Nice Model

A solar system that rearranged itself

For most of the twentieth century the working assumption was that the planets formed roughly where we find them today and have orbited there, near-circular and stately, for 4.5 billion years. The Nice model — introduced in a trio of Nature papers in 2005 by Kleomenis Tsiganis, Rodney Gomes, Alessandro Morbidelli and Harold Levison, and named for the French city where the group worked — overturns that picture for the outer solar system. It proposes that the four giant planets were born in a tightly packed configuration, spanning only about 5.5 to 17 AU, and that they migrated to their present, far more spread-out orbits by gravitationally flinging a massive disk of icy planetesimals around. The migration was slow and smooth for hundreds of millions of years — until Jupiter and Saturn crossed their mutual 2:1 mean-motion resonance, and the whole system went briefly, violently unstable.

That single instability does an astonishing amount of work. In one event it disperses the planetesimal disk, drives Uranus and Neptune outward to their current orbits, sculpts the resonant and excited populations of the Kuiper belt, captures the Trojan asteroids at the Lagrange points of Jupiter and Neptune, captures the giant planets' irregular satellites, and — in the model's original framing — rains debris on the inner solar system to produce the Late Heavy Bombardment recorded on the Moon around 3.9 billion years ago. The Nice model's appeal is precisely this economy: a list of seemingly unrelated solar-system features all turn out to be consequences of the same scattering-driven migration and its resonant trigger.

How it works: scattering, migration, resonance

The mechanism rests on a simple exchange of angular momentum. A planet embedded in a sea of small bodies repeatedly scatters them. When a planet flings a planetesimal inward, the planet itself recoils outward; when it flings one outward, the planet drifts inward. Because the four giants sit at different distances, the net effect is systematic: Jupiter, which dominates and ultimately ejects most of the scattered bodies from the solar system altogether, migrates slightly inward; Saturn, Uranus and Neptune migrate outward, with Neptune moving the farthest as it plows through the dense disk beyond it.

For a long time this migration is smooth and gentle, slowly widening the orbital spacings. The drama comes when the slowly separating Jupiter and Saturn reach the point where Saturn's orbital period is exactly twice Jupiter's — the 2:1 mean-motion resonance. A resonance is a configuration where orbital periods form a simple integer ratio, so the two planets line up at the same longitudes over and over, and their gravitational tugs reinforce instead of cancelling. Sweeping across the 2:1 resonance pumps the eccentricities of Jupiter and Saturn upward sharply. Those newly eccentric orbits perturb the ice giants, whose own orbits cross and destabilise. Uranus and Neptune are scattered outward into the planetesimal disk, dispersing it in a few million years — a blink compared with the hundreds of millions of years of quiet migration that preceded it.

Slow phase:   planetesimal scattering → smooth migration (10⁸ yr)
Trigger:      Jupiter-Saturn cross 2:1 resonance (P_Saturn = 2 × P_Jupiter)
Instability:  e_Jupiter, e_Saturn spike → ice giants scattered outward
Aftermath:    disk dispersed, Kuiper belt + Trojans sculpted, LHB delivered

Worked example: how far did Neptune travel?

Consider Neptune, today at a semi-major axis of about 30.1 AU on a nearly circular orbit. In the Nice model it formed much closer in — most simulations start it near 12 AU, just outside Uranus. To migrate outward by ~18 AU, Neptune has to absorb angular momentum from the planetesimals it scatters inward. The scale of the disk required follows from conservation: moving a ~17 Earth-mass planet outward by that much demands a comparable reservoir of mass to push against, which is why the standard setup uses a disk of about 35 Earth masses spread from ~16 AU to a sharp edge near 30–35 AU.

We can sanity-check the resonance timing with Kepler's third law, which relates orbital period to semi-major axis as P ∝ a^(3/2). For Jupiter at 5.4 AU and Saturn just outside it, the period ratio starts near 1:2.5 and decreases as Saturn migrates out and Jupiter in. The crossing of the 2:1 resonance occurs when:

P_Saturn / P_Jupiter = 2
(a_Saturn / a_Jupiter)^(3/2) = 2
a_Saturn / a_Jupiter = 2^(2/3) ≈ 1.587

So with Jupiter near 5.4 AU, the 2:1 crossing happens when
Saturn reaches a_Saturn ≈ 5.4 × 1.587 ≈ 8.6 AU.

That places the trigger when Saturn is still well inside its present 9.6 AU orbit — exactly the kind of configuration the compact initial conditions produce as the pair drifts apart. Once crossed, the eccentricity kick destabilises the system; the disk, originally ~35 Earth masses, is almost entirely lost. Today only about 0.01–0.1 Earth masses of Kuiper belt material survive: more than 99% of the primordial disk was ejected to interstellar space or stored in the distant Oort cloud. The same dispersal flings bodies inward across the giant-planet region, and the destabilised asteroid belt adds to the flux, delivering the impact spike interpreted as the Late Heavy Bombardment around 3.9 Gyr ago.

Variants: smooth migration, jumping Jupiter, and Nice 2

The 2005 model assumed Jupiter and Saturn separate smoothly. That smoothness turned out to be a liability: a gradually moving Jupiter sweeps its secular resonances slowly across the inner solar system, over-exciting the asteroid belt and pumping up the eccentricities of the terrestrial planets — Earth and Mars would end up on orbits far more elliptical than observed. The fix is the "jumping Jupiter" scenario (Morbidelli et al. 2009; Brasser et al. 2009): instead of sweeping smoothly, an ice giant has a close encounter with Jupiter that makes Jupiter's semi-major axis jump discontinuously, skipping over the dangerous resonances in a single step and leaving the terrestrial planets and asteroid belt intact.

To make such an encounter statistically likely, many modern simulations begin with five giant planets — an extra ice giant between Saturn and Uranus — one of which is ejected from the solar system during the instability (Nesvorný 2011; Nesvorný & Morbidelli 2012). The five-planet, jumping-Jupiter configurations succeed far more often at reproducing the present architecture than the original four-planet smooth case. The umbrella term Nice 2 (Levison et al. 2011) refers to refinements in which the instability arises naturally from the planets' interaction with the planetesimal disk and the gas-disk dispersal, rather than being tuned to a particular resonance crossing time.

Common pitfalls and misconceptions

• "The Nice model proves the Late Heavy Bombardment happened." It does not. The model can produce a delayed impact spike, but whether the lunar record truly shows a sharp cataclysm at 3.9 Gyr or a smoothly declining tail of accretion is actively debated. Some samples may be biased toward the single Imbrium basin-forming event. The bombardment is the model's most contested output, not a confirmed prediction.
• Confusing the Nice model with the Grand Tack. The Grand Tack describes Jupiter's early migration into ~1.5 AU and back out while the gas disk is still present, shaping the inner solar system and Mars's small mass. The Nice model describes a later, gas-free instability of the already-formed giants. They address different epochs and are largely complementary.
• Thinking the resonance crossing is itself the instability. The 2:1 crossing is the trigger; the instability is the cascade that follows — eccentricity growth, ice-giant scattering, disk dispersal. Crossing a resonance is necessary but the violence is in the aftermath.
• Assuming the original disk is the Kuiper belt we see. Today's Kuiper belt is the tiny surviving remnant — under 0.1% of the primordial ~35 Earth-mass disk. The belt's structure records the migration; its mass does not.
• Treating "smooth migration" as the final word. The original smooth-separation version is now largely superseded by jumping-Jupiter and five-planet variants precisely because smoothness wrecks the terrestrial planets. Citing the 2005 mechanism without the Nice 2 corrections is out of date.

Observational status and what it explains

The Nice model's standing rests less on any single observation than on the breadth of structures it ties together. Each of the following is a non-trivial feature of the solar system that the same instability reproduces:

• Resonant Kuiper belt. As Neptune migrates outward it sweeps bodies into its mean-motion resonances, populating the 3:2 "plutinos" (Pluto's family) and the 2:1 and other resonances — exactly the resonant clustering observed in the trans-Neptunian region.
• Hot vs. cold classical belt. The high-inclination, high-eccentricity "hot" population is implanted from the scattered disk; the dynamically "cold," low-inclination population near 42–45 AU likely formed in place, and the color dichotomy between them follows naturally.
• Trojan capture. Jupiter's and Neptune's Trojans are captured chaotically during the instability, when the co-orbital region opens and recloses. This explains the broad inclination spread of the Jupiter Trojans — impossible for in-situ formation but a direct prediction of chaotic capture.
• Irregular satellites. The distant, inclined, often retrograde moons of the giant planets are captured from the dispersing planetesimal disk during the same chaotic phase.
• Giant-planet eccentricities. The present small-but-nonzero eccentricities of Jupiter through Neptune require some excitation; the resonance crossing supplies it without over-exciting the system, provided Jupiter jumps.

Spacecraft data feed directly into this picture. NASA's New Horizons flyby of the cold-classical object Arrokoth in 2019 showed a gently accreted, undisturbed contact binary — consistent with in-situ formation of the cold belt — while the broader trans-Neptunian census from surveys like OSSOS continues to test the predicted resonant populations against reality.

Quantitative analysis: why a resonance destabilises

The reason a mean-motion resonance is dynamically dangerous comes from how perturbations accumulate. Off resonance, the relative geometry of two planets cycles through all longitudes, so their mutual gravitational perturbations average toward zero over many orbits. On a p:q resonance, the critical angle

φ = p·λ₁ − q·λ₂ − (p−q)·ϖ

(where λ are mean longitudes and ϖ is the longitude of perihelion) librates rather than circulating freely, so the perturbation no longer averages away — it acts coherently, orbit after orbit, and the eccentricity grows. For the Jupiter-Saturn 2:1 resonance the secular forcing on eccentricity is strong, and because the pair carries most of the system's mass, the eccentricities they acquire are transmitted to Uranus and Neptune, pushing those lighter planets onto crossing orbits.

The angular-momentum budget sets the migration scale. The specific orbital angular momentum is h = √(G M_☉ a (1 − e²)). Scattering a planetesimal of mass m changes the planet's semi-major axis by roughly Δa/a ≈ (m / M_planet) per encounter; to move Neptune from 12 to 30 AU requires summed encounters totalling a disk mass comparable to a planet's, hence the ~35 Earth-mass figure. The timing of the trigger is governed by how fast the disk is consumed — set by its mass and how close its inner edge sits to Neptune. In the original model these were tuned to delay the crossing to ~3.9 Gyr; in Nice 2 the delay emerges more naturally from the disk's self-stirring, and recent meteoritic and lunar constraints push many modern instabilities to within the first ~100 Myr instead.

Migration scenarios at a glance

Scenario	Epoch	Trigger / mechanism	Jupiter motion	Headline output	Status
Static (classical)	—	planets fixed in place	none	fails Kuiper belt structure	abandoned
Nice model (2005)	~3.9 Gyr ago	Jupiter-Saturn 2:1 crossing, smooth	smooth inward drift	Kuiper belt, Trojans, LHB	foundational
Jumping Jupiter	instability	ice-giant close encounter	discontinuous jump	spares terrestrial planets	preferred
Five-planet Nice	instability	extra ice giant ejected	jump, +1 planet lost	higher success rate	leading variant
Nice 2 (2011)	self-consistent	disk self-stirring / gas dispersal	jump or sweep	natural instability timing	active
Early instability	< ~100 Myr	same mechanism, sooner	jump	decouples from LHB	increasingly favored
Grand Tack	gas-disk era	Jupiter-Saturn gas-driven migration	in to ~1.5 AU then out	Mars's small mass	complementary

The bottom rows highlight the live debate: as evidence accumulates that the lunar "cataclysm" may not be a sharp spike, the instability that the Nice model describes is increasingly placed in the solar system's first hundred million years, decoupling the scattering mechanism (which is robust) from the Late Heavy Bombardment (whose reality is contested). The dynamical machinery stands; its connection to the bombardment is the part still being relitigated.