This Query is part of the Worldbuilding Resources Article.

Once the geography of the world is designed, it needs a planetary system to inhabit. But how should that solar system look? The only constraint is that the system needs the new world to be placed in a habitable zone for the suspiciously Earth-like creatures that evolved there. By placed, I only mean it's made up, the system needs to have naturally formed in all respects.

The naive solution would be to make it like our planetary system, the Solar System.

That is, arrange it so that the planets orbit the same direction in the ecliptic plane, there are some rocky planets close to the star followed by an asteroid belt and some gas giants, like this: SRRHRAGGG (incidentally, this is the sound the planetary system will make when it dies)

S- Star
R- Rocky Planet
H- Habitable Planet/Moon
A- Asteroid Belt
G- Gas Giant

Is this the most likely arrangement, RxAxGx (rocky planet[s], asteroid belt[s], gas giant[s])? Can a massive gas giant be orbiting near the star out of the ecliptic plane? Can the habitable world be alone with some comets and asteroids?

The main question:

What is the range of planetary configurations I can reasonably expect from a habitable system?


I'm interested in the ordering of planets (mass and type), the planet mass to star ratio, number of planets, orbital direction of planets (as in agreement between planets), ecliptic plane confinement, and the reasonable range of these aspects. Reasonable meaning very precisely "not, like, super rare among habitable systems".


The system must contain a planet which has evolved Earth-like life.

The system must have been formed by natural processes.

Magic, science fiction, and anecdotes need not apply, this is . We don't know much about other systems, let alone explicitly habitable systems, so inductive reasoning is allowed (if not required) but peer-reviewed papers should support any evidence used in that process.


This is related to a series of questions that tries to break down the process of creating a world from initial creation of the landmass through to erosion, weather patterns, biomes and every other related topics. Please restrict answers to this specific topic rather than branching on into other areas as other subjects will be covered by other questions.

See the other questions in this series here : Creating a realistic world Series

  • 1
    $\begingroup$ Any answer to this question is going to have to include at least one reference to Kepler data. For sure. $\endgroup$
    – Green
    Commented Jul 28, 2015 at 20:15
  • $\begingroup$ @Green Indeed, there are a lot of papers using it. $\endgroup$
    – Samuel
    Commented Jul 28, 2015 at 20:18
  • $\begingroup$ Does your habitable planet lies within Golilock's zone so that water can exist in liquid form? does it have molten metal core to erect magnetic field? greenhouse gases if outside habitable zone? gas giant sized companion planet to clear debris at least buy time until chicxulub impact? How do parent star(s) fare in anger management are they turning big and red and about to explode? $\endgroup$
    – user6760
    Commented Jul 28, 2015 at 22:54
  • 1
    $\begingroup$ @user6760 The planet is habitable and has evolved Earth-like life. Select answers to your questions appropriately. If the answer leads to "does not support the evolution of Earth-like life" then it's the wrong answer. $\endgroup$
    – Samuel
    Commented Jul 28, 2015 at 22:58
  • $\begingroup$ @Samuel, the string 'RxAxGx' reminded me of notation I've seen in relation to context free grammar used to define programming languages. A CRG to define solar systems would be incredibly cool. $\endgroup$
    – Green
    Commented Jul 28, 2015 at 23:01

2 Answers 2


Ordering of planets (mass and type)

Can I start out by jokingly complaining that you picked a rather complex system? We've found a lot of exoplanets, but there are not many that reside in complex systems like this. This is going to be a tough question. As Green predicted, Kepler data is useful here - Fang & Margo (2012) found that

75%–80% of planetary systems have one or two planets with orbital periods less than 200 days

They also were able to plot data from a variety of parameters to come up with some graphs that could be used to make distribution curves. You can extrapolate from that, if you wish.

Anyway, I'm off track here. Mass distributions were covered in Mazeh et al. (1998) (which is almost certainly outdated, but a good analysis nonetheless) and Malhotra (2015). Using some orbital spacing parameters (which you can adjust, if you want), Malhotra found that the peak value of $\log m/M_{\oplus}$ occurs at about 0.6-1.0, with a standard deviation of 1.1-1.2. Not the greatest accuracy, but still pretty good.

Llambay et al. (2011) were able to come up with a mass-period distribution for exoplanets close to the star, which you can then use to come up with a decent distribution of masses at a given radius:

Most smaller planets have orbital periods longer than P~2.5 days, while higher masses are found down to P~1 day.

In short, more massive planets are closer in, while less massive planets are further out. Still, Llambay et al. only considered planets extremely close to their parent stars. For the rest of the system (i.e. planets farther out), I refer you to Jiang et al. (2007). I can't copy the mass- and period- histograms they gave (relating each one to the total number observed), nor can I copy the scatter plots, but they are incredibly helpful, especially as they considered a sample size of 233 exoplanets.

This graph, complied on Wikipedia from the Open Exoplanet Catalogue, is also helpful for an at-a-glance reference:

Image in the public domain.

Something you must consider is planetary migration. I've written several answers on it across Stack Exchange (e.g. The Solar System Explosion in the Nice Model, Did Jupiter really make Earth (in)habitable, What gravitational impact would moving Jupiter to the inner solar system have on the outer?, etc. - the first focused on only one part, because Kyle Oman was already familiar with the it, hence the question), and others have written excellent answers elsewhere on Stack Exchange. By now, I'm sick and tired of typing the same thing up, so I refer you to the latter two posts I gave, as a starter. You need to include planetary migration because it will severely impact the orbits of the three gas giants in the system. Be careful that you have enough - my answer on Physics discusses just why a certain number is needed.

Planet mass to star ratio

No such ratio exists. You can have pretty much any (reasonable) combination you want. It all depends on the giant molecular cloud from which the star formed and the evolution of the protoplanetary disk. Anything can happen.

Number of planets

Fang & Margot are, once again, helpful. Weissbein et al. are also an excellent resource for this specific part. I once again wish I could figure out directly how to copy graphs and histogram without using imgur - I may use that later - but I can get around that. Unfortunately, they make three assumptions:

  1. All planets in a system are exactly aligned
  2. All of the stars and planets are identical
  3. The Occupancy distribution of a planet existing at a given distance from its stellar host, f(r), is the same for all the stars which are capable of producing planets and is given by equation (1).

The third is not a problem, but the first two are (see my section on ecliptic plane confinement for a discussion on the first). Luckily, as I show later, that criterion can easily be met. The second one is the problem.

Anyway, Weissbein et al. find the probability, $P$, that a star hosts $m$ planets to be $$P(m)=\int_0^{\infty}\left[\frac{F(r)^m}{r^2m!}e^{-F(r)}\right]dr$$ where $r$ is radius and $$F(r)\equiv \int_0^r f(r')dr'$$ where $f(r')$ is a modified form of the general occupation probability function.

They then used this to create a table of the results, which I will not include at the moment, as I am not good with tables in Stack Exchange. However, predictably, the number of systems went down as the number of planets increased.

Ecliptic plane confinement

"Ecliptic plane confinement" can be discussed in terms of orbital inclination, generally denoted by $i$. In the case of most systems, this is close to zero degrees for most of the bodies involved (although Pluto has a high inclination).

The planets in the Solar System orbit in one plane, because everything formed out of a protoplanetary disk. The planets tend to stay that way because of a conservation of angular momentum. This can change in some cases - notably, Kepler-452b has a high angle of inclination (90 degrees!). As I wrote in my answer there, this may have happened for several reasons:

  • The star's rotation axis was perturbed, just as Uranus's rotation axis was perturbed (although by different objects)
  • The planet was perturbed by another object, either in the system (e.g. a planetoid) or a companion star. The Sun was formed with many other stars in a cluster; this happens for many stars.

The relevant papers on this subject are Crida & Batygin (2014) and Xue et al. (2014). There are other reasons for the change in orbital inclination of one planet, notably the Lidov-Kozai mechanism (see Lidov (1962) and Kozai (1962)). The Lidov-Kozai mechanism basically states that the eccentricity of an object's orbit can be changed by interactions with another (more massive) object, which also changes the orbital eccentricity of the first object. The angular momentum in the $z$-axis must be conserved here; it is the quantity $$L_z=\sqrt{1-e^2}\cos i$$ You can play around with this a bit to see what happens when different parameters are changed (you should be able to apply the orbital formulas given here). However, the model assumes that the perturber is much more massive than the perturbed object (Kozai's original analysis applied to perturbations of asteroids by Jupiter!). For larger bodies being perturbed, you would need a larger perturber. This makes it very difficult for planets. This could happen in a binary system where one star is more massive than another star, and the second star perturbs a planet moving around the larger star. It is, however, unlikely, and does not fit your model of one star.

It makes sense that either most or the orbits have high orbital inclinations - a result of a perturbation of the star's rotation axis or the protoplanetary disk - or low orbital inclinations. The Lidov-Kozai mechanism is not good for large systems. It is also important to note that it is periodic in nature. Once again quoting Fang & Margot,

In addition, over 85% of planets have orbital inclinations less than 3◦ (relative to a common reference plane).

They used a Rayleigh distribution to describe this: $$P(k)=\frac{k}{\sigma^2}e^{-k^2/\sigma^2}$$ where $\sigma$ is the parameter that determines the distribution of $k$. Notice the difference between a Rayleigh distribution and a normal distribution. An distribution for orbital eccentricity can be found in Kane et al. (2012).

Bringing it all together.

There's the raw information that we need. Here's the synthesis.

Is this the most likely arrangement, RxAxGx?

Well, it's unlikely for that many planets to form around a star, so technically, no. Three gas giants implies orbital migration, which could push them outwards, as in our Solar System, but be prepared to have a fourth be in there at the beginning, as some variants of the Nice Model require (the "5th gas giant").

Can a massive gas giant be orbiting near the star out of the ecliptic plane?

I stated earlier that the perturber generally needs to be more massive than the perturbed object, in classical models of the Kozai effect. This means that such an arrangement is unlikely to happen. A gas giant could be close to the star, sure, but not out of the ecliptic, if it was with a system of other planets that stayed in the ecliptic.

Can the habitable would be alone with some comets and asteroids?

Asteroids? Sure. Well, the habitable could not be in the asteroid belt, because then it would not have cleared its orbit and would be prone to collisions, which would quickly make the planet not so habitable!

The arrangement, on the whole, could happen.

  • 2
    $\begingroup$ Very nice work. I am impressed. Some notes: 1) From a desktop you can snip images from PDFs and upload them here. 2) The system doesn't need to have only one star, it just needs to be habitable. 3) The 'RxAxGx' notation was meant to demonstrate rocky planet(s), asteroid belt(s), and gas giant(s) not necessarily one of each. I'll make that more clear. $\endgroup$
    – Samuel
    Commented Jul 28, 2015 at 22:28
  • $\begingroup$ @Samuel Thanks for #1, I didn't know that. #2 is very helpful. I understood #3, by the way, that didn't affect my answer. $\endgroup$
    – HDE 226868
    Commented Jul 28, 2015 at 22:30
  • 1
    $\begingroup$ I had developed my own empirical suite of formula to calculate this stuff for me. But I wrote mine before Kepler started returning data. I'm going to use your answer to reprogram my own algorithms. I know this board frowns on "thanks" but "Thanks!" (also +1 for question and answer). $\endgroup$
    – Jim2B
    Commented Aug 5, 2015 at 17:35
  • $\begingroup$ @Jim2B Cool, thanks! Let me know how it goes. $\endgroup$
    – HDE 226868
    Commented Aug 5, 2015 at 17:39
  • 1
    $\begingroup$ One thing I would like to see is that Kepler and other planet search programs favors spotting large planets over small planets and short period planets over long period planets. Somewhere there should be a statistical treatment of this applied backwards to give us an idea of what the over all population of planets should look like based upon what we've seen so far. But I haven't seen any and I don't have the statistical background to do this for myself. :( $\endgroup$
    – Jim2B
    Commented Aug 6, 2015 at 1:09

Rocky/Gas Configurations

Anything is possible but it is no coincidence that the solar system has its most dense planets at orbits closest to the sun and gas giants a lot further out. The greater temperatures and solar wind pressures nearer the star will be pushing lighter elements away more easily from inner orbits when the star's reactor fires up.

To quote https://en.wikipedia.org/wiki/Formation_and_evolution_of_the_Solar_System:

The inner Solar System, the region of the Solar System inside 4 AU, was too warm for volatile molecules like water and methane to condense, so the planetesimals that formed there could only form from compounds with high melting points, such as metals (like iron, nickel, and aluminium) and rocky silicates.

Heavy rocky planets nearer the star and gas giants further away are probably the most likely configuration with only peculiar formation events changing that.

Independent rocky planets in the zone of the gas giants don't exist in our solar system but rocky moons of those gas giants do. So, rocky things can and do exist at any distance but gas giants tend to be further away and capture or destroy anything in their path.

That gas giants don't exist beyond a certain point is probably simply a matter of the solar nebula being too thin after certain distances.

Binary to trinary systems could perhaps produce different situations. If Jupiter had been big enough to be a red dwarf then we would have a complex binary system: Still probably with rocky planets between the two stars but perhaps an even larger rocky system orbiting Jupiter. However, we expect multiple-star systems to be generally less conducive to stable orbits as close as the habitable zone.

Orbital Planes and Directions

All planets formed within the system will be orbiting in the same plane and in the same direction: they will be constructed in that orbit from the same rotating primordial mass.

However, collisions or gravitational interactions with a extra-stellar objects could feasibly knock a planet into a slightly different plane of orbit. It would have to be from an extra-stellar object, everything else has the same angular velocity vectors so collisions between system bodies just knock things into different motions in the same plane.

A solar system that moved too close to a particularly massive neighbor could end up with planets in orbital planes that slope and become more elliptical the further out they are. That is far more likely than a rogue extra-stellar planet crashing into one of the planets and producing a new planet orbiting in a different plane and/or direction.

Captured Planets

Capturing an object in orbit is extremely unlikely: they tend to follow parabolic or hyperbolic paths and leave the encounter with the same kinetic energy they arrived. Only collisions during the process can change that and they are highly unlikely to happen let alone happen in just the right way to cause the visitor to start orbiting.

However, it is possible: Neptune captured Triton (we know this because Triton orbits the opposite way to Neptune's rotation) so an extra-stellar planet could be captured and that could happen in any plane or any direction. It had to involve a collision so the result could be another planet also orbiting in a peculiar plane and direction.

The Star(s) Itself

Blue giants and super giants (O and B class) are almost certainly to be ruled out: they are too short lived.

Other giants are also likely to be ruled out as they tend to be dying stars: the habitable zone will have moved and that causes problems for evolution. Their lives as giants also tend to be short.

But anything else except dead stars (White Dwarfs, Neutron Stars and Black Holes) is feasible.

Only nice orange, yellow or white main sequence stars seem likely to produce habitable zones that are stable for long enough for life to evolve:

enter image description here


We live in one of hundreds of billions of galaxies each with hundreds of billions of stars. I expect every imaginable configuration exists at least around one of those stars.


You must log in to answer this question.

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