My question is fairly straightforward, but answering it is proving rather difficult.

The scenario is on a planet with some water. The surface is dominated by deserts, both hot and cold depending on latitude. However, the atmosphere must be breathable by humans.

So the question is: What is the minimum amount of surface water that a planet could have that would sustain sufficient photosynthesis to render the atmosphere breathable?

For reference - The planet has a mass and size similar to the Earth, and is also similar in geological composition. The major difference is that it received a smaller compliment of water, or has lost water somehow.

  • $\begingroup$ The primordial atmosphere would have been devoid of oxygen, yes. :) I envisage that the planet is geologically fairly Earth-like. It just has a significantly reduced compliment of water. $\endgroup$ Dec 30, 2017 at 5:48
  • $\begingroup$ It is similar in size and mass to the Earth. $\endgroup$ Dec 30, 2017 at 5:50
  • $\begingroup$ I am not familiar with where atmospheric hydrogen comes from, so it would depend on whether the process that produces it is able to operate on such a planet. $\endgroup$ Dec 30, 2017 at 5:54

4 Answers 4


Let's break this down in terms of human oxygen use v. plant oxygen production first.

The average human uses around 550 litres of O2 per day. To generate that much O2, you need a little less than 4Kgs of plant GROWTH per day. I don't have any charts at hand right now on how many plants that actually means, but assuming that the O2 levels in the atmosphere (I'll get to that) are already fit for humans, and assuming plant growth rates of 0.5% per day, you'd need nearly a metric tonne of plants per human just to keep the O2 levels as they are. That's a lot of plants, and it means that per human you're looking at a lot of water to keep the plants (not only alive but) growing, plus the water humans need to drink by themselves.

The other part of this is that you can't just have that much water in the hydrosphere; that's what has to be available to your plants. That means if you have a similar water cycle as we do on Earth (evaporation, condensation, precipitation, rinse and repeat) then you probably need a LOT more water than these minimums.

But for the sake of argument
Let's assume that you have isolated pockets of oases and that humans can live around these and the water doesn't evaporate (or if it does, it condenses very locally). How did O2 levels get up to breathable (around 20%) levels in the atmosphere?

The reality is that what makes O2 so essential for life is what also makes is very volatile. Oxidisation is one of the major results of any exothermic reaction that can occur in our universe. Essentially, O2 is a reserve of energy that can be used in a myriad of ways other than sustaining life. It can be burned in fire, it can be combined with Hydrogen to form water; it's what causes rust. What all this means is that without plants constantly (for billions of years) producing O2 and releasing it as a waste gas (Yes, plants use O2 themselves but they produce more than they need and can't store it because of its volatility) the atmosphere of any planet is highly unlikely to contain a lot of 'natural' O2 as it will have been used by many other processes on the planet over time.

So; in reality, you actually need a LOT of plants for a very long time, using heaps of water to grow and release excess O2 to get the atmosphere up to breathable levels.

Then there's the thermal mass problem.

Water by it's very nature is a stable thermal mass. Thermal mass means that it can absorb heat and release it over time, evening out temperature fluctuations on your planet.

On Earth, this is why coastal settlements have milder weather than inland, where there is less water. It's also why Earth deserts are hot during the day, and freezing at night. Even if you only have the minimum water needed to support O2 production, your humans are going to be very uncomfortable because of the extremes of temperature in your arid environment.

We haven't even spoken about food yet. Remember, that your plants are not just there to provide O2 (and shade from the blistering sun thanks to no thermal mass); they're also your food supply. I'm going to assume we're not meat eaters on this planet because then we have to feed and water the food animals, who also need O2, etc.

Let's assume that your humans only eat potatoes. (This is actually close to possible thanks to their chemical composition, including many trace nutritional elements, but not completely possible given poor protein levels as I understand it.) The average person is going to want to eat around 15-20 potatoes per day to be comfortable in a high physical workload environment. Let's say that on your planet, these take around 100 days to grow; that means you need around 2000 potatoes per person under cultivation at any point in time just to maintain a subsistence diet.

Yes, these potato plants also contribute to the O2 levels, but also remember that in this scenario, 1% of that growth is being consumed every day.

I don't have all the numbers at my disposal, but estimating based on this, I'd say that if you don't have around 10% of your surface under water in some form and preferably deeply, human sustainment becomes unviable due to the sheer lack of availability of useable water for plants unless you're VERY close to lakes, etc. Rivers become less likely, and storms are going to be less wet, more electrical in nature making your existence... well, unpleasant.

We haven't even touched on the evolutionary challenges of a planet without water and I won't as that is probably a bridge too far for the scope of this question.

All in all, using conventional water cycle mechanics, you still need a lot of water (and a lot of plants) to make your existence possible, and while that doesn't preclude deserts, there still has to be large lakes or small oceans from which that water cycle can draw to support plant life somewhere around your planet.

  • $\begingroup$ 10% water coverage is actually entirely within what I had in mind for the planet. I had assumed that there would need to bodies of water somewhere to allow plant life to exist. That leaves 90% of the planet dominated by inhospitable deserts. Do you think it is likely that the water bodies would be concentrated near the poles or not necessarily? $\endgroup$ Dec 30, 2017 at 7:53
  • 1
    $\begingroup$ Potatoes contain very very little protein and a lot of carbohydrates; to get enough protein from potatoes a person would need to eat about 2.5 kg of the stuff per day, which would exceed the required amount of calories by a factor of 5 or more! $\endgroup$
    – AlexP
    Dec 30, 2017 at 9:19
  • $\begingroup$ Hi @AlexP; you're right of course. I'm not suggesting that it would be a viable option, just that all the elements that are needed to support life are more or less there. They're certainly not in the correct proportion but it's not a bad vegetable analogue to use for food production and water requirement purposes. If we were doing this in real life, you'd want a genetically modified spud that was lower in carbs, higher in protein to make it viable as a singular food source. $\endgroup$
    – Tim B II
    Dec 30, 2017 at 9:24
  • $\begingroup$ Potatoes actually are not calorie dense. They run about 75 cal/100g 2.5 Kg spuds = 1875 cal. This fits closely with the history of the Irish potato famine, where they mentioned that the typical farmer ate 8 pounds of boiled spuds a day. $\endgroup$ Jan 1, 2018 at 15:09
  • $\begingroup$ Also it's easy to grow several tons of potatoes per acre. (With current farming techiniques about 20 t/ac. ) $\endgroup$ Jan 1, 2018 at 15:11

From OP:

The major difference is that it received a smaller compliment of water, or has lost water somehow.

You can have shallow seas over much of your planet in the geologically recent past. These seas are full of plant life and your planet's atmosphere is oxygen rich.

Then your planet's star heats up and the planet gets warmer. The oceans dry up, the water becoming vapor. The ocean beds become deserts.

The accumulated oxygen in the atmosphere will not disappear overnight. Demands on oxygen will be low. There is not much animal life and easily oxidizable (flammable) things on the surface will already have oxidized.

So: your world is dying. The oxygen it has will be gradually depleted over millions of years. Perhaps a bigger threat over this period is the star, if the process causing it to put out more heat continues to accelerate.

  • $\begingroup$ Not a bad thought! In the context of the setting the planet is already home to a sapient alien species, so they might perhaps be a remnant population from the planet's wetter past. $\endgroup$ Dec 30, 2017 at 7:55
  • 1
    $\begingroup$ Actually, this scenario sounds very similar to that used by Edgar Rice Burroughs in the Barsoom series to describe Mars. The only difference is that to prolong the life of the planet, there are several 'atmospheric generators' that release O2 back into the atmosphere. In the books, these are pretty much cargo cult artefacts that have been left behind by an ancient species from Mars. $\endgroup$
    – Tim B II
    Dec 30, 2017 at 9:20
  • $\begingroup$ @Tim B - I thought of Barsoom also while writing this answer. $\endgroup$
    – Willk
    Dec 30, 2017 at 20:30
  • $\begingroup$ Geology consumes significant oxygen. Search 'chemical weathering' $\endgroup$ Jan 1, 2018 at 15:20
  • 1
    $\begingroup$ @Sherwood Botsford: in the absence of water / rain, I suspect exposed external surfaces of oxidizable substances would oxidize and protect internal materials from further oxidation. There are no processes which would then expose these materials. $\endgroup$
    – Willk
    Jan 2, 2018 at 17:19

0% surface water

Lets take a page from the dying planet trope, a la A Princess of Mars.

What if an Earth-like planet is slowly dying out? Perhaps its magnetosphere has dissipated and allowed water to be stripped from its atmosphere (note that solar wind will remove water faster than either nitrogen or oxygen due it its lower molecular weight). Alternately, what if the sun is expanding into a red giant and slowly baking our planet dry?

Either way, what was once a verdant landscape could be turned into a hellish desert under a searing sun. The conditions that once allowed photosynthesizers to colonize the planet and create an oxygen atmosphere are no more, but the atmosphere of oxygen and nitrogen remain.

You don't have to vary the water content of your world to make life possible, you can simply change the time of the story until a point after which most of the water has been removed.

With this explanation, you can have as little water as you want, with life clinging on in the few areas the liquid of life is still found.

  • 2
    $\begingroup$ +1. Just a clarification: oxygen would be depleted in a short time (geologically speaking), mainly due to chemical rock weathering. It is estimated that on Earth chemical weathering of rocks uses about 4E11 kg of oxygen per year; the total amount of oxygen in the atmosphere is about 1E18 kg, so if oxygen production stopped we'd expect it to be mostly gone in a few million years. $\endgroup$
    – AlexP
    Dec 30, 2017 at 23:12

On earth the phytoplankton in the ocean are a big share of photosyntheses.

As an alternative, suppose that the process that produces tectonic plates is balanced toward more plates. You can have the same amount of water, but overall smaller deeper oceans. Current average ocean depth is about 12,000 feet. If you divided the area of the oceans by a factor of 3, you'd run 25% of the surface as water, 75% land, and you'd need 40-50 thousand foot deep oceans.

Your second factor to play with is the distribution of water: If you have many Black sea sized puddles, instead of a few north Atlantics, you end up with a somewhat more moderate climate. (I don't know if rock is strong enough to have a 50,000 foot deep Black Sea. Slope too steep?) Small puddles may require just less water.

You may want to make your world just somewhat drier instead of lots drier. Typically under 10 inches of rain per year is considered desert, but it will still support an ecology. Vegetation is sparse, and separated by patches of bare ground.

Much of the western U.S. is semi-desert with about 12-16 inches of rain a year. Grass and bush is present, but there are sp,e spaces of bare earth between plants. Water is the limiting factor. Somewhere between 16 and 20 inches a year you get tall grass prairie -- vegetation cover is continuous. At 25 inches per year depending on temperature you get forests.

Geology affects deserts: Look at Australia. At that latitude the prevailing wind is from the east. Hits the mountainous east coast, which wrings most of the water out of it, and behind that range you have an awful lot of not much. If you rotated the continent by 180 degrees, you would have a grassy interior. Or if you moved it south 15 degrees latitude to where the winds come from the north west.

Another factor in developing civilization: Water transport. Water is by far the cheaper way to move stuff, and to travel in a hurry. With more land than water, fast movement is relays of horses. The Pony Express could get a letter across the U.S. in 10 days. And it cost 10 bucks to do it -- a few weeks pay. Rome shipped tin from England, and grain from all over the Mediterranean.

Compare the early settlement of Canada vs the U.S. The Hudson Bay Co. had a network of fur trading posts that spread from York Factory on the Bay to the mouth of the Columbia. The trade cycle took 4 years between goods shipped from England to furs from the remote regions arriving back. A York Boat could carry 6 tons of cargo and had a crew of 6-8. Their competition, the North West company, used large (up to 36 feet) canoes. They could carry about half the cargo at with about half again more crew -- overall about 1/3 the tonnage per crew. They were MUCH easier to portage however. Tons per crew man year favoured the canoe except on the Hayes, Saskatchewan, Athabasca and Slave Rivers, where the lack of portages and where the river size allowed the use of sail at times.

George Simpson had a crack crew of Iroquois when he was chief Factor. He could travel from the depot at Thunder Bay to Lower Fort Garry, a distance of about 700 miles in 11 days. Freight canoes took about twice this.

In the U.S. with the only really comparable east-west river being the upper Missouri, along with the Great Lakes. Most goods moved by wagon. Tonnage per person was on par with the northern methods, but where the typical York boat or Canot du Maitre did 40 miles a day, wagons typically did 10. Where trails were rough in the mountains, early traders used pack horses or mules. This isn't efficient: each animal can carry only a couple hundred pounds, and needs hours to graze. It also requires considerable time harnessing and unharnesing.


Later, in eastern America there was a flurry of canal building. One horse pulling a canal boat moved many times the freight of a horse and wagon.

Having limited water doesn't mean no water transport: Canals moving water from where it does rain to where it's used. Then boats on the canals. Wars could be fought over the control of locks. Conventions that you don't damage the lock itself

Continental U.S. wheel and foot transport.

Northern Canada -- water transport.

The rainfall regimes are't much different. However much of Canada is granite shield -- not very water absorbent. And with every glaciation the hydrology gets messed up, so what water that is there sits in puddles. Cooler temperatures reduce evaporation.

So: You could have wet or dry depending on the geologic history of the place.

Rivers in deserts and steppes are also possible. Look at the Jordan River in the middle east, and the Tigris-Euphrates. In both cases the water sources are in the mountains with higher precipitation. There are similar rivers on coastal Peru, one of the driest deserts in the world.

A world is a big place. Lots of room for variations.


You must log in to answer this question.

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