If suddenly "knocked" or perturbed from its orbit, would gravity eventually return the Earth to its original orbit?

I am curious as to whether this is even possible.

It seems to me that since the orbit is based on gravity, the same gravity will pull it back to correct orbit eventually. Is this reasoning right?

  • 3
    $\begingroup$ Welcome to Astronomy! I've updated your title so that it better fits this site. Can you double check that it still matches the intent of your question? Thanks! $\endgroup$
    – uhoh
    Nov 17, 2020 at 13:11
  • 4
    $\begingroup$ I can't explain, But one-word Ans is No. Actually solar system as a whole is a complex phenomenon $\endgroup$
    – crabNebula
    Nov 17, 2020 at 14:27
  • 3
    $\begingroup$ Do you mean literally knocked by something huge colliding with it, or just perturbed by the gravity of a large rogue body passing nearby? In the former case, anything with enough momentum to make a significant difference to Earth's velocity is likely to do significant damage. And in the latter case, I'd also expect it to do enough damage to make Earth uninhabitable (at the very least). $\endgroup$
    – PM 2Ring
    Nov 17, 2020 at 15:18
  • 1
    $\begingroup$ I think your reasoning is wholly wrong. "The same gravity…" will not pull it back, because "the same gravity" will be applied to an Earth in a different situation. It's the situation - IE, the Earth's relationship to the other bodies - that matters. That is so much true, there isn't real any such thing as "the same gravity…" $\endgroup$ Nov 18, 2020 at 1:38
  • 1
    $\begingroup$ @Toby Good point. It'd be interesting to calculate how much Earth's orbit was perturbed by the impact with Theia (assuming that theory is correct). $\endgroup$
    – PM 2Ring
    Nov 18, 2020 at 12:18

4 Answers 4


There's a few parts to this question so there's more than one answer.

Earth gets knocked a little bit out of its orbit all the time by gravitational influence of other planets in our solar system. Jupiter and Venus are the primary two, but all the planets have some effect. These are called orbital perturbations and they tend to alternate, not add up. They're the causes of Earth's Milankovich cycles and the effect isn't negligible. These variations in Earth's orbit, which take thousands of years to move back and forth can cause the formation and recession of glaciation, sometimes called ice ages.

These orbital perturbations, in addition to being cyclical, as a rule, don't effect Earth's semi-major axis which is very consistent, and may only change slightly as the Sun loses mass.

Orbital perturbations can to lead to much greater variations of eccentricity and axial tilt. Mars, for example, undergoes much bigger variations than Earth, but it still undergoes a cyclical back and forth that leads to general stability within a range. All eight known planets in our solar system are thought to be relatively long term stable.

The exception to this cyclical back and forth is if there's a resonance where the effect can grow over time. The closest example of this in our current solar system is Jupiter and Mercury, where they're not in, but they're close to resonance and it's possible that Mercury will be tossed out of its orbit in a few billion years. It's the most unstable planet in our solar system.

This article isn't published, but I still think it's a good summary of both the stability of orbits in our solar system and the possible (but perhaps unlikely) destabilization of Mercury in a few billion years.

Planetary migration is another means for tossing a smaller planet out of its orbit. This is thought to be fairly common, based on observations of other solar systems and some uncertainty that gas giant planets could form close to their suns. (there may be room for debate on that), but migrating planets is thought to be fairly common, if somewhat slow. It's never been observed but it can be modeled. It's been suggested that when our solar system was young, Jupiter moved inwards, perhaps tossing Uranus and Neptune outwards and reducing the material available that would eventually become Mars and when migrating back outwards, leading to the late heavy bombardment which brought water to Earth. This is called the grand tack hypothesis.

An inwardly migrating Jupiter could certainly toss Earth about easily into a completely different orbit, but there's no evidence that suggests Jupiter is likely to migrate that much in the future.

A massive enough object from beyond the solar system could toss Earth into a different orbit and in doing so, not only change Earth's eccentricity and orbital plane, but change Earth's semi-major axis as well and that change would effectively be permanent.

A gravity assist of that kind could give Earth both a new eccentricity, a new orbital plane and a new semi major axis and perhaps a new set of Milankovich cycles, though I suspect over time, the orbit could re-circularize and the Milankovich cycles, largely reset to where they were, so there could be some correction not by the sun, but by the other planets orbiting the sun, but not everything would go back to where it was. The Semi-major axis change would likely be permanent and there would also be the question of whether the new orbit was in near resonance with other planets which would lead to further changes over time.

A higher eccentricity could create stronger seasonal variations and perhaps trigger a new ice age, or maybe fix man made global warming. :-)

The effects of a change in semi-major axis would change the length of a year, so, we'd need new calendars, and could warm or cool if it was big enough.

There would also be tidal concerns if a massive object passed close enough to significantly change Earth's orbit and perhaps even, pulling the Moon out of its stable orbit.

The good news is, space is big and quite empty and an object massive enough to change Earth's orbit passing close enough to do so is extremely unlikely.

  • $\begingroup$ This and other answers have the planet changed to a different orbit. But, if something were to pull earth further away from the sun, without changing its “forward” speed, wouldn’t that allow it to continue moving further away? Because it would be faster than an orbit at that distance requires? $\endgroup$
    – WGroleau
    Nov 18, 2020 at 3:46
  • 1
    $\begingroup$ @WGroleau that's an extremely unlikely scenario - the perturbing body would have to come it with the exact right speed, angle, and mass to get that effect. If it does happen, Earth is "too fast" for a circular orbit, so it will got into an ellipse that reaches out to whatever distance its velocity will carry it. $\endgroup$
    – user132372
    Nov 18, 2020 at 11:49
  • $\begingroup$ @WGroleau In a simple Kepler orbit scenario (one body orbiting a much more massive central body), if the orbiting body receives a single impulse at point X it gets perturbed into a new orbit, but that new orbit still passes through point X. Things are more complicated in the real Solar System, but the Keplerian analysis is still a good first approximation. $\endgroup$
    – PM 2Ring
    Nov 18, 2020 at 16:28
  • 1
    $\begingroup$ No mention of angular momentum - which is the key quantity. $\endgroup$
    – ProfRob
    Nov 18, 2020 at 16:50
  • 2
    $\begingroup$ You're missing the obligatory HHGTTG quote. :) $\endgroup$
    – Barmar
    Nov 18, 2020 at 17:22

In a 2 body scenario: Lets say there is one large body (sun) and a smaller body (asteroid)

The orbit of the asteroid is an ellipse around the sun and completely stable. If you knock the asteroid it will change the orbit to a different ellipse, and that new orbit is also completely stable. The asteroid won't change back to its orginal orbit unless you give it another knock that reverses the effects of the first.

The same is true of any two body system such as Sun-planet (except a knock that is big enough to change the orbit of the Earth by a large amount is probably enough to wipe out life on Earth)

Gravity would return the asteroid to the same position, but with different velocity (the new velocity that it had after the knock) and so the asteroid would be in a new orbit.

In more complex systems with three or more bodies, resonant orbits can exist whereby an asteroid has an orbit which is has the exact ratio of orbital periods with a planet. In these orbits a gentle enough knock on the asteroid will be perturbed back towards the resonant orbit. Note that the Earth is not in a resonant orbit with any larger body. So this doesn't apply to the Earth.


There is no "correct" orbit. In a two body system, no orbit is special; each is equally valid. If the Earth is perturbed into a different orbit, the new orbit is "just as good" as the original one.

  • 5
    $\begingroup$ For humans, I think "correct" means "the one we're used to". ;) $\endgroup$ Nov 17, 2020 at 21:31
  • $\begingroup$ @LawnmowerMan I think most people would include "circular" in the definition of "correct". Would the orbit of a comet be considered incorrect even if it's well known? $\endgroup$ Nov 20, 2020 at 3:49

In general, the answer is no, but I give a possible exception below.

The earth's orbit is pretty darned Keplerian, since the Sun is much more massive than the earth, and since gravitational forces exerted on the earth from other bodies are very small compared to the forces exerted by the sun. For Keplerian orbits, Kepler's orbital elements are stable without perturbations. So, if you change the orbital elements through a perturbation, they won't change back unless another perfect perturbation occurs that resets the elements to their original values.

But it's statistically impossible for an orbit to be perturbed into a new orbit and then perturbed back to it's original orbit, right? Nope, the retrograde asteroid 2015 BZ 509 orbits the sun and has it's orbit perturbed to a new orbit and then perturbed back every orbit by Jupiter. Per the article in Nature:

The retrograde motion of 2015 BZ509 relative to Jupiter is also depicted in Fig. 1. Two relatively close passes to Jupiter take place each orbit, and for stability the effects on the orbit due to each pass must cancel out.

Of course, this means Jupiter's orbit also gets changed slightly each pass. If Earth had a hugely massive 1:1 resonant retrograde asteroid like this, then the earth's orbit would continually be perturbed into a new orbit and then perturbed back, once every year!

  • $\begingroup$ The first paragraph might be a bit much for a layperson. Maybe clarify? "The orbit is an oval, and any bumps will usually simply shift earth to a different oval shape." $\endgroup$ Nov 17, 2020 at 19:14
  • 3
    $\begingroup$ But Earth does have a 1:1 resonant asteroid, so its orbit does indeed change, but not noticeably because of the vast difference in mass. en.wikipedia.org/wiki/3753_Cruithne $\endgroup$
    – Mike Scott
    Nov 17, 2020 at 19:16
  • 4
    $\begingroup$ @MikeScott Gosh, I did not know we had a 1:1 resonant asteroid to the earth. Your comment leaves me mildly perturbed. $\endgroup$
    – Connor Garcia
    Nov 17, 2020 at 19:58
  • 1
    $\begingroup$ @MooingDuck It gets kind of tricky since a perturbation may affect only some of the orbital elements. For example, if an asteroid strikes the earth orthogonal to the ecliptic, for example, the orbit could change, but the oval shape (eccentricity) and size (semi-major axis) could stay the same. Also, I think it's ok to have some heavy jargon answers to this question. There are several other excellent answers, so the OP and voters can certainly choose one that they think explains things better! $\endgroup$
    – Connor Garcia
    Nov 17, 2020 at 20:19
  • $\begingroup$ I'd consider your last paragraph, if technically not quite wrong, at least highly misleading. The effects that even a closely passing asteroid could have on the Earth's (much less Jupiter's) orbit are so vanishingly small that they're surely lost in the noise from other random perturbations. To a first approximation, compared to actual planets, asteroids are massless. $\endgroup$ Nov 18, 2020 at 18:25

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .