposes is that comets lose mass every time they swing past the sun. The loss rate per passage varies — Halley loses about 3×10¹⁰ grams per perihelion, of an estimated nucleus mass of 2×10¹⁷ grams, or about \(10^{-4}\) of itself per pass² — but it is steady, and over many perihelion passages it adds up. Periodic outbursts, splittings, and collisions add to the attrition.

The lifetime arithmetic

A typical short-period comet — orbital period under 200 years — loses something like 0.01% to 0.1% of its nucleus per perihelion passage. At one passage every couple of decades, the half-life of the nucleus is on the order of 10,000 to 100,000 years. Some comets last longer; many disintegrate suddenly. The upper bound on the lifetime of an individual short-period comet, from straightforward mass-loss arithmetic combined with the observed disruption rate, is somewhere around a hundred thousand years.³ After a few hundred thousand years, the nucleus is gone — either spread along the orbit as a meteor stream, ejected from the system, or fallen into the sun.

We have direct observational support for this. Comet Biela split in 1846 and disintegrated entirely by 1852. Comet West fragmented in 1976. Comet 73P/Schwassmann–Wachmann broke into more than sixty pieces in 1995. Comet Shoemaker–Levy 9, captured by Jupiter, fragmented and impacted in 1994 in front of every telescope on earth. Comets do not survive forever.

The number of currently active short-period comets we know about is in the few hundreds. Long-period comets (orbital periods greater than 200 years) are more numerous in the historical record — thousands have been observed across human history — but each is, by definition, much rarer per unit time at perihelion, and the long-period population is supplied (on the standard view) from a different proposed source.

If the solar system is 4.6 billion years old, and short-period comets last on the order of \(10^5\) years before depletion, then — with no replenishment — the original population of short-period comets would have had to be about forty thousand times the present population to leave today’s residue. That is not a serious initial condition. It is not the standard assumption.

The standard assumption is replenishment.

The proposed reservoirs

Two reservoirs have been proposed:

The Kuiper Belt, a disc of icy bodies in the region beyond Neptune (roughly 30 to 50 AU from the sun), was hypothesized by Kenneth Edgeworth in 1943 and independently by Gerard Kuiper in 1951. The belt was directly observed starting in 1992, when (15760) Albion was discovered by David Jewitt and Jane Luu. We now know of more than three thousand Kuiper Belt objects (KBOs). It is a real, observed population. The standard proposal is that gravitational perturbations (chiefly from Neptune) occasionally nudge KBOs and scattered-disc objects into Neptune-crossing orbits, from which they can be passed inward to become Jupiter-family comets — the short-period, low-inclination class.

The Oort cloud, a roughly spherical shell of icy bodies extending from about 2,000 AU to 100,000 AU, was hypothesized by Jan Oort in 1950 specifically to explain the orbital distribution of long-period comets.⁴ Long-period comets appear to come from random directions, which suggests an isotropic source very far out. The proposal is that stellar passages and galactic tides occasionally perturb Oort-cloud objects into the inner solar system, where they appear to us as long-period comets.

These are sensible hypotheses, and the Kuiper Belt one was vindicated empirically in 1992. The Oort cloud one has not been.

What we actually know about the reservoirs

The Kuiper Belt is observed. We have direct images, orbital data, taxonomic classifications, and spectral compositions for thousands of its members. The New Horizons spacecraft flew past 486958 Arrokoth on 1 January 2019 and returned the first close-up images of a primordial KBO. Arrokoth is a contact binary, almost certainly never warmed since formation, and not — in composition or structure — much like a typical Jupiter-family comet. The KBO-to-short-period- comet dynamical pipeline does function: Levison and Duncan have shown that a fraction of scattered-disc objects can be perturbed into Jupiter-crossing orbits over megayear timescales.⁵ The remaining question is whether the rate is high enough to keep the observed short-period comet population in steady state for billions of years. This is a matter of active debate, and the honest answer in the literature is “perhaps, with assumptions.”

The Oort cloud is, as of 2026, unobserved. The most distant solar system object directly observed is Sedna (perihelion ~76 AU, aphelion ~937 AU), and Sedna is generally described as an “inner Oort cloud” candidate rather than a member of the canonical Oort cloud at thousands of AU. The cloud itself — the entity Oort proposed — has been searched for at infrared wavelengths by IRAS, Spitzer, and WISE, and at optical wavelengths by multiple sky-survey campaigns. It has not been seen. The most stringent upper limits set so far still leave room for the population Oort needed, but they do not confirm it.⁶

That is not the same as saying the Oort cloud does not exist. It might. But it is honest to point out the structure of the argument:

1. Long-period comets are observed; they cannot have been around for 4.6 billion years at the present rate of attrition.
2. Therefore there must be a reservoir feeding them.
3. The reservoir has been searched for and not seen.
4. We assume it exists anyway, because the alternative is that long-period comets have not been around for 4.6 billion years.

That is a reasonable hypothesis at the level of theoretical astrophysics. It is a weak response to the empirical question “what does the observed depletion of comets tell us about how long the solar system has been depleting them?” The empirical answer, without an observed reservoir, is: not very long. Tens of thousands of years, plausibly. Not billions.

Where the comet argument lands

The comet argument is the most contested of the five in this section, because of the existence of the Kuiper Belt — a real, observed reservoir that does supply short-period comets at some rate. A young-earth case that uses the comet argument has to engage the Kuiper Belt seriously and explain why it cannot do the job; it cannot pretend the Kuiper Belt does not exist.

The argument retains its force, however, for long-period comets, because their proposed source — the Oort cloud — remains unobserved seventy-five years after it was hypothesized. The proposal was ad hoc in the technical sense: it was put forward to solve a specific dynamical problem, namely the persistence of long-period comets in a universe assumed to be billions of years old. If the universe is not assumed to be billions of years old, the problem the Oort cloud was invented to solve does not exist, and the long-period comets are simply primordial — left over from creation, slowly being whittled away exactly as we observe.

Cumulative force

This is the fifth of five arguments in this section. Each, taken alone, can be dismissed by an old-earth reader as an isolated anomaly with some imaginable explanation. Five together are harder to dismiss.

• Lunar recession says the moon cannot have been receding for 4.5 billion years at anything like the present rate.
• Ocean sodium says the oceans cannot have been accumulating salt for 3 billion years at anything like the present rate.
• Dinosaur collagen says the bones cannot have been preserved for 65 million years at anything like the present rate of protein decay.
• Carbon-14 says coal and diamonds cannot be older than tens of thousands of years given the half-life and the measured atomic abundance.
• Comets say the solar system cannot have been losing comets for 4.6 billion years at anything like the present rate, unless an unobserved reservoir is supplying them.

Five independent clocks. Five different fields — astronomy, oceanography, palaeontology, nuclear physics, planetary dynamics. Five converging upper bounds, none of them remotely compatible with the standard chronology.

The honest reading, in my view, is that the clocks are telling us something true. The earth — and the solar system around it — on the evidence of what we can actually measure today, is not billions of years old.

How much younger it is than billions is a separate question. The biblical chronology of roughly six thousand years is consistent with the upper bounds these methods give, but it is more specific than any of them require. The serious young-earth literature does not pretend the five arguments above prove a six-thousand-year-old earth on their own. What they do prove — strongly — is that the deep-time chronology is in deep trouble.

Whether you take the next step and accept the biblical chronology specifically depends on how you weigh the historical evidence for the Bible’s reliability. That is a question for the history section, not for this one.