Late Heavy Bombardment

A wave of fire on a quiet Moon

Look at the Moon through binoculars and you are reading a calendar. The dark, smooth plains — the maria — are flood basalts that filled the largest impact basins. The bright, rugged highlands are older, saturated with overlapping craters. Counting and dating those features, planetary scientists noticed something strange in the Apollo samples: instead of impact ages spread evenly across the Moon's history, a large fraction clustered tightly around 3.9 billion years ago, with surprisingly few older than about 4.1 Gyr.

One reading of that pattern is dramatic. After the solar system formed and the impact rate from leftover building material had largely died down, something restarted the bombardment — a comparatively brief, intense pulse of large impactors that carved the youngest giant basins and reset the clocks in lunar rock. That hypothesised pulse is the Late Heavy Bombardment, often called the lunar cataclysm. Whether it was a true spike or just the tail of a long decline is one of the liveliest arguments in planetary science, and the rest of this article walks through both the evidence for it and the case against.

Reading clocks in shocked rock

The bombardment is dated, not seen. An impact large enough to melt rock resets radiometric clocks: shock and heat drive out accumulated argon, so the ⁴⁰Ar/³⁹Ar age of an impact-melt breccia records the moment of the impact, not the formation of the original crust. When Apollo melt rocks were dated through the early 1970s, the ages piled up:

Apollo impact-melt ⁴⁰Ar/³⁹Ar ages:
  many cluster near        3.8 – 4.0 Gyr
  Imbrium basin            ≈ 3.92 Gyr
  Serenitatis              ≈ 3.89 Gyr (debated; may be Imbrium ejecta)
  Nectaris                 ≈ 3.92 – 4.1 Gyr (model-dependent)
  very few melt rocks      > 4.1 Gyr

In 1974 Tera, Papanastassiou and Wasserburg argued that the sharp pile-up plus the scarcity of older melt could not be the smooth tail of accretion — it looked like a discrete, terminal cataclysm. The ⁴⁰Ar/³⁹Ar system is the workhorse here because it can date individual melt clasts and reveals partial resetting through its release spectrum. The same era shows up faintly in meteorites: some HED and ordinary-chondrite samples carry shock ages clustered near 3.5–4.1 Gyr, hinting that whatever happened was not confined to the Earth–Moon system.

The Nice model: a resonance pulls the trigger

A spike needs a delay mechanism. Why would the impact rate stay low for ~600 million years and then surge? The most influential answer is the Nice model (named for Nice, France), developed by Gomes, Levison, Tsiganis and Morbidelli in a set of 2005 papers. Its key idea is that the giant planets did not form in their present orbits.

In the Nice picture, Jupiter, Saturn, Uranus and Neptune began in a more compact configuration, embedded just outside a massive disk of leftover planetesimals (tens of Earth masses) extending out to roughly 30–35 AU. Gravitational scattering of those planetesimals slowly changed the planets' orbits — Jupiter creeping inward, the others outward — over hundreds of millions of years. The detonator is a mean-motion resonance: as Saturn migrated outward, the orbital-period ratio of Jupiter to Saturn drifted toward 2:1.

Mean-motion resonance condition (j:k):
  P_Saturn / P_Jupiter → 2 / 1     (the crossing)

Kepler's third law links period and semi-major axis:
  P² = (4π² / G M_⊙) a³     →     P ∝ a^(3/2)

When the ratio passed through 2:1, the two giant planets exchanged angular momentum coherently every conjunction, pumping their eccentricities. Their now-eccentric orbits swept through the disk and the asteroid belt like a stirring rod, raising eccentricities and inclinations across the inner edge of the planetesimal disk and the outer asteroid belt. A large fraction of both reservoirs was suddenly thrown onto planet-crossing orbits — and a burst of asteroids and comets rained inward over a few tens of millions of years. Tuned to put the crossing at ~3.9 Gyr, the model reproduces the lunar timing, the capture of Jupiter's Trojans, the irregular satellites, and the structure of the Kuiper belt in one stroke.

How much mass, how many craters

Estimating the delivered mass turns clusters of craters into kilograms. A crater's diameter scales with impactor energy through crater-scaling laws; for the gravity regime the transient crater diameter goes roughly as

D_crater ∝ (E)^(~0.28) ∝ (½ m v²)^(0.28)
final basin diameter ≳ 20 × impactor diameter  (very rough, large basins)

Inverting the size–frequency distribution of lunar basins and craters gives an order-of-magnitude bookkeeping. Estimates for the total mass delivered to the Moon during the proposed spike are of order 10¹⁹–10²¹ kg — equivalent to a few thousand impactors larger than ~10 km, and a few dozen large enough (~100–200 km diameter) to excavate basins hundreds of kilometres across. Scaled by gravitational cross-section, the Earth would have received on the order of 20× the lunar flux. The numbers below summarise the scale of the largest scars.

Basin (Moon)	Diameter	Estimated age	Impactor size (rough)	Note
South Pole–Aitken	~2,500 km	~4.2–4.3 Gyr	~170 km	Oldest/largest; pre-spike
Nectaris	~860 km	~3.92–4.1 Gyr	~50 km	Defines Nectarian period
Imbrium	~1,150 km	≈ 3.92 Gyr	~70–100 km	Dominates sample ejecta
Orientale	~930 km	~3.8 Gyr	~60 km	Youngest large basin
Schrödinger	~320 km	~3.8 Gyr	~25 km	Well-preserved peak ring
For comparison — Earth's largest preserved crater
Vredefort (Earth)	~160–300 km	~2.02 Gyr	~10–15 km	Far younger than LHB

Quantified figures to anchor the era

• Timing. Solar system formed ~4.567 Gyr ago; Moon-forming giant impact ~4.5 Gyr ago; proposed bombardment peak ~3.9 Gyr ago — a delay of ~600 Myr.
• Flux contrast. Cataclysm models require the impact rate near 3.9 Gyr to be at least an order of magnitude (10×–100×) above the smoothly declining background extrapolated from the post-3.5-Gyr cratering record.
• Reservoirs. Primordial trans-Neptunian disk ~20–35 Earth masses; primordial asteroid belt perhaps tens of times more massive than today's ~5 × 10⁻⁴ M⊕ before depletion.
• Impact speeds. Asteroidal impactors strike the Moon at ~12–20 km/s; long-period comets can exceed 50 km/s, releasing far more energy per kilogram (E = ½mv²).
• Sample reach. All six Apollo and three Luna sample sites lie on the lunar near side, most within reach of Imbrium ejecta — a footprint that biases the age statistics, as discussed below.

Fingerprints across the inner solar system

• The Moon. The primary archive. Saturated highlands, the youngest giant basins (Imbrium, Orientale), and the clustered melt ages are the canonical dataset. Lacking plate tectonics and weather, the Moon froze its early impact record in place.
• Mercury. The Caloris basin (~1,550 km) and heavily cratered terrain echo the same heavy early flux, though Mercury's ages are not directly sampled.
• Mars. The giant Hellas (~2,300 km) and Argyre basins, and the ancient southern highlands, date to roughly the same window. Some martian meteorites carry ~4.0 Gyr shock signatures.
• Earth. The crust from this era is almost entirely gone, but 4.4-Gyr Jack Hills zircons and the late-veneer siderophile-element budget of the mantle provide indirect constraints. Asteroid 4 Vesta's HED meteorites record shock ages overlapping the LHB.
• The outer system. Jupiter's Trojan asteroids, the giant planets' irregular satellites, and the dynamically excited 'hot' Kuiper belt are interpreted by Nice-model advocates as collateral signatures of the same instability that drove the inner-planet bombardment.

Cataclysm vs. accretion tail

The deepest disagreement is whether the lunar data show a true spike or a smoothly declining flux mis-read as one. The competing pictures make different predictions:

Feature	Cataclysm (true spike)	Accretion tail / sawtooth
Flux vs. time	Low, then sharp surge at ~3.9 Gyr	Monotonic decline from 4.5 Gyr
Trigger	Delayed giant-planet instability	Leftover planetesimals, no delay
Melt-age clustering	Real, reflects the burst	Artifact of Imbrium sample bias
Pre-4.1-Gyr basins	Genuinely few formed	Many formed, then erased/overprinted
Nice-model timing	Instability at ~3.9 Gyr	Instability within first ~100 Myr
Key test	Wide spread of basin ages	Older melt ages once Imbrium removed

Recent dynamical work has shifted opinion toward an early giant-planet instability — within ~100 Myr of gas-disk dispersal — partly because a late instability tends to over-excite the terrestrial planets and destabilise their orbits. If the instability was early, the apparent 3.9-Gyr "spike" is more naturally the tail of a long, steeply declining accretion flux, with the lunar age clustering produced by a single dominant late basin (Imbrium) resetting most sampled rocks.

Common misconceptions and edge cases

• "It is an established fact." It is a hypothesis with strong but contested support. The clustered ages are real; their interpretation as a discrete cataclysm is what remains open.
• "All the lunar craters formed during the LHB." No. South Pole–Aitken, the largest basin, predates the proposed spike (~4.2–4.3 Gyr), and cratering has continued ever since (e.g. the ~1-Gyr-old crater Copernicus). The LHB, if real, is one episode in a long history.
• The Imbrium sampling bias. The single most important caveat: because nearly all returned samples sit within Imbrium's enormous ejecta blanket, one late, colossal impact can dominate the age statistics and manufacture a false spike. Breaking this bias needs samples from far-flung basins.
• "Cataclysm" doesn't mean instantaneous. Even in spike models the intense phase lasts tens of millions of years — geologically brief, but tens of thousands of human civilisations long. It is a pulse, not a single day.
• Comets vs. asteroids. The mix matters. A comet-dominated bombardment (volatile-rich, high-speed) implies different impact energies, volatile delivery and crater morphologies than an asteroid-dominated one; siderophile and isotopic signatures in lunar and terrestrial rocks are used to constrain the ratio, and they favour an asteroid-dominated late flux.
• Confusing LHB with the Moon-forming impact. The Theia giant impact (~4.5 Gyr) that formed the Moon is a separate, far earlier and far larger event; the LHB came hundreds of millions of years later onto an already-solid Moon.