This is a follow up question to Sky of Earth and Seas of Sky

I have a medieval fantasy setting where people live in a vast expanse of giant caves contained in a sort of giant Stanford torus. This is not quite a true Stanford torus though; it was created from a rapidly spinning disk shaped planetoid about 100,000-200,000 years ago. Rather than building a whole mega structure themselves, the "gods" who built this habitat took this pre-existing planetoid and just encased it in a "firmament" which is an unobtanium clear shell that keeps the oceans and atmosphere from flying away.

The result is that the outermost level of these caves have many lakes and oceans through which you can see a sun, moon, and stars, and the water has a day night cycle. The sun shines through bright enough to provide light for a dense forest of plant life on the ceiling of the cave. The apparent gravity is about 1G and points outward toward the seas due to the planetoid's spin imparting a greater than gravity outward force. Most of the forests are suspended hundreds of meters above the water.

My original thought was to feed this canopy of trees with water from the clouds that form off of the lakes and oceans similar to a Cloud Forest, but as I've read more into this I don't think this will work. Cloud forests have the water vapor condense on the leaves and then fall to the ground as rain where it moistens the soil. But, in this case, the condensation would fall away from the roots. I'm also concerned about soil fertility since dead plants would presumably fall away from the ground both choking up the seas by blocking out the sunlight and depriving the ceiling of nutrients needed to sustain the forest.

This question: How can I water/soil upside-down trees and grass? is similar, but in their scenario, they were not really looking for a particularly fertile or ancient upside down ecosystem; so, they were able to handwave away a lot of my concerns that it was a mediocre ecosystem for plant life.

How would I create a stable ecosystem where this forest gets enough water, nutrients, and sun light to be sustainable?

A best answer will be one that does not require any of the builders' technology/magic to still be doing things to maintain the cycle, but this is not strictly necessary.

The trees themselves are assumed to be evolved or engineered for this environment; so, while I need them to be photosynthetic tree like organisms, they do not necessarily need to be any existing species of tree. A best answer will stick to adaptations that already exist in plants, but this is also not strictly necessary.

I need the seas to remain clear enough to see the sun, moon, and stars; so, a best answer also needs to avoid situations where the water gets all mucked up with plantlife and debris over time; so, if a method does not actively recycle the nutrients back up to the plants, it should explain how the water is kept clear, where the plants get more nutrients from, and why such a method would be sustainable over hundreds of thousands of years.

enter image description here

enter image description here

  • $\begingroup$ Does the cave become flooded as tide rises? $\endgroup$
    – user6760
    Commented May 22, 2020 at 8:30
  • $\begingroup$ i dont quite get your question but maybe try checking banyan tree and mangrove ? for your tree to get the water, like some stalactite type of roots, i dont know how to fertilize the soil, but maybe create something to create burning reaction to get ash or something to fertilize the soil. $\endgroup$
    – Li Jun
    Commented May 22, 2020 at 10:21
  • $\begingroup$ @user6760 no, it does not. $\endgroup$
    – Nosajimiki
    Commented May 22, 2020 at 13:06
  • $\begingroup$ The difficulty here is understanding why the trees don't fall down to the surface? $\endgroup$
    – Slarty
    Commented May 22, 2020 at 15:33
  • $\begingroup$ @Slarty Many kinds of plants including trees grow into solid rock, just not usually upside down. I don't think seeds will necessarily be a good means of reproductions for them, but I should be able to solve this by making them runner plants. Such plants obviously don't exist on Earth; so, for the purposes of this question I expect certain liberties need to be taken that assume they evolved or were engineered for this purpose. $\endgroup$
    – Nosajimiki
    Commented May 22, 2020 at 16:09

7 Answers 7


You need a way to remove sediment from the sea floor and move it to the roof of the cave

As far as I can tell the biggest problem here is keeping the floor clean. All life in the ecosystem is going to produce some sort of waste that will fall to the floor.

This also means that nutrients will over time migrate from the cave ceiling, and organisms would evolve to make use of the nutrients on the lit floor and break them down into a fine sediment, so essentially soil. Very quickly the sea floor would become completely covered, and even if you were to make the surface ridged, the valleys would quickly overflow with sediment.

On top of that, the trees wouldn’t have any reason to be trees and grow tall if they aren’t competing for sunlight. This will naturally create a canopy as tall as they can possibly can grow, meaning that the ocean floor would be covered in a dense canopy.

So what we need is someway to move the sediment from the sea floor to the rocks above, and to make it difficult to for life to exist on the sea floor.

My first thought that wouldn’t fit your needs super well is just some sort of hand wavium powered paddle that sweeps the dirt away every ten years or so and deposits it in a cavern in the ceiling, to prevent it from falling back down where the trees can get to it. This will always work so long as the machinery is working and sufficient caverns can be dug.

Another solution I had is a biologically engineered bottom feeder that lives on the sea floor, and very quickly gobbles up the sediment and grows fat. It’d preferably have no anus and instead a large bladder that collects it’s waste. You could Also have them just obsessively eat their own feces and have an extremely slow digestive system.

I picture this as an army of shrimp like creatures and living on the sea floor, with a sucker mouth looking that have have a ever growing butt, similar to honeypot ant. Then once it reaches the maximum weight it can transport to the ceiling, it’d crawl up the rock walls in your diagram to the roof. To us this would look incredibly strange but to the inhabitants of this world it’s just be a natural part of the ecosystem like migrating birds. The organism would need to be extremely effective and also stay on the floor it’s entire life. It could even be so effective that any plants growing to close to the floor would be immediately eaten

It’d have a very quick life cycle that ends with the creature flying or crawling to the roof of the cave, burrowing deep into the rock, reproducing and then immediately dying. Then, once born the offspring would immediately burrow out and fall to the ground.

Biggest issue with this guys would be preventing evolution. Over time natural selection would lead to these creatures simply staying on the sea floor and gaining a more efficient digestive system. Even if their reproduction was asexual and produced perfect and they had a crazy effective immune system that prevents all disease, mutations would still arise after enough time

The only real solution I could think of to prevent this adding some sort of technological solution to this. Perhaps each one of them contains a some sort of robotic core that independently prints an embroyo, another robotic core, and a encases this in an egg.

This certainly isn’t the most elegant solution and turns them into robots essentially. They’d also need to get enough metal/silicon/unobtonium throughout their life cycle to make more robotic cores. Maybe add another organism living on the roof, smelting the metal and dropping it the sea floor.

Another solution would be to to make them immortal, and have insane regenerative tissue. Instead of dying on the ceiling they’d just fly up to bury their poop. This probably wouldn’t last as long as if you stick one of them into a blender, they’d have to either die or each little bit of tissue would regenerate, and could possibly lead to mutations arising from the separate clumps. Maybe some combination of the two would work.

Another possibility for an engineered organism would be a tree like creature that produces a large hallow trunk leading to the sea floor, and it would produce some sort of suction and vacume to the sediment to the cave ceiling.

Bottom line though, the only evolution would lead to the floor being covered pretty quickly. If not tree leaves or algae than dirt.

Creating a engineered crature would likely fit the medieval setting better than a machine, but a handwavium mechanism would likely last a lot longer but might look strange in this setting

  • $\begingroup$ If anyone has any feed back I’d appreciate it $\endgroup$
    – Alex
    Commented May 29, 2020 at 22:10
  • $\begingroup$ Tough call between this and redbud's water pump idea. I think they both answer the question equally well as asked, but I'm going to go with this one since I can show the process in action while making it appear a "part of nature" to the locals. $\endgroup$
    – Nosajimiki
    Commented May 31, 2020 at 17:19
  • $\begingroup$ I realize this isn’t what comments are for, But I just had to thank you choosing my answer. After lurking for years on this sit it put a huge smile on my face. $\endgroup$
    – Alex
    Commented Jun 1, 2020 at 6:50
  • $\begingroup$ why deposit in the cave, maybe it deposits on the inner side of the ring. this will be alot like earth in practice, in both nutrients and sediment are transported long distance on earth nutrients flow downstream to the ocean. $\endgroup$
    – John
    Commented Feb 19 at 16:28

This idea might create some new problems, but perhaps to aid the integrity of the torus mechanically, and to aid with some of the problems with debris covering the outer portion of the torus where people have pointed out the incoming light could be blocked.

Figure not drawn to scale. The green dot could be a tree... The arrows are where light can leave the light guide. This could be done where the thin lines are.

Torus with light guiding for interior illumination

The poorly drawn diagram is meant to show that there could be light guides capture the light and direct it to the interior the structure. This could be pretty efficient since total internal reflection is not lossy. In the region with the darker lines the totally internal reflection layer is protected on the outside. In the thinner lines of the light pipe the light can be allowed to leave the light guide in a controlled direction.

This is similar to how an optical fiber works. One way you can let the light out is through the aperture at the top of the pillar (like the end of an optical fiber), or you could also design in patterns that either frustrate the total internal reflection letting the light out, or scatter the light out by changing its direction so it escapes the critical angle. These structures are also big, so you could also have internal mirrors or other optical structures. By the way, optical fibers these days do not have to be solid glass, there are photonic crystal, and "holey" optical fibers. You can also make sensors where you pipe the light to an area, let the light interact with the surface and what you want to sense on the surface etc.

To the people and trees in the interior of the torus these "sunlight pillars" would periodically rise up out of the landscape providing illumination. The upside down trees might be genetically engineered to have a very strong photo-tropic response so they want to grow away from the soil and to the light. To sprout and grow against gravity there might also have to be genetic engineering since normally sprouts want to grow away from the center of the earth.

The light piping structures could also be used other ways. There potentially could be other connections that could pipe the light to other places in the structure that would otherwise be dark. The could penetrate all the way into the soil layer and provide some illumination into the interior of the caves.

The shape of the pillars rising out of the sunlight side could be sculpted to direct the flow of water and debris that might fall to the transparent surface. This could direct the water to areas were the light was guided and concentrated and turn it into vapor so it could be carried back up to the tree and soil/cavern layers. Perhaps it could even burn or vaporize organic materials to help with the cleaning of the optical side of the torus.

If you can control the index of refraction as a function of position, you can also focus the light and make graded index lens this kind of technology could be very versatile in moving the optical energy around.

For the vaporization of materials part, the main concern is having enough optical energy on the outside of the structure, since the brightness and how much you can focus the light is limited by what eventually boils down to the violating the second law of thermodynamics. However, I am assuming the Torus is pretty huge.

In general, being able to control where the optical energy is turned into heat should also give you a lot of engineering flexibility, and maybe help with the climate control.

  • $\begingroup$ This is a really ingenious solution that would require basically no maintenance! Very nice $\endgroup$
    – Alex
    Commented May 29, 2020 at 22:12

i dont quite get your question since you ignore my comment, and i know i probably get backlash from this since i dont know much regarding plant,science, or physic but i try to help......

based on my question on other site i suggest to make hanging/aerial roots or vines like banyan tree or mangrove types of roots where it develop like stalactite type of roots.

enter image description here

to collect water its up to you how long you want the roots is.... maybe it can take the water droplet if it only reach sunlite caves, but i suggest it to surpass the water at least so the roots can be a substitute or assist to do photosynthesis like assimilitory roots to collect sunlight, get water and nutrients from the dead plants and maybe it can help clearing it out, or make it work like carnivore plants by making the roots or vines that can constrict and entrap animal that touch it assuming the water or the sunlite caves has animals in it to get nutrients or to stuck it to the soil.

  • 1
    $\begingroup$ Sorry, I did not really understand what part you did not understand so I was assuming you were having the same issue as Slarty with the order of layers. This was one of my first thoughts too, but would not necessarily be a good option for plants that are 100s of meters above the water. $\endgroup$
    – Nosajimiki
    Commented May 23, 2020 at 19:02
  • 1
    $\begingroup$ Although looking back at your diagram, you may have been confused when I called the "firmament" unobtanium. On this site this is a general term used to refer to any super material that has unachievable physical properties according to known sciences. In this case, it is a thin clear crust with the ability to contain the weight all the crazy forces working against it. Most likely roots could not burrow through such a substance. $\endgroup$
    – Nosajimiki
    Commented May 26, 2020 at 14:18

What a fascinating question.

You have many questions rolled up in your premise, so I will try to answer the most pertinent ones, though that would still be just scratching the surface of what would be a fantastically intricate ecosystem.

Keeping the sea "floor" clean, and water cycle

I will start with this because the other two areas reference my thoughts in this section.

For your firmament's glass-like surface, I would suggest either a hydrophobic or self cleaning hydrophilic surface. I know that the hydrophobic surface can be made permanent by micro-structures etched into the surface. This would keep the water, and the things contained within it, separate from the surface due to surface tension. Due to internal reflection, the water's interface with the glass would look silvery at oblique angles, while being more or less clear when viewed straight on. The hydrophilic coating is another option that actually binds the water and particles to it, and uses UV light to break down the molecules. This is used in some self-cleaning windshields. It might have less of an internal reflection issue than the other one.

In my opinion, the best solution to gather debris dropping into the ocean is to use the currents in the water itself. You have a number of ways to induce currents. One of the best forms would be eddy currents, or vortices, because they have a tendency to gather particles inwards to the center. This could be partially due to structural members protruding like pillars in the water, and the overall motion could be provided by the Coriolis effect and temperature convection, since you are using a rotating habitat.

  • Option 1

Of course now you need a way to get that organic debris back up to the ceiling, so one option, since in your graphic that you have pillars in the cave like area, that you make these the center of the eddy like currents with hollow central tube structures. These can act like straws that will suck debris up from the bottom and redistribute it through the ceiling via percolation. As for the power source, you can use a heat differential pump powered by the sun. The fluidyne (https://en.wikipedia.org/wiki/Fluidyne_engine) is a type of Stirling engine with no moving parts that can be used to move water and last practically indefinitely. You could use a lens built into the firmament to create a hot spot and use the water to cool the "cold" liquid piston.

  • Option 2

Using mist and condensation to give the plants moisture seems to be feasible in my opinion, as you can have structures that would absorb and wick water upwards. Plants already have a mechanism for transporting water upwards via capillary action and negative pressure caused by respiration. Essentially, allowing the water at the top end to evaporate to draw up water behind it by creating a decrease in pressure to fill.

However, this would not solve the problem of losing organic matter to gravity, which ends up in the water. Bits and pieces will fall over time, and that's not something that can be raised into the air via water droplets very easily unless you have something breaking down organic matter in the ocean. If you did, that would not make the ocean area very hospitable to life. I'm not sure whether or not that's a good or bad thing in your case, but that way, you might be able to aerosolize the dissolved organic matter along with the water to get it back up to the plants without using a direct fluid pump.

You might need to do something to purify water that slowly breaks down organic materials though. Maybe filtering or boiling it would be enough?

As for how to aerosolize the water, I feel I should at least throw some idea out there. Of course there's evaporation, but to supplement it, perhaps solar heated areas like geysers that have nozzle or sponge like mineral formations at the top that would produce steam or mist? Perhaps the heat of the geyser would deactivate the organic dissolving compounds, allowing vapor and condensed water to be safe for humans and plants to consume.


[Unrelated, but interesting thought - You might consider other possibilities than a seamless clear "sky". It might be interesting, for instance to have a cellular mesh like firmament that resembles something like a dragonfly's wings, or a hexagon lattice. Just my two cents.]

The issue of structural integrity of the ceiling and root systems

One could imagine an ecosystem where no soil is left anymore, but is made of recycled plant matter held in place by a constantly regenerating matrix of plant fibers.There could be a multi generational tangled organic mat of roots, fungi, and bacteria built on the decaying framework of the old, using it for material, nutrition, and as a scaffolding framework for building upon.

Some plants might not even have life cycles like earth plants, but bud asexually, or be colonies of smaller organisms that are periodically replaced. Maybe even they could be similar to coral. Or perhaps there's a tunnel digging ant like creatures that open up new tunnels, that when they migrate, they abandon that nest, and it provides new avenues for plant growth.

Porous rock or sponge like organic growth medium with strong fiber reinforcement could make up the ceiling while allowing water and nutrients to percolate down from above. There are tube like calcium formations and things akin to natural fiberglass you might be interested in.

If adhesion to the ceiling is an issue, tap roots could act as main supporting elements, which the rest of the plants can focus around and use for support.

Nutrient recycling

Fungi, bacteria and nutritionally complementing plants sharing by-products via root systems. Certain plants benefit from one another's by products, which prevents the area from becoming nutritionally depleted. You might look up the term "companion planting" for ideas on how this might work.

The fungi will be especially important for recycling dead plant matter. Plants practially always exist side by side with fungi and bacteria.

The bacteria fulfills some important roles too. One would be nitrogen fixation which transfers the nitrogen in the air into organic compounds that can be used by plants as a sort of fertilizer.

[Bacteria could potentially evolve into disease and blight, which would be a problem for asexually budding plants. Perhaps they could communicate with their roots chemically and kill off a buffer zone barrier to create a wall to contain/quarantine the disease. Like a fire break in a forest. There is actually some evidence that trees can use fungal networks to share nutrients and for threat signaling.] http://www.bbc.com/earth/story/20141111-plants-have-a-hidden-internet

At least for Earth plants, nitrogen, phosphorus, potassium, magnesium, sulfur, and calcium, and hydrocarbons, are particularly important for their growth. Notice that these elements are opposite sides of the periodic table. Having a concentration of one side more than the other will result in an overly acidic or basic environment. Plants need a certain balance of these elements to grow properly. Not sure if you wanted to get that detailed, but I thought I'd put it out there.


Anyway, that's just about all I have on the matter!

(Although one could potentially write books on all the possibilities here. These kind of questions are why I come to this site! So thank you for the fun question!)


The most likely outcome would be aquatic plants that spread upside down leaves across the water surface. And / or plants that grew across the unobtainium envelope at the bottom of the sea, this could drastically reduce the light levels in the cave.

But assuming there is something in the water to prevent plant growth there then plants might anchor outside the cave entrance and grow in and down.

Assuming the cave is very large and still has plant coverage away from the entrance then they would have to cling to the ceiling. If there were many irregularities, cracks and ridges this would help and ledges in the wall especially so. Water and nutrients might reach such plants through the rock itself if it was porous. Such plants would probably produce long lightweight stems that hung downward full of leaves. Any substantial weight might over stress their tenuous hold to the ceiling.

One other possibility depending on the toxicity of the sea and the height of the cave would be for the plants to grow down from the ceiling until reaching the surface and then create some form of gas bladder (like some seaweed) or a wooden floating platform just below the surface. This could help stabilise the trunk by supporting some of the weight and allow it to grow to a greater diameter. Branches with leaves could then grow out all around the sides of this trunk.

One problem would be the constant rain of debris from dead bits of plant to crumbling fragments of cave wall that fell to the water and sank to the unobtainium shell obscuring the light source. This might be solved if the unobtainium was sloped into ridges. These would tend to collect the debris. Another option would be some form of heat driven circulation that swept those debris away.


Could the root system be extremely extensive? An ancient root system through the rock face that could have eaten through aquifers (assuming that, like on earth, there can be water pockets in rock formations) and eventually, these roots extended down to the source. Maybe the roots even appear on the sea floor and grow like seaweed. Since this is right at the firmament, maybe it's these that really get the majority of sunlight, and the trees you see at the top of the cave are merely for reproduction/seeding. I can imagine this could either be natural, or as you suggested could be the case, engineered by humans, as two different plant species transplanted together long ago.

As for how the tree re-seeds, I imagine some kind of bird or climbing mammal eats it, makes a nest or dwelling in the rock above, where the tree grows from, and poops out the seeds there continuing the cycle.

What a cool idea for a world, by the way. I'm very intrigued!


We could assume that Schistostega was down there to, and that the tree was anchored to solid rock and holding on to some dirt. Water from pools at the bottom of the cave would evaporate, drench the soil, and boom! Now you have light and water to support your trees. We could also assume that some bats lived down there and left droppings in these pools, nd yay nutrients!


You must log in to answer this question.

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