This is my first question here, so any help would be greatly appreciated. I am trying to figure out how my world's day and night cycle works, or if something like what I'm hoping for is even possible. I am building this world around multiple random small-scale ideas, so I've been having trouble making them work together with my limited knowledge of astronomy.

The world is an inhabited planet that is smaller than Earth, though I don't know the exact size. Possibly around the size of Mars. There is a small archipelago somewhere on the planet (I don't have a definitive map yet), of which the largest island is the only inhabited one, where the main idea is that it is night for a large portion of the year. I know that some places on Earth are like this, but if possible I would like this area to spend more of the year experiencing nighttime than daylight. For example, 9 months of darkness and 3 months of light in summer (I don't actually know what the proportions would be, basically anything that would actually work would be fine as I'm still in early stages). The rest of the planet, or just a section of it large enough to hold the main continent (it has one main continent, that archipelago, and several small continents and many islands that don't really matter) would ideally have either a "regular" day/night cycle or possibly the opposite of the archipelago area. It needs to have a fair amount of daylight, and the archipelago's long nights should seem special. That is what I'm hoping for, but it is still very flexible at this point.

For other information on the planet and context, this is a fantasy world (magic exists). Magic could be part of the explanation if need be, however, the magic system actually depends more on this cycle, which is why I'm trying to figure this out first. There are some other planets in the solar system, which has one sun to keep things simple. The focus planet has at least one moon, though it might have more. At some point in the past (a long time ago) the moon was hit by an asteroid or something-again astronomy is not my strong suit- resulting in pieces of debris from the moon still in orbit around the planet. Those are all the possibly relevant details I can think of.

Basically, I am wondering if something like this could theoretically work, and if so, how?


  • $\begingroup$ Please clarify your specific problem or provide additional details to highlight exactly what you need. As it's currently written, it's hard to tell exactly what you're asking. $\endgroup$
    – Community Bot
    Sep 22, 2022 at 1:22
  • 1
    $\begingroup$ Does the night-time of roughly 9 months have to be consecutive or would every night being several hours longer (but not 24 hours) than everywhere else also count? Because if it does, I got a potential idea with a tidally locked moon that has a debris cloud from said Asteroid impact. $\endgroup$ Sep 22, 2022 at 14:12
  • $\begingroup$ You can lock a planet tidally... one side then experiences 100% darkness. That's unhelpful. You can then make the system a binary star system, a much smaller star orbits the larger outside the orbit of the planet. The period could be such that it experiences very infrequent light. Good luck with parameters where a) the dark side isn't a frozen wasteland b) the planet isn't so close to the star that its dayside fries Mercury-style, c) the second star provides enough light that it resembles day when it is visible, and d) enough light that plants can grow and you can have a biological cycle. $\endgroup$
    – John O
    Sep 22, 2022 at 14:17
  • $\begingroup$ As an FYI, the longest "night" on Earth is not quite 6 months long. Then there will be a fairly short period of time while the Sun bobs above and behind the horizon, then will not set again for another not quite 6 months, when it will bob about a bit, and then set again for another almost half year. $\endgroup$ Sep 22, 2022 at 16:46
  • $\begingroup$ Although it won't reproduce what you've described, if you're interested in chaotic day/night cycles, for which the length of day is unpredictable, there is real-world precedent: en.wikipedia.org/wiki/Chaotic_rotation Those moons' chaotic rotations are driven by uneven mass distribution and gravitational interactions with multiple nearby bodies. $\endgroup$ Sep 22, 2022 at 18:23

9 Answers 9


Have your planet tidally locked to one star in a high-eccentricity binary system.

For most of the year, the main continent faces its parent star and is in total sunlight. The center of this region is a scorching desert, but as you reach the edges (especially the regions that have daily twilight due to the slight eccentricity of the planet's orbit), the climate becomes pleasant. As you venture further into the dark zone, however, convection currents pick up and the weather becomes dangerous. Constant hurricane-force winds block access to the mysterious island, and although they bring warmth and moisture to the region, the clouds they bring block the dim light from the sister star:

Far Orbit

But once per year, the small but bright sister star plunges quickly towards the planet and primary star. The main continent gets brighter, to be sure, but the dark island begins to receive enough light that the winds die down, the clouds clear, and the region experiences several weeks of what we'd consider a "normal" day/night cycle:

Near Orbit

But this doesn't last. As quickly as the sister sun arrived, it leaves again, vanishing into the distant reaches of space. The island cools, the wind picks up, and clouds darken the skies once again.

Additional Info:

The planet orbits very close to its parent star. This is necessary for the orbit to remain stable in the presence of the sister star, and for its rotational period to be on the order of a day. This means the star needs to be dense but dim. Universe Sandbox has "carbon stars" that meet this criteria, though it required a bit of tweaking. A perfectly circular orbit would mean the star remains fixed in the sky, but a slightly elliptical one means it "wobbles".

enter image description here

The planet+star orbit a much more massive star in a highly elliptical orbit. This is necessary to get the seasonal aspects of the dark island. When the stars are far apart, the distant one contributes very little energy to the planet. When they are close, however, its brightness rivals that of the main star.

enter image description here

The surface can be made habitable by tweaking values; I was able to achieve a temperate climate by melting the ice caps. I don't know how stable such a configuration would be, but in the sim it lasted at least 50 years without exploding or turning into a black hole, which means it's probably close enough to be plausible in fiction.

enter image description here

  • $\begingroup$ The weather effects you described sound really interesting. Having storms like that around the island would work for my story. To clarify though, because astronomy is not my strong suit- the planet is tidally locked to one of the stars giving a side of it constant daylight. And the stars have a highly eccentric orbit around each other. Or one of them does? Would the orbit of the planet be inside or outside the sister star's orbit? Also, how much would this affect average temperature of the different areas? $\endgroup$ Sep 23, 2022 at 23:20
  • $\begingroup$ I think you must have modelled your scenario yourself! Big upvote for the sweet gifs! That 500 bounty is destined for you, Moose. More modelling please! $\endgroup$
    – Willk
    Sep 24, 2022 at 15:08

One possibility is with a highly eccentric (elliptical) orbit, plus axial tilt.

The orbits of the planets are mostly circular, so the periods of constant night at the poles are roughly equal to the periods of constant day. But if the orbit was highly eccentric, then one of the poles would spend more time per year pointed away from the star, and and therefore have more night than day. The opposite would be true on the opposite pole. In contrast, places near the equator would continue to have a regular day/night cycle.

One major caveat - an orbit with a high eccentricity would probably not be ideal for habitability, with the amount of solar radiation varying significantly throughout the year. So the temperature ranges will be much higher.

  • $\begingroup$ This works, but tends to make the night-time region very arctic. If anyone wants to do anything outside without the air freezing solid in their lungs, it's probably a no-go. It's a good answer, it's just the constraints of the problem are impossible to satisfy well. $\endgroup$
    – John O
    Sep 22, 2022 at 14:11
  • 2
    $\begingroup$ OP allows solution to be magical so maybe some magical lifeform [say, warmth magic trees] could be compensating for temperature [and other] problems arising from this answer? $\endgroup$ Sep 22, 2022 at 14:56
  • $\begingroup$ I probably should have mentioned that the inhabitants aren't humans. I think I could figure out something with magic like Failus Maximus said. I'll look into eccentric orbits, I think that sounds like it could work. Thank you! $\endgroup$ Sep 23, 2022 at 23:01
  • $\begingroup$ The high thermal capacity and conductivity of your planet (large ocean) will help moderate the temperatures, but you aren't going to get away from the cold with long nights.. How much colder is it after a mere 12 hours of darkness where you live? Perhaps the small planet has some magic keeping ocean currents circulating. $\endgroup$
    – user121330
    Sep 24, 2022 at 12:50
  • $\begingroup$ You might be able to have a particularly thick atmosphere (he says the size is as big as Mars, but maybe with a much higher density than Earth) or maybe there's some biological process that puts a thicker atmosphere in place relative to Earth. Either way, that would mitigate the extremes to a decent extent $\endgroup$ Sep 25, 2022 at 19:08

Persistent eclipse

Your planet has a ring system. The ring's plane is very closely aligned with the rotation plane of the planet around its star, presumably because they all derive from accretion of material from the same disc[1]. As a result, the ring casts a persistent shadow over the planet's equator.

One side of the ring is denser than the other, or there's a very slight tilt to the planet's axis (and therefore the ring is very slightly angled in relation to the sun), so that the ring's shadow has cyclical wobbles that leave your equatorial archipelago more exposed for parts of the year, and in deeper shadow for other parts. Landmasses away from the equator would experience no ring eclipses at all.

[1] not sure why the ring would remain in ring form if this is the case, maybe it's a former satellite, astronomy people please weigh in

  • $\begingroup$ A ring is unlikely to cast a persistent shadow, even if it was aligned with the sun. Reason being that rings tend to be thin (Saturn tops out at ~1km thick for instance.). The umbra of a 1km thick ring at 1AU would only be ~108km from the inner edge of the ring - and if you have enough stuff at 108km up to form even a 1km thick ring, you're shortly going to have far worse problems than a little shade. (Zero altitude in <4 days for a 1t body with 1m^2 of surface area at 108km up... short enough that this calculator is probably inaccurate.) $\endgroup$
    – TLW
    Sep 24, 2022 at 2:42

Are you looking for realistic or believable?

I'm not a fan of realistic. It's boring. Believable, on the other hand. That's how you get Tolkien's elves walking on top of snow without leaving footprints. Nimble little minxes.

Earth's days can be thought of as fairly equal (at the equator and ignoring seasons) because its rotation speed is substantially faster than its orbital period. In other words, the ratio is more-or-less 365:1.

But what if your planet was almost tidally locked in that the rotation of the planet is just faster (or slower, I'm sticking with faster) than the orbital period. Now the rotation is 1.05:1 Simplifying things a bit, you have a night that's one year long followed by a day that's 0.05 year long (in Earth terms, it's 365 days of dark followed by 18 days of light).

It should be obvious that you can't get an arbitrarily long night followed by an arbitrarily long day. Using the method I just proposed to rationalize your world's astronomy, you can only deal with the ratio of rotation speed and orbital period.

Whether or not this abides by Real World celestial mechanics is IMO irrelevant. I'm proposing the concept is believable as a consistently applicable rule of your world.

  • 8
    $\begingroup$ Wouldn't an almost tidally locked planet just result in both long days and (equally) long nights? Otherwise, I am all for believable models over strictly realistic ones, this one just doesn't strike me as particularly believable. $\endgroup$
    – Surpriser
    Sep 22, 2022 at 10:30
  • $\begingroup$ @Surpriser combining this idea with a highly elliptical orbit such that the point of interest faces the sun at the closest point, you'd get an awful lot of twilight all over that side of the planet. Some suitably-placed mountains would provide further local darkening $\endgroup$
    – Chris H
    Sep 22, 2022 at 12:06
  • $\begingroup$ Why would the rotation speed up during that 0.05? You have the Sun sprinting across for half its rotation, and then spending almost all of its time during the other half. $\endgroup$ Sep 22, 2022 at 19:18
  • $\begingroup$ @Surpriser You know, you just might be right, but you've missed the point about worldbuilding on a site that has as its primary goal the creation of imaginary worlds (help center). Real Life cannot be an overriding limitation on any question unless specifically requested and the OP did not tag the question science-based or hard-science. All I did was suggest a way to achieve the OP's goal in a way that would satisfy suspension-of-disbelief. $\endgroup$
    – JBH
    Sep 24, 2022 at 1:36
  • $\begingroup$ @Surpriser Now, your next comment is likely to be along the lines of, "I don't believe it... I even said that in my comment." As I've said to many users of this Stack, astronomy and/or physics experts are not their target audience. The folks who haunt this Stack tend to be very science-heavy people. In other words, you've done a poor job of putting yourself into the shoes of the people most likely to read the OP's story. $\endgroup$
    – JBH
    Sep 24, 2022 at 1:38

Not a full answer, but points to consider, too long for a comment:

On earth every place has 50:50 of light and darkness within the period of one year. Differences of day/night length between summer and winter are caused by the slight pitch of earth' rotation axis. If you are on the northern half of earth, you'll experience summer during the time when the north pole end of the rotation axis looks toward the sun.

To have another light/darkness ratio than 50:50 during the year, the rotation axis would have to change. This is likely quite unstable, yet I cannot think of a stable orbit around the sun with a changing rotation axis of the planet.

Currently I cannot think of a stable setup that would provide 75% daylight during one year.

A special case is our moon that looks at the earth with the same side all the year. Its own rotation fits its orbit around the earth. A planet that looks at the sun the same way like the moon looks at the earth would have daylight throughout the year on one side and everlasting night on the other side.

Whatever your planets stable rotation axis would be, a change of the rotation axis would be needed to change its day/night cycles. This, however, would need a really hard impact that likely changes more than the day/night cycle. For example, an asteroid hitting angulary the equator might cause a quicker or slower rotation. This in turn might change the planet's orbit around the sun (though, I am not entirely sure.) An asteroid hitting angulary the north pole might change the rotation axis.

  • 1
    $\begingroup$ The 50/50 ratio is a key data point for a circular orbit. Though for an asteroid impact of such magnitude as to alter rotation speed and/or orbit would result in returning the planet's crust to a molten state. $\endgroup$ Sep 22, 2022 at 16:54

Daily solar eclipses

On Earth, there are two cases, when the sun is completely obstructed: nights and solar eclipses. So let's extend the night by eclipses. For most practical reasons, it is pretty much the same thing - it is dark outside, people need torches, vampires can leave their coffins, etc.

Let's have a planet with zero inclination (relative to the star) and a moon on a geosynchronous orbit with nonzero inclination (relative to the planet). There is a place on the equator (let's call it the Eclipse Island), where the moon covers the sun twice a day - in the morning and in the evening.

Surface view on geosynchronous moon with inclination

This results in multiple distinct time periods on the Eclipse Island:

  1. Proper night (both the moon and the sun still under the horizon)
  2. Moon dawn (the moon rises over the horizon, still eclipsing the sun)
  3. Sun dawn (the moon leaves the sun, letting in some of its light)
  4. Day (the moon travelling West side by side with the sun)
  5. Sun dusk (the moon comes back to the sun disc, obscuring some of its light)
  6. Moon night (the moon eclipsed the sun again, but is still over the horizon)
  7. Proper night follows again, when the moon falls under the horizon

The length of dawns and dusks would depend on the size of the moon and its inclination, so you can fine-tune them to your liking.

Side effects

The same daily eclipses would happen also in some places around the Eclipse island. If you sailed North from there, you would experience both eclipses closer and closer to the noon. Go North enough and you will have three short "days" - one before the first eclipse, one after the second eclipse, and one in between them. There will be a small spot, where these three short days have the same length (followed by a "proper" long night). Go even more North and the two eclipses will be so close to the noon, that they become one. Go Norther and there will be no eclipses at all.

Apart of the single area, that experiences the shadow of the moon, there will be no eclipses anywhere else. From most places on the planet, the moon will be always before the sun, or always chasing after it. If you want to bind the day-night cycle with your magic, the Eclipse Island and its surrounding "shadow zones" would definitely feel very magical.

Significantly shorter day would also mean the Eclipse Island would be significantly colder, than the rest of the equator. The longer the nights, the greater the temperature difference. Make it big enough and it can lead to some pretty extreme weather.

Which may be a good reason, why there is nobody living on the rest of the archipelago.

  • 1
    $\begingroup$ This is where I started too, but couldn't work out a way to make the moon occlude the sun for a non-trivial amount of time - good solution! One effect is that the sun would look crescent-shaped for most of the time, which would be very visually striking. Another is that there would be no perceptible "moon", or more precisely, the moon would always be new. $\endgroup$
    – Ottie
    Sep 23, 2022 at 16:46

A highly eccentric orbit tied to a 3:2 rotation lock (like what Mercury has) could give one very long day and one very long night on alternating orbits, but that's not what you asked for.

A 2:1 resonance, on the other hand, with the inhabited island on the light side at apoapsis, could give a fairly short "day" (relative to year length) once a year in which the sun may show some retrograde motion for a time between sunrise and sunset. Depending on the eccentricity of the orbit, there might be other shenanigans (like multiple sunrises and sunsets each "morning" and similar each "evening") as the rotation rate interacts with orbital position (I'd have to set up a simulation, I'm not getting a good visualization).


The planet's axial tilt determines the length of day/night cycles along with the latitude of a location. It is possible for a planet to have locations with days/nights lasting 6 consecutive months. Such a planet would have an axial tilt of 90 degrees. An axial tilt of 90 degrees puts the axis in line with the planet's orbital plane. We have a planet close to this in our solar system, with Uranus having an axial tilt of 98 degrees. An explanation for how this works is that the closer to 90 degrees the axial tilt is, the larger the artic and antarctic circles are. With 90 degrees the polar circles extend to the equator giving every location on ethe planet at least a day without sunlight. This also increases the amount of time the poles experience night and day with each pole experiencing 6 months of night and day as mentioned previously. Locations on the equator would have somewhat standard day night lengths. For a 24 hour rotation like Earth, a location on the equator would have a 12 hour day/night on the equinoxes and a 24 hour night on the solstices.

It is neat that a planet could have a location with days and nights as long as 6 months, but you want nights as long as 9 months. This would be possible with axial precession. Axial precession is when a planet's axis wobbles and changes the amount of the planet's axial tilt. If the planet's axis changes its tilt on a regular period throughout the year a 9 month night and 3 month year would be possible. I haven't looked into the proper amount the tilt would need to change but I'll give an example for how this could work. Let's say the planet fluctuates between 45 degrees tilt and 135 degrees tilt throughout the year and we place our 9 month night location on the north pole. At the winter solstice we will start with a 45 degree tilt. Our location will be in darkness since it is winter for our pole. Now with axial precession our axis will change toward 0 degrees as the spring equinox approaches. But this axial precession not only moves 45 degrees against the orbital plane but also moves the axis 45 degrees with the orbital plane. This tilts the pole away from the sun keeping it in the dark. At the summer solstice the pole will be facing the sun with a 135 degree tilt in the middle of a long sunny period. The autumn equinox would be the same as the spring equinox with the pole facing away from the sun. Now this would not make for a full 9 month night and 3 month day if the planet rotates on a regular period such as 24 hours. For example in the early spring there would be short amounts of daylight that would lengthen as the season went on until day would last longer than 24 hours until there would be short periods of night later in summer. If the planet does not rotate on an axis it would be possible for an exactly 9 month night and 3 month day. In the case of 24 hour rotation around the equator locations would have days and nights of regular lengtha.


Imperfect Peaks of Eternal Light and Craters of Eternal Darkness are potentially ways of getting interestingly asymmetrical day/night cycles.

Imagine a planet that has no axial spin, so the stars remain fixed in the sky, never rising nor setting, and no axial tilt. At the northern-most point, there is a humongous crater, whose mountaneous rim has a single gigantic chasm in it. As the planet orbits the star, the angle the sun makes with the North Pole changes, and once per year it shines through the gigantic chasm, allowing the inside of the crater to recieve light for a few days.

(Or perhaps there is a small axial tilt, and at the height of summer, the sun pokes above the crater's rim)

This crater contains water, and the archipelago in question.

Quite how the crater walls are so high is suspicious -- I'd suspect some sort of engineering, because the crater walls are going to need to be high enough to be visible from over the horizon (how else do they block the sun?) and that height grows very fast as the angular size of the archipelago grows...

Potentially relevant to your interests: a town in a valley in Norway, which is dark during the middle of the day...

  • $\begingroup$ This could also work for a planet with a very long "day" relative to its year like Venus or Mercury -- and a planet with slow rotation or a tidal resonance is much more believable than one that's has an infinite sidereal day. $\endgroup$
    – Zeiss Ikon
    Sep 22, 2022 at 16:04
  • 1
    $\begingroup$ Very true. I've drastically oversimplified to demonstrate the point. $\endgroup$ Sep 23, 2022 at 14:01

You must log in to answer this question.

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