What Is Planetary Migration? How Planets Wander Across Solar Systems

The solar system looks orderly: small rocky planets near the Sun, giants farther out, everyone in near-circular lanes. Yet the planets we find around other stars tell a messier story. Jupiter-sized worlds circle their stars in a few days, and tightly packed families of planets march in lockstep. The best explanation is planetary migration, the shift of a planet's orbit after it forms. Our own planets may have moved too, and that movement may have shaped Mars, the asteroid belt, and the outer system.

Why planets move at all

A newborn star is surrounded by a protoplanetary disk, a rotating pancake of gas and dust from which planets form. Our article on the protoplanetary disk covers how the pieces come together. Here, the important point is that the disk is not an empty stage. A planet embedded in it pulls on the gas, the gas pulls back, and the exchange of angular momentum changes the planet's orbit.

A useful analogy is a boat in a river with an uneven current. A boat does not stay where it is launched: the current, the eddies the boat itself creates, and the banks all push it around. A planet in a gas disk is pushed by the wakes it makes in the gas. The difference is that the result can be a drift of hundreds of millions of kilometers.

Migration is not one mechanism. It is a family of effects, and which one dominates depends on the planet's mass, whether the gas disk is still present, and what else is nearby. The three main drivers are the gas disk, other planets (through resonances and close encounters), and leftover small bodies called planetesimals.

The clock: gas disks do not last

Gas disks live a few to several million years before they are cleared, according to a review cited in the Wikipedia article on planetary migration (Ercolano and Pascucci 2017). That limit matters. Disk-driven migration can only proceed while the gas is there, so it must happen quickly and then stop. After the gas is gone, planets keep moving only through slower effects: encounters with each other, interactions with planetesimals, and tides.

The three types of disk migration

When a planet moves through gas, it launches spiral density waves, ripples in the disk that trail and lead the planet. The gravitational pull of those spirals on the planet creates a net torque (a twisting force). What happens next depends on the planet's mass.

Type	Planet mass	What the planet does to the disk	Typical result
Type I	Small, below about Saturn's mass	Cannot open a gap; excites waves in the gas	Usually inward drift; fast compared with the disk's life
Type II	Large, roughly Saturn's mass and above	Clears a gap in the disk	Moves with the disk's viscous evolution; slower
Type III	Special cases	Co-orbital gas torques in a partly cleared region	Can be fast; direction depends on conditions

The transition from Type I to Type II happens at roughly Saturn's mass, according to D'Angelo and colleagues (2003), as summarized in the same Wikipedia review.

Type I: the small-planet drift

For a planet too small to carve a gap, the torques from the spiral waves it raises generally push it inward, and the rate scales with the planet's mass and the local gas density. Because the rate is fast relative to the disk lifetime, theorists have long worried that Type I migration would drag small planets straight into their stars. This is the so-called migration problem, and part of the research of the last two decades has been to find out why that does not happen as often as the simplest models predict.

Type II: the gap-opener

A planet massive enough to open a gap sits in a clear lane, like a boulder dividing a river into two streams. It no longer migrates independently and instead rides along with the gas as the disk slowly spirals onto the star. Type II is often modeled as a modified version of Type I. The result is typically inward, but if the outer disk is spreading outward, a gap-opening planet can in principle be carried outward too.

Why migration can stop or reverse

Migration does not have to end at the star. Disk models show several "halting zones," places where torques cancel or a boundary stops further motion:

• The disk's inner edge. Where the gas disk ends near the star, planets can pile up. This is thought to be one way tightly packed inner systems form (Cossou and colleagues, 2014).
• The ice line. Where Type I torques cancel near the distance at which water freezes, migration can stall (Cossou and colleagues, 2013).
• Steep gradients. In regions where density or temperature changes sharply, Type I torques can reverse and push a planet outward.

Outward migration has another form that works after the gas is gone. In planetesimal-driven migration, a planet scatters small bodies, and the exchange of angular momentum nudges the planet in the opposite direction. Neptune is the textbook case. As it scattered icy planetesimals inward, it drifted outward, and in doing so it is thought to have swept Pluto and other plutinos into the 3:2 resonance with Neptune. Pluto and its relatives are covered in the Related Reading list below.

Hot Jupiters: the clue that started it

In 1995 astronomers announced 51 Pegasi b, the first planet found orbiting a Sun-like star. It had roughly Jupiter's mass but circled its star in about four days. Today a hot Jupiter is defined as a Jupiter-like gas giant with an orbital period under about 10 days. They are most common around F- and G-type stars and rare around red dwarfs.

They are a puzzle because of how giant planets are thought to form. In the standard core accretion picture, a giant planet starts as a solid core, which then pulls in gas. Building a big core requires a lot of solid material, and the solid supply is far richer beyond the ice line (also called the frost line), the distance where water freezes and the solid mass in the disk jumps. Close to the star, where it is hot, there is much less solid material available. For a long time the simplest reading was that hot Jupiters formed farther out and then moved in.

That is not the only possibility, and it is not clear that all hot Jupiters formed the same way. The main candidates are:

Channel	How it works	What it predicts
In situ formation	A core forms close to the star and grows a gas envelope there	No migration needed; planets should often have close neighbors
Disk migration	The planet forms beyond the ice line and moves in through Type II migration (Armitage 2007)	Circular orbits aligned with the star's spin
Planet-planet scattering	Gravitational encounters throw the planet onto an eccentric orbit; tides then shrink it (Ford and Rasio 2008)	Eccentric or misaligned orbits
Kozai cycles with tides	A distant companion tilts the orbit and drives eccentricity; tidal friction circularizes it (Fabrycky and Tremaine 2007)	Misaligned orbits; a distant massive companion

The in situ idea was traditionally disfavored but is taken more seriously now that super-Earths are known to be common in inner systems. For the others, there is observational evidence. About half of hot Jupiters are reported to have distant, massive companions, which is what the Kozai route would need. And several hot Jupiters have orbits tilted relative to their star's spin, including HAT-P-14b, which runs backward (retrograde), and the strongly misaligned HD 80606 b. A smooth migration through an aligned disk would not produce that, so something else must have tilted at least some of them.

The honest summary is that migration, in the broad sense of a planet moving a long way from its birthplace, is the leading explanation for most hot Jupiters, but "disk migration" is only one of several ways it could have happened. Astronomers often use the orientation of the orbit as a fingerprint to tell them apart. For how those planets are found in the first place, see how we find exoplanets and the TESS exoplanet survey.

The Grand Tack: did Jupiter sail in and out?

Our own solar system presents a different puzzle. The inner planets look too small. In computer simulations of rocky planet formation, a planet of 0.5 to 1.0 Earth masses often forms around Mars's distance from the Sun. The actual Mars is about 0.107 Earth masses. And the asteroid belt has a tiny fraction of the mass it should have had.

In 2011, Kevin Walsh, Alessandro Morbidelli, Sean Raymond, David O'Brien, and Avi Mandell published "A low mass for Mars from Jupiter's early gas-driven migration" in Nature. The idea became known as the Grand Tack (a sailing term for a sharp change of course). In the original scenario:

1. Jupiter forms near the ice line, at about 3.5 AU (an AU is the Earth-Sun distance).
2. Jupiter migrates inward by Type II migration, as far as about 1.5 AU.
3. Saturn, which migrates faster, catches up and is captured in a 2:3 mean-motion resonance with Jupiter, meaning Jupiter orbits three times for every two of Saturn's orbits.
4. The pair's combined torque turns positive, and they migrate outward together.
5. The outward motion stops near Jupiter's present orbit at about 5.2 AU, when the gas disk disappears.

Why does this help? As Jupiter plows inward it clears most of the material beyond about 1 AU, leaving a narrow ring of solids from which Earth and Venus grow and a starved region where Mars forms small. The asteroid belt gets emptied and then partially refilled. In the simulations, about 15% of inner asteroids are scattered outward during the inward leg, and a small share (about 0.5% of the original population) is scattered back in after the reversal, with a similar fraction of primitive asteroids landing in the belt from beyond Jupiter. The result is a low-mass belt that mixes rocky material from inside Jupiter's original orbit with primitive material from outside. See our overview of asteroids, comets, and meteors for how those two kinds differ.

Problems with the Grand Tack

The Grand Tack is a hypothesis, and it has real critics:

• Capturing Saturn in the 3:2 resonance seems unlikely in a fading gas disk. Later studies found that disk conditions would more likely produce a 2:1 resonance, and 2:1 capture usually does not reverse migration.
• In the 2:1 versions, the eccentricities of the giant planets come out two to three times larger than the real ones.
• The scenario ignores ongoing gas accretion onto Jupiter and Saturn, which would change their growth and mass ratio.
• If Mars's composition differs from that of Earth and Venus, only about 2% of simulated outcomes reproduce it.

Other ideas can also make a small Mars. These include less efficient pebble accretion at larger distances, an early Nice-model instability that stripped mass from the Mars region, convergence of embryos toward 1 AU, and sweeping secular resonances as the gas clears. The scientific position is that the Grand Tack is a serious, testable candidate, not a settled history. For the standard timeline, see how the solar system formed.

The Nice model: a late shake-up of the outer planets

A second migration story concerns what happened after the gas was gone. The Nice model, named for the Observatoire de la Cote d'Azur in Nice, France, was published in 2005 in three papers in Nature by Rodney Gomes, Hal Levison, Kleomenis Tsiganis, and Alessandro Morbidelli. In its original form, the four giant planets began on near-circular, compact orbits roughly between 5.5 and 17 AU, surrounded by a dense disk of icy planetesimals of about 35 Earth masses.

Encounters with the planetesimals gradually shifted the planets, pushing Uranus and Neptune outward and Jupiter inward. After hundreds of millions of years, Jupiter and Saturn reached a period ratio of 2:1, meaning that Jupiter orbited twice for each Saturn orbit. That crossing raised the eccentricities of both planets and triggered a rapid, chaotic rearrangement in which Neptune and Uranus were thrown onto eccentric orbits and scattered most of the planetesimal disk. The model also accounts for several outer solar system populations:

• Trojan asteroids. The resonance crossing destabilized the Trojan regions of Jupiter and Neptune and allowed new objects to be captured, with a wide range of orbital inclinations, as observed.
• The Kuiper Belt. Neptune's outward scattering helps explain the plutinos and the belt's outer edge, though the model's eccentricities are too high and it struggles with the cold population's colors and binary objects. See our article on the Oort Cloud and Kuiper Belt.
• Irregular moons. The "jumping Jupiter" variant, in which an ice giant is scattered by Saturn and then by Jupiter, helps explain Jupiter's irregular satellites and addresses problems with terrestrial-planet eccentricities.

The Late Heavy Bombardment problem

The Nice model was originally built to explain a supposed spike in impacts on the Moon about 600 million years after formation, called the Late Heavy Bombardment. That idea has weakened. Newer lunar crater studies find no clear peak, only a decay in crater numbers, and evidence from the asteroid Vesta conflicts with a sharp spike. The Wikipedia article on the Nice model says the Late Heavy Bombardment is now widely doubted, which undermines the original motivation for a late instability.

That does not kill the idea. Variants such as a five-giant-planet start with one planet ejected, and a "Nice 2" version in which a Pluto-mass planetesimal population triggers a late instability, have been proposed. A 2015 study found that an early instability is statistically unlikely to reproduce the terrestrial planets' orbits, so the timing of the instability remains a live question. Together, the Grand Tack and the Nice model show a solar system that was reorganized twice, once while gas was present and once after.

A moon that was captured

Reshuffled orbits are not limited to planets. Neptune's moon Triton orbits backward relative to Neptune's spin, a sign that it was captured rather than formed in place. Read more in Triton: Neptune's captured moon.

Super-Earths and the missing planets in our system

Among the planets that the Kepler mission found, super-Earths and mini-Neptunes (planets between roughly Earth's size and Neptune's) are among the most common. Super-Earths are generally defined as having masses above Earth's and well below those of the ice giants, with an upper bound near 10 Earth masses. Kepler also showed that planets between about 1.5 and 2.0 Earth radii are relatively rare, a gap called the radius valley, first measured by Fulton and colleagues in 2017. It separates rocky super-Earths below about 1.75 Earth radii from sub-Neptunes with thick gas envelopes above.

Our solar system has none of these. Earth is the largest terrestrial planet, and every larger planet is a gas or ice giant of at least 14 Earth masses. Why?

Migration offers several candidate answers:

• Jupiter as a barrier. Jupiter's growth may have blocked the inward migration of super-Earths that formed farther out.
• The Grand Tack. Jupiter's inward sweep may have cleared the region where such planets would have lodged.
• Engulfment. A super-Earth may once have existed in the inner system and been swallowed by the Sun. A 2026 study cited on the Wikipedia article on super-Earths links this idea to the Sun's lithium content.

These are alternatives, not a consensus. Hints from the radius valley and from planet counts around other stars are helping to narrow them down, but no single account has won.

Resonant chains: migration frozen in the orbits

When two planets migrate and one catches up to the other, they can be caught in a mean-motion resonance, an orbital period ratio of small whole numbers such as 3:2. In our solar system, Neptune and Pluto are locked in 3:2, and Jupiter's moons Io, Europa, and Ganymede are in a chain called the Laplace resonance (see our article on Io). A resonance keeps the planets in step the way two people walking in time stay in step.

The showcase example is TRAPPIST-1, a small red star about 40.66 light-years from Earth. Its seven planets were found in two steps: two in 2016 and five more in 2017. Their orbital periods are roughly 1.51, 2.42, 4.05, 6.10, 9.21, 12.35, and 18.77 days. The ratios between neighbors are 8:5, 5:3, 3:2, 3:2, 4:3, and 3:2, and each set of three planets is in a Laplace resonance. Chains like this are what you expect if the planets migrated inward in the gas disk and were captured one after another. The data suggest the planets most likely formed farther from the star and moved in, although they may have formed where they are.

Most planetary systems do not look like TRAPPIST-1. Most exoplanets are not in resonance at all, and one explanation, from Izidoro and colleagues (2017), is that resonant chains are often broken after the gas disk disperses, when planets scatter into new orbits. Seen that way, TRAPPIST-1 is a rare survivor: a system whose original arrangement was never disturbed.

Why migration matters for habitability

It is tempting to treat migration as a nuisance for life. A Jupiter that migrates inward can sweep through the inner system, so systems with hot Jupiters are unlikely to hold intact Earth-like planets in the habitable zone (the range of distances where liquid water could exist on a surface). Our article on the habitable zone explains that concept.

But migration can also help. In the Grand Tack picture, Jupiter's movements scatter water-rich primitive asteroids inward, which could be one route by which Earth got its water. That link is a hypothesis, not an established fact.

TRAPPIST-1 is where this is being tested. Planets e, f, and g are usually cited as being in the habitable zone. The planets' densities are lower than Earth's, which hints at substantial volatile material, and one formation model suggests about 10% water by mass. If the planets formed farther out and migrated in, that would explain why they may be water-rich. JWST has begun to probe them: the day side of TRAPPIST-1 b has a brightness temperature of about 508 K and shows either no atmosphere or a very thin one, and a thick Venus-like carbon dioxide atmosphere is ruled out for TRAPPIST-1 c, although a bare surface or other atmospheres are still possible. For more on how astronomers read those worlds, see the exoplanet atmosphere revolution.

Planets that orbit two stars

Migration is also a possible answer to a puzzle about planets in binary systems. Planets that orbit both stars of a pair ("circumbinary" planets) tend to be found just outside the zone where orbits would be unstable. One reading is that they formed farther out and migrated in until they approached the unstable edge. For more on those systems, see our article on binary star systems.

How astronomers test migration

No one has watched a planet migrate through a disk, so astronomers rely on indirect tests:

• Orbital architecture. Resonant chains, period ratios, and the spacing of planets in multi-planet systems.
• Orbital orientation. Whether a hot Jupiter's orbit is aligned with its star's spin, which separates smooth disk migration from violent scattering.
• Companion statistics. How often hot Jupiters have distant massive companions, or close neighbors.
• Composition. The density and atmosphere of a planet can show whether it formed in a volatile-rich region, which is part of the TRAPPIST-1 work.
• Young systems. Observing planets while the disk is still present, and comparing the population of young and old stars.
• Computer simulations. Models that run thousands of versions of a system and ask how often the observed arrangement appears. The 2% figure for a Mars-like planet in some Grand Tack versions is a result of this kind.

Common misconceptions

"Migration means planets spiral into their stars." Sometimes planets are lost that way, but migration often stops at a halting zone, the inner disk edge, or a resonance, which is why compact systems like TRAPPIST-1 exist.

"Hot Jupiters prove disk migration." They show that something moved giant planets or built them in place, but the tilts and companions suggest that at least some arrived by scattering or Kozai cycles.

"The Grand Tack is settled." It is a hypothesis with clear weak points, and alternatives can explain Mars's small mass.

"The Late Heavy Bombardment proves the Nice model." The bombardment itself is now doubted, so the Nice model has to be tested against other evidence such as Trojans and the Kuiper Belt.

"Planets move only while they are forming." Tides, planetesimal scattering, and gravitational encounters continue to change orbits for billions of years.

FAQ

Q: What is planetary migration? A: Planetary migration is a change in a planet's orbital distance after it forms. It is driven by gravitational exchange with the gas disk (Type I and Type II migration), with planetesimals, or with other planets. It explains hot Jupiters, resonant chains such as TRAPPIST-1, and possibly features of our own solar system.

Q: What is the difference between Type I and Type II migration? A: Type I applies to planets below about Saturn's mass, which cannot clear a gap. They exchange torque with spiral waves in the gas and generally drift inward quickly. Type II applies to planets massive enough to open a gap, which then move with the disk's slower viscous evolution.

Q: Did Jupiter really move in the solar system? A: Many models say yes, but the details are debated. The Grand Tack hypothesis of Walsh and colleagues (2011) has Jupiter forming near 3.5 AU, migrating to about 1.5 AU, and returning to about 5.2 AU. It explains Mars's small mass but has weak points, including doubts about the 3:2 resonance capture.

Q: What is a hot Jupiter? A: A hot Jupiter is a Jupiter-like gas giant with an orbital period under about 10 days. The first, 51 Pegasi b, was announced in 1995 and has a period of about four days. Most are thought to have formed farther out and moved in, though in situ formation is also being considered.

Q: What is the Nice model? A: The Nice model, published in 2005, proposes that the giant planets began on compact orbits and that Jupiter and Saturn crossed a 2:1 resonance after hundreds of millions of years, scattering Uranus, Neptune, and a planetesimal disk. It was built to explain a Late Heavy Bombardment of the Moon, an event now widely doubted.

Q: Why does our solar system have no super-Earths? A: It is unknown. Proposed explanations include Jupiter blocking inward migration, the Grand Tack clearing the inner system, or a past super-Earth being engulfed by the Sun. None has been confirmed.

Q: Is TRAPPIST-1 proof of migration? A: It is strong circumstantial evidence. Its seven planets sit in a chain of near-resonant period ratios (8:5, 5:3, 3:2, 3:2, 4:3, 3:2), which is what migration and capture would produce. The planets most likely formed farther out, but they might have formed in place.

What this means

For most of the history of planetary science, planets were assumed to stay where they were born, and our solar system was treated as the template. Exoplanet surveys broke that assumption: a template drawn from one system could not explain the hot Jupiters, the super-Earths, or the resonant chains found elsewhere.

Migration did not replace the older picture so much as make it dynamic. The planets we see are a snapshot of a long process in which orbits were rearranged, locked, broken, and sometimes lost. Our own system may have gone through at least two such episodes, and the evidence for them, from Mars's small mass to the asteroid belt and the outer moons, is still being weighed. The orderly look of the solar system is not a sign that it never changed. It may be what a system looks like after the changing is over.

Sources

• Planetary migration (Wikipedia review with Type I, II, III definitions and cited timescales)
• Grand tack hypothesis, after Walsh et al. (2011), Nature (Wikipedia)
• Nice model (Wikipedia)
• Hot Jupiter (Wikipedia)
• Super-Earth (Wikipedia)
• TRAPPIST-1 (Wikipedia)
• Circumbinary planet (Wikipedia)

---

Related Reading

• What Are Exoplanets and Could Any of Them Host Life?
• What Is Planetary Differentiation?
• Pluto and the Kuiper Belt
• How Does the Moon Affect Earth?